summaryrefslogtreecommitdiffstats
path: root/contrib/diff/lib
diff options
context:
space:
mode:
authordelphij <delphij@FreeBSD.org>2007-06-15 07:06:13 +0000
committerdelphij <delphij@FreeBSD.org>2007-06-15 07:06:13 +0000
commit40ec4fdc9a74bfdb83f13672acdb88af5c91ab46 (patch)
treefa574500c8a91fba34c02a5fbe177ac9ef1890ea /contrib/diff/lib
parentf75ef5d3d4da5cc670cb41c22a97279e60c6cb0c (diff)
downloadFreeBSD-src-40ec4fdc9a74bfdb83f13672acdb88af5c91ab46.zip
FreeBSD-src-40ec4fdc9a74bfdb83f13672acdb88af5c91ab46.tar.gz
Virgin import of diffutils 2.8.7.
Diffstat (limited to 'contrib/diff/lib')
-rw-r--r--contrib/diff/lib/Makefile.am35
-rw-r--r--contrib/diff/lib/alloca.c489
-rw-r--r--contrib/diff/lib/alloca_.h68
-rw-r--r--contrib/diff/lib/basename.c79
-rw-r--r--contrib/diff/lib/c-stack.c311
-rw-r--r--contrib/diff/lib/c-stack.h19
-rw-r--r--contrib/diff/lib/cmpbuf.c147
-rw-r--r--contrib/diff/lib/cmpbuf.h21
-rw-r--r--contrib/diff/lib/dirname.c119
-rw-r--r--contrib/diff/lib/dirname.h43
-rw-r--r--contrib/diff/lib/error.c310
-rw-r--r--contrib/diff/lib/error.h66
-rw-r--r--contrib/diff/lib/exclude.c263
-rw-r--r--contrib/diff/lib/exclude.h43
-rw-r--r--contrib/diff/lib/exit.h32
-rw-r--r--contrib/diff/lib/exitfail.c27
-rw-r--r--contrib/diff/lib/exitfail.h20
-rw-r--r--contrib/diff/lib/file-type.c75
-rw-r--r--contrib/diff/lib/file-type.h166
-rw-r--r--contrib/diff/lib/fnmatch.c403
-rw-r--r--contrib/diff/lib/fnmatch_.h63
-rw-r--r--contrib/diff/lib/fnmatch_loop.c1192
-rw-r--r--contrib/diff/lib/getopt.c1254
-rw-r--r--contrib/diff/lib/getopt.h176
-rw-r--r--contrib/diff/lib/getopt1.c191
-rw-r--r--contrib/diff/lib/getopt_int.h129
-rw-r--r--contrib/diff/lib/gettext.h68
-rw-r--r--contrib/diff/lib/gettimeofday.c121
-rw-r--r--contrib/diff/lib/gnulib.mk90
-rw-r--r--contrib/diff/lib/hard-locale.c74
-rw-r--r--contrib/diff/lib/hard-locale.h24
-rw-r--r--contrib/diff/lib/imaxtostr.c3
-rw-r--r--contrib/diff/lib/inttostr.c49
-rw-r--r--contrib/diff/lib/inttostr.h47
-rw-r--r--contrib/diff/lib/malloc.c36
-rw-r--r--contrib/diff/lib/mkstemp.c42
-rw-r--r--contrib/diff/lib/offtostr.c3
-rw-r--r--contrib/diff/lib/posixver.c59
-rw-r--r--contrib/diff/lib/posixver.h1
-rw-r--r--contrib/diff/lib/prepargs.c93
-rw-r--r--contrib/diff/lib/prepargs.h3
-rw-r--r--contrib/diff/lib/quotesys.c125
-rw-r--r--contrib/diff/lib/quotesys.h9
-rw-r--r--contrib/diff/lib/realloc.c39
-rw-r--r--contrib/diff/lib/regex.c8298
-rw-r--r--contrib/diff/lib/regex.h556
-rw-r--r--contrib/diff/lib/setmode.c67
-rw-r--r--contrib/diff/lib/setmode.h27
-rw-r--r--contrib/diff/lib/stdbool_.h93
-rw-r--r--contrib/diff/lib/strcase.h35
-rw-r--r--contrib/diff/lib/strcasecmp.c66
-rw-r--r--contrib/diff/lib/strftime.c1330
-rw-r--r--contrib/diff/lib/stripslash.c39
-rw-r--r--contrib/diff/lib/strncasecmp.c2
-rw-r--r--contrib/diff/lib/strtoimax.c80
-rw-r--r--contrib/diff/lib/strtol.c432
-rw-r--r--contrib/diff/lib/strtoll.c33
-rw-r--r--contrib/diff/lib/strtoul.c20
-rw-r--r--contrib/diff/lib/strtoull.c27
-rw-r--r--contrib/diff/lib/strtoumax.c2
-rw-r--r--contrib/diff/lib/tempname.c344
-rw-r--r--contrib/diff/lib/time_r.c69
-rw-r--r--contrib/diff/lib/time_r.h46
-rw-r--r--contrib/diff/lib/umaxtostr.c3
-rw-r--r--contrib/diff/lib/unlocked-io.h132
-rw-r--r--contrib/diff/lib/version-etc.c176
-rw-r--r--contrib/diff/lib/version-etc.h37
-rw-r--r--contrib/diff/lib/waitpid.c72
-rw-r--r--contrib/diff/lib/xalloc.h87
-rw-r--r--contrib/diff/lib/xmalloc.c255
-rw-r--r--contrib/diff/lib/xstrdup.c33
-rw-r--r--contrib/diff/lib/xstrtol.c295
-rw-r--r--contrib/diff/lib/xstrtol.h91
-rw-r--r--contrib/diff/lib/xstrtoul.c6
-rw-r--r--contrib/diff/lib/xstrtoumax.c37
75 files changed, 19417 insertions, 0 deletions
diff --git a/contrib/diff/lib/Makefile.am b/contrib/diff/lib/Makefile.am
new file mode 100644
index 0000000..dbf8b0f
--- /dev/null
+++ b/contrib/diff/lib/Makefile.am
@@ -0,0 +1,35 @@
+# Automakefile for GNU Diffutils library.
+
+# Copyright (C) 2001, 2002, 2004 Free Software Foundation, Inc.
+
+## This program is free software; you can redistribute it and/or modify
+## it under the terms of the GNU General Public License as published by
+## the Free Software Foundation; either version 2, or (at your option)
+## any later version.
+
+## This program is distributed in the hope that it will be useful,
+## but WITHOUT ANY WARRANTY; without even the implied warranty of
+## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+## GNU General Public License for more details.
+
+## You should have received a copy of the GNU General Public License
+## along with this program; if not, write to the Free Software
+## Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+## 02111-1307, USA.
+
+noinst_LIBRARIES = libdiffutils.a
+
+noinst_HEADERS = cmpbuf.h prepargs.h quotesys.h
+
+libdiffutils_a_SOURCES = $(lib_SOURCES)
+lib_SOURCES = cmpbuf.c prepargs.c quotesys.c
+
+libdiffutils_a_LIBADD = @ALLOCA@ @LIBOBJS@
+libdiffutils_a_DEPENDENCIES = $(libdiffutils_a_LIBADD)
+
+BUILT_SOURCES =
+DISTCLEANFILES =
+EXTRA_DIST = setmode.h waitpid.c
+MOSTLYCLEANFILES =
+
+include gnulib.mk
diff --git a/contrib/diff/lib/alloca.c b/contrib/diff/lib/alloca.c
new file mode 100644
index 0000000..a580078
--- /dev/null
+++ b/contrib/diff/lib/alloca.c
@@ -0,0 +1,489 @@
+/* alloca.c -- allocate automatically reclaimed memory
+ (Mostly) portable public-domain implementation -- D A Gwyn
+
+ This implementation of the PWB library alloca function,
+ which is used to allocate space off the run-time stack so
+ that it is automatically reclaimed upon procedure exit,
+ was inspired by discussions with J. Q. Johnson of Cornell.
+ J.Otto Tennant <jot@cray.com> contributed the Cray support.
+
+ There are some preprocessor constants that can
+ be defined when compiling for your specific system, for
+ improved efficiency; however, the defaults should be okay.
+
+ The general concept of this implementation is to keep
+ track of all alloca-allocated blocks, and reclaim any
+ that are found to be deeper in the stack than the current
+ invocation. This heuristic does not reclaim storage as
+ soon as it becomes invalid, but it will do so eventually.
+
+ As a special case, alloca(0) reclaims storage without
+ allocating any. It is a good idea to use alloca(0) in
+ your main control loop, etc. to force garbage collection. */
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <string.h>
+#include <stdlib.h>
+
+#ifdef emacs
+# include "lisp.h"
+# include "blockinput.h"
+# ifdef EMACS_FREE
+# undef free
+# define free EMACS_FREE
+# endif
+#else
+# define memory_full() abort ()
+#endif
+
+/* If compiling with GCC 2, this file's not needed. */
+#if !defined (__GNUC__) || __GNUC__ < 2
+
+/* If someone has defined alloca as a macro,
+ there must be some other way alloca is supposed to work. */
+# ifndef alloca
+
+# ifdef emacs
+# ifdef static
+/* actually, only want this if static is defined as ""
+ -- this is for usg, in which emacs must undefine static
+ in order to make unexec workable
+ */
+# ifndef STACK_DIRECTION
+you
+lose
+-- must know STACK_DIRECTION at compile-time
+/* Using #error here is not wise since this file should work for
+ old and obscure compilers. */
+# endif /* STACK_DIRECTION undefined */
+# endif /* static */
+# endif /* emacs */
+
+/* If your stack is a linked list of frames, you have to
+ provide an "address metric" ADDRESS_FUNCTION macro. */
+
+# if defined (CRAY) && defined (CRAY_STACKSEG_END)
+long i00afunc ();
+# define ADDRESS_FUNCTION(arg) (char *) i00afunc (&(arg))
+# else
+# define ADDRESS_FUNCTION(arg) &(arg)
+# endif
+
+/* Define STACK_DIRECTION if you know the direction of stack
+ growth for your system; otherwise it will be automatically
+ deduced at run-time.
+
+ STACK_DIRECTION > 0 => grows toward higher addresses
+ STACK_DIRECTION < 0 => grows toward lower addresses
+ STACK_DIRECTION = 0 => direction of growth unknown */
+
+# ifndef STACK_DIRECTION
+# define STACK_DIRECTION 0 /* Direction unknown. */
+# endif
+
+# if STACK_DIRECTION != 0
+
+# define STACK_DIR STACK_DIRECTION /* Known at compile-time. */
+
+# else /* STACK_DIRECTION == 0; need run-time code. */
+
+static int stack_dir; /* 1 or -1 once known. */
+# define STACK_DIR stack_dir
+
+static void
+find_stack_direction (void)
+{
+ static char *addr = NULL; /* Address of first `dummy', once known. */
+ auto char dummy; /* To get stack address. */
+
+ if (addr == NULL)
+ { /* Initial entry. */
+ addr = ADDRESS_FUNCTION (dummy);
+
+ find_stack_direction (); /* Recurse once. */
+ }
+ else
+ {
+ /* Second entry. */
+ if (ADDRESS_FUNCTION (dummy) > addr)
+ stack_dir = 1; /* Stack grew upward. */
+ else
+ stack_dir = -1; /* Stack grew downward. */
+ }
+}
+
+# endif /* STACK_DIRECTION == 0 */
+
+/* An "alloca header" is used to:
+ (a) chain together all alloca'ed blocks;
+ (b) keep track of stack depth.
+
+ It is very important that sizeof(header) agree with malloc
+ alignment chunk size. The following default should work okay. */
+
+# ifndef ALIGN_SIZE
+# define ALIGN_SIZE sizeof(double)
+# endif
+
+typedef union hdr
+{
+ char align[ALIGN_SIZE]; /* To force sizeof(header). */
+ struct
+ {
+ union hdr *next; /* For chaining headers. */
+ char *deep; /* For stack depth measure. */
+ } h;
+} header;
+
+static header *last_alloca_header = NULL; /* -> last alloca header. */
+
+/* Return a pointer to at least SIZE bytes of storage,
+ which will be automatically reclaimed upon exit from
+ the procedure that called alloca. Originally, this space
+ was supposed to be taken from the current stack frame of the
+ caller, but that method cannot be made to work for some
+ implementations of C, for example under Gould's UTX/32. */
+
+void *
+alloca (size_t size)
+{
+ auto char probe; /* Probes stack depth: */
+ register char *depth = ADDRESS_FUNCTION (probe);
+
+# if STACK_DIRECTION == 0
+ if (STACK_DIR == 0) /* Unknown growth direction. */
+ find_stack_direction ();
+# endif
+
+ /* Reclaim garbage, defined as all alloca'd storage that
+ was allocated from deeper in the stack than currently. */
+
+ {
+ register header *hp; /* Traverses linked list. */
+
+# ifdef emacs
+ BLOCK_INPUT;
+# endif
+
+ for (hp = last_alloca_header; hp != NULL;)
+ if ((STACK_DIR > 0 && hp->h.deep > depth)
+ || (STACK_DIR < 0 && hp->h.deep < depth))
+ {
+ register header *np = hp->h.next;
+
+ free (hp); /* Collect garbage. */
+
+ hp = np; /* -> next header. */
+ }
+ else
+ break; /* Rest are not deeper. */
+
+ last_alloca_header = hp; /* -> last valid storage. */
+
+# ifdef emacs
+ UNBLOCK_INPUT;
+# endif
+ }
+
+ if (size == 0)
+ return NULL; /* No allocation required. */
+
+ /* Allocate combined header + user data storage. */
+
+ {
+ /* Address of header. */
+ register header *new;
+
+ size_t combined_size = sizeof (header) + size;
+ if (combined_size < sizeof (header))
+ memory_full ();
+
+ new = malloc (combined_size);
+
+ if (! new)
+ memory_full ();
+
+ new->h.next = last_alloca_header;
+ new->h.deep = depth;
+
+ last_alloca_header = new;
+
+ /* User storage begins just after header. */
+
+ return (void *) (new + 1);
+ }
+}
+
+# if defined (CRAY) && defined (CRAY_STACKSEG_END)
+
+# ifdef DEBUG_I00AFUNC
+# include <stdio.h>
+# endif
+
+# ifndef CRAY_STACK
+# define CRAY_STACK
+# ifndef CRAY2
+/* Stack structures for CRAY-1, CRAY X-MP, and CRAY Y-MP */
+struct stack_control_header
+ {
+ long shgrow:32; /* Number of times stack has grown. */
+ long shaseg:32; /* Size of increments to stack. */
+ long shhwm:32; /* High water mark of stack. */
+ long shsize:32; /* Current size of stack (all segments). */
+ };
+
+/* The stack segment linkage control information occurs at
+ the high-address end of a stack segment. (The stack
+ grows from low addresses to high addresses.) The initial
+ part of the stack segment linkage control information is
+ 0200 (octal) words. This provides for register storage
+ for the routine which overflows the stack. */
+
+struct stack_segment_linkage
+ {
+ long ss[0200]; /* 0200 overflow words. */
+ long sssize:32; /* Number of words in this segment. */
+ long ssbase:32; /* Offset to stack base. */
+ long:32;
+ long sspseg:32; /* Offset to linkage control of previous
+ segment of stack. */
+ long:32;
+ long sstcpt:32; /* Pointer to task common address block. */
+ long sscsnm; /* Private control structure number for
+ microtasking. */
+ long ssusr1; /* Reserved for user. */
+ long ssusr2; /* Reserved for user. */
+ long sstpid; /* Process ID for pid based multi-tasking. */
+ long ssgvup; /* Pointer to multitasking thread giveup. */
+ long sscray[7]; /* Reserved for Cray Research. */
+ long ssa0;
+ long ssa1;
+ long ssa2;
+ long ssa3;
+ long ssa4;
+ long ssa5;
+ long ssa6;
+ long ssa7;
+ long sss0;
+ long sss1;
+ long sss2;
+ long sss3;
+ long sss4;
+ long sss5;
+ long sss6;
+ long sss7;
+ };
+
+# else /* CRAY2 */
+/* The following structure defines the vector of words
+ returned by the STKSTAT library routine. */
+struct stk_stat
+ {
+ long now; /* Current total stack size. */
+ long maxc; /* Amount of contiguous space which would
+ be required to satisfy the maximum
+ stack demand to date. */
+ long high_water; /* Stack high-water mark. */
+ long overflows; /* Number of stack overflow ($STKOFEN) calls. */
+ long hits; /* Number of internal buffer hits. */
+ long extends; /* Number of block extensions. */
+ long stko_mallocs; /* Block allocations by $STKOFEN. */
+ long underflows; /* Number of stack underflow calls ($STKRETN). */
+ long stko_free; /* Number of deallocations by $STKRETN. */
+ long stkm_free; /* Number of deallocations by $STKMRET. */
+ long segments; /* Current number of stack segments. */
+ long maxs; /* Maximum number of stack segments so far. */
+ long pad_size; /* Stack pad size. */
+ long current_address; /* Current stack segment address. */
+ long current_size; /* Current stack segment size. This
+ number is actually corrupted by STKSTAT to
+ include the fifteen word trailer area. */
+ long initial_address; /* Address of initial segment. */
+ long initial_size; /* Size of initial segment. */
+ };
+
+/* The following structure describes the data structure which trails
+ any stack segment. I think that the description in 'asdef' is
+ out of date. I only describe the parts that I am sure about. */
+
+struct stk_trailer
+ {
+ long this_address; /* Address of this block. */
+ long this_size; /* Size of this block (does not include
+ this trailer). */
+ long unknown2;
+ long unknown3;
+ long link; /* Address of trailer block of previous
+ segment. */
+ long unknown5;
+ long unknown6;
+ long unknown7;
+ long unknown8;
+ long unknown9;
+ long unknown10;
+ long unknown11;
+ long unknown12;
+ long unknown13;
+ long unknown14;
+ };
+
+# endif /* CRAY2 */
+# endif /* not CRAY_STACK */
+
+# ifdef CRAY2
+/* Determine a "stack measure" for an arbitrary ADDRESS.
+ I doubt that "lint" will like this much. */
+
+static long
+i00afunc (long *address)
+{
+ struct stk_stat status;
+ struct stk_trailer *trailer;
+ long *block, size;
+ long result = 0;
+
+ /* We want to iterate through all of the segments. The first
+ step is to get the stack status structure. We could do this
+ more quickly and more directly, perhaps, by referencing the
+ $LM00 common block, but I know that this works. */
+
+ STKSTAT (&status);
+
+ /* Set up the iteration. */
+
+ trailer = (struct stk_trailer *) (status.current_address
+ + status.current_size
+ - 15);
+
+ /* There must be at least one stack segment. Therefore it is
+ a fatal error if "trailer" is null. */
+
+ if (trailer == 0)
+ abort ();
+
+ /* Discard segments that do not contain our argument address. */
+
+ while (trailer != 0)
+ {
+ block = (long *) trailer->this_address;
+ size = trailer->this_size;
+ if (block == 0 || size == 0)
+ abort ();
+ trailer = (struct stk_trailer *) trailer->link;
+ if ((block <= address) && (address < (block + size)))
+ break;
+ }
+
+ /* Set the result to the offset in this segment and add the sizes
+ of all predecessor segments. */
+
+ result = address - block;
+
+ if (trailer == 0)
+ {
+ return result;
+ }
+
+ do
+ {
+ if (trailer->this_size <= 0)
+ abort ();
+ result += trailer->this_size;
+ trailer = (struct stk_trailer *) trailer->link;
+ }
+ while (trailer != 0);
+
+ /* We are done. Note that if you present a bogus address (one
+ not in any segment), you will get a different number back, formed
+ from subtracting the address of the first block. This is probably
+ not what you want. */
+
+ return (result);
+}
+
+# else /* not CRAY2 */
+/* Stack address function for a CRAY-1, CRAY X-MP, or CRAY Y-MP.
+ Determine the number of the cell within the stack,
+ given the address of the cell. The purpose of this
+ routine is to linearize, in some sense, stack addresses
+ for alloca. */
+
+static long
+i00afunc (long address)
+{
+ long stkl = 0;
+
+ long size, pseg, this_segment, stack;
+ long result = 0;
+
+ struct stack_segment_linkage *ssptr;
+
+ /* Register B67 contains the address of the end of the
+ current stack segment. If you (as a subprogram) store
+ your registers on the stack and find that you are past
+ the contents of B67, you have overflowed the segment.
+
+ B67 also points to the stack segment linkage control
+ area, which is what we are really interested in. */
+
+ stkl = CRAY_STACKSEG_END ();
+ ssptr = (struct stack_segment_linkage *) stkl;
+
+ /* If one subtracts 'size' from the end of the segment,
+ one has the address of the first word of the segment.
+
+ If this is not the first segment, 'pseg' will be
+ nonzero. */
+
+ pseg = ssptr->sspseg;
+ size = ssptr->sssize;
+
+ this_segment = stkl - size;
+
+ /* It is possible that calling this routine itself caused
+ a stack overflow. Discard stack segments which do not
+ contain the target address. */
+
+ while (!(this_segment <= address && address <= stkl))
+ {
+# ifdef DEBUG_I00AFUNC
+ fprintf (stderr, "%011o %011o %011o\n", this_segment, address, stkl);
+# endif
+ if (pseg == 0)
+ break;
+ stkl = stkl - pseg;
+ ssptr = (struct stack_segment_linkage *) stkl;
+ size = ssptr->sssize;
+ pseg = ssptr->sspseg;
+ this_segment = stkl - size;
+ }
+
+ result = address - this_segment;
+
+ /* If you subtract pseg from the current end of the stack,
+ you get the address of the previous stack segment's end.
+ This seems a little convoluted to me, but I'll bet you save
+ a cycle somewhere. */
+
+ while (pseg != 0)
+ {
+# ifdef DEBUG_I00AFUNC
+ fprintf (stderr, "%011o %011o\n", pseg, size);
+# endif
+ stkl = stkl - pseg;
+ ssptr = (struct stack_segment_linkage *) stkl;
+ size = ssptr->sssize;
+ pseg = ssptr->sspseg;
+ result += size;
+ }
+ return (result);
+}
+
+# endif /* not CRAY2 */
+# endif /* CRAY */
+
+# endif /* no alloca */
+#endif /* not GCC version 2 */
diff --git a/contrib/diff/lib/alloca_.h b/contrib/diff/lib/alloca_.h
new file mode 100644
index 0000000..8ed34fd
--- /dev/null
+++ b/contrib/diff/lib/alloca_.h
@@ -0,0 +1,68 @@
+/* Memory allocation on the stack.
+ Copyright (C) 1995, 1999, 2001-2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published
+ by the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+
+ You should have received a copy of the GNU General Public
+ License along with this program; if not, write to the Free Software
+ Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
+ USA. */
+
+/* When this file is included, it may be preceded only by preprocessor
+ declarations. Thanks to AIX. Therefore we include it right after
+ "config.h", not later. */
+
+#ifndef _ALLOCA_H
+# define _ALLOCA_H
+
+/* alloca(N) returns a pointer (void* or char*) to N bytes of memory
+ allocated on the stack, and which will last until the function returns.
+ Use of alloca should be avoided:
+ - inside arguments of function calls - undefined behaviour,
+ - in inline functions - the allocation may actually last until the
+ calling function returns,
+ - for huge N (say, N >= 65536) - you never know how large (or small)
+ the stack is, and when the stack cannot fulfill the memory allocation
+ request, the program just crashes.
+ */
+
+# ifdef __GNUC__
+# ifndef alloca
+# define alloca __builtin_alloca
+# endif
+# else
+# ifdef _MSC_VER
+# include <malloc.h>
+# define alloca _alloca
+# else
+# if HAVE_ALLOCA_H
+# include <alloca.h>
+# else
+# ifdef _AIX
+ # pragma alloca
+# else
+# ifdef __hpux /* This section must match that of bison generated files. */
+# ifdef __cplusplus
+extern "C" void *alloca (unsigned int);
+# else /* not __cplusplus */
+extern void *alloca ();
+# endif /* not __cplusplus */
+# else /* not __hpux */
+# ifndef alloca
+extern char *alloca ();
+# endif
+# endif /* __hpux */
+# endif
+# endif
+# endif
+# endif
+
+#endif /* _ALLOCA_H */
diff --git a/contrib/diff/lib/basename.c b/contrib/diff/lib/basename.c
new file mode 100644
index 0000000..5ff2989
--- /dev/null
+++ b/contrib/diff/lib/basename.c
@@ -0,0 +1,79 @@
+/* basename.c -- return the last element in a path
+
+ Copyright (C) 1990, 1998, 1999, 2000, 2001, 2003 Free Software
+ Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "dirname.h"
+#include <string.h>
+
+/* In general, we can't use the builtin `basename' function if available,
+ since it has different meanings in different environments.
+ In some environments the builtin `basename' modifies its argument.
+
+ Return the address of the last file name component of NAME. If
+ NAME has no file name components because it is all slashes, return
+ NAME if it is empty, the address of its last slash otherwise. */
+
+char *
+base_name (char const *name)
+{
+ char const *base = name + FILESYSTEM_PREFIX_LEN (name);
+ char const *p;
+
+ for (p = base; *p; p++)
+ {
+ if (ISSLASH (*p))
+ {
+ /* Treat multiple adjacent slashes like a single slash. */
+ do p++;
+ while (ISSLASH (*p));
+
+ /* If the file name ends in slash, use the trailing slash as
+ the basename if no non-slashes have been found. */
+ if (! *p)
+ {
+ if (ISSLASH (*base))
+ base = p - 1;
+ break;
+ }
+
+ /* *P is a non-slash preceded by a slash. */
+ base = p;
+ }
+ }
+
+ return (char *) base;
+}
+
+/* Return the length of of the basename NAME. Typically NAME is the
+ value returned by base_name. Act like strlen (NAME), except omit
+ redundant trailing slashes. */
+
+size_t
+base_len (char const *name)
+{
+ size_t len;
+
+ for (len = strlen (name); 1 < len && ISSLASH (name[len - 1]); len--)
+ continue;
+
+ return len;
+}
diff --git a/contrib/diff/lib/c-stack.c b/contrib/diff/lib/c-stack.c
new file mode 100644
index 0000000..a018e51
--- /dev/null
+++ b/contrib/diff/lib/c-stack.c
@@ -0,0 +1,311 @@
+/* Stack overflow handling.
+
+ Copyright (C) 2002, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+/* NOTES:
+
+ A program that uses alloca, dynamic arrays, or large local
+ variables may extend the stack by more than a page at a time. If
+ so, when the stack overflows the operating system may not detect
+ the overflow until the program uses the array, and this module may
+ incorrectly report a program error instead of a stack overflow.
+
+ To avoid this problem, allocate only small objects on the stack; a
+ program should be OK if it limits single allocations to a page or
+ less. Allocate larger arrays in static storage, or on the heap
+ (e.g., with malloc). Yes, this is a pain, but we don't know of any
+ better solution that is portable.
+
+ No attempt has been made to deal with multithreaded applications. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#ifndef __attribute__
+# if __GNUC__ < 3 || __STRICT_ANSI__
+# define __attribute__(x)
+# endif
+#endif
+
+#include "gettext.h"
+#define _(msgid) gettext (msgid)
+
+#include <errno.h>
+#ifndef ENOTSUP
+# define ENOTSUP EINVAL
+#endif
+#ifndef EOVERFLOW
+# define EOVERFLOW EINVAL
+#endif
+
+#include <signal.h>
+#if ! HAVE_STACK_T && ! defined stack_t
+typedef struct sigaltstack stack_t;
+#endif
+
+#include <stdlib.h>
+#include <string.h>
+
+#if HAVE_SYS_RESOURCE_H
+/* Include sys/time.h here, because...
+ SunOS-4.1.x <sys/resource.h> fails to include <sys/time.h>.
+ This gives "incomplete type" errors for ru_utime and tu_stime. */
+# if HAVE_SYS_TIME_H
+# include <sys/time.h>
+# endif
+# include <sys/resource.h>
+#endif
+
+#if HAVE_UCONTEXT_H
+# include <ucontext.h>
+#endif
+
+#if HAVE_UNISTD_H
+# include <unistd.h>
+#endif
+#ifndef STDERR_FILENO
+# define STDERR_FILENO 2
+#endif
+
+#if DEBUG
+# include <stdio.h>
+#endif
+
+#include "c-stack.h"
+#include "exitfail.h"
+
+#if (HAVE_STRUCT_SIGACTION_SA_SIGACTION && defined SA_NODEFER \
+ && defined SA_ONSTACK && defined SA_RESETHAND && defined SA_SIGINFO)
+# define SIGACTION_WORKS 1
+#else
+# define SIGACTION_WORKS 0
+#endif
+
+extern char *program_name;
+
+/* The user-specified action to take when a SEGV-related program error
+ or stack overflow occurs. */
+static void (* volatile segv_action) (int);
+
+/* Translated messages for program errors and stack overflow. Do not
+ translate them in the signal handler, since gettext is not
+ async-signal-safe. */
+static char const * volatile program_error_message;
+static char const * volatile stack_overflow_message;
+
+/* Output an error message, then exit with status EXIT_FAILURE if it
+ appears to have been a stack overflow, or with a core dump
+ otherwise. This function is async-signal-safe. */
+
+static void die (int) __attribute__ ((noreturn));
+static void
+die (int signo)
+{
+ char const *message;
+ segv_action (signo);
+ message = signo ? program_error_message : stack_overflow_message;
+ write (STDERR_FILENO, program_name, strlen (program_name));
+ write (STDERR_FILENO, ": ", 2);
+ write (STDERR_FILENO, message, strlen (message));
+ write (STDERR_FILENO, "\n", 1);
+ if (! signo)
+ _exit (exit_failure);
+ kill (getpid (), signo);
+ abort ();
+}
+
+#if HAVE_SIGALTSTACK && HAVE_DECL_SIGALTSTACK
+
+/* Direction of the C runtime stack. This function is
+ async-signal-safe. */
+
+# if STACK_DIRECTION
+# define find_stack_direction(ptr) STACK_DIRECTION
+# else
+static int
+find_stack_direction (char const *addr)
+{
+ char dummy;
+ return ! addr ? find_stack_direction (&dummy) : addr < &dummy ? 1 : -1;
+}
+# endif
+
+/* Storage for the alternate signal stack. */
+static union
+{
+ char buffer[SIGSTKSZ];
+
+ /* These other members are for proper alignment. There's no
+ standard way to guarantee stack alignment, but this seems enough
+ in practice. */
+ long double ld;
+ long l;
+ void *p;
+} alternate_signal_stack;
+
+# if SIGACTION_WORKS
+
+/* Handle a segmentation violation and exit. This function is
+ async-signal-safe. */
+
+static void segv_handler (int, siginfo_t *, void *) __attribute__((noreturn));
+static void
+segv_handler (int signo, siginfo_t *info,
+ void *context __attribute__ ((unused)))
+{
+ /* Clear SIGNO if it seems to have been a stack overflow. */
+ if (0 < info->si_code)
+ {
+# if ! HAVE_XSI_STACK_OVERFLOW_HEURISTIC
+ /* We can't easily determine whether it is a stack overflow; so
+ assume that the rest of our program is perfect (!) and that
+ this segmentation violation is a stack overflow. */
+ signo = 0;
+# else
+ /* If the faulting address is within the stack, or within one
+ page of the stack end, assume that it is a stack
+ overflow. */
+ ucontext_t const *user_context = context;
+ char const *stack_base = user_context->uc_stack.ss_sp;
+ size_t stack_size = user_context->uc_stack.ss_size;
+ char const *faulting_address = info->si_addr;
+ size_t s = faulting_address - stack_base;
+ size_t page_size = sysconf (_SC_PAGESIZE);
+ if (find_stack_direction (0) < 0)
+ s += page_size;
+ if (s < stack_size + page_size)
+ signo = 0;
+
+# if DEBUG
+ {
+ char buf[1024];
+ sprintf (buf,
+ "segv_handler fault=%p base=%p size=%lx page=%lx signo=%d\n",
+ faulting_address, stack_base, (unsigned long) stack_size,
+ (unsigned long) page_size, signo);
+ write (STDERR_FILENO, buf, strlen (buf));
+ }
+# endif
+# endif
+ }
+
+ die (signo);
+}
+# endif
+
+static void
+null_action (int signo __attribute__ ((unused)))
+{
+}
+
+/* Set up ACTION so that it is invoked on C stack overflow. Return -1
+ (setting errno) if this cannot be done.
+
+ When ACTION is called, it is passed an argument equal to SIGSEGV
+ for a segmentation violation that does not appear related to stack
+ overflow, and is passed zero otherwise. On many platforms it is
+ hard to tell; when in doubt, zero is passed.
+
+ A null ACTION acts like an action that does nothing.
+
+ ACTION must be async-signal-safe. ACTION together with its callees
+ must not require more than SIGSTKSZ bytes of stack space. */
+
+int
+c_stack_action (void (*action) (int))
+{
+ int r;
+ stack_t st;
+ st.ss_flags = 0;
+ st.ss_sp = alternate_signal_stack.buffer;
+ st.ss_size = sizeof alternate_signal_stack.buffer;
+ r = sigaltstack (&st, 0);
+ if (r != 0)
+ return r;
+
+ segv_action = action ? action : null_action;
+ program_error_message = _("program error");
+ stack_overflow_message = _("stack overflow");
+
+ {
+# if SIGACTION_WORKS
+ struct sigaction act;
+ sigemptyset (&act.sa_mask);
+
+ /* POSIX 1003.1-2001 says SA_RESETHAND implies SA_NODEFER, but
+ this is not true on Solaris 8 at least. It doesn't hurt to use
+ SA_NODEFER here, so leave it in. */
+ act.sa_flags = SA_NODEFER | SA_ONSTACK | SA_RESETHAND | SA_SIGINFO;
+
+ act.sa_sigaction = segv_handler;
+
+ return sigaction (SIGSEGV, &act, 0);
+# else
+ return signal (SIGSEGV, die) == SIG_ERR ? -1 : 0;
+# endif
+ }
+}
+
+#else /* ! (HAVE_SIGALTSTACK && HAVE_DECL_SIGALTSTACK) */
+
+int
+c_stack_action (void (*action) (int) __attribute__ ((unused)))
+{
+ errno = ENOTSUP;
+ return -1;
+}
+
+#endif
+
+
+
+#if DEBUG
+
+int volatile exit_failure;
+
+static long
+recurse (char *p)
+{
+ char array[500];
+ array[0] = 1;
+ return *p + recurse (array);
+}
+
+char *program_name;
+
+int
+main (int argc __attribute__ ((unused)), char **argv)
+{
+ program_name = argv[0];
+ fprintf (stderr,
+ "The last output line should contain \"stack overflow\".\n");
+ if (c_stack_action (0) == 0)
+ return recurse ("\1");
+ perror ("c_stack_action");
+ return 1;
+}
+
+#endif /* DEBUG */
+
+/*
+Local Variables:
+compile-command: "gcc -DDEBUG -DHAVE_CONFIG_H -I.. -g -O -Wall -W c-stack.c"
+End:
+*/
diff --git a/contrib/diff/lib/c-stack.h b/contrib/diff/lib/c-stack.h
new file mode 100644
index 0000000..968fd52
--- /dev/null
+++ b/contrib/diff/lib/c-stack.h
@@ -0,0 +1,19 @@
+/* Stack overflow handling.
+
+ Copyright (C) 2002, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+int c_stack_action (void (*) (int));
diff --git a/contrib/diff/lib/cmpbuf.c b/contrib/diff/lib/cmpbuf.c
new file mode 100644
index 0000000..858aa39
--- /dev/null
+++ b/contrib/diff/lib/cmpbuf.c
@@ -0,0 +1,147 @@
+/* Buffer primitives for comparison operations.
+
+ Copyright (C) 1993, 1995, 1998, 2001, 2002 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <errno.h>
+#include <limits.h>
+
+#include <signal.h>
+#ifndef SA_RESTART
+# ifdef SA_INTERRUPT /* e.g. SunOS 4.1.x */
+# define SA_RESTART SA_INTERRUPT
+# else
+# define SA_RESTART 0
+# endif
+#endif
+
+#if HAVE_UNISTD_H
+# include <unistd.h>
+#endif
+
+#if HAVE_INTTYPES_H
+# include <inttypes.h>
+#endif
+
+#include <sys/types.h>
+#include "cmpbuf.h"
+
+/* Determine whether an integer type is signed, and its bounds.
+ This code assumes two's (or one's!) complement with no holes. */
+
+/* The extra casts work around common compiler bugs,
+ e.g. Cray C 5.0.3.0 when t == time_t. */
+#ifndef TYPE_SIGNED
+# define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
+#endif
+#ifndef TYPE_MINIMUM
+# define TYPE_MINIMUM(t) ((t) (TYPE_SIGNED (t) \
+ ? ~ (t) 0 << (sizeof (t) * CHAR_BIT - 1) \
+ : (t) 0))
+#endif
+#ifndef TYPE_MAXIMUM
+# define TYPE_MAXIMUM(t) ((t) (~ (t) 0 - TYPE_MINIMUM (t)))
+#endif
+
+#ifndef PTRDIFF_MAX
+# define PTRDIFF_MAX TYPE_MAXIMUM (ptrdiff_t)
+#endif
+#ifndef SIZE_MAX
+# define SIZE_MAX TYPE_MAXIMUM (size_t)
+#endif
+#ifndef SSIZE_MAX
+# define SSIZE_MAX TYPE_MAXIMUM (ssize_t)
+#endif
+
+#undef MIN
+#define MIN(a, b) ((a) <= (b) ? (a) : (b))
+
+/* Read NBYTES bytes from descriptor FD into BUF.
+ NBYTES must not be SIZE_MAX.
+ Return the number of characters successfully read.
+ On error, return SIZE_MAX, setting errno.
+ The number returned is always NBYTES unless end-of-file or error. */
+
+size_t
+block_read (int fd, char *buf, size_t nbytes)
+{
+ char *bp = buf;
+ char const *buflim = buf + nbytes;
+ size_t readlim = SSIZE_MAX;
+
+ do
+ {
+ size_t bytes_to_read = MIN (buflim - bp, readlim);
+ ssize_t nread = read (fd, bp, bytes_to_read);
+ if (nread <= 0)
+ {
+ if (nread == 0)
+ break;
+
+ /* Accommodate Tru64 5.1, which can't read more than INT_MAX
+ bytes at a time. They call that a 64-bit OS? */
+ if (errno == EINVAL && INT_MAX < bytes_to_read)
+ {
+ readlim = INT_MAX;
+ continue;
+ }
+
+ /* This is needed for programs that have signal handlers on
+ older hosts without SA_RESTART. It also accommodates
+ ancient AIX hosts that set errno to EINTR after uncaught
+ SIGCONT. See <news:1r77ojINN85n@ftp.UU.NET>
+ (1993-04-22). */
+ if (! SA_RESTART && errno == EINTR)
+ continue;
+
+ return SIZE_MAX;
+ }
+ bp += nread;
+ }
+ while (bp < buflim);
+
+ return bp - buf;
+}
+
+/* Least common multiple of two buffer sizes A and B. However, if
+ either A or B is zero, or if the multiple is greater than LCM_MAX,
+ return a reasonable buffer size. */
+
+size_t
+buffer_lcm (size_t a, size_t b, size_t lcm_max)
+{
+ size_t lcm, m, n, q, r;
+
+ /* Yield reasonable values if buffer sizes are zero. */
+ if (!a)
+ return b ? b : 8 * 1024;
+ if (!b)
+ return a;
+
+ /* n = gcd (a, b) */
+ for (m = a, n = b; (r = m % n) != 0; m = n, n = r)
+ continue;
+
+ /* Yield a if there is an overflow. */
+ q = a / n;
+ lcm = q * b;
+ return lcm <= lcm_max && lcm / b == q ? lcm : a;
+}
diff --git a/contrib/diff/lib/cmpbuf.h b/contrib/diff/lib/cmpbuf.h
new file mode 100644
index 0000000..9b29cd0
--- /dev/null
+++ b/contrib/diff/lib/cmpbuf.h
@@ -0,0 +1,21 @@
+/* Buffer primitives for comparison operations.
+
+ Copyright (C) 2002 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+size_t block_read (int, char *, size_t);
+size_t buffer_lcm (size_t, size_t, size_t);
diff --git a/contrib/diff/lib/dirname.c b/contrib/diff/lib/dirname.c
new file mode 100644
index 0000000..815192a
--- /dev/null
+++ b/contrib/diff/lib/dirname.c
@@ -0,0 +1,119 @@
+/* dirname.c -- return all but the last element in a path
+ Copyright (C) 1990, 1998, 2000, 2001, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "dirname.h"
+
+#include <string.h>
+#include "xalloc.h"
+
+/* Return the length of `dirname (PATH)', or zero if PATH is
+ in the working directory. Works properly even if
+ there are trailing slashes (by effectively ignoring them). */
+size_t
+dir_len (char const *path)
+{
+ size_t prefix_length = FILESYSTEM_PREFIX_LEN (path);
+ size_t length;
+
+ /* Strip the basename and any redundant slashes before it. */
+ for (length = base_name (path) - path; prefix_length < length; length--)
+ if (! ISSLASH (path[length - 1]))
+ return length;
+
+ /* But don't strip the only slash from "/". */
+ return prefix_length + ISSLASH (path[prefix_length]);
+}
+
+/* Return the leading directories part of PATH,
+ allocated with xmalloc.
+ Works properly even if there are trailing slashes
+ (by effectively ignoring them). */
+
+char *
+dir_name (char const *path)
+{
+ size_t length = dir_len (path);
+ int append_dot = (length == FILESYSTEM_PREFIX_LEN (path));
+ char *newpath = xmalloc (length + append_dot + 1);
+ memcpy (newpath, path, length);
+ if (append_dot)
+ newpath[length++] = '.';
+ newpath[length] = 0;
+ return newpath;
+}
+
+#ifdef TEST_DIRNAME
+/*
+
+Run the test like this (expect no output):
+ gcc -DHAVE_CONFIG_H -DTEST_DIRNAME -I.. -O -Wall \
+ basename.c dirname.c xmalloc.c error.c
+ sed -n '/^BEGIN-DATA$/,/^END-DATA$/p' dirname.c|grep -v DATA|./a.out
+
+If it's been built on a DOS or Windows platforms, run another test like
+this (again, expect no output):
+ sed -n '/^BEGIN-DOS-DATA$/,/^END-DOS-DATA$/p' dirname.c|grep -v DATA|./a.out
+
+BEGIN-DATA
+foo//// .
+bar/foo//// bar
+foo/ .
+/ /
+. .
+a .
+END-DATA
+
+BEGIN-DOS-DATA
+c:///// c:/
+c:/ c:/
+c:/. c:/
+c:foo c:.
+c:foo/bar c:foo
+END-DOS-DATA
+
+*/
+
+# define MAX_BUFF_LEN 1024
+# include <stdio.h>
+
+char *program_name;
+
+int
+main (int argc, char *argv[])
+{
+ char buff[MAX_BUFF_LEN + 1];
+
+ program_name = argv[0];
+
+ buff[MAX_BUFF_LEN] = 0;
+ while (fgets (buff, MAX_BUFF_LEN, stdin) && buff[0])
+ {
+ char path[MAX_BUFF_LEN];
+ char expected_result[MAX_BUFF_LEN];
+ char const *result;
+ sscanf (buff, "%s %s", path, expected_result);
+ result = dir_name (path);
+ if (strcmp (result, expected_result))
+ printf ("%s: got %s, expected %s\n", path, result, expected_result);
+ }
+ return 0;
+}
+#endif
diff --git a/contrib/diff/lib/dirname.h b/contrib/diff/lib/dirname.h
new file mode 100644
index 0000000..62da37d
--- /dev/null
+++ b/contrib/diff/lib/dirname.h
@@ -0,0 +1,43 @@
+/* Take file names apart into directory and base names.
+
+ Copyright (C) 1998, 2001, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef DIRNAME_H_
+# define DIRNAME_H_ 1
+
+# include <stddef.h>
+
+# ifndef DIRECTORY_SEPARATOR
+# define DIRECTORY_SEPARATOR '/'
+# endif
+
+# ifndef ISSLASH
+# define ISSLASH(C) ((C) == DIRECTORY_SEPARATOR)
+# endif
+
+# ifndef FILESYSTEM_PREFIX_LEN
+# define FILESYSTEM_PREFIX_LEN(Filename) 0
+# endif
+
+char *base_name (char const *path);
+char *dir_name (char const *path);
+size_t base_len (char const *path);
+size_t dir_len (char const *path);
+
+int strip_trailing_slashes (char *path);
+
+#endif /* not DIRNAME_H_ */
diff --git a/contrib/diff/lib/error.c b/contrib/diff/lib/error.c
new file mode 100644
index 0000000..1149235
--- /dev/null
+++ b/contrib/diff/lib/error.c
@@ -0,0 +1,310 @@
+/* Error handler for noninteractive utilities
+ Copyright (C) 1990-1998, 2000-2002, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by David MacKenzie <djm@gnu.ai.mit.edu>. */
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "error.h"
+
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef _LIBC
+# include <libintl.h>
+#else
+# include "gettext.h"
+#endif
+
+#ifdef _LIBC
+# include <wchar.h>
+# define mbsrtowcs __mbsrtowcs
+#endif
+
+#if !_LIBC
+# include "unlocked-io.h"
+#endif
+
+#ifndef _
+# define _(String) String
+#endif
+
+/* If NULL, error will flush stdout, then print on stderr the program
+ name, a colon and a space. Otherwise, error will call this
+ function without parameters instead. */
+void (*error_print_progname) (void);
+
+/* This variable is incremented each time `error' is called. */
+unsigned int error_message_count;
+
+#ifdef _LIBC
+/* In the GNU C library, there is a predefined variable for this. */
+
+# define program_name program_invocation_name
+# include <errno.h>
+# include <libio/libioP.h>
+
+/* In GNU libc we want do not want to use the common name `error' directly.
+ Instead make it a weak alias. */
+extern void __error (int status, int errnum, const char *message, ...)
+ __attribute__ ((__format__ (__printf__, 3, 4)));
+extern void __error_at_line (int status, int errnum, const char *file_name,
+ unsigned int line_number, const char *message,
+ ...)
+ __attribute__ ((__format__ (__printf__, 5, 6)));;
+# define error __error
+# define error_at_line __error_at_line
+
+# include <libio/iolibio.h>
+# define fflush(s) INTUSE(_IO_fflush) (s)
+# undef putc
+# define putc(c, fp) INTUSE(_IO_putc) (c, fp)
+
+# include <bits/libc-lock.h>
+
+#else /* not _LIBC */
+
+# if !HAVE_DECL_STRERROR_R && STRERROR_R_CHAR_P
+# ifndef HAVE_DECL_STRERROR_R
+"this configure-time declaration test was not run"
+# endif
+char *strerror_r ();
+# endif
+
+# ifndef SIZE_MAX
+# define SIZE_MAX ((size_t) -1)
+# endif
+
+/* The calling program should define program_name and set it to the
+ name of the executing program. */
+extern char *program_name;
+
+# if HAVE_STRERROR_R || defined strerror_r
+# define __strerror_r strerror_r
+# endif
+#endif /* not _LIBC */
+
+static void
+print_errno_message (int errnum)
+{
+ char const *s;
+
+#if defined HAVE_STRERROR_R || _LIBC
+ char errbuf[1024];
+# if STRERROR_R_CHAR_P || _LIBC
+ s = __strerror_r (errnum, errbuf, sizeof errbuf);
+# else
+ if (__strerror_r (errnum, errbuf, sizeof errbuf) == 0)
+ s = errbuf;
+ else
+ s = 0;
+# endif
+#else
+ s = strerror (errnum);
+#endif
+
+#if !_LIBC
+ if (! s)
+ s = _("Unknown system error");
+#endif
+
+#if _LIBC
+ if (_IO_fwide (stderr, 0) > 0)
+ {
+ __fwprintf (stderr, L": %s", s);
+ return;
+ }
+#endif
+
+ fprintf (stderr, ": %s", s);
+}
+
+static void
+error_tail (int status, int errnum, const char *message, va_list args)
+{
+#if _LIBC
+ if (_IO_fwide (stderr, 0) > 0)
+ {
+# define ALLOCA_LIMIT 2000
+ size_t len = strlen (message) + 1;
+ const wchar_t *wmessage = L"out of memory";
+ wchar_t *wbuf = (len < ALLOCA_LIMIT
+ ? alloca (len * sizeof *wbuf)
+ : len <= SIZE_MAX / sizeof *wbuf
+ ? malloc (len * sizeof *wbuf)
+ : NULL);
+
+ if (wbuf)
+ {
+ size_t res;
+ mbstate_t st;
+ const char *tmp = message;
+ memset (&st, '\0', sizeof (st));
+ res = mbsrtowcs (wbuf, &tmp, len, &st);
+ wmessage = res == (size_t) -1 ? L"???" : wbuf;
+ }
+
+ __vfwprintf (stderr, wmessage, args);
+ if (! (len < ALLOCA_LIMIT))
+ free (wbuf);
+ }
+ else
+#endif
+ vfprintf (stderr, message, args);
+ va_end (args);
+
+ ++error_message_count;
+ if (errnum)
+ print_errno_message (errnum);
+#if _LIBC
+ if (_IO_fwide (stderr, 0) > 0)
+ putwc (L'\n', stderr);
+ else
+#endif
+ putc ('\n', stderr);
+ fflush (stderr);
+ if (status)
+ exit (status);
+}
+
+
+/* Print the program name and error message MESSAGE, which is a printf-style
+ format string with optional args.
+ If ERRNUM is nonzero, print its corresponding system error message.
+ Exit with status STATUS if it is nonzero. */
+void
+error (int status, int errnum, const char *message, ...)
+{
+ va_list args;
+
+#if defined _LIBC && defined __libc_ptf_call
+ /* We do not want this call to be cut short by a thread
+ cancellation. Therefore disable cancellation for now. */
+ int state = PTHREAD_CANCEL_ENABLE;
+ __libc_ptf_call (pthread_setcancelstate, (PTHREAD_CANCEL_DISABLE, &state),
+ 0);
+#endif
+
+ fflush (stdout);
+#ifdef _LIBC
+ _IO_flockfile (stderr);
+#endif
+ if (error_print_progname)
+ (*error_print_progname) ();
+ else
+ {
+#if _LIBC
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s: ", program_name);
+ else
+#endif
+ fprintf (stderr, "%s: ", program_name);
+ }
+
+ va_start (args, message);
+ error_tail (status, errnum, message, args);
+
+#ifdef _LIBC
+ _IO_funlockfile (stderr);
+# ifdef __libc_ptf_call
+ __libc_ptf_call (pthread_setcancelstate, (state, NULL), 0);
+# endif
+#endif
+}
+
+/* Sometimes we want to have at most one error per line. This
+ variable controls whether this mode is selected or not. */
+int error_one_per_line;
+
+void
+error_at_line (int status, int errnum, const char *file_name,
+ unsigned int line_number, const char *message, ...)
+{
+ va_list args;
+
+ if (error_one_per_line)
+ {
+ static const char *old_file_name;
+ static unsigned int old_line_number;
+
+ if (old_line_number == line_number
+ && (file_name == old_file_name
+ || strcmp (old_file_name, file_name) == 0))
+ /* Simply return and print nothing. */
+ return;
+
+ old_file_name = file_name;
+ old_line_number = line_number;
+ }
+
+#if defined _LIBC && defined __libc_ptf_call
+ /* We do not want this call to be cut short by a thread
+ cancellation. Therefore disable cancellation for now. */
+ int state = PTHREAD_CANCEL_ENABLE;
+ __libc_ptf_call (pthread_setcancelstate, (PTHREAD_CANCEL_DISABLE, &state),
+ 0);
+#endif
+
+ fflush (stdout);
+#ifdef _LIBC
+ _IO_flockfile (stderr);
+#endif
+ if (error_print_progname)
+ (*error_print_progname) ();
+ else
+ {
+#if _LIBC
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s: ", program_name);
+ else
+#endif
+ fprintf (stderr, "%s:", program_name);
+ }
+
+ if (file_name != NULL)
+ {
+#if _LIBC
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s:%d: ", file_name, line_number);
+ else
+#endif
+ fprintf (stderr, "%s:%d: ", file_name, line_number);
+ }
+
+ va_start (args, message);
+ error_tail (status, errnum, message, args);
+
+#ifdef _LIBC
+ _IO_funlockfile (stderr);
+# ifdef __libc_ptf_call
+ __libc_ptf_call (pthread_setcancelstate, (state, NULL), 0);
+# endif
+#endif
+}
+
+#ifdef _LIBC
+/* Make the weak alias. */
+# undef error
+# undef error_at_line
+weak_alias (__error, error)
+weak_alias (__error_at_line, error_at_line)
+#endif
diff --git a/contrib/diff/lib/error.h b/contrib/diff/lib/error.h
new file mode 100644
index 0000000..8ed6359
--- /dev/null
+++ b/contrib/diff/lib/error.h
@@ -0,0 +1,66 @@
+/* Declaration for error-reporting function
+ Copyright (C) 1995, 1996, 1997, 2003 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _ERROR_H
+#define _ERROR_H 1
+
+#ifndef __attribute__
+/* This feature is available in gcc versions 2.5 and later. */
+# if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 5)
+# define __attribute__(Spec) /* empty */
+# endif
+/* The __-protected variants of `format' and `printf' attributes
+ are accepted by gcc versions 2.6.4 (effectively 2.7) and later. */
+# if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 7)
+# define __format__ format
+# define __printf__ printf
+# endif
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Print a message with `fprintf (stderr, FORMAT, ...)';
+ if ERRNUM is nonzero, follow it with ": " and strerror (ERRNUM).
+ If STATUS is nonzero, terminate the program with `exit (STATUS)'. */
+
+extern void error (int __status, int __errnum, const char *__format, ...)
+ __attribute__ ((__format__ (__printf__, 3, 4)));
+
+extern void error_at_line (int __status, int __errnum, const char *__fname,
+ unsigned int __lineno, const char *__format, ...)
+ __attribute__ ((__format__ (__printf__, 5, 6)));
+
+/* If NULL, error will flush stdout, then print on stderr the program
+ name, a colon and a space. Otherwise, error will call this
+ function without parameters instead. */
+extern void (*error_print_progname) (void);
+
+/* This variable is incremented each time `error' is called. */
+extern unsigned int error_message_count;
+
+/* Sometimes we want to have at most one error per line. This
+ variable controls whether this mode is selected or not. */
+extern int error_one_per_line;
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* error.h */
diff --git a/contrib/diff/lib/exclude.c b/contrib/diff/lib/exclude.c
new file mode 100644
index 0000000..55f1a39
--- /dev/null
+++ b/contrib/diff/lib/exclude.c
@@ -0,0 +1,263 @@
+/* exclude.c -- exclude file names
+
+ Copyright (C) 1992, 1993, 1994, 1997, 1999, 2000, 2001, 2002, 2003 Free
+ Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert <eggert@twinsun.com> */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <stdbool.h>
+
+#include <ctype.h>
+#include <errno.h>
+#ifndef errno
+extern int errno;
+#endif
+#include <stddef.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "exclude.h"
+#include "fnmatch.h"
+#include "unlocked-io.h"
+#include "xalloc.h"
+
+#if STDC_HEADERS || (! defined isascii && ! HAVE_ISASCII)
+# define IN_CTYPE_DOMAIN(c) true
+#else
+# define IN_CTYPE_DOMAIN(c) isascii (c)
+#endif
+
+static inline bool
+is_space (unsigned char c)
+{
+ return IN_CTYPE_DOMAIN (c) && isspace (c);
+}
+
+/* Verify a requirement at compile-time (unlike assert, which is runtime). */
+#define verify(name, assertion) struct name { char a[(assertion) ? 1 : -1]; }
+
+/* Non-GNU systems lack these options, so we don't need to check them. */
+#ifndef FNM_CASEFOLD
+# define FNM_CASEFOLD 0
+#endif
+#ifndef FNM_LEADING_DIR
+# define FNM_LEADING_DIR 0
+#endif
+
+verify (EXCLUDE_macros_do_not_collide_with_FNM_macros,
+ (((EXCLUDE_ANCHORED | EXCLUDE_INCLUDE | EXCLUDE_WILDCARDS)
+ & (FNM_PATHNAME | FNM_NOESCAPE | FNM_PERIOD | FNM_LEADING_DIR
+ | FNM_CASEFOLD))
+ == 0));
+
+/* An exclude pattern-options pair. The options are fnmatch options
+ ORed with EXCLUDE_* options. */
+
+struct patopts
+ {
+ char const *pattern;
+ int options;
+ };
+
+/* An exclude list, of pattern-options pairs. */
+
+struct exclude
+ {
+ struct patopts *exclude;
+ size_t exclude_alloc;
+ size_t exclude_count;
+ };
+
+/* Return a newly allocated and empty exclude list. */
+
+struct exclude *
+new_exclude (void)
+{
+ return xzalloc (sizeof *new_exclude ());
+}
+
+/* Free the storage associated with an exclude list. */
+
+void
+free_exclude (struct exclude *ex)
+{
+ free (ex->exclude);
+ free (ex);
+}
+
+/* Return zero if PATTERN matches F, obeying OPTIONS, except that
+ (unlike fnmatch) wildcards are disabled in PATTERN. */
+
+static int
+fnmatch_no_wildcards (char const *pattern, char const *f, int options)
+{
+ if (! (options & FNM_LEADING_DIR))
+ return ((options & FNM_CASEFOLD)
+ ? strcasecmp (pattern, f)
+ : strcmp (pattern, f));
+ else
+ {
+ size_t patlen = strlen (pattern);
+ int r = ((options & FNM_CASEFOLD)
+ ? strncasecmp (pattern, f, patlen)
+ : strncmp (pattern, f, patlen));
+ if (! r)
+ {
+ r = f[patlen];
+ if (r == '/')
+ r = 0;
+ }
+ return r;
+ }
+}
+
+/* Return true if EX excludes F. */
+
+bool
+excluded_filename (struct exclude const *ex, char const *f)
+{
+ size_t exclude_count = ex->exclude_count;
+
+ /* If no options are given, the default is to include. */
+ if (exclude_count == 0)
+ return false;
+ else
+ {
+ struct patopts const *exclude = ex->exclude;
+ size_t i;
+
+ /* Otherwise, the default is the opposite of the first option. */
+ bool excluded = !! (exclude[0].options & EXCLUDE_INCLUDE);
+
+ /* Scan through the options, seeing whether they change F from
+ excluded to included or vice versa. */
+ for (i = 0; i < exclude_count; i++)
+ {
+ char const *pattern = exclude[i].pattern;
+ int options = exclude[i].options;
+ if (excluded == !! (options & EXCLUDE_INCLUDE))
+ {
+ int (*matcher) (char const *, char const *, int) =
+ (options & EXCLUDE_WILDCARDS
+ ? fnmatch
+ : fnmatch_no_wildcards);
+ bool matched = ((*matcher) (pattern, f, options) == 0);
+ char const *p;
+
+ if (! (options & EXCLUDE_ANCHORED))
+ for (p = f; *p && ! matched; p++)
+ if (*p == '/' && p[1] != '/')
+ matched = ((*matcher) (pattern, p + 1, options) == 0);
+
+ excluded ^= matched;
+ }
+ }
+
+ return excluded;
+ }
+}
+
+/* Append to EX the exclusion PATTERN with OPTIONS. */
+
+void
+add_exclude (struct exclude *ex, char const *pattern, int options)
+{
+ struct patopts *patopts;
+
+ if (ex->exclude_count == ex->exclude_alloc)
+ ex->exclude = x2nrealloc (ex->exclude, &ex->exclude_alloc,
+ sizeof *ex->exclude);
+
+ patopts = &ex->exclude[ex->exclude_count++];
+ patopts->pattern = pattern;
+ patopts->options = options;
+}
+
+/* Use ADD_FUNC to append to EX the patterns in FILENAME, each with
+ OPTIONS. LINE_END terminates each pattern in the file. If
+ LINE_END is a space character, ignore trailing spaces and empty
+ lines in FILE. Return -1 on failure, 0 on success. */
+
+int
+add_exclude_file (void (*add_func) (struct exclude *, char const *, int),
+ struct exclude *ex, char const *filename, int options,
+ char line_end)
+{
+ bool use_stdin = filename[0] == '-' && !filename[1];
+ FILE *in;
+ char *buf = NULL;
+ char *p;
+ char const *pattern;
+ char const *lim;
+ size_t buf_alloc = 0;
+ size_t buf_count = 0;
+ int c;
+ int e = 0;
+
+ if (use_stdin)
+ in = stdin;
+ else if (! (in = fopen (filename, "r")))
+ return -1;
+
+ while ((c = getc (in)) != EOF)
+ {
+ if (buf_count == buf_alloc)
+ buf = x2realloc (buf, &buf_alloc);
+ buf[buf_count++] = c;
+ }
+
+ if (ferror (in))
+ e = errno;
+
+ if (!use_stdin && fclose (in) != 0)
+ e = errno;
+
+ buf = xrealloc (buf, buf_count + 1);
+ buf[buf_count] = line_end;
+ lim = buf + buf_count + ! (buf_count == 0 || buf[buf_count - 1] == line_end);
+ pattern = buf;
+
+ for (p = buf; p < lim; p++)
+ if (*p == line_end)
+ {
+ char *pattern_end = p;
+
+ if (is_space (line_end))
+ {
+ for (; ; pattern_end--)
+ if (pattern_end == pattern)
+ goto next_pattern;
+ else if (! is_space (pattern_end[-1]))
+ break;
+ }
+
+ *pattern_end = '\0';
+ (*add_func) (ex, pattern, options);
+
+ next_pattern:
+ pattern = p + 1;
+ }
+
+ errno = e;
+ return e ? -1 : 0;
+}
diff --git a/contrib/diff/lib/exclude.h b/contrib/diff/lib/exclude.h
new file mode 100644
index 0000000..a7e2971
--- /dev/null
+++ b/contrib/diff/lib/exclude.h
@@ -0,0 +1,43 @@
+/* exclude.h -- declarations for excluding file names
+
+ Copyright (C) 1992, 1993, 1994, 1997, 1999, 2001, 2002, 2003 Free
+ Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert <eggert@twinsun.com> */
+
+/* Exclude options, which can be ORed with fnmatch options. */
+
+/* Patterns must match the start of file names, instead of matching
+ anywhere after a '/'. */
+#define EXCLUDE_ANCHORED (1 << 30)
+
+/* Include instead of exclude. */
+#define EXCLUDE_INCLUDE (1 << 29)
+
+/* '?', '*', '[', and '\\' are special in patterns. Without this
+ option, these characters are ordinary and fnmatch is not used. */
+#define EXCLUDE_WILDCARDS (1 << 28)
+
+struct exclude;
+
+struct exclude *new_exclude (void);
+void free_exclude (struct exclude *);
+void add_exclude (struct exclude *, char const *, int);
+int add_exclude_file (void (*) (struct exclude *, char const *, int),
+ struct exclude *, char const *, int, char);
+bool excluded_filename (struct exclude const *, char const *);
diff --git a/contrib/diff/lib/exit.h b/contrib/diff/lib/exit.h
new file mode 100644
index 0000000..4e8d465
--- /dev/null
+++ b/contrib/diff/lib/exit.h
@@ -0,0 +1,32 @@
+/* exit() function.
+ Copyright (C) 1995, 2001 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _EXIT_H
+#define _EXIT_H
+
+/* Get exit() declaration. */
+#include <stdlib.h>
+
+/* Some systems do not define EXIT_*, even with STDC_HEADERS. */
+#ifndef EXIT_SUCCESS
+# define EXIT_SUCCESS 0
+#endif
+#ifndef EXIT_FAILURE
+# define EXIT_FAILURE 1
+#endif
+
+#endif /* _EXIT_H */
diff --git a/contrib/diff/lib/exitfail.c b/contrib/diff/lib/exitfail.c
new file mode 100644
index 0000000..2ae5f69
--- /dev/null
+++ b/contrib/diff/lib/exitfail.c
@@ -0,0 +1,27 @@
+/* Failure exit status
+
+ Copyright (C) 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "exitfail.h"
+#include "exit.h"
+
+int volatile exit_failure = EXIT_FAILURE;
diff --git a/contrib/diff/lib/exitfail.h b/contrib/diff/lib/exitfail.h
new file mode 100644
index 0000000..cf5ab71
--- /dev/null
+++ b/contrib/diff/lib/exitfail.h
@@ -0,0 +1,20 @@
+/* Failure exit status
+
+ Copyright (C) 2002 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+extern int volatile exit_failure;
diff --git a/contrib/diff/lib/file-type.c b/contrib/diff/lib/file-type.c
new file mode 100644
index 0000000..3c58c7f
--- /dev/null
+++ b/contrib/diff/lib/file-type.c
@@ -0,0 +1,75 @@
+/* Return a string describing the type of a file.
+
+ Copyright (C) 1993, 1994, 2001, 2002, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <sys/types.h>
+#include <sys/stat.h>
+#include "file-type.h"
+
+#include <gettext.h>
+#define _(text) gettext (text)
+
+char const *
+file_type (struct stat const *st)
+{
+ /* See POSIX 1003.1-2001 XCU Table 4-8 lines 17093-17107 for some of
+ these formats.
+
+ To keep diagnostics grammatical in English, the returned string
+ must start with a consonant. */
+
+ if (S_ISREG (st->st_mode))
+ return st->st_size == 0 ? _("regular empty file") : _("regular file");
+
+ if (S_ISDIR (st->st_mode))
+ return _("directory");
+
+ if (S_ISBLK (st->st_mode))
+ return _("block special file");
+
+ if (S_ISCHR (st->st_mode))
+ return _("character special file");
+
+ if (S_ISFIFO (st->st_mode))
+ return _("fifo");
+
+ if (S_ISLNK (st->st_mode))
+ return _("symbolic link");
+
+ if (S_ISSOCK (st->st_mode))
+ return _("socket");
+
+ if (S_TYPEISMQ (st))
+ return _("message queue");
+
+ if (S_TYPEISSEM (st))
+ return _("semaphore");
+
+ if (S_TYPEISSHM (st))
+ return _("shared memory object");
+
+ if (S_TYPEISTMO (st))
+ return _("typed memory object");
+
+ return _("weird file");
+}
diff --git a/contrib/diff/lib/file-type.h b/contrib/diff/lib/file-type.h
new file mode 100644
index 0000000..502f091
--- /dev/null
+++ b/contrib/diff/lib/file-type.h
@@ -0,0 +1,166 @@
+/* Return a string describing the type of a file.
+
+ Copyright (C) 1993, 1994, 2001, 2002, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert and Jim Meyering. */
+
+#ifndef FILE_TYPE_H
+# define FILE_TYPE_H 1
+
+# if ! defined S_ISREG && ! defined S_IFREG
+you must include <sys/stat.h> before including this file
+# endif
+
+char const *file_type (struct stat const *);
+
+# ifndef S_IFMT
+# define S_IFMT 0170000
+# endif
+
+# if STAT_MACROS_BROKEN
+# undef S_ISBLK
+# undef S_ISCHR
+# undef S_ISDIR
+# undef S_ISDOOR
+# undef S_ISFIFO
+# undef S_ISLNK
+# undef S_ISNAM
+# undef S_ISMPB
+# undef S_ISMPC
+# undef S_ISNWK
+# undef S_ISREG
+# undef S_ISSOCK
+# endif
+
+
+# ifndef S_ISBLK
+# ifdef S_IFBLK
+# define S_ISBLK(m) (((m) & S_IFMT) == S_IFBLK)
+# else
+# define S_ISBLK(m) 0
+# endif
+# endif
+
+# ifndef S_ISCHR
+# ifdef S_IFCHR
+# define S_ISCHR(m) (((m) & S_IFMT) == S_IFCHR)
+# else
+# define S_ISCHR(m) 0
+# endif
+# endif
+
+# ifndef S_ISDIR
+# ifdef S_IFDIR
+# define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR)
+# else
+# define S_ISDIR(m) 0
+# endif
+# endif
+
+# ifndef S_ISDOOR /* Solaris 2.5 and up */
+# ifdef S_IFDOOR
+# define S_ISDOOR(m) (((m) & S_IFMT) == S_IFDOOR)
+# else
+# define S_ISDOOR(m) 0
+# endif
+# endif
+
+# ifndef S_ISFIFO
+# ifdef S_IFIFO
+# define S_ISFIFO(m) (((m) & S_IFMT) == S_IFIFO)
+# else
+# define S_ISFIFO(m) 0
+# endif
+# endif
+
+# ifndef S_ISLNK
+# ifdef S_IFLNK
+# define S_ISLNK(m) (((m) & S_IFMT) == S_IFLNK)
+# else
+# define S_ISLNK(m) 0
+# endif
+# endif
+
+# ifndef S_ISMPB /* V7 */
+# ifdef S_IFMPB
+# define S_ISMPB(m) (((m) & S_IFMT) == S_IFMPB)
+# define S_ISMPC(m) (((m) & S_IFMT) == S_IFMPC)
+# else
+# define S_ISMPB(m) 0
+# define S_ISMPC(m) 0
+# endif
+# endif
+
+# ifndef S_ISNAM /* Xenix */
+# ifdef S_IFNAM
+# define S_ISNAM(m) (((m) & S_IFMT) == S_IFNAM)
+# else
+# define S_ISNAM(m) 0
+# endif
+# endif
+
+# ifndef S_ISNWK /* HP/UX */
+# ifdef S_IFNWK
+# define S_ISNWK(m) (((m) & S_IFMT) == S_IFNWK)
+# else
+# define S_ISNWK(m) 0
+# endif
+# endif
+
+# ifndef S_ISREG
+# ifdef S_IFREG
+# define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
+# else
+# define S_ISREG(m) 0
+# endif
+# endif
+
+# ifndef S_ISSOCK
+# ifdef S_IFSOCK
+# define S_ISSOCK(m) (((m) & S_IFMT) == S_IFSOCK)
+# else
+# define S_ISSOCK(m) 0
+# endif
+# endif
+
+
+# ifndef S_TYPEISMQ
+# define S_TYPEISMQ(p) 0
+# endif
+
+# ifndef S_TYPEISTMO
+# define S_TYPEISTMO(p) 0
+# endif
+
+
+# ifndef S_TYPEISSEM
+# ifdef S_INSEM
+# define S_TYPEISSEM(p) (S_ISNAM ((p)->st_mode) && (p)->st_rdev == S_INSEM)
+# else
+# define S_TYPEISSEM(p) 0
+# endif
+# endif
+
+# ifndef S_TYPEISSHM
+# ifdef S_INSHD
+# define S_TYPEISSHM(p) (S_ISNAM ((p)->st_mode) && (p)->st_rdev == S_INSHD)
+# else
+# define S_TYPEISSHM(p) 0
+# endif
+# endif
+
+#endif /* FILE_TYPE_H */
diff --git a/contrib/diff/lib/fnmatch.c b/contrib/diff/lib/fnmatch.c
new file mode 100644
index 0000000..d0cd8cd
--- /dev/null
+++ b/contrib/diff/lib/fnmatch.c
@@ -0,0 +1,403 @@
+/* Copyright (C) 1991,1992,1993,1996,1997,1998,1999,2000,2001,2002,2003,2004
+ Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+/* Enable GNU extensions in fnmatch.h. */
+#ifndef _GNU_SOURCE
+# define _GNU_SOURCE 1
+#endif
+
+#ifdef __GNUC__
+# define alloca __builtin_alloca
+# define HAVE_ALLOCA 1
+#else
+# if defined HAVE_ALLOCA_H || defined _LIBC
+# include <alloca.h>
+# else
+# ifdef _AIX
+ # pragma alloca
+# else
+# ifndef alloca
+char *alloca ();
+# endif
+# endif
+# endif
+#endif
+
+#if ! defined __builtin_expect && __GNUC__ < 3
+# define __builtin_expect(expr, expected) (expr)
+#endif
+
+#include <fnmatch.h>
+
+#include <assert.h>
+#include <ctype.h>
+#include <errno.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define WIDE_CHAR_SUPPORT (HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_BTOWC)
+
+/* For platform which support the ISO C amendement 1 functionality we
+ support user defined character classes. */
+#if defined _LIBC || WIDE_CHAR_SUPPORT
+/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */
+# include <wchar.h>
+# include <wctype.h>
+#endif
+
+/* We need some of the locale data (the collation sequence information)
+ but there is no interface to get this information in general. Therefore
+ we support a correct implementation only in glibc. */
+#ifdef _LIBC
+# include "../locale/localeinfo.h"
+# include "../locale/elem-hash.h"
+# include "../locale/coll-lookup.h"
+# include <shlib-compat.h>
+
+# define CONCAT(a,b) __CONCAT(a,b)
+# define mbsrtowcs __mbsrtowcs
+# define fnmatch __fnmatch
+extern int fnmatch (const char *pattern, const char *string, int flags);
+#endif
+
+#ifndef SIZE_MAX
+# define SIZE_MAX ((size_t) -1)
+#endif
+
+/* We often have to test for FNM_FILE_NAME and FNM_PERIOD being both set. */
+#define NO_LEADING_PERIOD(flags) \
+ ((flags & (FNM_FILE_NAME | FNM_PERIOD)) == (FNM_FILE_NAME | FNM_PERIOD))
+
+/* Comment out all this code if we are using the GNU C Library, and are not
+ actually compiling the library itself, and have not detected a bug
+ in the library. This code is part of the GNU C
+ Library, but also included in many other GNU distributions. Compiling
+ and linking in this code is a waste when using the GNU C library
+ (especially if it is a shared library). Rather than having every GNU
+ program understand `configure --with-gnu-libc' and omit the object files,
+ it is simpler to just do this in the source for each such file. */
+
+#if defined _LIBC || !defined __GNU_LIBRARY__ || !HAVE_FNMATCH_GNU
+
+
+# if defined STDC_HEADERS || !defined isascii
+# define ISASCII(c) 1
+# else
+# define ISASCII(c) isascii(c)
+# endif
+
+# ifdef isblank
+# define ISBLANK(c) (ISASCII (c) && isblank (c))
+# else
+# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
+# endif
+# ifdef isgraph
+# define ISGRAPH(c) (ISASCII (c) && isgraph (c))
+# else
+# define ISGRAPH(c) (ISASCII (c) && isprint (c) && !isspace (c))
+# endif
+
+# define ISPRINT(c) (ISASCII (c) && isprint (c))
+# define ISDIGIT(c) (ISASCII (c) && isdigit (c))
+# define ISALNUM(c) (ISASCII (c) && isalnum (c))
+# define ISALPHA(c) (ISASCII (c) && isalpha (c))
+# define ISCNTRL(c) (ISASCII (c) && iscntrl (c))
+# define ISLOWER(c) (ISASCII (c) && islower (c))
+# define ISPUNCT(c) (ISASCII (c) && ispunct (c))
+# define ISSPACE(c) (ISASCII (c) && isspace (c))
+# define ISUPPER(c) (ISASCII (c) && isupper (c))
+# define ISXDIGIT(c) (ISASCII (c) && isxdigit (c))
+
+# define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
+
+# if defined _LIBC || WIDE_CHAR_SUPPORT
+/* The GNU C library provides support for user-defined character classes
+ and the functions from ISO C amendement 1. */
+# ifdef CHARCLASS_NAME_MAX
+# define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
+# else
+/* This shouldn't happen but some implementation might still have this
+ problem. Use a reasonable default value. */
+# define CHAR_CLASS_MAX_LENGTH 256
+# endif
+
+# ifdef _LIBC
+# define IS_CHAR_CLASS(string) __wctype (string)
+# else
+# define IS_CHAR_CLASS(string) wctype (string)
+# endif
+
+# ifdef _LIBC
+# define ISWCTYPE(WC, WT) __iswctype (WC, WT)
+# else
+# define ISWCTYPE(WC, WT) iswctype (WC, WT)
+# endif
+
+# if (HAVE_MBSTATE_T && HAVE_MBSRTOWCS) || _LIBC
+/* In this case we are implementing the multibyte character handling. */
+# define HANDLE_MULTIBYTE 1
+# endif
+
+# else
+# define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
+
+# define IS_CHAR_CLASS(string) \
+ (STREQ (string, "alpha") || STREQ (string, "upper") \
+ || STREQ (string, "lower") || STREQ (string, "digit") \
+ || STREQ (string, "alnum") || STREQ (string, "xdigit") \
+ || STREQ (string, "space") || STREQ (string, "print") \
+ || STREQ (string, "punct") || STREQ (string, "graph") \
+ || STREQ (string, "cntrl") || STREQ (string, "blank"))
+# endif
+
+/* Avoid depending on library functions or files
+ whose names are inconsistent. */
+
+# ifndef errno
+extern int errno;
+# endif
+
+/* Global variable. */
+static int posixly_correct;
+
+# ifndef internal_function
+/* Inside GNU libc we mark some function in a special way. In other
+ environments simply ignore the marking. */
+# define internal_function
+# endif
+
+/* Note that this evaluates C many times. */
+# ifdef _LIBC
+# define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
+# else
+# define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower (c) : (c))
+# endif
+# define CHAR char
+# define UCHAR unsigned char
+# define INT int
+# define FCT internal_fnmatch
+# define EXT ext_match
+# define END end_pattern
+# define L(CS) CS
+# ifdef _LIBC
+# define BTOWC(C) __btowc (C)
+# else
+# define BTOWC(C) btowc (C)
+# endif
+# define STRLEN(S) strlen (S)
+# define STRCAT(D, S) strcat (D, S)
+# ifdef _LIBC
+# define MEMPCPY(D, S, N) __mempcpy (D, S, N)
+# else
+# if HAVE_MEMPCPY
+# define MEMPCPY(D, S, N) mempcpy (D, S, N)
+# else
+# define MEMPCPY(D, S, N) ((void *) ((char *) memcpy (D, S, N) + (N)))
+# endif
+# endif
+# define MEMCHR(S, C, N) memchr (S, C, N)
+# define STRCOLL(S1, S2) strcoll (S1, S2)
+# include "fnmatch_loop.c"
+
+
+# if HANDLE_MULTIBYTE
+# define FOLD(c) ((flags & FNM_CASEFOLD) ? towlower (c) : (c))
+# define CHAR wchar_t
+# define UCHAR wint_t
+# define INT wint_t
+# define FCT internal_fnwmatch
+# define EXT ext_wmatch
+# define END end_wpattern
+# define L(CS) L##CS
+# define BTOWC(C) (C)
+# ifdef _LIBC
+# define STRLEN(S) __wcslen (S)
+# define STRCAT(D, S) __wcscat (D, S)
+# define MEMPCPY(D, S, N) __wmempcpy (D, S, N)
+# else
+# define STRLEN(S) wcslen (S)
+# define STRCAT(D, S) wcscat (D, S)
+# if HAVE_WMEMPCPY
+# define MEMPCPY(D, S, N) wmempcpy (D, S, N)
+# else
+# define MEMPCPY(D, S, N) (wmemcpy (D, S, N) + (N))
+# endif
+# endif
+# define MEMCHR(S, C, N) wmemchr (S, C, N)
+# define STRCOLL(S1, S2) wcscoll (S1, S2)
+# define WIDE_CHAR_VERSION 1
+
+# undef IS_CHAR_CLASS
+/* We have to convert the wide character string in a multibyte string. But
+ we know that the character class names consist of alphanumeric characters
+ from the portable character set, and since the wide character encoding
+ for a member of the portable character set is the same code point as
+ its single-byte encoding, we can use a simplified method to convert the
+ string to a multibyte character string. */
+static wctype_t
+is_char_class (const wchar_t *wcs)
+{
+ char s[CHAR_CLASS_MAX_LENGTH + 1];
+ char *cp = s;
+
+ do
+ {
+ /* Test for a printable character from the portable character set. */
+# ifdef _LIBC
+ if (*wcs < 0x20 || *wcs > 0x7e
+ || *wcs == 0x24 || *wcs == 0x40 || *wcs == 0x60)
+ return (wctype_t) 0;
+# else
+ switch (*wcs)
+ {
+ case L' ': case L'!': case L'"': case L'#': case L'%':
+ case L'&': case L'\'': case L'(': case L')': case L'*':
+ case L'+': case L',': case L'-': case L'.': case L'/':
+ case L'0': case L'1': case L'2': case L'3': case L'4':
+ case L'5': case L'6': case L'7': case L'8': case L'9':
+ case L':': case L';': case L'<': case L'=': case L'>':
+ case L'?':
+ case L'A': case L'B': case L'C': case L'D': case L'E':
+ case L'F': case L'G': case L'H': case L'I': case L'J':
+ case L'K': case L'L': case L'M': case L'N': case L'O':
+ case L'P': case L'Q': case L'R': case L'S': case L'T':
+ case L'U': case L'V': case L'W': case L'X': case L'Y':
+ case L'Z':
+ case L'[': case L'\\': case L']': case L'^': case L'_':
+ case L'a': case L'b': case L'c': case L'd': case L'e':
+ case L'f': case L'g': case L'h': case L'i': case L'j':
+ case L'k': case L'l': case L'm': case L'n': case L'o':
+ case L'p': case L'q': case L'r': case L's': case L't':
+ case L'u': case L'v': case L'w': case L'x': case L'y':
+ case L'z': case L'{': case L'|': case L'}': case L'~':
+ break;
+ default:
+ return (wctype_t) 0;
+ }
+# endif
+
+ /* Avoid overrunning the buffer. */
+ if (cp == s + CHAR_CLASS_MAX_LENGTH)
+ return (wctype_t) 0;
+
+ *cp++ = (char) *wcs++;
+ }
+ while (*wcs != L'\0');
+
+ *cp = '\0';
+
+# ifdef _LIBC
+ return __wctype (s);
+# else
+ return wctype (s);
+# endif
+}
+# define IS_CHAR_CLASS(string) is_char_class (string)
+
+# include "fnmatch_loop.c"
+# endif
+
+
+int
+fnmatch (const char *pattern, const char *string, int flags)
+{
+# if HANDLE_MULTIBYTE
+# define ALLOCA_LIMIT 2000
+ if (__builtin_expect (MB_CUR_MAX, 1) != 1)
+ {
+ mbstate_t ps;
+ size_t patsize;
+ size_t strsize;
+ size_t totsize;
+ wchar_t *wpattern;
+ wchar_t *wstring;
+ int res;
+
+ /* Calculate the size needed to convert the strings to
+ wide characters. */
+ memset (&ps, '\0', sizeof (ps));
+ patsize = mbsrtowcs (NULL, &pattern, 0, &ps) + 1;
+ if (__builtin_expect (patsize == 0, 0))
+ /* Something wrong.
+ XXX Do we have to set `errno' to something which mbsrtows hasn't
+ already done? */
+ return -1;
+ assert (mbsinit (&ps));
+ strsize = mbsrtowcs (NULL, &string, 0, &ps) + 1;
+ if (__builtin_expect (strsize == 0, 0))
+ /* Something wrong.
+ XXX Do we have to set `errno' to something which mbsrtows hasn't
+ already done? */
+ return -1;
+ assert (mbsinit (&ps));
+ totsize = patsize + strsize;
+ if (__builtin_expect (! (patsize <= totsize
+ && totsize <= SIZE_MAX / sizeof (wchar_t)),
+ 0))
+ {
+ errno = ENOMEM;
+ return -1;
+ }
+
+ /* Allocate room for the wide characters. */
+ if (__builtin_expect (totsize < ALLOCA_LIMIT, 1))
+ wpattern = (wchar_t *) alloca (totsize * sizeof (wchar_t));
+ else
+ {
+ wpattern = malloc (totsize * sizeof (wchar_t));
+ if (__builtin_expect (! wpattern, 0))
+ {
+ errno = ENOMEM;
+ return -1;
+ }
+ }
+ wstring = wpattern + patsize;
+
+ /* Convert the strings into wide characters. */
+ mbsrtowcs (wpattern, &pattern, patsize, &ps);
+ assert (mbsinit (&ps));
+ mbsrtowcs (wstring, &string, strsize, &ps);
+
+ res = internal_fnwmatch (wpattern, wstring, wstring + strsize - 1,
+ flags & FNM_PERIOD, flags);
+
+ if (__builtin_expect (! (totsize < ALLOCA_LIMIT), 0))
+ free (wpattern);
+ return res;
+ }
+# endif /* HANDLE_MULTIBYTE */
+
+ return internal_fnmatch (pattern, string, string + strlen (string),
+ flags & FNM_PERIOD, flags);
+}
+
+# ifdef _LIBC
+# undef fnmatch
+versioned_symbol (libc, __fnmatch, fnmatch, GLIBC_2_2_3);
+# if SHLIB_COMPAT(libc, GLIBC_2_0, GLIBC_2_2_3)
+strong_alias (__fnmatch, __fnmatch_old)
+compat_symbol (libc, __fnmatch_old, fnmatch, GLIBC_2_0);
+# endif
+libc_hidden_ver (__fnmatch, fnmatch)
+# endif
+
+#endif /* _LIBC or not __GNU_LIBRARY__. */
diff --git a/contrib/diff/lib/fnmatch_.h b/contrib/diff/lib/fnmatch_.h
new file mode 100644
index 0000000..87e661f
--- /dev/null
+++ b/contrib/diff/lib/fnmatch_.h
@@ -0,0 +1,63 @@
+/* Copyright (C) 1991, 1992, 1993, 1996, 1997, 1998, 1999, 2001, 2002, 2003
+ Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _FNMATCH_H
+# define _FNMATCH_H 1
+
+# ifdef __cplusplus
+extern "C" {
+# endif
+
+/* We #undef these before defining them because some losing systems
+ (HP-UX A.08.07 for example) define these in <unistd.h>. */
+# undef FNM_PATHNAME
+# undef FNM_NOESCAPE
+# undef FNM_PERIOD
+
+/* Bits set in the FLAGS argument to `fnmatch'. */
+# define FNM_PATHNAME (1 << 0) /* No wildcard can ever match `/'. */
+# define FNM_NOESCAPE (1 << 1) /* Backslashes don't quote special chars. */
+# define FNM_PERIOD (1 << 2) /* Leading `.' is matched only explicitly. */
+
+# if !defined _POSIX_C_SOURCE || _POSIX_C_SOURCE < 2 || defined _GNU_SOURCE
+# define FNM_FILE_NAME FNM_PATHNAME /* Preferred GNU name. */
+# define FNM_LEADING_DIR (1 << 3) /* Ignore `/...' after a match. */
+# define FNM_CASEFOLD (1 << 4) /* Compare without regard to case. */
+# define FNM_EXTMATCH (1 << 5) /* Use ksh-like extended matching. */
+# endif
+
+/* Value returned by `fnmatch' if STRING does not match PATTERN. */
+# define FNM_NOMATCH 1
+
+/* This value is returned if the implementation does not support
+ `fnmatch'. Since this is not the case here it will never be
+ returned but the conformance test suites still require the symbol
+ to be defined. */
+# ifdef _XOPEN_SOURCE
+# define FNM_NOSYS (-1)
+# endif
+
+/* Match NAME against the filename pattern PATTERN,
+ returning zero if it matches, FNM_NOMATCH if not. */
+extern int fnmatch (const char *__pattern, const char *__name,
+ int __flags);
+
+# ifdef __cplusplus
+}
+# endif
+
+#endif /* fnmatch.h */
diff --git a/contrib/diff/lib/fnmatch_loop.c b/contrib/diff/lib/fnmatch_loop.c
new file mode 100644
index 0000000..7798b59
--- /dev/null
+++ b/contrib/diff/lib/fnmatch_loop.c
@@ -0,0 +1,1192 @@
+/* Copyright (C) 1991,1992,1993,1996,1997,1998,1999,2000,2001,2002,2003,2004
+ Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Match STRING against the filename pattern PATTERN, returning zero if
+ it matches, nonzero if not. */
+static int EXT (INT opt, const CHAR *pattern, const CHAR *string,
+ const CHAR *string_end, int no_leading_period, int flags)
+ internal_function;
+static const CHAR *END (const CHAR *patternp) internal_function;
+
+static int
+internal_function
+FCT (const CHAR *pattern, const CHAR *string, const CHAR *string_end,
+ int no_leading_period, int flags)
+{
+ register const CHAR *p = pattern, *n = string;
+ register UCHAR c;
+#ifdef _LIBC
+# if WIDE_CHAR_VERSION
+ const char *collseq = (const char *)
+ _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQWC);
+# else
+ const UCHAR *collseq = (const UCHAR *)
+ _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQMB);
+# endif
+#endif
+
+ while ((c = *p++) != L('\0'))
+ {
+ int new_no_leading_period = 0;
+ c = FOLD (c);
+
+ switch (c)
+ {
+ case L('?'):
+ if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
+ {
+ int res;
+
+ res = EXT (c, p, n, string_end, no_leading_period,
+ flags);
+ if (res != -1)
+ return res;
+ }
+
+ if (n == string_end)
+ return FNM_NOMATCH;
+ else if (*n == L('/') && (flags & FNM_FILE_NAME))
+ return FNM_NOMATCH;
+ else if (*n == L('.') && no_leading_period)
+ return FNM_NOMATCH;
+ break;
+
+ case L('\\'):
+ if (!(flags & FNM_NOESCAPE))
+ {
+ c = *p++;
+ if (c == L('\0'))
+ /* Trailing \ loses. */
+ return FNM_NOMATCH;
+ c = FOLD (c);
+ }
+ if (n == string_end || FOLD ((UCHAR) *n) != c)
+ return FNM_NOMATCH;
+ break;
+
+ case L('*'):
+ if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
+ {
+ int res;
+
+ res = EXT (c, p, n, string_end, no_leading_period,
+ flags);
+ if (res != -1)
+ return res;
+ }
+
+ if (n != string_end && *n == L('.') && no_leading_period)
+ return FNM_NOMATCH;
+
+ for (c = *p++; c == L('?') || c == L('*'); c = *p++)
+ {
+ if (*p == L('(') && (flags & FNM_EXTMATCH) != 0)
+ {
+ const CHAR *endp = END (p);
+ if (endp != p)
+ {
+ /* This is a pattern. Skip over it. */
+ p = endp;
+ continue;
+ }
+ }
+
+ if (c == L('?'))
+ {
+ /* A ? needs to match one character. */
+ if (n == string_end)
+ /* There isn't another character; no match. */
+ return FNM_NOMATCH;
+ else if (*n == L('/')
+ && __builtin_expect (flags & FNM_FILE_NAME, 0))
+ /* A slash does not match a wildcard under
+ FNM_FILE_NAME. */
+ return FNM_NOMATCH;
+ else
+ /* One character of the string is consumed in matching
+ this ? wildcard, so *??? won't match if there are
+ less than three characters. */
+ ++n;
+ }
+ }
+
+ if (c == L('\0'))
+ /* The wildcard(s) is/are the last element of the pattern.
+ If the name is a file name and contains another slash
+ this means it cannot match, unless the FNM_LEADING_DIR
+ flag is set. */
+ {
+ int result = (flags & FNM_FILE_NAME) == 0 ? 0 : FNM_NOMATCH;
+
+ if (flags & FNM_FILE_NAME)
+ {
+ if (flags & FNM_LEADING_DIR)
+ result = 0;
+ else
+ {
+ if (MEMCHR (n, L('/'), string_end - n) == NULL)
+ result = 0;
+ }
+ }
+
+ return result;
+ }
+ else
+ {
+ const CHAR *endp;
+
+ endp = MEMCHR (n, (flags & FNM_FILE_NAME) ? L('/') : L('\0'),
+ string_end - n);
+ if (endp == NULL)
+ endp = string_end;
+
+ if (c == L('[')
+ || (__builtin_expect (flags & FNM_EXTMATCH, 0) != 0
+ && (c == L('@') || c == L('+') || c == L('!'))
+ && *p == L('(')))
+ {
+ int flags2 = ((flags & FNM_FILE_NAME)
+ ? flags : (flags & ~FNM_PERIOD));
+ int no_leading_period2 = no_leading_period;
+
+ for (--p; n < endp; ++n, no_leading_period2 = 0)
+ if (FCT (p, n, string_end, no_leading_period2, flags2)
+ == 0)
+ return 0;
+ }
+ else if (c == L('/') && (flags & FNM_FILE_NAME))
+ {
+ while (n < string_end && *n != L('/'))
+ ++n;
+ if (n < string_end && *n == L('/')
+ && (FCT (p, n + 1, string_end, flags & FNM_PERIOD, flags)
+ == 0))
+ return 0;
+ }
+ else
+ {
+ int flags2 = ((flags & FNM_FILE_NAME)
+ ? flags : (flags & ~FNM_PERIOD));
+ int no_leading_period2 = no_leading_period;
+
+ if (c == L('\\') && !(flags & FNM_NOESCAPE))
+ c = *p;
+ c = FOLD (c);
+ for (--p; n < endp; ++n, no_leading_period2 = 0)
+ if (FOLD ((UCHAR) *n) == c
+ && (FCT (p, n, string_end, no_leading_period2, flags2)
+ == 0))
+ return 0;
+ }
+ }
+
+ /* If we come here no match is possible with the wildcard. */
+ return FNM_NOMATCH;
+
+ case L('['):
+ {
+ /* Nonzero if the sense of the character class is inverted. */
+ register int not;
+ CHAR cold;
+ UCHAR fn;
+
+ if (posixly_correct == 0)
+ posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;
+
+ if (n == string_end)
+ return FNM_NOMATCH;
+
+ if (*n == L('.') && no_leading_period)
+ return FNM_NOMATCH;
+
+ if (*n == L('/') && (flags & FNM_FILE_NAME))
+ /* `/' cannot be matched. */
+ return FNM_NOMATCH;
+
+ not = (*p == L('!') || (posixly_correct < 0 && *p == L('^')));
+ if (not)
+ ++p;
+
+ fn = FOLD ((UCHAR) *n);
+
+ c = *p++;
+ for (;;)
+ {
+ if (!(flags & FNM_NOESCAPE) && c == L('\\'))
+ {
+ if (*p == L('\0'))
+ return FNM_NOMATCH;
+ c = FOLD ((UCHAR) *p);
+ ++p;
+
+ if (c == fn)
+ goto matched;
+ }
+ else if (c == L('[') && *p == L(':'))
+ {
+ /* Leave room for the null. */
+ CHAR str[CHAR_CLASS_MAX_LENGTH + 1];
+ size_t c1 = 0;
+#if defined _LIBC || WIDE_CHAR_SUPPORT
+ wctype_t wt;
+#endif
+ const CHAR *startp = p;
+
+ for (;;)
+ {
+ if (c1 == CHAR_CLASS_MAX_LENGTH)
+ /* The name is too long and therefore the pattern
+ is ill-formed. */
+ return FNM_NOMATCH;
+
+ c = *++p;
+ if (c == L(':') && p[1] == L(']'))
+ {
+ p += 2;
+ break;
+ }
+ if (c < L('a') || c >= L('z'))
+ {
+ /* This cannot possibly be a character class name.
+ Match it as a normal range. */
+ p = startp;
+ c = L('[');
+ goto normal_bracket;
+ }
+ str[c1++] = c;
+ }
+ str[c1] = L('\0');
+
+#if defined _LIBC || WIDE_CHAR_SUPPORT
+ wt = IS_CHAR_CLASS (str);
+ if (wt == 0)
+ /* Invalid character class name. */
+ return FNM_NOMATCH;
+
+# if defined _LIBC && ! WIDE_CHAR_VERSION
+ /* The following code is glibc specific but does
+ there a good job in speeding up the code since
+ we can avoid the btowc() call. */
+ if (_ISCTYPE ((UCHAR) *n, wt))
+ goto matched;
+# else
+ if (ISWCTYPE (BTOWC ((UCHAR) *n), wt))
+ goto matched;
+# endif
+#else
+ if ((STREQ (str, L("alnum")) && ISALNUM ((UCHAR) *n))
+ || (STREQ (str, L("alpha")) && ISALPHA ((UCHAR) *n))
+ || (STREQ (str, L("blank")) && ISBLANK ((UCHAR) *n))
+ || (STREQ (str, L("cntrl")) && ISCNTRL ((UCHAR) *n))
+ || (STREQ (str, L("digit")) && ISDIGIT ((UCHAR) *n))
+ || (STREQ (str, L("graph")) && ISGRAPH ((UCHAR) *n))
+ || (STREQ (str, L("lower")) && ISLOWER ((UCHAR) *n))
+ || (STREQ (str, L("print")) && ISPRINT ((UCHAR) *n))
+ || (STREQ (str, L("punct")) && ISPUNCT ((UCHAR) *n))
+ || (STREQ (str, L("space")) && ISSPACE ((UCHAR) *n))
+ || (STREQ (str, L("upper")) && ISUPPER ((UCHAR) *n))
+ || (STREQ (str, L("xdigit")) && ISXDIGIT ((UCHAR) *n)))
+ goto matched;
+#endif
+ c = *p++;
+ }
+#ifdef _LIBC
+ else if (c == L('[') && *p == L('='))
+ {
+ UCHAR str[1];
+ uint32_t nrules =
+ _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+ const CHAR *startp = p;
+
+ c = *++p;
+ if (c == L('\0'))
+ {
+ p = startp;
+ c = L('[');
+ goto normal_bracket;
+ }
+ str[0] = c;
+
+ c = *++p;
+ if (c != L('=') || p[1] != L(']'))
+ {
+ p = startp;
+ c = L('[');
+ goto normal_bracket;
+ }
+ p += 2;
+
+ if (nrules == 0)
+ {
+ if ((UCHAR) *n == str[0])
+ goto matched;
+ }
+ else
+ {
+ const int32_t *table;
+# if WIDE_CHAR_VERSION
+ const int32_t *weights;
+ const int32_t *extra;
+# else
+ const unsigned char *weights;
+ const unsigned char *extra;
+# endif
+ const int32_t *indirect;
+ int32_t idx;
+ const UCHAR *cp = (const UCHAR *) str;
+
+ /* This #include defines a local function! */
+# if WIDE_CHAR_VERSION
+# include <locale/weightwc.h>
+# else
+# include <locale/weight.h>
+# endif
+
+# if WIDE_CHAR_VERSION
+ table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC);
+ weights = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC);
+ extra = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC);
+ indirect = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC);
+# else
+ table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
+ weights = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
+ extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
+ indirect = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
+# endif
+
+ idx = findidx (&cp);
+ if (idx != 0)
+ {
+ /* We found a table entry. Now see whether the
+ character we are currently at has the same
+ equivalance class value. */
+ int len = weights[idx];
+ int32_t idx2;
+ const UCHAR *np = (const UCHAR *) n;
+
+ idx2 = findidx (&np);
+ if (idx2 != 0 && len == weights[idx2])
+ {
+ int cnt = 0;
+
+ while (cnt < len
+ && (weights[idx + 1 + cnt]
+ == weights[idx2 + 1 + cnt]))
+ ++cnt;
+
+ if (cnt == len)
+ goto matched;
+ }
+ }
+ }
+
+ c = *p++;
+ }
+#endif
+ else if (c == L('\0'))
+ /* [ (unterminated) loses. */
+ return FNM_NOMATCH;
+ else
+ {
+ int is_range = 0;
+
+#ifdef _LIBC
+ int is_seqval = 0;
+
+ if (c == L('[') && *p == L('.'))
+ {
+ uint32_t nrules =
+ _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+ const CHAR *startp = p;
+ size_t c1 = 0;
+
+ while (1)
+ {
+ c = *++p;
+ if (c == L('.') && p[1] == L(']'))
+ {
+ p += 2;
+ break;
+ }
+ if (c == '\0')
+ return FNM_NOMATCH;
+ ++c1;
+ }
+
+ /* We have to handling the symbols differently in
+ ranges since then the collation sequence is
+ important. */
+ is_range = *p == L('-') && p[1] != L('\0');
+
+ if (nrules == 0)
+ {
+ /* There are no names defined in the collation
+ data. Therefore we only accept the trivial
+ names consisting of the character itself. */
+ if (c1 != 1)
+ return FNM_NOMATCH;
+
+ if (!is_range && *n == startp[1])
+ goto matched;
+
+ cold = startp[1];
+ c = *p++;
+ }
+ else
+ {
+ int32_t table_size;
+ const int32_t *symb_table;
+# ifdef WIDE_CHAR_VERSION
+ char str[c1];
+ unsigned int strcnt;
+# else
+# define str (startp + 1)
+# endif
+ const unsigned char *extra;
+ int32_t idx;
+ int32_t elem;
+ int32_t second;
+ int32_t hash;
+
+# ifdef WIDE_CHAR_VERSION
+ /* We have to convert the name to a single-byte
+ string. This is possible since the names
+ consist of ASCII characters and the internal
+ representation is UCS4. */
+ for (strcnt = 0; strcnt < c1; ++strcnt)
+ str[strcnt] = startp[1 + strcnt];
+# endif
+
+ table_size =
+ _NL_CURRENT_WORD (LC_COLLATE,
+ _NL_COLLATE_SYMB_HASH_SIZEMB);
+ symb_table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_TABLEMB);
+ extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_EXTRAMB);
+
+ /* Locate the character in the hashing table. */
+ hash = elem_hash (str, c1);
+
+ idx = 0;
+ elem = hash % table_size;
+ second = hash % (table_size - 2);
+ while (symb_table[2 * elem] != 0)
+ {
+ /* First compare the hashing value. */
+ if (symb_table[2 * elem] == hash
+ && c1 == extra[symb_table[2 * elem + 1]]
+ && memcmp (str,
+ &extra[symb_table[2 * elem + 1]
+ + 1], c1) == 0)
+ {
+ /* Yep, this is the entry. */
+ idx = symb_table[2 * elem + 1];
+ idx += 1 + extra[idx];
+ break;
+ }
+
+ /* Next entry. */
+ elem += second;
+ }
+
+ if (symb_table[2 * elem] != 0)
+ {
+ /* Compare the byte sequence but only if
+ this is not part of a range. */
+# ifdef WIDE_CHAR_VERSION
+ int32_t *wextra;
+
+ idx += 1 + extra[idx];
+ /* Adjust for the alignment. */
+ idx = (idx + 3) & ~3;
+
+ wextra = (int32_t *) &extra[idx + 4];
+# endif
+
+ if (! is_range)
+ {
+# ifdef WIDE_CHAR_VERSION
+ for (c1 = 0;
+ (int32_t) c1 < wextra[idx];
+ ++c1)
+ if (n[c1] != wextra[1 + c1])
+ break;
+
+ if ((int32_t) c1 == wextra[idx])
+ goto matched;
+# else
+ for (c1 = 0; c1 < extra[idx]; ++c1)
+ if (n[c1] != extra[1 + c1])
+ break;
+
+ if (c1 == extra[idx])
+ goto matched;
+# endif
+ }
+
+ /* Get the collation sequence value. */
+ is_seqval = 1;
+# ifdef WIDE_CHAR_VERSION
+ cold = wextra[1 + wextra[idx]];
+# else
+ /* Adjust for the alignment. */
+ idx += 1 + extra[idx];
+ idx = (idx + 3) & ~4;
+ cold = *((int32_t *) &extra[idx]);
+# endif
+
+ c = *p++;
+ }
+ else if (c1 == 1)
+ {
+ /* No valid character. Match it as a
+ single byte. */
+ if (!is_range && *n == str[0])
+ goto matched;
+
+ cold = str[0];
+ c = *p++;
+ }
+ else
+ return FNM_NOMATCH;
+ }
+ }
+ else
+# undef str
+#endif
+ {
+ c = FOLD (c);
+ normal_bracket:
+
+ /* We have to handling the symbols differently in
+ ranges since then the collation sequence is
+ important. */
+ is_range = (*p == L('-') && p[1] != L('\0')
+ && p[1] != L(']'));
+
+ if (!is_range && c == fn)
+ goto matched;
+
+ cold = c;
+ c = *p++;
+ }
+
+ if (c == L('-') && *p != L(']'))
+ {
+#if _LIBC
+ /* We have to find the collation sequence
+ value for C. Collation sequence is nothing
+ we can regularly access. The sequence
+ value is defined by the order in which the
+ definitions of the collation values for the
+ various characters appear in the source
+ file. A strange concept, nowhere
+ documented. */
+ uint32_t fcollseq;
+ uint32_t lcollseq;
+ UCHAR cend = *p++;
+
+# ifdef WIDE_CHAR_VERSION
+ /* Search in the `names' array for the characters. */
+ fcollseq = __collseq_table_lookup (collseq, fn);
+ if (fcollseq == ~((uint32_t) 0))
+ /* XXX We don't know anything about the character
+ we are supposed to match. This means we are
+ failing. */
+ goto range_not_matched;
+
+ if (is_seqval)
+ lcollseq = cold;
+ else
+ lcollseq = __collseq_table_lookup (collseq, cold);
+# else
+ fcollseq = collseq[fn];
+ lcollseq = is_seqval ? cold : collseq[(UCHAR) cold];
+# endif
+
+ is_seqval = 0;
+ if (cend == L('[') && *p == L('.'))
+ {
+ uint32_t nrules =
+ _NL_CURRENT_WORD (LC_COLLATE,
+ _NL_COLLATE_NRULES);
+ const CHAR *startp = p;
+ size_t c1 = 0;
+
+ while (1)
+ {
+ c = *++p;
+ if (c == L('.') && p[1] == L(']'))
+ {
+ p += 2;
+ break;
+ }
+ if (c == '\0')
+ return FNM_NOMATCH;
+ ++c1;
+ }
+
+ if (nrules == 0)
+ {
+ /* There are no names defined in the
+ collation data. Therefore we only
+ accept the trivial names consisting
+ of the character itself. */
+ if (c1 != 1)
+ return FNM_NOMATCH;
+
+ cend = startp[1];
+ }
+ else
+ {
+ int32_t table_size;
+ const int32_t *symb_table;
+# ifdef WIDE_CHAR_VERSION
+ char str[c1];
+ unsigned int strcnt;
+# else
+# define str (startp + 1)
+# endif
+ const unsigned char *extra;
+ int32_t idx;
+ int32_t elem;
+ int32_t second;
+ int32_t hash;
+
+# ifdef WIDE_CHAR_VERSION
+ /* We have to convert the name to a single-byte
+ string. This is possible since the names
+ consist of ASCII characters and the internal
+ representation is UCS4. */
+ for (strcnt = 0; strcnt < c1; ++strcnt)
+ str[strcnt] = startp[1 + strcnt];
+# endif
+
+ table_size =
+ _NL_CURRENT_WORD (LC_COLLATE,
+ _NL_COLLATE_SYMB_HASH_SIZEMB);
+ symb_table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_TABLEMB);
+ extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_EXTRAMB);
+
+ /* Locate the character in the hashing
+ table. */
+ hash = elem_hash (str, c1);
+
+ idx = 0;
+ elem = hash % table_size;
+ second = hash % (table_size - 2);
+ while (symb_table[2 * elem] != 0)
+ {
+ /* First compare the hashing value. */
+ if (symb_table[2 * elem] == hash
+ && (c1
+ == extra[symb_table[2 * elem + 1]])
+ && memcmp (str,
+ &extra[symb_table[2 * elem + 1]
+ + 1], c1) == 0)
+ {
+ /* Yep, this is the entry. */
+ idx = symb_table[2 * elem + 1];
+ idx += 1 + extra[idx];
+ break;
+ }
+
+ /* Next entry. */
+ elem += second;
+ }
+
+ if (symb_table[2 * elem] != 0)
+ {
+ /* Compare the byte sequence but only if
+ this is not part of a range. */
+# ifdef WIDE_CHAR_VERSION
+ int32_t *wextra;
+
+ idx += 1 + extra[idx];
+ /* Adjust for the alignment. */
+ idx = (idx + 3) & ~4;
+
+ wextra = (int32_t *) &extra[idx + 4];
+# endif
+ /* Get the collation sequence value. */
+ is_seqval = 1;
+# ifdef WIDE_CHAR_VERSION
+ cend = wextra[1 + wextra[idx]];
+# else
+ /* Adjust for the alignment. */
+ idx += 1 + extra[idx];
+ idx = (idx + 3) & ~4;
+ cend = *((int32_t *) &extra[idx]);
+# endif
+ }
+ else if (symb_table[2 * elem] != 0 && c1 == 1)
+ {
+ cend = str[0];
+ c = *p++;
+ }
+ else
+ return FNM_NOMATCH;
+ }
+# undef str
+ }
+ else
+ {
+ if (!(flags & FNM_NOESCAPE) && cend == L('\\'))
+ cend = *p++;
+ if (cend == L('\0'))
+ return FNM_NOMATCH;
+ cend = FOLD (cend);
+ }
+
+ /* XXX It is not entirely clear to me how to handle
+ characters which are not mentioned in the
+ collation specification. */
+ if (
+# ifdef WIDE_CHAR_VERSION
+ lcollseq == 0xffffffff ||
+# endif
+ lcollseq <= fcollseq)
+ {
+ /* We have to look at the upper bound. */
+ uint32_t hcollseq;
+
+ if (is_seqval)
+ hcollseq = cend;
+ else
+ {
+# ifdef WIDE_CHAR_VERSION
+ hcollseq =
+ __collseq_table_lookup (collseq, cend);
+ if (hcollseq == ~((uint32_t) 0))
+ {
+ /* Hum, no information about the upper
+ bound. The matching succeeds if the
+ lower bound is matched exactly. */
+ if (lcollseq != fcollseq)
+ goto range_not_matched;
+
+ goto matched;
+ }
+# else
+ hcollseq = collseq[cend];
+# endif
+ }
+
+ if (lcollseq <= hcollseq && fcollseq <= hcollseq)
+ goto matched;
+ }
+# ifdef WIDE_CHAR_VERSION
+ range_not_matched:
+# endif
+#else
+ /* We use a boring value comparison of the character
+ values. This is better than comparing using
+ `strcoll' since the latter would have surprising
+ and sometimes fatal consequences. */
+ UCHAR cend = *p++;
+
+ if (!(flags & FNM_NOESCAPE) && cend == L('\\'))
+ cend = *p++;
+ if (cend == L('\0'))
+ return FNM_NOMATCH;
+
+ /* It is a range. */
+ if (cold <= fn && fn <= cend)
+ goto matched;
+#endif
+
+ c = *p++;
+ }
+ }
+
+ if (c == L(']'))
+ break;
+ }
+
+ if (!not)
+ return FNM_NOMATCH;
+ break;
+
+ matched:
+ /* Skip the rest of the [...] that already matched. */
+ do
+ {
+ ignore_next:
+ c = *p++;
+
+ if (c == L('\0'))
+ /* [... (unterminated) loses. */
+ return FNM_NOMATCH;
+
+ if (!(flags & FNM_NOESCAPE) && c == L('\\'))
+ {
+ if (*p == L('\0'))
+ return FNM_NOMATCH;
+ /* XXX 1003.2d11 is unclear if this is right. */
+ ++p;
+ }
+ else if (c == L('[') && *p == L(':'))
+ {
+ int c1 = 0;
+ const CHAR *startp = p;
+
+ while (1)
+ {
+ c = *++p;
+ if (++c1 == CHAR_CLASS_MAX_LENGTH)
+ return FNM_NOMATCH;
+
+ if (*p == L(':') && p[1] == L(']'))
+ break;
+
+ if (c < L('a') || c >= L('z'))
+ {
+ p = startp;
+ goto ignore_next;
+ }
+ }
+ p += 2;
+ c = *p++;
+ }
+ else if (c == L('[') && *p == L('='))
+ {
+ c = *++p;
+ if (c == L('\0'))
+ return FNM_NOMATCH;
+ c = *++p;
+ if (c != L('=') || p[1] != L(']'))
+ return FNM_NOMATCH;
+ p += 2;
+ c = *p++;
+ }
+ else if (c == L('[') && *p == L('.'))
+ {
+ ++p;
+ while (1)
+ {
+ c = *++p;
+ if (c == '\0')
+ return FNM_NOMATCH;
+
+ if (*p == L('.') && p[1] == L(']'))
+ break;
+ }
+ p += 2;
+ c = *p++;
+ }
+ }
+ while (c != L(']'));
+ if (not)
+ return FNM_NOMATCH;
+ }
+ break;
+
+ case L('+'):
+ case L('@'):
+ case L('!'):
+ if (__builtin_expect (flags & FNM_EXTMATCH, 0) && *p == '(')
+ {
+ int res;
+
+ res = EXT (c, p, n, string_end, no_leading_period, flags);
+ if (res != -1)
+ return res;
+ }
+ goto normal_match;
+
+ case L('/'):
+ if (NO_LEADING_PERIOD (flags))
+ {
+ if (n == string_end || c != (UCHAR) *n)
+ return FNM_NOMATCH;
+
+ new_no_leading_period = 1;
+ break;
+ }
+ /* FALLTHROUGH */
+ default:
+ normal_match:
+ if (n == string_end || c != FOLD ((UCHAR) *n))
+ return FNM_NOMATCH;
+ }
+
+ no_leading_period = new_no_leading_period;
+ ++n;
+ }
+
+ if (n == string_end)
+ return 0;
+
+ if ((flags & FNM_LEADING_DIR) && n != string_end && *n == L('/'))
+ /* The FNM_LEADING_DIR flag says that "foo*" matches "foobar/frobozz". */
+ return 0;
+
+ return FNM_NOMATCH;
+}
+
+
+static const CHAR *
+internal_function
+END (const CHAR *pattern)
+{
+ const CHAR *p = pattern;
+
+ while (1)
+ if (*++p == L('\0'))
+ /* This is an invalid pattern. */
+ return pattern;
+ else if (*p == L('['))
+ {
+ /* Handle brackets special. */
+ if (posixly_correct == 0)
+ posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;
+
+ /* Skip the not sign. We have to recognize it because of a possibly
+ following ']'. */
+ if (*++p == L('!') || (posixly_correct < 0 && *p == L('^')))
+ ++p;
+ /* A leading ']' is recognized as such. */
+ if (*p == L(']'))
+ ++p;
+ /* Skip over all characters of the list. */
+ while (*p != L(']'))
+ if (*p++ == L('\0'))
+ /* This is no valid pattern. */
+ return pattern;
+ }
+ else if ((*p == L('?') || *p == L('*') || *p == L('+') || *p == L('@')
+ || *p == L('!')) && p[1] == L('('))
+ p = END (p + 1);
+ else if (*p == L(')'))
+ break;
+
+ return p + 1;
+}
+
+
+static int
+internal_function
+EXT (INT opt, const CHAR *pattern, const CHAR *string, const CHAR *string_end,
+ int no_leading_period, int flags)
+{
+ const CHAR *startp;
+ int level;
+ struct patternlist
+ {
+ struct patternlist *next;
+ CHAR str[1];
+ } *list = NULL;
+ struct patternlist **lastp = &list;
+ size_t pattern_len = STRLEN (pattern);
+ const CHAR *p;
+ const CHAR *rs;
+ enum { ALLOCA_LIMIT = 8000 };
+
+ /* Parse the pattern. Store the individual parts in the list. */
+ level = 0;
+ for (startp = p = pattern + 1; level >= 0; ++p)
+ if (*p == L('\0'))
+ /* This is an invalid pattern. */
+ return -1;
+ else if (*p == L('['))
+ {
+ /* Handle brackets special. */
+ if (posixly_correct == 0)
+ posixly_correct = getenv ("POSIXLY_CORRECT") != NULL ? 1 : -1;
+
+ /* Skip the not sign. We have to recognize it because of a possibly
+ following ']'. */
+ if (*++p == L('!') || (posixly_correct < 0 && *p == L('^')))
+ ++p;
+ /* A leading ']' is recognized as such. */
+ if (*p == L(']'))
+ ++p;
+ /* Skip over all characters of the list. */
+ while (*p != L(']'))
+ if (*p++ == L('\0'))
+ /* This is no valid pattern. */
+ return -1;
+ }
+ else if ((*p == L('?') || *p == L('*') || *p == L('+') || *p == L('@')
+ || *p == L('!')) && p[1] == L('('))
+ /* Remember the nesting level. */
+ ++level;
+ else if (*p == L(')'))
+ {
+ if (level-- == 0)
+ {
+ /* This means we found the end of the pattern. */
+#define NEW_PATTERN \
+ struct patternlist *newp; \
+ size_t plen; \
+ size_t plensize; \
+ size_t newpsize; \
+ \
+ plen = (opt == L('?') || opt == L('@') \
+ ? pattern_len \
+ : p - startp + 1); \
+ plensize = plen * sizeof (CHAR); \
+ newpsize = offsetof (struct patternlist, str) + plensize; \
+ if ((size_t) -1 / sizeof (CHAR) < plen \
+ || newpsize < offsetof (struct patternlist, str) \
+ || ALLOCA_LIMIT <= newpsize) \
+ return -1; \
+ newp = (struct patternlist *) alloca (newpsize); \
+ *((CHAR *) MEMPCPY (newp->str, startp, p - startp)) = L('\0'); \
+ newp->next = NULL; \
+ *lastp = newp; \
+ lastp = &newp->next
+ NEW_PATTERN;
+ }
+ }
+ else if (*p == L('|'))
+ {
+ if (level == 0)
+ {
+ NEW_PATTERN;
+ startp = p + 1;
+ }
+ }
+ assert (list != NULL);
+ assert (p[-1] == L(')'));
+#undef NEW_PATTERN
+
+ switch (opt)
+ {
+ case L('*'):
+ if (FCT (p, string, string_end, no_leading_period, flags) == 0)
+ return 0;
+ /* FALLTHROUGH */
+
+ case L('+'):
+ do
+ {
+ for (rs = string; rs <= string_end; ++rs)
+ /* First match the prefix with the current pattern with the
+ current pattern. */
+ if (FCT (list->str, string, rs, no_leading_period,
+ flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0
+ /* This was successful. Now match the rest with the rest
+ of the pattern. */
+ && (FCT (p, rs, string_end,
+ rs == string
+ ? no_leading_period
+ : rs[-1] == '/' && NO_LEADING_PERIOD (flags) ? 1 : 0,
+ flags & FNM_FILE_NAME
+ ? flags : flags & ~FNM_PERIOD) == 0
+ /* This didn't work. Try the whole pattern. */
+ || (rs != string
+ && FCT (pattern - 1, rs, string_end,
+ rs == string
+ ? no_leading_period
+ : (rs[-1] == '/' && NO_LEADING_PERIOD (flags)
+ ? 1 : 0),
+ flags & FNM_FILE_NAME
+ ? flags : flags & ~FNM_PERIOD) == 0)))
+ /* It worked. Signal success. */
+ return 0;
+ }
+ while ((list = list->next) != NULL);
+
+ /* None of the patterns lead to a match. */
+ return FNM_NOMATCH;
+
+ case L('?'):
+ if (FCT (p, string, string_end, no_leading_period, flags) == 0)
+ return 0;
+ /* FALLTHROUGH */
+
+ case L('@'):
+ do
+ /* I cannot believe it but `strcat' is actually acceptable
+ here. Match the entire string with the prefix from the
+ pattern list and the rest of the pattern following the
+ pattern list. */
+ if (FCT (STRCAT (list->str, p), string, string_end,
+ no_leading_period,
+ flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0)
+ /* It worked. Signal success. */
+ return 0;
+ while ((list = list->next) != NULL);
+
+ /* None of the patterns lead to a match. */
+ return FNM_NOMATCH;
+
+ case L('!'):
+ for (rs = string; rs <= string_end; ++rs)
+ {
+ struct patternlist *runp;
+
+ for (runp = list; runp != NULL; runp = runp->next)
+ if (FCT (runp->str, string, rs, no_leading_period,
+ flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD) == 0)
+ break;
+
+ /* If none of the patterns matched see whether the rest does. */
+ if (runp == NULL
+ && (FCT (p, rs, string_end,
+ rs == string
+ ? no_leading_period
+ : rs[-1] == '/' && NO_LEADING_PERIOD (flags) ? 1 : 0,
+ flags & FNM_FILE_NAME ? flags : flags & ~FNM_PERIOD)
+ == 0))
+ /* This is successful. */
+ return 0;
+ }
+
+ /* None of the patterns together with the rest of the pattern
+ lead to a match. */
+ return FNM_NOMATCH;
+
+ default:
+ assert (! "Invalid extended matching operator");
+ break;
+ }
+
+ return -1;
+}
+
+
+#undef FOLD
+#undef CHAR
+#undef UCHAR
+#undef INT
+#undef FCT
+#undef EXT
+#undef END
+#undef MEMPCPY
+#undef MEMCHR
+#undef STRCOLL
+#undef STRLEN
+#undef STRCAT
+#undef L
+#undef BTOWC
diff --git a/contrib/diff/lib/getopt.c b/contrib/diff/lib/getopt.c
new file mode 100644
index 0000000..8f017f8
--- /dev/null
+++ b/contrib/diff/lib/getopt.c
@@ -0,0 +1,1254 @@
+/* Getopt for GNU.
+ NOTE: getopt is now part of the C library, so if you don't know what
+ "Keep this file name-space clean" means, talk to drepper@gnu.org
+ before changing it!
+ Copyright (C) 1987,88,89,90,91,92,93,94,95,96,98,99,2000,2001,2002,2003,2004
+ Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* This tells Alpha OSF/1 not to define a getopt prototype in <stdio.h>.
+ Ditto for AIX 3.2 and <stdlib.h>. */
+#ifndef _NO_PROTO
+# define _NO_PROTO
+#endif
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <stdio.h>
+
+/* Comment out all this code if we are using the GNU C Library, and are not
+ actually compiling the library itself. This code is part of the GNU C
+ Library, but also included in many other GNU distributions. Compiling
+ and linking in this code is a waste when using the GNU C library
+ (especially if it is a shared library). Rather than having every GNU
+ program understand `configure --with-gnu-libc' and omit the object files,
+ it is simpler to just do this in the source for each such file. */
+
+#define GETOPT_INTERFACE_VERSION 2
+#if !defined _LIBC && defined __GLIBC__ && __GLIBC__ >= 2
+# include <gnu-versions.h>
+# if _GNU_GETOPT_INTERFACE_VERSION == GETOPT_INTERFACE_VERSION
+# define ELIDE_CODE
+# endif
+#endif
+
+#ifndef ELIDE_CODE
+
+
+/* This needs to come after some library #include
+ to get __GNU_LIBRARY__ defined. */
+#ifdef __GNU_LIBRARY__
+/* Don't include stdlib.h for non-GNU C libraries because some of them
+ contain conflicting prototypes for getopt. */
+# include <stdlib.h>
+# include <unistd.h>
+#endif /* GNU C library. */
+
+#include <string.h>
+
+#ifdef VMS
+# include <unixlib.h>
+#endif
+
+#ifdef _LIBC
+# include <libintl.h>
+#else
+# include "gettext.h"
+# define _(msgid) gettext (msgid)
+#endif
+
+#if defined _LIBC && defined USE_IN_LIBIO
+# include <wchar.h>
+#endif
+
+#ifndef attribute_hidden
+# define attribute_hidden
+#endif
+
+/* This version of `getopt' appears to the caller like standard Unix `getopt'
+ but it behaves differently for the user, since it allows the user
+ to intersperse the options with the other arguments.
+
+ As `getopt' works, it permutes the elements of ARGV so that,
+ when it is done, all the options precede everything else. Thus
+ all application programs are extended to handle flexible argument order.
+
+ Setting the environment variable POSIXLY_CORRECT disables permutation.
+ Then the behavior is completely standard.
+
+ GNU application programs can use a third alternative mode in which
+ they can distinguish the relative order of options and other arguments. */
+
+#include "getopt.h"
+#include "getopt_int.h"
+
+/* For communication from `getopt' to the caller.
+ When `getopt' finds an option that takes an argument,
+ the argument value is returned here.
+ Also, when `ordering' is RETURN_IN_ORDER,
+ each non-option ARGV-element is returned here. */
+
+char *optarg;
+
+/* Index in ARGV of the next element to be scanned.
+ This is used for communication to and from the caller
+ and for communication between successive calls to `getopt'.
+
+ On entry to `getopt', zero means this is the first call; initialize.
+
+ When `getopt' returns -1, this is the index of the first of the
+ non-option elements that the caller should itself scan.
+
+ Otherwise, `optind' communicates from one call to the next
+ how much of ARGV has been scanned so far. */
+
+/* 1003.2 says this must be 1 before any call. */
+int optind = 1;
+
+/* Callers store zero here to inhibit the error message
+ for unrecognized options. */
+
+int opterr = 1;
+
+/* Set to an option character which was unrecognized.
+ This must be initialized on some systems to avoid linking in the
+ system's own getopt implementation. */
+
+int optopt = '?';
+
+/* Keep a global copy of all internal members of getopt_data. */
+
+static struct _getopt_data getopt_data;
+
+
+#ifndef __GNU_LIBRARY__
+
+/* Avoid depending on library functions or files
+ whose names are inconsistent. */
+
+#ifndef getenv
+extern char *getenv ();
+#endif
+
+#endif /* not __GNU_LIBRARY__ */
+
+#ifdef _LIBC
+/* Stored original parameters.
+ XXX This is no good solution. We should rather copy the args so
+ that we can compare them later. But we must not use malloc(3). */
+extern int __libc_argc;
+extern char **__libc_argv;
+
+/* Bash 2.0 gives us an environment variable containing flags
+ indicating ARGV elements that should not be considered arguments. */
+
+# ifdef USE_NONOPTION_FLAGS
+/* Defined in getopt_init.c */
+extern char *__getopt_nonoption_flags;
+# endif
+
+# ifdef USE_NONOPTION_FLAGS
+# define SWAP_FLAGS(ch1, ch2) \
+ if (d->__nonoption_flags_len > 0) \
+ { \
+ char __tmp = __getopt_nonoption_flags[ch1]; \
+ __getopt_nonoption_flags[ch1] = __getopt_nonoption_flags[ch2]; \
+ __getopt_nonoption_flags[ch2] = __tmp; \
+ }
+# else
+# define SWAP_FLAGS(ch1, ch2)
+# endif
+#else /* !_LIBC */
+# define SWAP_FLAGS(ch1, ch2)
+#endif /* _LIBC */
+
+/* Exchange two adjacent subsequences of ARGV.
+ One subsequence is elements [first_nonopt,last_nonopt)
+ which contains all the non-options that have been skipped so far.
+ The other is elements [last_nonopt,optind), which contains all
+ the options processed since those non-options were skipped.
+
+ `first_nonopt' and `last_nonopt' are relocated so that they describe
+ the new indices of the non-options in ARGV after they are moved. */
+
+static void
+exchange (char **argv, struct _getopt_data *d)
+{
+ int bottom = d->__first_nonopt;
+ int middle = d->__last_nonopt;
+ int top = d->optind;
+ char *tem;
+
+ /* Exchange the shorter segment with the far end of the longer segment.
+ That puts the shorter segment into the right place.
+ It leaves the longer segment in the right place overall,
+ but it consists of two parts that need to be swapped next. */
+
+#if defined _LIBC && defined USE_NONOPTION_FLAGS
+ /* First make sure the handling of the `__getopt_nonoption_flags'
+ string can work normally. Our top argument must be in the range
+ of the string. */
+ if (d->__nonoption_flags_len > 0 && top >= d->__nonoption_flags_max_len)
+ {
+ /* We must extend the array. The user plays games with us and
+ presents new arguments. */
+ char *new_str = malloc (top + 1);
+ if (new_str == NULL)
+ d->__nonoption_flags_len = d->__nonoption_flags_max_len = 0;
+ else
+ {
+ memset (__mempcpy (new_str, __getopt_nonoption_flags,
+ d->__nonoption_flags_max_len),
+ '\0', top + 1 - d->__nonoption_flags_max_len);
+ d->__nonoption_flags_max_len = top + 1;
+ __getopt_nonoption_flags = new_str;
+ }
+ }
+#endif
+
+ while (top > middle && middle > bottom)
+ {
+ if (top - middle > middle - bottom)
+ {
+ /* Bottom segment is the short one. */
+ int len = middle - bottom;
+ register int i;
+
+ /* Swap it with the top part of the top segment. */
+ for (i = 0; i < len; i++)
+ {
+ tem = argv[bottom + i];
+ argv[bottom + i] = argv[top - (middle - bottom) + i];
+ argv[top - (middle - bottom) + i] = tem;
+ SWAP_FLAGS (bottom + i, top - (middle - bottom) + i);
+ }
+ /* Exclude the moved bottom segment from further swapping. */
+ top -= len;
+ }
+ else
+ {
+ /* Top segment is the short one. */
+ int len = top - middle;
+ register int i;
+
+ /* Swap it with the bottom part of the bottom segment. */
+ for (i = 0; i < len; i++)
+ {
+ tem = argv[bottom + i];
+ argv[bottom + i] = argv[middle + i];
+ argv[middle + i] = tem;
+ SWAP_FLAGS (bottom + i, middle + i);
+ }
+ /* Exclude the moved top segment from further swapping. */
+ bottom += len;
+ }
+ }
+
+ /* Update records for the slots the non-options now occupy. */
+
+ d->__first_nonopt += (d->optind - d->__last_nonopt);
+ d->__last_nonopt = d->optind;
+}
+
+/* Initialize the internal data when the first call is made. */
+
+static const char *
+_getopt_initialize (int argc, char *const *argv, const char *optstring,
+ struct _getopt_data *d)
+{
+ /* Start processing options with ARGV-element 1 (since ARGV-element 0
+ is the program name); the sequence of previously skipped
+ non-option ARGV-elements is empty. */
+
+ d->__first_nonopt = d->__last_nonopt = d->optind;
+
+ d->__nextchar = NULL;
+
+ d->__posixly_correct = !!getenv ("POSIXLY_CORRECT");
+
+ /* Determine how to handle the ordering of options and nonoptions. */
+
+ if (optstring[0] == '-')
+ {
+ d->__ordering = RETURN_IN_ORDER;
+ ++optstring;
+ }
+ else if (optstring[0] == '+')
+ {
+ d->__ordering = REQUIRE_ORDER;
+ ++optstring;
+ }
+ else if (d->__posixly_correct)
+ d->__ordering = REQUIRE_ORDER;
+ else
+ d->__ordering = PERMUTE;
+
+#if defined _LIBC && defined USE_NONOPTION_FLAGS
+ if (!d->__posixly_correct
+ && argc == __libc_argc && argv == __libc_argv)
+ {
+ if (d->__nonoption_flags_max_len == 0)
+ {
+ if (__getopt_nonoption_flags == NULL
+ || __getopt_nonoption_flags[0] == '\0')
+ d->__nonoption_flags_max_len = -1;
+ else
+ {
+ const char *orig_str = __getopt_nonoption_flags;
+ int len = d->__nonoption_flags_max_len = strlen (orig_str);
+ if (d->__nonoption_flags_max_len < argc)
+ d->__nonoption_flags_max_len = argc;
+ __getopt_nonoption_flags =
+ (char *) malloc (d->__nonoption_flags_max_len);
+ if (__getopt_nonoption_flags == NULL)
+ d->__nonoption_flags_max_len = -1;
+ else
+ memset (__mempcpy (__getopt_nonoption_flags, orig_str, len),
+ '\0', d->__nonoption_flags_max_len - len);
+ }
+ }
+ d->__nonoption_flags_len = d->__nonoption_flags_max_len;
+ }
+ else
+ d->__nonoption_flags_len = 0;
+#endif
+
+ return optstring;
+}
+
+/* Scan elements of ARGV (whose length is ARGC) for option characters
+ given in OPTSTRING.
+
+ If an element of ARGV starts with '-', and is not exactly "-" or "--",
+ then it is an option element. The characters of this element
+ (aside from the initial '-') are option characters. If `getopt'
+ is called repeatedly, it returns successively each of the option characters
+ from each of the option elements.
+
+ If `getopt' finds another option character, it returns that character,
+ updating `optind' and `nextchar' so that the next call to `getopt' can
+ resume the scan with the following option character or ARGV-element.
+
+ If there are no more option characters, `getopt' returns -1.
+ Then `optind' is the index in ARGV of the first ARGV-element
+ that is not an option. (The ARGV-elements have been permuted
+ so that those that are not options now come last.)
+
+ OPTSTRING is a string containing the legitimate option characters.
+ If an option character is seen that is not listed in OPTSTRING,
+ return '?' after printing an error message. If you set `opterr' to
+ zero, the error message is suppressed but we still return '?'.
+
+ If a char in OPTSTRING is followed by a colon, that means it wants an arg,
+ so the following text in the same ARGV-element, or the text of the following
+ ARGV-element, is returned in `optarg'. Two colons mean an option that
+ wants an optional arg; if there is text in the current ARGV-element,
+ it is returned in `optarg', otherwise `optarg' is set to zero.
+
+ If OPTSTRING starts with `-' or `+', it requests different methods of
+ handling the non-option ARGV-elements.
+ See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above.
+
+ Long-named options begin with `--' instead of `-'.
+ Their names may be abbreviated as long as the abbreviation is unique
+ or is an exact match for some defined option. If they have an
+ argument, it follows the option name in the same ARGV-element, separated
+ from the option name by a `=', or else the in next ARGV-element.
+ When `getopt' finds a long-named option, it returns 0 if that option's
+ `flag' field is nonzero, the value of the option's `val' field
+ if the `flag' field is zero.
+
+ The elements of ARGV aren't really const, because we permute them.
+ But we pretend they're const in the prototype to be compatible
+ with other systems.
+
+ LONGOPTS is a vector of `struct option' terminated by an
+ element containing a name which is zero.
+
+ LONGIND returns the index in LONGOPT of the long-named option found.
+ It is only valid when a long-named option has been found by the most
+ recent call.
+
+ If LONG_ONLY is nonzero, '-' as well as '--' can introduce
+ long-named options. */
+
+int
+_getopt_internal_r (int argc, char *const *argv, const char *optstring,
+ const struct option *longopts, int *longind,
+ int long_only, struct _getopt_data *d)
+{
+ int print_errors = d->opterr;
+ if (optstring[0] == ':')
+ print_errors = 0;
+
+ if (argc < 1)
+ return -1;
+
+ d->optarg = NULL;
+
+ if (d->optind == 0 || !d->__initialized)
+ {
+ if (d->optind == 0)
+ d->optind = 1; /* Don't scan ARGV[0], the program name. */
+ optstring = _getopt_initialize (argc, argv, optstring, d);
+ d->__initialized = 1;
+ }
+
+ /* Test whether ARGV[optind] points to a non-option argument.
+ Either it does not have option syntax, or there is an environment flag
+ from the shell indicating it is not an option. The later information
+ is only used when the used in the GNU libc. */
+#if defined _LIBC && defined USE_NONOPTION_FLAGS
+# define NONOPTION_P (argv[d->optind][0] != '-' || argv[d->optind][1] == '\0' \
+ || (d->optind < d->__nonoption_flags_len \
+ && __getopt_nonoption_flags[d->optind] == '1'))
+#else
+# define NONOPTION_P (argv[d->optind][0] != '-' || argv[d->optind][1] == '\0')
+#endif
+
+ if (d->__nextchar == NULL || *d->__nextchar == '\0')
+ {
+ /* Advance to the next ARGV-element. */
+
+ /* Give FIRST_NONOPT & LAST_NONOPT rational values if OPTIND has been
+ moved back by the user (who may also have changed the arguments). */
+ if (d->__last_nonopt > d->optind)
+ d->__last_nonopt = d->optind;
+ if (d->__first_nonopt > d->optind)
+ d->__first_nonopt = d->optind;
+
+ if (d->__ordering == PERMUTE)
+ {
+ /* If we have just processed some options following some non-options,
+ exchange them so that the options come first. */
+
+ if (d->__first_nonopt != d->__last_nonopt
+ && d->__last_nonopt != d->optind)
+ exchange ((char **) argv, d);
+ else if (d->__last_nonopt != d->optind)
+ d->__first_nonopt = d->optind;
+
+ /* Skip any additional non-options
+ and extend the range of non-options previously skipped. */
+
+ while (d->optind < argc && NONOPTION_P)
+ d->optind++;
+ d->__last_nonopt = d->optind;
+ }
+
+ /* The special ARGV-element `--' means premature end of options.
+ Skip it like a null option,
+ then exchange with previous non-options as if it were an option,
+ then skip everything else like a non-option. */
+
+ if (d->optind != argc && !strcmp (argv[d->optind], "--"))
+ {
+ d->optind++;
+
+ if (d->__first_nonopt != d->__last_nonopt
+ && d->__last_nonopt != d->optind)
+ exchange ((char **) argv, d);
+ else if (d->__first_nonopt == d->__last_nonopt)
+ d->__first_nonopt = d->optind;
+ d->__last_nonopt = argc;
+
+ d->optind = argc;
+ }
+
+ /* If we have done all the ARGV-elements, stop the scan
+ and back over any non-options that we skipped and permuted. */
+
+ if (d->optind == argc)
+ {
+ /* Set the next-arg-index to point at the non-options
+ that we previously skipped, so the caller will digest them. */
+ if (d->__first_nonopt != d->__last_nonopt)
+ d->optind = d->__first_nonopt;
+ return -1;
+ }
+
+ /* If we have come to a non-option and did not permute it,
+ either stop the scan or describe it to the caller and pass it by. */
+
+ if (NONOPTION_P)
+ {
+ if (d->__ordering == REQUIRE_ORDER)
+ return -1;
+ d->optarg = argv[d->optind++];
+ return 1;
+ }
+
+ /* We have found another option-ARGV-element.
+ Skip the initial punctuation. */
+
+ d->__nextchar = (argv[d->optind] + 1
+ + (longopts != NULL && argv[d->optind][1] == '-'));
+ }
+
+ /* Decode the current option-ARGV-element. */
+
+ /* Check whether the ARGV-element is a long option.
+
+ If long_only and the ARGV-element has the form "-f", where f is
+ a valid short option, don't consider it an abbreviated form of
+ a long option that starts with f. Otherwise there would be no
+ way to give the -f short option.
+
+ On the other hand, if there's a long option "fubar" and
+ the ARGV-element is "-fu", do consider that an abbreviation of
+ the long option, just like "--fu", and not "-f" with arg "u".
+
+ This distinction seems to be the most useful approach. */
+
+ if (longopts != NULL
+ && (argv[d->optind][1] == '-'
+ || (long_only && (argv[d->optind][2]
+ || !strchr (optstring, argv[d->optind][1])))))
+ {
+ char *nameend;
+ const struct option *p;
+ const struct option *pfound = NULL;
+ int exact = 0;
+ int ambig = 0;
+ int indfound = -1;
+ int option_index;
+
+ for (nameend = d->__nextchar; *nameend && *nameend != '='; nameend++)
+ /* Do nothing. */ ;
+
+ /* Test all long options for either exact match
+ or abbreviated matches. */
+ for (p = longopts, option_index = 0; p->name; p++, option_index++)
+ if (!strncmp (p->name, d->__nextchar, nameend - d->__nextchar))
+ {
+ if ((unsigned int) (nameend - d->__nextchar)
+ == (unsigned int) strlen (p->name))
+ {
+ /* Exact match found. */
+ pfound = p;
+ indfound = option_index;
+ exact = 1;
+ break;
+ }
+ else if (pfound == NULL)
+ {
+ /* First nonexact match found. */
+ pfound = p;
+ indfound = option_index;
+ }
+ else if (long_only
+ || pfound->has_arg != p->has_arg
+ || pfound->flag != p->flag
+ || pfound->val != p->val)
+ /* Second or later nonexact match found. */
+ ambig = 1;
+ }
+
+ if (ambig && !exact)
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf, _("%s: option `%s' is ambiguous\n"),
+ argv[0], argv[d->optind]) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2 |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr, _("%s: option `%s' is ambiguous\n"),
+ argv[0], argv[d->optind]);
+#endif
+ }
+ d->__nextchar += strlen (d->__nextchar);
+ d->optind++;
+ d->optopt = 0;
+ return '?';
+ }
+
+ if (pfound != NULL)
+ {
+ option_index = indfound;
+ d->optind++;
+ if (*nameend)
+ {
+ /* Don't test has_arg with >, because some C compilers don't
+ allow it to be used on enums. */
+ if (pfound->has_arg)
+ d->optarg = nameend + 1;
+ else
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+ int n;
+#endif
+
+ if (argv[d->optind - 1][1] == '-')
+ {
+ /* --option */
+#if defined _LIBC && defined USE_IN_LIBIO
+ n = __asprintf (&buf, _("\
+%s: option `--%s' doesn't allow an argument\n"),
+ argv[0], pfound->name);
+#else
+ fprintf (stderr, _("\
+%s: option `--%s' doesn't allow an argument\n"),
+ argv[0], pfound->name);
+#endif
+ }
+ else
+ {
+ /* +option or -option */
+#if defined _LIBC && defined USE_IN_LIBIO
+ n = __asprintf (&buf, _("\
+%s: option `%c%s' doesn't allow an argument\n"),
+ argv[0], argv[d->optind - 1][0],
+ pfound->name);
+#else
+ fprintf (stderr, _("\
+%s: option `%c%s' doesn't allow an argument\n"),
+ argv[0], argv[d->optind - 1][0],
+ pfound->name);
+#endif
+ }
+
+#if defined _LIBC && defined USE_IN_LIBIO
+ if (n >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2
+ |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#endif
+ }
+
+ d->__nextchar += strlen (d->__nextchar);
+
+ d->optopt = pfound->val;
+ return '?';
+ }
+ }
+ else if (pfound->has_arg == 1)
+ {
+ if (d->optind < argc)
+ d->optarg = argv[d->optind++];
+ else
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf, _("\
+%s: option `%s' requires an argument\n"),
+ argv[0], argv[d->optind - 1]) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2
+ |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr,
+ _("%s: option `%s' requires an argument\n"),
+ argv[0], argv[d->optind - 1]);
+#endif
+ }
+ d->__nextchar += strlen (d->__nextchar);
+ d->optopt = pfound->val;
+ return optstring[0] == ':' ? ':' : '?';
+ }
+ }
+ d->__nextchar += strlen (d->__nextchar);
+ if (longind != NULL)
+ *longind = option_index;
+ if (pfound->flag)
+ {
+ *(pfound->flag) = pfound->val;
+ return 0;
+ }
+ return pfound->val;
+ }
+
+ /* Can't find it as a long option. If this is not getopt_long_only,
+ or the option starts with '--' or is not a valid short
+ option, then it's an error.
+ Otherwise interpret it as a short option. */
+ if (!long_only || argv[d->optind][1] == '-'
+ || strchr (optstring, *d->__nextchar) == NULL)
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+ int n;
+#endif
+
+ if (argv[d->optind][1] == '-')
+ {
+ /* --option */
+#if defined _LIBC && defined USE_IN_LIBIO
+ n = __asprintf (&buf, _("%s: unrecognized option `--%s'\n"),
+ argv[0], d->__nextchar);
+#else
+ fprintf (stderr, _("%s: unrecognized option `--%s'\n"),
+ argv[0], d->__nextchar);
+#endif
+ }
+ else
+ {
+ /* +option or -option */
+#if defined _LIBC && defined USE_IN_LIBIO
+ n = __asprintf (&buf, _("%s: unrecognized option `%c%s'\n"),
+ argv[0], argv[d->optind][0], d->__nextchar);
+#else
+ fprintf (stderr, _("%s: unrecognized option `%c%s'\n"),
+ argv[0], argv[d->optind][0], d->__nextchar);
+#endif
+ }
+
+#if defined _LIBC && defined USE_IN_LIBIO
+ if (n >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2 |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#endif
+ }
+ d->__nextchar = (char *) "";
+ d->optind++;
+ d->optopt = 0;
+ return '?';
+ }
+ }
+
+ /* Look at and handle the next short option-character. */
+
+ {
+ char c = *d->__nextchar++;
+ char *temp = strchr (optstring, c);
+
+ /* Increment `optind' when we start to process its last character. */
+ if (*d->__nextchar == '\0')
+ ++d->optind;
+
+ if (temp == NULL || c == ':')
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+ int n;
+#endif
+
+ if (d->__posixly_correct)
+ {
+ /* 1003.2 specifies the format of this message. */
+#if defined _LIBC && defined USE_IN_LIBIO
+ n = __asprintf (&buf, _("%s: illegal option -- %c\n"),
+ argv[0], c);
+#else
+ fprintf (stderr, _("%s: illegal option -- %c\n"), argv[0], c);
+#endif
+ }
+ else
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ n = __asprintf (&buf, _("%s: invalid option -- %c\n"),
+ argv[0], c);
+#else
+ fprintf (stderr, _("%s: invalid option -- %c\n"), argv[0], c);
+#endif
+ }
+
+#if defined _LIBC && defined USE_IN_LIBIO
+ if (n >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2 |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#endif
+ }
+ d->optopt = c;
+ return '?';
+ }
+ /* Convenience. Treat POSIX -W foo same as long option --foo */
+ if (temp[0] == 'W' && temp[1] == ';')
+ {
+ char *nameend;
+ const struct option *p;
+ const struct option *pfound = NULL;
+ int exact = 0;
+ int ambig = 0;
+ int indfound = 0;
+ int option_index;
+
+ /* This is an option that requires an argument. */
+ if (*d->__nextchar != '\0')
+ {
+ d->optarg = d->__nextchar;
+ /* If we end this ARGV-element by taking the rest as an arg,
+ we must advance to the next element now. */
+ d->optind++;
+ }
+ else if (d->optind == argc)
+ {
+ if (print_errors)
+ {
+ /* 1003.2 specifies the format of this message. */
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf,
+ _("%s: option requires an argument -- %c\n"),
+ argv[0], c) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2 |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr, _("%s: option requires an argument -- %c\n"),
+ argv[0], c);
+#endif
+ }
+ d->optopt = c;
+ if (optstring[0] == ':')
+ c = ':';
+ else
+ c = '?';
+ return c;
+ }
+ else
+ /* We already incremented `d->optind' once;
+ increment it again when taking next ARGV-elt as argument. */
+ d->optarg = argv[d->optind++];
+
+ /* optarg is now the argument, see if it's in the
+ table of longopts. */
+
+ for (d->__nextchar = nameend = d->optarg; *nameend && *nameend != '=';
+ nameend++)
+ /* Do nothing. */ ;
+
+ /* Test all long options for either exact match
+ or abbreviated matches. */
+ for (p = longopts, option_index = 0; p->name; p++, option_index++)
+ if (!strncmp (p->name, d->__nextchar, nameend - d->__nextchar))
+ {
+ if ((unsigned int) (nameend - d->__nextchar) == strlen (p->name))
+ {
+ /* Exact match found. */
+ pfound = p;
+ indfound = option_index;
+ exact = 1;
+ break;
+ }
+ else if (pfound == NULL)
+ {
+ /* First nonexact match found. */
+ pfound = p;
+ indfound = option_index;
+ }
+ else
+ /* Second or later nonexact match found. */
+ ambig = 1;
+ }
+ if (ambig && !exact)
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf, _("%s: option `-W %s' is ambiguous\n"),
+ argv[0], argv[d->optind]) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2 |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr, _("%s: option `-W %s' is ambiguous\n"),
+ argv[0], argv[d->optind]);
+#endif
+ }
+ d->__nextchar += strlen (d->__nextchar);
+ d->optind++;
+ return '?';
+ }
+ if (pfound != NULL)
+ {
+ option_index = indfound;
+ if (*nameend)
+ {
+ /* Don't test has_arg with >, because some C compilers don't
+ allow it to be used on enums. */
+ if (pfound->has_arg)
+ d->optarg = nameend + 1;
+ else
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf, _("\
+%s: option `-W %s' doesn't allow an argument\n"),
+ argv[0], pfound->name) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2
+ |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr, _("\
+%s: option `-W %s' doesn't allow an argument\n"),
+ argv[0], pfound->name);
+#endif
+ }
+
+ d->__nextchar += strlen (d->__nextchar);
+ return '?';
+ }
+ }
+ else if (pfound->has_arg == 1)
+ {
+ if (d->optind < argc)
+ d->optarg = argv[d->optind++];
+ else
+ {
+ if (print_errors)
+ {
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf, _("\
+%s: option `%s' requires an argument\n"),
+ argv[0], argv[d->optind - 1]) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2
+ |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr,
+ _("%s: option `%s' requires an argument\n"),
+ argv[0], argv[d->optind - 1]);
+#endif
+ }
+ d->__nextchar += strlen (d->__nextchar);
+ return optstring[0] == ':' ? ':' : '?';
+ }
+ }
+ d->__nextchar += strlen (d->__nextchar);
+ if (longind != NULL)
+ *longind = option_index;
+ if (pfound->flag)
+ {
+ *(pfound->flag) = pfound->val;
+ return 0;
+ }
+ return pfound->val;
+ }
+ d->__nextchar = NULL;
+ return 'W'; /* Let the application handle it. */
+ }
+ if (temp[1] == ':')
+ {
+ if (temp[2] == ':')
+ {
+ /* This is an option that accepts an argument optionally. */
+ if (*d->__nextchar != '\0')
+ {
+ d->optarg = d->__nextchar;
+ d->optind++;
+ }
+ else
+ d->optarg = NULL;
+ d->__nextchar = NULL;
+ }
+ else
+ {
+ /* This is an option that requires an argument. */
+ if (*d->__nextchar != '\0')
+ {
+ d->optarg = d->__nextchar;
+ /* If we end this ARGV-element by taking the rest as an arg,
+ we must advance to the next element now. */
+ d->optind++;
+ }
+ else if (d->optind == argc)
+ {
+ if (print_errors)
+ {
+ /* 1003.2 specifies the format of this message. */
+#if defined _LIBC && defined USE_IN_LIBIO
+ char *buf;
+
+ if (__asprintf (&buf, _("\
+%s: option requires an argument -- %c\n"),
+ argv[0], c) >= 0)
+ {
+ _IO_flockfile (stderr);
+
+ int old_flags2 = ((_IO_FILE *) stderr)->_flags2;
+ ((_IO_FILE *) stderr)->_flags2 |= _IO_FLAGS2_NOTCANCEL;
+
+ if (_IO_fwide (stderr, 0) > 0)
+ __fwprintf (stderr, L"%s", buf);
+ else
+ fputs (buf, stderr);
+
+ ((_IO_FILE *) stderr)->_flags2 = old_flags2;
+ _IO_funlockfile (stderr);
+
+ free (buf);
+ }
+#else
+ fprintf (stderr,
+ _("%s: option requires an argument -- %c\n"),
+ argv[0], c);
+#endif
+ }
+ d->optopt = c;
+ if (optstring[0] == ':')
+ c = ':';
+ else
+ c = '?';
+ }
+ else
+ /* We already incremented `optind' once;
+ increment it again when taking next ARGV-elt as argument. */
+ d->optarg = argv[d->optind++];
+ d->__nextchar = NULL;
+ }
+ }
+ return c;
+ }
+}
+
+int
+_getopt_internal (int argc, char *const *argv, const char *optstring,
+ const struct option *longopts, int *longind, int long_only)
+{
+ int result;
+
+ getopt_data.optind = optind;
+ getopt_data.opterr = opterr;
+
+ result = _getopt_internal_r (argc, argv, optstring, longopts,
+ longind, long_only, &getopt_data);
+
+ optind = getopt_data.optind;
+ optarg = getopt_data.optarg;
+ optopt = getopt_data.optopt;
+
+ return result;
+}
+
+int
+getopt (int argc, char *const *argv, const char *optstring)
+{
+ return _getopt_internal (argc, argv, optstring,
+ (const struct option *) 0,
+ (int *) 0,
+ 0);
+}
+
+#endif /* Not ELIDE_CODE. */
+
+#ifdef TEST
+
+/* Compile with -DTEST to make an executable for use in testing
+ the above definition of `getopt'. */
+
+int
+main (int argc, char **argv)
+{
+ int c;
+ int digit_optind = 0;
+
+ while (1)
+ {
+ int this_option_optind = optind ? optind : 1;
+
+ c = getopt (argc, argv, "abc:d:0123456789");
+ if (c == -1)
+ break;
+
+ switch (c)
+ {
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9':
+ if (digit_optind != 0 && digit_optind != this_option_optind)
+ printf ("digits occur in two different argv-elements.\n");
+ digit_optind = this_option_optind;
+ printf ("option %c\n", c);
+ break;
+
+ case 'a':
+ printf ("option a\n");
+ break;
+
+ case 'b':
+ printf ("option b\n");
+ break;
+
+ case 'c':
+ printf ("option c with value `%s'\n", optarg);
+ break;
+
+ case '?':
+ break;
+
+ default:
+ printf ("?? getopt returned character code 0%o ??\n", c);
+ }
+ }
+
+ if (optind < argc)
+ {
+ printf ("non-option ARGV-elements: ");
+ while (optind < argc)
+ printf ("%s ", argv[optind++]);
+ printf ("\n");
+ }
+
+ exit (0);
+}
+
+#endif /* TEST */
diff --git a/contrib/diff/lib/getopt.h b/contrib/diff/lib/getopt.h
new file mode 100644
index 0000000..c61768c
--- /dev/null
+++ b/contrib/diff/lib/getopt.h
@@ -0,0 +1,176 @@
+/* Declarations for getopt.
+ Copyright (C) 1989-1994,1996-1999,2001,2003,2004
+ Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _GETOPT_H
+
+#ifndef __need_getopt
+# define _GETOPT_H 1
+#endif
+
+/* If __GNU_LIBRARY__ is not already defined, either we are being used
+ standalone, or this is the first header included in the source file.
+ If we are being used with glibc, we need to include <features.h>, but
+ that does not exist if we are standalone. So: if __GNU_LIBRARY__ is
+ not defined, include <ctype.h>, which will pull in <features.h> for us
+ if it's from glibc. (Why ctype.h? It's guaranteed to exist and it
+ doesn't flood the namespace with stuff the way some other headers do.) */
+#if !defined __GNU_LIBRARY__
+# include <ctype.h>
+#endif
+
+#ifndef __THROW
+# ifndef __GNUC_PREREQ
+# define __GNUC_PREREQ(maj, min) (0)
+# endif
+# if defined __cplusplus && __GNUC_PREREQ (2,8)
+# define __THROW throw ()
+# else
+# define __THROW
+# endif
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* For communication from `getopt' to the caller.
+ When `getopt' finds an option that takes an argument,
+ the argument value is returned here.
+ Also, when `ordering' is RETURN_IN_ORDER,
+ each non-option ARGV-element is returned here. */
+
+extern char *optarg;
+
+/* Index in ARGV of the next element to be scanned.
+ This is used for communication to and from the caller
+ and for communication between successive calls to `getopt'.
+
+ On entry to `getopt', zero means this is the first call; initialize.
+
+ When `getopt' returns -1, this is the index of the first of the
+ non-option elements that the caller should itself scan.
+
+ Otherwise, `optind' communicates from one call to the next
+ how much of ARGV has been scanned so far. */
+
+extern int optind;
+
+/* Callers store zero here to inhibit the error message `getopt' prints
+ for unrecognized options. */
+
+extern int opterr;
+
+/* Set to an option character which was unrecognized. */
+
+extern int optopt;
+
+#ifndef __need_getopt
+/* Describe the long-named options requested by the application.
+ The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector
+ of `struct option' terminated by an element containing a name which is
+ zero.
+
+ The field `has_arg' is:
+ no_argument (or 0) if the option does not take an argument,
+ required_argument (or 1) if the option requires an argument,
+ optional_argument (or 2) if the option takes an optional argument.
+
+ If the field `flag' is not NULL, it points to a variable that is set
+ to the value given in the field `val' when the option is found, but
+ left unchanged if the option is not found.
+
+ To have a long-named option do something other than set an `int' to
+ a compiled-in constant, such as set a value from `optarg', set the
+ option's `flag' field to zero and its `val' field to a nonzero
+ value (the equivalent single-letter option character, if there is
+ one). For long options that have a zero `flag' field, `getopt'
+ returns the contents of the `val' field. */
+
+struct option
+{
+ const char *name;
+ /* has_arg can't be an enum because some compilers complain about
+ type mismatches in all the code that assumes it is an int. */
+ int has_arg;
+ int *flag;
+ int val;
+};
+
+/* Names for the values of the `has_arg' field of `struct option'. */
+
+# define no_argument 0
+# define required_argument 1
+# define optional_argument 2
+#endif /* need getopt */
+
+
+/* Get definitions and prototypes for functions to process the
+ arguments in ARGV (ARGC of them, minus the program name) for
+ options given in OPTS.
+
+ Return the option character from OPTS just read. Return -1 when
+ there are no more options. For unrecognized options, or options
+ missing arguments, `optopt' is set to the option letter, and '?' is
+ returned.
+
+ The OPTS string is a list of characters which are recognized option
+ letters, optionally followed by colons, specifying that that letter
+ takes an argument, to be placed in `optarg'.
+
+ If a letter in OPTS is followed by two colons, its argument is
+ optional. This behavior is specific to the GNU `getopt'.
+
+ The argument `--' causes premature termination of argument
+ scanning, explicitly telling `getopt' that there are no more
+ options.
+
+ If OPTS begins with `--', then non-option arguments are treated as
+ arguments to the option '\0'. This behavior is specific to the GNU
+ `getopt'. */
+
+#ifdef __GNU_LIBRARY__
+/* Many other libraries have conflicting prototypes for getopt, with
+ differences in the consts, in stdlib.h. To avoid compilation
+ errors, only prototype getopt for the GNU C library. */
+extern int getopt (int ___argc, char *const *___argv, const char *__shortopts)
+ __THROW;
+#else /* not __GNU_LIBRARY__ */
+extern int getopt ();
+#endif /* __GNU_LIBRARY__ */
+
+#ifndef __need_getopt
+extern int getopt_long (int ___argc, char *const *___argv,
+ const char *__shortopts,
+ const struct option *__longopts, int *__longind)
+ __THROW;
+extern int getopt_long_only (int ___argc, char *const *___argv,
+ const char *__shortopts,
+ const struct option *__longopts, int *__longind)
+ __THROW;
+
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+/* Make sure we later can get all the definitions and declarations. */
+#undef __need_getopt
+
+#endif /* getopt.h */
diff --git a/contrib/diff/lib/getopt1.c b/contrib/diff/lib/getopt1.c
new file mode 100644
index 0000000..7671eba
--- /dev/null
+++ b/contrib/diff/lib/getopt1.c
@@ -0,0 +1,191 @@
+/* getopt_long and getopt_long_only entry points for GNU getopt.
+ Copyright (C) 1987,88,89,90,91,92,93,94,96,97,98,2004
+ Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifdef HAVE_CONFIG_H
+#include <config.h>
+#endif
+
+#ifdef _LIBC
+# include <getopt.h>
+#else
+# include "getopt.h"
+#endif
+#include "getopt_int.h"
+
+#include <stdio.h>
+
+/* Comment out all this code if we are using the GNU C Library, and are not
+ actually compiling the library itself. This code is part of the GNU C
+ Library, but also included in many other GNU distributions. Compiling
+ and linking in this code is a waste when using the GNU C library
+ (especially if it is a shared library). Rather than having every GNU
+ program understand `configure --with-gnu-libc' and omit the object files,
+ it is simpler to just do this in the source for each such file. */
+
+#define GETOPT_INTERFACE_VERSION 2
+#if !defined _LIBC && defined __GLIBC__ && __GLIBC__ >= 2
+#include <gnu-versions.h>
+#if _GNU_GETOPT_INTERFACE_VERSION == GETOPT_INTERFACE_VERSION
+#define ELIDE_CODE
+#endif
+#endif
+
+#ifndef ELIDE_CODE
+
+
+/* This needs to come after some library #include
+ to get __GNU_LIBRARY__ defined. */
+#ifdef __GNU_LIBRARY__
+#include <stdlib.h>
+#endif
+
+#ifndef NULL
+#define NULL 0
+#endif
+
+int
+getopt_long (int argc, char *const *argv, const char *options,
+ const struct option *long_options, int *opt_index)
+{
+ return _getopt_internal (argc, argv, options, long_options, opt_index, 0);
+}
+
+int
+_getopt_long_r (int argc, char *const *argv, const char *options,
+ const struct option *long_options, int *opt_index,
+ struct _getopt_data *d)
+{
+ return _getopt_internal_r (argc, argv, options, long_options, opt_index,
+ 0, d);
+}
+
+/* Like getopt_long, but '-' as well as '--' can indicate a long option.
+ If an option that starts with '-' (not '--') doesn't match a long option,
+ but does match a short option, it is parsed as a short option
+ instead. */
+
+int
+getopt_long_only (int argc, char *const *argv, const char *options,
+ const struct option *long_options, int *opt_index)
+{
+ return _getopt_internal (argc, argv, options, long_options, opt_index, 1);
+}
+
+int
+_getopt_long_only_r (int argc, char *const *argv, const char *options,
+ const struct option *long_options, int *opt_index,
+ struct _getopt_data *d)
+{
+ return _getopt_internal_r (argc, argv, options, long_options, opt_index,
+ 1, d);
+}
+
+#endif /* Not ELIDE_CODE. */
+
+#ifdef TEST
+
+#include <stdio.h>
+
+int
+main (int argc, char **argv)
+{
+ int c;
+ int digit_optind = 0;
+
+ while (1)
+ {
+ int this_option_optind = optind ? optind : 1;
+ int option_index = 0;
+ static struct option long_options[] =
+ {
+ {"add", 1, 0, 0},
+ {"append", 0, 0, 0},
+ {"delete", 1, 0, 0},
+ {"verbose", 0, 0, 0},
+ {"create", 0, 0, 0},
+ {"file", 1, 0, 0},
+ {0, 0, 0, 0}
+ };
+
+ c = getopt_long (argc, argv, "abc:d:0123456789",
+ long_options, &option_index);
+ if (c == -1)
+ break;
+
+ switch (c)
+ {
+ case 0:
+ printf ("option %s", long_options[option_index].name);
+ if (optarg)
+ printf (" with arg %s", optarg);
+ printf ("\n");
+ break;
+
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9':
+ if (digit_optind != 0 && digit_optind != this_option_optind)
+ printf ("digits occur in two different argv-elements.\n");
+ digit_optind = this_option_optind;
+ printf ("option %c\n", c);
+ break;
+
+ case 'a':
+ printf ("option a\n");
+ break;
+
+ case 'b':
+ printf ("option b\n");
+ break;
+
+ case 'c':
+ printf ("option c with value `%s'\n", optarg);
+ break;
+
+ case 'd':
+ printf ("option d with value `%s'\n", optarg);
+ break;
+
+ case '?':
+ break;
+
+ default:
+ printf ("?? getopt returned character code 0%o ??\n", c);
+ }
+ }
+
+ if (optind < argc)
+ {
+ printf ("non-option ARGV-elements: ");
+ while (optind < argc)
+ printf ("%s ", argv[optind++]);
+ printf ("\n");
+ }
+
+ exit (0);
+}
+
+#endif /* TEST */
diff --git a/contrib/diff/lib/getopt_int.h b/contrib/diff/lib/getopt_int.h
new file mode 100644
index 0000000..0c5edde
--- /dev/null
+++ b/contrib/diff/lib/getopt_int.h
@@ -0,0 +1,129 @@
+/* Internal declarations for getopt.
+ Copyright (C) 1989-1994,1996-1999,2001,2003,2004
+ Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _GETOPT_INT_H
+#define _GETOPT_INT_H 1
+
+extern int _getopt_internal (int ___argc, char *const *___argv,
+ const char *__shortopts,
+ const struct option *__longopts, int *__longind,
+ int __long_only);
+
+
+/* Reentrant versions which can handle parsing multiple argument
+ vectors at the same time. */
+
+/* Data type for reentrant functions. */
+struct _getopt_data
+{
+ /* These have exactly the same meaning as the corresponding global
+ variables, except that they are used for the reentrant
+ versions of getopt. */
+ int optind;
+ int opterr;
+ int optopt;
+ char *optarg;
+
+ /* Internal members. */
+
+ /* True if the internal members have been initialized. */
+ int __initialized;
+
+ /* The next char to be scanned in the option-element
+ in which the last option character we returned was found.
+ This allows us to pick up the scan where we left off.
+
+ If this is zero, or a null string, it means resume the scan
+ by advancing to the next ARGV-element. */
+ char *__nextchar;
+
+ /* Describe how to deal with options that follow non-option ARGV-elements.
+
+ If the caller did not specify anything,
+ the default is REQUIRE_ORDER if the environment variable
+ POSIXLY_CORRECT is defined, PERMUTE otherwise.
+
+ REQUIRE_ORDER means don't recognize them as options;
+ stop option processing when the first non-option is seen.
+ This is what Unix does.
+ This mode of operation is selected by either setting the environment
+ variable POSIXLY_CORRECT, or using `+' as the first character
+ of the list of option characters.
+
+ PERMUTE is the default. We permute the contents of ARGV as we
+ scan, so that eventually all the non-options are at the end.
+ This allows options to be given in any order, even with programs
+ that were not written to expect this.
+
+ RETURN_IN_ORDER is an option available to programs that were
+ written to expect options and other ARGV-elements in any order
+ and that care about the ordering of the two. We describe each
+ non-option ARGV-element as if it were the argument of an option
+ with character code 1. Using `-' as the first character of the
+ list of option characters selects this mode of operation.
+
+ The special argument `--' forces an end of option-scanning regardless
+ of the value of `ordering'. In the case of RETURN_IN_ORDER, only
+ `--' can cause `getopt' to return -1 with `optind' != ARGC. */
+
+ enum
+ {
+ REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER
+ } __ordering;
+
+ /* If the POSIXLY_CORRECT environment variable is set. */
+ int __posixly_correct;
+
+
+ /* Handle permutation of arguments. */
+
+ /* Describe the part of ARGV that contains non-options that have
+ been skipped. `first_nonopt' is the index in ARGV of the first
+ of them; `last_nonopt' is the index after the last of them. */
+
+ int __first_nonopt;
+ int __last_nonopt;
+
+#if defined _LIBC && defined USE_NONOPTION_FLAGS
+ int __nonoption_flags_max_len;
+ int __nonoption_flags_len;
+# endif
+};
+
+/* The initializer is necessary to set OPTIND and OPTERR to their
+ default values and to clear the initialization flag. */
+#define _GETOPT_DATA_INITIALIZER { 1, 1 }
+
+extern int _getopt_internal_r (int ___argc, char *const *___argv,
+ const char *__shortopts,
+ const struct option *__longopts, int *__longind,
+ int __long_only, struct _getopt_data *__data);
+
+extern int _getopt_long_r (int ___argc, char *const *___argv,
+ const char *__shortopts,
+ const struct option *__longopts, int *__longind,
+ struct _getopt_data *__data);
+
+extern int _getopt_long_only_r (int ___argc, char *const *___argv,
+ const char *__shortopts,
+ const struct option *__longopts,
+ int *__longind,
+ struct _getopt_data *__data);
+
+#endif /* getopt_int.h */
diff --git a/contrib/diff/lib/gettext.h b/contrib/diff/lib/gettext.h
new file mode 100644
index 0000000..835732e
--- /dev/null
+++ b/contrib/diff/lib/gettext.h
@@ -0,0 +1,68 @@
+/* Convenience header for conditional use of GNU <libintl.h>.
+ Copyright (C) 1995-1998, 2000-2002 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _LIBGETTEXT_H
+#define _LIBGETTEXT_H 1
+
+/* NLS can be disabled through the configure --disable-nls option. */
+#if ENABLE_NLS
+
+/* Get declarations of GNU message catalog functions. */
+# include <libintl.h>
+
+#else
+
+/* Solaris /usr/include/locale.h includes /usr/include/libintl.h, which
+ chokes if dcgettext is defined as a macro. So include it now, to make
+ later inclusions of <locale.h> a NOP. We don't include <libintl.h>
+ as well because people using "gettext.h" will not include <libintl.h>,
+ and also including <libintl.h> would fail on SunOS 4, whereas <locale.h>
+ is OK. */
+#if defined(__sun)
+# include <locale.h>
+#endif
+
+/* Disabled NLS.
+ The casts to 'const char *' serve the purpose of producing warnings
+ for invalid uses of the value returned from these functions.
+ On pre-ANSI systems without 'const', the config.h file is supposed to
+ contain "#define const". */
+# define gettext(Msgid) ((const char *) (Msgid))
+# define dgettext(Domainname, Msgid) ((const char *) (Msgid))
+# define dcgettext(Domainname, Msgid, Category) ((const char *) (Msgid))
+# define ngettext(Msgid1, Msgid2, N) \
+ ((N) == 1 ? (const char *) (Msgid1) : (const char *) (Msgid2))
+# define dngettext(Domainname, Msgid1, Msgid2, N) \
+ ((N) == 1 ? (const char *) (Msgid1) : (const char *) (Msgid2))
+# define dcngettext(Domainname, Msgid1, Msgid2, N, Category) \
+ ((N) == 1 ? (const char *) (Msgid1) : (const char *) (Msgid2))
+# define textdomain(Domainname) ((const char *) (Domainname))
+# define bindtextdomain(Domainname, Dirname) ((const char *) (Dirname))
+# define bind_textdomain_codeset(Domainname, Codeset) ((const char *) (Codeset))
+
+#endif
+
+/* A pseudo function call that serves as a marker for the automated
+ extraction of messages, but does not call gettext(). The run-time
+ translation is done at a different place in the code.
+ The argument, String, should be a literal string. Concatenated strings
+ and other string expressions won't work.
+ The macro's expansion is not parenthesized, so that it is suitable as
+ initializer for static 'char[]' or 'const char[]' variables. */
+#define gettext_noop(String) String
+
+#endif /* _LIBGETTEXT_H */
diff --git a/contrib/diff/lib/gettimeofday.c b/contrib/diff/lib/gettimeofday.c
new file mode 100644
index 0000000..1ebe4d0
--- /dev/null
+++ b/contrib/diff/lib/gettimeofday.c
@@ -0,0 +1,121 @@
+/* Work around the bug in some systems whereby gettimeofday clobbers the
+ static buffer that localtime uses for it's return value. The gettimeofday
+ function from Mac OS X 10.0.4, i.e. Darwin 1.3.7 has this problem.
+ The tzset replacement is necessary for at least Solaris 2.5, 2.5.1, and 2.6.
+ Copyright (C) 2001, 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* written by Jim Meyering */
+
+#include <config.h>
+
+/* Disable the definitions of these functions (from config.h)
+ so we can use the library versions here. */
+#undef gettimeofday
+#undef gmtime
+#undef localtime
+#undef tzset
+
+#include <sys/types.h>
+
+#if TIME_WITH_SYS_TIME
+# include <sys/time.h>
+# include <time.h>
+#else
+# if HAVE_SYS_TIME_H
+# include <sys/time.h>
+# else
+# include <time.h>
+# endif
+#endif
+
+#include <stdlib.h>
+
+static struct tm *localtime_buffer_addr;
+
+/* This is a wrapper for localtime. It is used only on systems for which
+ gettimeofday clobbers the static buffer used for localtime's result.
+
+ On the first call, record the address of the static buffer that
+ localtime uses for its result. */
+
+struct tm *
+rpl_localtime (const time_t *timep)
+{
+ struct tm *tm = localtime (timep);
+
+ if (! localtime_buffer_addr)
+ localtime_buffer_addr = tm;
+
+ return tm;
+}
+
+/* Same as above, since gmtime and localtime use the same buffer. */
+struct tm *
+rpl_gmtime (const time_t *timep)
+{
+ struct tm *tm = gmtime (timep);
+
+ if (! localtime_buffer_addr)
+ localtime_buffer_addr = tm;
+
+ return tm;
+}
+
+/* This is a wrapper for gettimeofday. It is used only on systems for which
+ gettimeofday clobbers the static buffer used for localtime's result.
+
+ Save and restore the contents of the buffer used for localtime's result
+ around the call to gettimeofday. */
+
+int
+rpl_gettimeofday (struct timeval *tv, struct timezone *tz)
+{
+ struct tm save;
+ int result;
+
+ if (! localtime_buffer_addr)
+ {
+ time_t t = 0;
+ localtime_buffer_addr = localtime (&t);
+ }
+
+ save = *localtime_buffer_addr;
+ result = gettimeofday (tv, tz);
+ *localtime_buffer_addr = save;
+
+ return result;
+}
+
+/* This is a wrapper for tzset. It is used only on systems for which
+ tzset may clobber the static buffer used for localtime's result.
+ Save and restore the contents of the buffer used for localtime's
+ result around the call to tzset. */
+void
+rpl_tzset (void)
+{
+ struct tm save;
+
+ if (! localtime_buffer_addr)
+ {
+ time_t t = 0;
+ localtime_buffer_addr = localtime (&t);
+ }
+
+ save = *localtime_buffer_addr;
+ tzset ();
+ *localtime_buffer_addr = save;
+}
diff --git a/contrib/diff/lib/gnulib.mk b/contrib/diff/lib/gnulib.mk
new file mode 100644
index 0000000..6e7f05a
--- /dev/null
+++ b/contrib/diff/lib/gnulib.mk
@@ -0,0 +1,90 @@
+# This file is generated automatically by "bootstrap".
+BUILT_SOURCES += $(ALLOCA_H)
+EXTRA_DIST += alloca_.h
+
+# We need the following in order to create an <alloca.h> when the system
+# doesn't have one that works with the given compiler.
+all-local $(lib_OBJECTS): $(ALLOCA_H)
+alloca.h: alloca_.h
+ cp $(srcdir)/alloca_.h $@-t
+ mv $@-t $@
+MOSTLYCLEANFILES += alloca.h alloca.h-t
+
+lib_SOURCES += c-stack.h c-stack.c
+
+lib_SOURCES += dirname.h dirname.c basename.c stripslash.c
+
+
+lib_SOURCES += exclude.h exclude.c
+
+lib_SOURCES += exit.h
+
+lib_SOURCES += exitfail.h exitfail.c
+
+
+lib_SOURCES += file-type.h file-type.c
+
+BUILT_SOURCES += $(FNMATCH_H)
+EXTRA_DIST += fnmatch_.h fnmatch_loop.c
+
+# We need the following in order to create an <fnmatch.h> when the system
+# doesn't have one that supports the required API.
+all-local $(lib_OBJECTS): $(FNMATCH_H)
+fnmatch.h: fnmatch_.h
+ cp $(srcdir)/fnmatch_.h $@-t
+ mv $@-t $@
+MOSTLYCLEANFILES += fnmatch.h fnmatch.h-t
+
+
+lib_SOURCES += getopt.h getopt.c getopt1.c getopt_int.h
+
+lib_SOURCES += gettext.h
+
+
+lib_SOURCES += hard-locale.h hard-locale.c
+
+EXTRA_DIST += inttostr.c
+lib_SOURCES += imaxtostr.c inttostr.h offtostr.c umaxtostr.c
+
+
+
+lib_SOURCES += posixver.h posixver.c
+
+
+lib_SOURCES += regex.h
+
+
+BUILT_SOURCES += $(STDBOOL_H)
+EXTRA_DIST += stdbool_.h
+
+# We need the following in order to create an <stdbool.h> when the system
+# doesn't have one that works.
+all-local $(lib_OBJECTS): $(STDBOOL_H)
+stdbool.h: stdbool_.h
+ sed -e 's/@''HAVE__BOOL''@/$(HAVE__BOOL)/g' < $(srcdir)/stdbool_.h > $@-t
+ mv $@-t $@
+MOSTLYCLEANFILES += stdbool.h stdbool.h-t
+
+lib_SOURCES += strcase.h
+
+lib_SOURCES += strftime.c
+
+
+
+
+
+
+
+lib_SOURCES += time_r.h
+
+
+lib_SOURCES += unlocked-io.h
+
+lib_SOURCES += version-etc.h version-etc.c
+
+lib_SOURCES += xalloc.h xmalloc.c xstrdup.c
+
+lib_SOURCES += xstrtol.h xstrtol.c xstrtoul.c
+
+lib_SOURCES += xstrtoumax.c
+
diff --git a/contrib/diff/lib/hard-locale.c b/contrib/diff/lib/hard-locale.c
new file mode 100644
index 0000000..8c0ee8c
--- /dev/null
+++ b/contrib/diff/lib/hard-locale.c
@@ -0,0 +1,74 @@
+/* hard-locale.c -- Determine whether a locale is hard.
+
+ Copyright (C) 1997, 1998, 1999, 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "hard-locale.h"
+
+#if HAVE_LOCALE_H
+# include <locale.h>
+#endif
+
+#include <stdlib.h>
+#include <string.h>
+
+/* Return nonzero if the current CATEGORY locale is hard, i.e. if you
+ can't get away with assuming traditional C or POSIX behavior. */
+int
+hard_locale (int category)
+{
+#if ! HAVE_SETLOCALE
+ return 0;
+#else
+
+ int hard = 1;
+ char const *p = setlocale (category, 0);
+
+ if (p)
+ {
+# if defined __GLIBC__ && 2 <= __GLIBC__
+ if (strcmp (p, "C") == 0 || strcmp (p, "POSIX") == 0)
+ hard = 0;
+# else
+ char *locale = malloc (strlen (p) + 1);
+ if (locale)
+ {
+ strcpy (locale, p);
+
+ /* Temporarily set the locale to the "C" and "POSIX" locales
+ to find their names, so that we can determine whether one
+ or the other is the caller's locale. */
+ if (((p = setlocale (category, "C"))
+ && strcmp (p, locale) == 0)
+ || ((p = setlocale (category, "POSIX"))
+ && strcmp (p, locale) == 0))
+ hard = 0;
+
+ /* Restore the caller's locale. */
+ setlocale (category, locale);
+ free (locale);
+ }
+# endif
+ }
+
+ return hard;
+
+#endif
+}
diff --git a/contrib/diff/lib/hard-locale.h b/contrib/diff/lib/hard-locale.h
new file mode 100644
index 0000000..ddc15d0
--- /dev/null
+++ b/contrib/diff/lib/hard-locale.h
@@ -0,0 +1,24 @@
+/* Determine whether a locale is hard.
+
+ Copyright (C) 1999, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef HARD_LOCALE_H_
+# define HARD_LOCALE_H_ 1
+
+int hard_locale (int);
+
+#endif /* HARD_LOCALE_H_ */
diff --git a/contrib/diff/lib/imaxtostr.c b/contrib/diff/lib/imaxtostr.c
new file mode 100644
index 0000000..5e87ad5
--- /dev/null
+++ b/contrib/diff/lib/imaxtostr.c
@@ -0,0 +1,3 @@
+#define inttostr imaxtostr
+#define inttype intmax_t
+#include "inttostr.c"
diff --git a/contrib/diff/lib/inttostr.c b/contrib/diff/lib/inttostr.c
new file mode 100644
index 0000000..78a48af
--- /dev/null
+++ b/contrib/diff/lib/inttostr.c
@@ -0,0 +1,49 @@
+/* inttostr.c -- convert integers to printable strings
+
+ Copyright (C) 2001 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert */
+
+#include "inttostr.h"
+
+/* Convert I to a printable string in BUF, which must be at least
+ INT_BUFSIZE_BOUND (INTTYPE) bytes long. Return the address of the
+ printable string, which need not start at BUF. */
+
+char *
+inttostr (inttype i, char *buf)
+{
+ char *p = buf + INT_STRLEN_BOUND (inttype);
+ *p = 0;
+
+ if (i < 0)
+ {
+ do
+ *--p = '0' - i % 10;
+ while ((i /= 10) != 0);
+
+ *--p = '-';
+ }
+ else
+ {
+ do
+ *--p = '0' + i % 10;
+ while ((i /= 10) != 0);
+ }
+
+ return p;
+}
diff --git a/contrib/diff/lib/inttostr.h b/contrib/diff/lib/inttostr.h
new file mode 100644
index 0000000..6f2416b
--- /dev/null
+++ b/contrib/diff/lib/inttostr.h
@@ -0,0 +1,47 @@
+/* inttostr.h -- convert integers to printable strings
+
+ Copyright (C) 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#if HAVE_INTTYPES_H
+# include <inttypes.h>
+#endif
+#if HAVE_STDINT_H
+# include <stdint.h>
+#endif
+
+#include <limits.h>
+
+#if HAVE_SYS_TYPES_H
+# include <sys/types.h>
+#endif
+
+/* Upper bound on the string length of an integer converted to string.
+ 302 / 1000 is ceil (log10 (2.0)). Subtract 1 for the sign bit;
+ add 1 for integer division truncation; add 1 more for a minus sign. */
+#define INT_STRLEN_BOUND(t) ((sizeof (t) * CHAR_BIT - 1) * 302 / 1000 + 2)
+
+#define INT_BUFSIZE_BOUND(t) (INT_STRLEN_BOUND (t) + 1)
+
+char *offtostr (off_t, char *);
+char *imaxtostr (intmax_t, char *);
+char *umaxtostr (uintmax_t, char *);
diff --git a/contrib/diff/lib/malloc.c b/contrib/diff/lib/malloc.c
new file mode 100644
index 0000000..a43d169
--- /dev/null
+++ b/contrib/diff/lib/malloc.c
@@ -0,0 +1,36 @@
+/* Work around bug on some systems where malloc (0) fails.
+ Copyright (C) 1997, 1998 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* written by Jim Meyering */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+#undef malloc
+
+#include <stdlib.h>
+
+/* Allocate an N-byte block of memory from the heap.
+ If N is zero, allocate a 1-byte block. */
+
+void *
+rpl_malloc (size_t n)
+{
+ if (n == 0)
+ n = 1;
+ return malloc (n);
+}
diff --git a/contrib/diff/lib/mkstemp.c b/contrib/diff/lib/mkstemp.c
new file mode 100644
index 0000000..f6312b6
--- /dev/null
+++ b/contrib/diff/lib/mkstemp.c
@@ -0,0 +1,42 @@
+/* Copyright (C) 1998, 1999, 2001 Free Software Foundation, Inc.
+ This file is derived from the one in the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#include <config.h>
+
+/* Disable the definition of mkstemp to rpl_mkstemp (from config.h) in this
+ file. Otherwise, we'd get conflicting prototypes for rpl_mkstemp on
+ most systems. */
+#undef mkstemp
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#ifndef __GT_FILE
+# define __GT_FILE 0
+#endif
+
+int __gen_tempname ();
+
+/* Generate a unique temporary file name from TEMPLATE.
+ The last six characters of TEMPLATE must be "XXXXXX";
+ they are replaced with a string that makes the filename unique.
+ Then open the file and return a fd. */
+int
+rpl_mkstemp (char *template)
+{
+ return __gen_tempname (template, __GT_FILE);
+}
diff --git a/contrib/diff/lib/offtostr.c b/contrib/diff/lib/offtostr.c
new file mode 100644
index 0000000..45196e2
--- /dev/null
+++ b/contrib/diff/lib/offtostr.c
@@ -0,0 +1,3 @@
+#define inttostr offtostr
+#define inttype off_t
+#include "inttostr.c"
diff --git a/contrib/diff/lib/posixver.c b/contrib/diff/lib/posixver.c
new file mode 100644
index 0000000..754d7ac
--- /dev/null
+++ b/contrib/diff/lib/posixver.c
@@ -0,0 +1,59 @@
+/* Which POSIX version to conform to, for utilities.
+
+ Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "posixver.h"
+
+#include <limits.h>
+#include <stdlib.h>
+
+#if HAVE_UNISTD_H
+# include <unistd.h>
+#endif
+#ifndef _POSIX2_VERSION
+# define _POSIX2_VERSION 0
+#endif
+
+#ifndef DEFAULT_POSIX2_VERSION
+# define DEFAULT_POSIX2_VERSION _POSIX2_VERSION
+#endif
+
+/* The POSIX version that utilities should conform to. The default is
+ specified by the system. */
+
+int
+posix2_version (void)
+{
+ long int v = DEFAULT_POSIX2_VERSION;
+ char const *s = getenv ("_POSIX2_VERSION");
+
+ if (s && *s)
+ {
+ char *e;
+ long int i = strtol (s, &e, 10);
+ if (! *e)
+ v = i;
+ }
+
+ return v < INT_MIN ? INT_MIN : v < INT_MAX ? v : INT_MAX;
+}
diff --git a/contrib/diff/lib/posixver.h b/contrib/diff/lib/posixver.h
new file mode 100644
index 0000000..b64f6a2
--- /dev/null
+++ b/contrib/diff/lib/posixver.h
@@ -0,0 +1 @@
+int posix2_version (void);
diff --git a/contrib/diff/lib/prepargs.c b/contrib/diff/lib/prepargs.c
new file mode 100644
index 0000000..f5735ac
--- /dev/null
+++ b/contrib/diff/lib/prepargs.c
@@ -0,0 +1,93 @@
+/* Parse arguments from a string and prepend them to an argv.
+
+ Copyright (C) 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software
+ Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
+ 02111-1307, USA. */
+
+/* Written by Paul Eggert <eggert@twinsun.com>. */
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+#include "prepargs.h"
+#include <string.h>
+#include <sys/types.h>
+#include <xalloc.h>
+
+#include <ctype.h>
+
+/* IN_CTYPE_DOMAIN (C) is nonzero if the unsigned char C can safely be given
+ as an argument to <ctype.h> macros like "isspace". */
+#ifdef STDC_HEADERS
+# define IN_CTYPE_DOMAIN(c) 1
+#else
+# define IN_CTYPE_DOMAIN(c) ((c) <= 0177)
+#endif
+
+#define ISSPACE(c) (IN_CTYPE_DOMAIN (c) && isspace (c))
+
+/* Find the white-space-separated options specified by OPTIONS, and
+ using BUF to store copies of these options, set ARGV[0], ARGV[1],
+ etc. to the option copies. Return the number N of options found.
+ Do not set ARGV[N]. If ARGV is zero, do not store ARGV[0] etc.
+ Backslash can be used to escape whitespace (and backslashes). */
+static int
+prepend_args (char const *options, char *buf, char **argv)
+{
+ char const *o = options;
+ char *b = buf;
+ int n = 0;
+
+ for (;;)
+ {
+ while (ISSPACE ((unsigned char) *o))
+ o++;
+ if (!*o)
+ return n;
+ if (argv)
+ argv[n] = b;
+ n++;
+
+ do
+ if ((*b++ = *o++) == '\\' && *o)
+ b[-1] = *o++;
+ while (*o && ! ISSPACE ((unsigned char) *o));
+
+ *b++ = '\0';
+ }
+}
+
+/* Prepend the whitespace-separated options in OPTIONS to the argument
+ vector of a main program with argument count *PARGC and argument
+ vector *PARGV. */
+void
+prepend_default_options (char const *options, int *pargc, char ***pargv)
+{
+ if (options)
+ {
+ char *buf = xmalloc (strlen (options) + 1);
+ int prepended = prepend_args (options, buf, (char **) 0);
+ int argc = *pargc;
+ char * const *argv = *pargv;
+ char **pp = (char **) xmalloc ((prepended + argc + 1) * sizeof *pp);
+ *pargc = prepended + argc;
+ *pargv = pp;
+ *pp++ = *argv++;
+ pp += prepend_args (options, buf, pp);
+ while ((*pp++ = *argv++))
+ continue;
+ }
+}
diff --git a/contrib/diff/lib/prepargs.h b/contrib/diff/lib/prepargs.h
new file mode 100644
index 0000000..ce93ea8
--- /dev/null
+++ b/contrib/diff/lib/prepargs.h
@@ -0,0 +1,3 @@
+/* Parse arguments from a string and prepend them to an argv. */
+
+void prepend_default_options (char const *, int *, char ***);
diff --git a/contrib/diff/lib/quotesys.c b/contrib/diff/lib/quotesys.c
new file mode 100644
index 0000000..528f382
--- /dev/null
+++ b/contrib/diff/lib/quotesys.c
@@ -0,0 +1,125 @@
+/* Shell command argument quoting.
+ Copyright (C) 1994, 1995, 1997 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert <eggert@twinsun.com> */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <sys/types.h>
+#include <quotesys.h>
+
+/* Place into QUOTED a quoted version of ARG suitable for `system'.
+ Return the length of the resulting string (which is not null-terminated).
+ If QUOTED is null, return the length without any side effects. */
+
+size_t
+quote_system_arg (quoted, arg)
+ char *quoted;
+ char const *arg;
+{
+ char const *a;
+ size_t len = 0;
+
+ /* Scan ARG, copying it to QUOTED if QUOTED is not null,
+ looking for shell metacharacters. */
+
+ for (a = arg; ; a++)
+ {
+ char c = *a;
+ switch (c)
+ {
+ case 0:
+ /* ARG has no shell metacharacters. */
+ return len;
+
+ case '=':
+ if (*arg == '-')
+ break;
+ /* Fall through. */
+ case '\t': case '\n': case ' ':
+ case '!': case '"': case '#': case '$': case '%': case '&': case '\'':
+ case '(': case ')': case '*': case ';':
+ case '<': case '>': case '?': case '[': case '\\':
+ case '^': case '`': case '|': case '~':
+ {
+ /* ARG has a shell metacharacter.
+ Start over, quoting it this time. */
+
+ len = 0;
+ c = *arg++;
+
+ /* If ARG is an option, quote just its argument.
+ This is not necessary, but it looks nicer. */
+ if (c == '-' && arg < a)
+ {
+ c = *arg++;
+
+ if (quoted)
+ {
+ quoted[len] = '-';
+ quoted[len + 1] = c;
+ }
+ len += 2;
+
+ if (c == '-')
+ while (arg < a)
+ {
+ c = *arg++;
+ if (quoted)
+ quoted[len] = c;
+ len++;
+ if (c == '=')
+ break;
+ }
+ c = *arg++;
+ }
+
+ if (quoted)
+ quoted[len] = '\'';
+ len++;
+
+ for (; c; c = *arg++)
+ {
+ if (c == '\'')
+ {
+ if (quoted)
+ {
+ quoted[len] = '\'';
+ quoted[len + 1] = '\\';
+ quoted[len + 2] = '\'';
+ }
+ len += 3;
+ }
+ if (quoted)
+ quoted[len] = c;
+ len++;
+ }
+
+ if (quoted)
+ quoted[len] = '\'';
+ return len + 1;
+ }
+ }
+
+ if (quoted)
+ quoted[len] = c;
+ len++;
+ }
+}
diff --git a/contrib/diff/lib/quotesys.h b/contrib/diff/lib/quotesys.h
new file mode 100644
index 0000000..0d0d825
--- /dev/null
+++ b/contrib/diff/lib/quotesys.h
@@ -0,0 +1,9 @@
+/* quotesys.h -- declarations for quoting system arguments */
+
+#if defined __STDC__ || __GNUC__
+# define __QUOTESYS_P(args) args
+#else
+# define __QUOTESYS_P(args) ()
+#endif
+
+size_t quote_system_arg __QUOTESYS_P ((char *, char const *));
diff --git a/contrib/diff/lib/realloc.c b/contrib/diff/lib/realloc.c
new file mode 100644
index 0000000..ccbf991
--- /dev/null
+++ b/contrib/diff/lib/realloc.c
@@ -0,0 +1,39 @@
+/* Work around bug on some systems where realloc (NULL, 0) fails.
+ Copyright (C) 1997, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* written by Jim Meyering */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+#undef realloc
+
+#include <stdlib.h>
+
+/* Change the size of an allocated block of memory P to N bytes,
+ with error checking. If N is zero, change it to 1. If P is NULL,
+ use malloc. */
+
+void *
+rpl_realloc (void *p, size_t n)
+{
+ if (n == 0)
+ n = 1;
+ if (p == 0)
+ return malloc (n);
+ return realloc (p, n);
+}
diff --git a/contrib/diff/lib/regex.c b/contrib/diff/lib/regex.c
new file mode 100644
index 0000000..ac70b90
--- /dev/null
+++ b/contrib/diff/lib/regex.c
@@ -0,0 +1,8298 @@
+/* Extended regular expression matching and search library,
+ version 0.12.
+ (Implements POSIX draft P1003.2/D11.2, except for some of the
+ internationalization features.)
+
+ Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
+ 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* AIX requires this to be the first thing in the file. */
+#if defined _AIX && !defined REGEX_MALLOC
+ #pragma alloca
+#endif
+
+#undef _GNU_SOURCE
+#define _GNU_SOURCE
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#ifndef INSIDE_RECURSION
+
+# include <stddef.h>
+
+# define WIDE_CHAR_SUPPORT (HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_BTOWC)
+
+/* For platform which support the ISO C amendement 1 functionality we
+ support user defined character classes. */
+# if defined _LIBC || WIDE_CHAR_SUPPORT
+/* Solaris 2.5 has a bug: <wchar.h> must be included before <wctype.h>. */
+# include <wchar.h>
+# include <wctype.h>
+# endif
+
+# ifdef _LIBC
+/* We have to keep the namespace clean. */
+# define regfree(preg) __regfree (preg)
+# define regexec(pr, st, nm, pm, ef) __regexec (pr, st, nm, pm, ef)
+# define regcomp(preg, pattern, cflags) __regcomp (preg, pattern, cflags)
+# define regerror(errcode, preg, errbuf, errbuf_size) \
+ __regerror(errcode, preg, errbuf, errbuf_size)
+# define re_set_registers(bu, re, nu, st, en) \
+ __re_set_registers (bu, re, nu, st, en)
+# define re_match_2(bufp, string1, size1, string2, size2, pos, regs, stop) \
+ __re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop)
+# define re_match(bufp, string, size, pos, regs) \
+ __re_match (bufp, string, size, pos, regs)
+# define re_search(bufp, string, size, startpos, range, regs) \
+ __re_search (bufp, string, size, startpos, range, regs)
+# define re_compile_pattern(pattern, length, bufp) \
+ __re_compile_pattern (pattern, length, bufp)
+# define re_set_syntax(syntax) __re_set_syntax (syntax)
+# define re_search_2(bufp, st1, s1, st2, s2, startpos, range, regs, stop) \
+ __re_search_2 (bufp, st1, s1, st2, s2, startpos, range, regs, stop)
+# define re_compile_fastmap(bufp) __re_compile_fastmap (bufp)
+
+# define btowc __btowc
+# define iswctype __iswctype
+# define mbrtowc __mbrtowc
+# define wcslen __wcslen
+# define wcscoll __wcscoll
+# define wcrtomb __wcrtomb
+
+/* We are also using some library internals. */
+# include <locale/localeinfo.h>
+# include <locale/elem-hash.h>
+# include <langinfo.h>
+# include <locale/coll-lookup.h>
+# endif
+
+# ifdef _LIBC
+# include <libintl.h>
+# undef gettext
+# define gettext(msgid) __dcgettext ("libc", msgid, LC_MESSAGES)
+ /* This define is so xgettext can find the internationalizable strings. */
+# define gettext_noop(msgid) msgid
+# else
+/* This is for other GNU distributions with internationalized messages. */
+# include "gettext.h"
+# endif
+
+/* Support for bounded pointers. */
+# if !defined _LIBC && !defined __BOUNDED_POINTERS__
+# define __bounded /* nothing */
+# define __unbounded /* nothing */
+# define __ptrvalue /* nothing */
+# endif
+
+/* The `emacs' switch turns on certain matching commands
+ that make sense only in Emacs. */
+# ifdef emacs
+
+# include "lisp.h"
+# include "buffer.h"
+# include "syntax.h"
+
+# else /* not emacs */
+
+/* If we are not linking with Emacs proper,
+ we can't use the relocating allocator
+ even if config.h says that we can. */
+# undef REL_ALLOC
+
+# include <stdlib.h>
+
+/* When used in Emacs's lib-src, we need to get bzero and bcopy somehow.
+ If nothing else has been done, use the method below. */
+# ifdef INHIBIT_STRING_HEADER
+# if !(defined HAVE_BZERO && defined HAVE_BCOPY)
+# if !defined bzero && !defined bcopy
+# undef INHIBIT_STRING_HEADER
+# endif
+# endif
+# endif
+
+/* This is the normal way of making sure we have a bcopy and a bzero.
+ This is used in most programs--a few other programs avoid this
+ by defining INHIBIT_STRING_HEADER. */
+# ifndef INHIBIT_STRING_HEADER
+# include <string.h>
+# ifndef bzero
+# ifndef _LIBC
+# define bzero(s, n) (memset (s, '\0', n), (s))
+# else
+# define bzero(s, n) __bzero (s, n)
+# endif
+# endif
+# endif
+
+/* Define the syntax stuff for \<, \>, etc. */
+
+/* This must be nonzero for the wordchar and notwordchar pattern
+ commands in re_match_2. */
+# ifndef Sword
+# define Sword 1
+# endif
+
+# ifdef SWITCH_ENUM_BUG
+# define SWITCH_ENUM_CAST(x) ((int)(x))
+# else
+# define SWITCH_ENUM_CAST(x) (x)
+# endif
+
+# endif /* not emacs */
+
+# include <limits.h>
+
+# ifndef MB_LEN_MAX
+# define MB_LEN_MAX 1
+# endif
+
+/* Get the interface, including the syntax bits. */
+# include <regex.h>
+
+/* isalpha etc. are used for the character classes. */
+# include <ctype.h>
+
+/* Jim Meyering writes:
+
+ "... Some ctype macros are valid only for character codes that
+ isascii says are ASCII (SGI's IRIX-4.0.5 is one such system --when
+ using /bin/cc or gcc but without giving an ansi option). So, all
+ ctype uses should be through macros like ISPRINT... If
+ STDC_HEADERS is defined, then autoconf has verified that the ctype
+ macros don't need to be guarded with references to isascii. ...
+ Defining isascii to 1 should let any compiler worth its salt
+ eliminate the && through constant folding."
+ Solaris defines some of these symbols so we must undefine them first. */
+
+# if defined STDC_HEADERS || (!defined isascii && !defined HAVE_ISASCII)
+# define IN_CTYPE_DOMAIN(c) 1
+# else
+# define IN_CTYPE_DOMAIN(c) isascii(c)
+# endif
+
+# ifdef isblank
+# define ISBLANK(c) (IN_CTYPE_DOMAIN (c) && isblank (c))
+# else
+# define ISBLANK(c) ((c) == ' ' || (c) == '\t')
+# endif
+# ifdef isgraph
+# define ISGRAPH(c) (IN_CTYPE_DOMAIN (c) && isgraph (c))
+# else
+# define ISGRAPH(c) (IN_CTYPE_DOMAIN (c) && isprint (c) && !isspace (c))
+# endif
+
+# undef ISPRINT
+# define ISPRINT(c) (IN_CTYPE_DOMAIN (c) && isprint (c))
+# define ISDIGIT(c) (IN_CTYPE_DOMAIN (c) && isdigit (c))
+# define ISALNUM(c) (IN_CTYPE_DOMAIN (c) && isalnum (c))
+# define ISALPHA(c) (IN_CTYPE_DOMAIN (c) && isalpha (c))
+# define ISCNTRL(c) (IN_CTYPE_DOMAIN (c) && iscntrl (c))
+# define ISLOWER(c) (IN_CTYPE_DOMAIN (c) && islower (c))
+# define ISPUNCT(c) (IN_CTYPE_DOMAIN (c) && ispunct (c))
+# define ISSPACE(c) (IN_CTYPE_DOMAIN (c) && isspace (c))
+# define ISUPPER(c) (IN_CTYPE_DOMAIN (c) && isupper (c))
+# define ISXDIGIT(c) (IN_CTYPE_DOMAIN (c) && isxdigit (c))
+
+# ifdef _tolower
+# define TOLOWER(c) _tolower(c)
+# else
+# define TOLOWER(c) tolower(c)
+# endif
+
+# ifndef emacs
+/* How many characters in the character set. */
+# define CHAR_SET_SIZE 256
+
+# ifdef SYNTAX_TABLE
+
+extern char *re_syntax_table;
+
+# else /* not SYNTAX_TABLE */
+
+static char re_syntax_table[CHAR_SET_SIZE];
+
+static void
+init_syntax_once (void)
+{
+ register int c;
+ static int done = 0;
+
+ if (done)
+ return;
+ bzero (re_syntax_table, sizeof re_syntax_table);
+
+ for (c = 0; c < CHAR_SET_SIZE; ++c)
+ if (ISALNUM (c))
+ re_syntax_table[c] = Sword;
+
+ re_syntax_table['_'] = Sword;
+
+ done = 1;
+}
+
+# endif /* not SYNTAX_TABLE */
+
+# define SYNTAX(c) re_syntax_table[(unsigned char) (c)]
+
+# endif /* emacs */
+
+/* Should we use malloc or alloca? If REGEX_MALLOC is not defined, we
+ use `alloca' instead of `malloc'. This is because using malloc in
+ re_search* or re_match* could cause memory leaks when C-g is used in
+ Emacs; also, malloc is slower and causes storage fragmentation. On
+ the other hand, malloc is more portable, and easier to debug.
+
+ Because we sometimes use alloca, some routines have to be macros,
+ not functions -- `alloca'-allocated space disappears at the end of the
+ function it is called in. */
+
+# ifdef REGEX_MALLOC
+
+# define REGEX_ALLOCATE malloc
+# define REGEX_REALLOCATE(source, osize, nsize) realloc (source, nsize)
+# define REGEX_FREE free
+
+# else /* not REGEX_MALLOC */
+
+/* Emacs already defines alloca, sometimes. */
+# ifndef alloca
+
+/* Make alloca work the best possible way. */
+# ifdef __GNUC__
+# define alloca __builtin_alloca
+# else /* not __GNUC__ */
+# if HAVE_ALLOCA_H
+# include <alloca.h>
+# endif /* HAVE_ALLOCA_H */
+# endif /* not __GNUC__ */
+
+# endif /* not alloca */
+
+# define REGEX_ALLOCATE alloca
+
+/* Assumes a `char *destination' variable. */
+# define REGEX_REALLOCATE(source, osize, nsize) \
+ (destination = (char *) alloca (nsize), \
+ memcpy (destination, source, osize))
+
+/* No need to do anything to free, after alloca. */
+# define REGEX_FREE(arg) ((void)0) /* Do nothing! But inhibit gcc warning. */
+
+# endif /* not REGEX_MALLOC */
+
+/* Define how to allocate the failure stack. */
+
+# if defined REL_ALLOC && defined REGEX_MALLOC
+
+# define REGEX_ALLOCATE_STACK(size) \
+ r_alloc (&failure_stack_ptr, (size))
+# define REGEX_REALLOCATE_STACK(source, osize, nsize) \
+ r_re_alloc (&failure_stack_ptr, (nsize))
+# define REGEX_FREE_STACK(ptr) \
+ r_alloc_free (&failure_stack_ptr)
+
+# else /* not using relocating allocator */
+
+# ifdef REGEX_MALLOC
+
+# define REGEX_ALLOCATE_STACK malloc
+# define REGEX_REALLOCATE_STACK(source, osize, nsize) realloc (source, nsize)
+# define REGEX_FREE_STACK free
+
+# else /* not REGEX_MALLOC */
+
+# define REGEX_ALLOCATE_STACK alloca
+
+# define REGEX_REALLOCATE_STACK(source, osize, nsize) \
+ REGEX_REALLOCATE (source, osize, nsize)
+/* No need to explicitly free anything. */
+# define REGEX_FREE_STACK(arg)
+
+# endif /* not REGEX_MALLOC */
+# endif /* not using relocating allocator */
+
+
+/* True if `size1' is non-NULL and PTR is pointing anywhere inside
+ `string1' or just past its end. This works if PTR is NULL, which is
+ a good thing. */
+# define FIRST_STRING_P(ptr) \
+ (size1 && string1 <= (ptr) && (ptr) <= string1 + size1)
+
+/* (Re)Allocate N items of type T using malloc, or fail. */
+# define TALLOC(n, t) ((t *) malloc ((n) * sizeof (t)))
+# define RETALLOC(addr, n, t) ((addr) = (t *) realloc (addr, (n) * sizeof (t)))
+# define RETALLOC_IF(addr, n, t) \
+ if (addr) RETALLOC((addr), (n), t); else (addr) = TALLOC ((n), t)
+# define REGEX_TALLOC(n, t) ((t *) REGEX_ALLOCATE ((n) * sizeof (t)))
+
+# define BYTEWIDTH 8 /* In bits. */
+
+# define STREQ(s1, s2) ((strcmp (s1, s2) == 0))
+
+# undef MAX
+# undef MIN
+# define MAX(a, b) ((a) > (b) ? (a) : (b))
+# define MIN(a, b) ((a) < (b) ? (a) : (b))
+
+typedef char boolean;
+# define false 0
+# define true 1
+
+static reg_errcode_t byte_regex_compile (const char *pattern, size_t size,
+ reg_syntax_t syntax,
+ struct re_pattern_buffer *bufp);
+
+static int byte_re_match_2_internal (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int pos,
+ struct re_registers *regs,
+ int stop);
+static int byte_re_search_2 (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int startpos, int range,
+ struct re_registers *regs, int stop);
+static int byte_re_compile_fastmap (struct re_pattern_buffer *bufp);
+
+#ifdef MBS_SUPPORT
+static reg_errcode_t wcs_regex_compile (const char *pattern, size_t size,
+ reg_syntax_t syntax,
+ struct re_pattern_buffer *bufp);
+
+
+static int wcs_re_match_2_internal (struct re_pattern_buffer *bufp,
+ const char *cstring1, int csize1,
+ const char *cstring2, int csize2,
+ int pos,
+ struct re_registers *regs,
+ int stop,
+ wchar_t *string1, int size1,
+ wchar_t *string2, int size2,
+ int *mbs_offset1, int *mbs_offset2);
+static int wcs_re_search_2 (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int startpos, int range,
+ struct re_registers *regs, int stop);
+static int wcs_re_compile_fastmap (struct re_pattern_buffer *bufp);
+#endif
+
+/* These are the command codes that appear in compiled regular
+ expressions. Some opcodes are followed by argument bytes. A
+ command code can specify any interpretation whatsoever for its
+ arguments. Zero bytes may appear in the compiled regular expression. */
+
+typedef enum
+{
+ no_op = 0,
+
+ /* Succeed right away--no more backtracking. */
+ succeed,
+
+ /* Followed by one byte giving n, then by n literal bytes. */
+ exactn,
+
+# ifdef MBS_SUPPORT
+ /* Same as exactn, but contains binary data. */
+ exactn_bin,
+# endif
+
+ /* Matches any (more or less) character. */
+ anychar,
+
+ /* Matches any one char belonging to specified set. First
+ following byte is number of bitmap bytes. Then come bytes
+ for a bitmap saying which chars are in. Bits in each byte
+ are ordered low-bit-first. A character is in the set if its
+ bit is 1. A character too large to have a bit in the map is
+ automatically not in the set. */
+ /* ifdef MBS_SUPPORT, following element is length of character
+ classes, length of collating symbols, length of equivalence
+ classes, length of character ranges, and length of characters.
+ Next, character class element, collating symbols elements,
+ equivalence class elements, range elements, and character
+ elements follow.
+ See regex_compile function. */
+ charset,
+
+ /* Same parameters as charset, but match any character that is
+ not one of those specified. */
+ charset_not,
+
+ /* Start remembering the text that is matched, for storing in a
+ register. Followed by one byte with the register number, in
+ the range 0 to one less than the pattern buffer's re_nsub
+ field. Then followed by one byte with the number of groups
+ inner to this one. (This last has to be part of the
+ start_memory only because we need it in the on_failure_jump
+ of re_match_2.) */
+ start_memory,
+
+ /* Stop remembering the text that is matched and store it in a
+ memory register. Followed by one byte with the register
+ number, in the range 0 to one less than `re_nsub' in the
+ pattern buffer, and one byte with the number of inner groups,
+ just like `start_memory'. (We need the number of inner
+ groups here because we don't have any easy way of finding the
+ corresponding start_memory when we're at a stop_memory.) */
+ stop_memory,
+
+ /* Match a duplicate of something remembered. Followed by one
+ byte containing the register number. */
+ duplicate,
+
+ /* Fail unless at beginning of line. */
+ begline,
+
+ /* Fail unless at end of line. */
+ endline,
+
+ /* Succeeds if at beginning of buffer (if emacs) or at beginning
+ of string to be matched (if not). */
+ begbuf,
+
+ /* Analogously, for end of buffer/string. */
+ endbuf,
+
+ /* Followed by two byte relative address to which to jump. */
+ jump,
+
+ /* Same as jump, but marks the end of an alternative. */
+ jump_past_alt,
+
+ /* Followed by two-byte relative address of place to resume at
+ in case of failure. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ on_failure_jump,
+
+ /* Like on_failure_jump, but pushes a placeholder instead of the
+ current string position when executed. */
+ on_failure_keep_string_jump,
+
+ /* Throw away latest failure point and then jump to following
+ two-byte relative address. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ pop_failure_jump,
+
+ /* Change to pop_failure_jump if know won't have to backtrack to
+ match; otherwise change to jump. This is used to jump
+ back to the beginning of a repeat. If what follows this jump
+ clearly won't match what the repeat does, such that we can be
+ sure that there is no use backtracking out of repetitions
+ already matched, then we change it to a pop_failure_jump.
+ Followed by two-byte address. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ maybe_pop_jump,
+
+ /* Jump to following two-byte address, and push a dummy failure
+ point. This failure point will be thrown away if an attempt
+ is made to use it for a failure. A `+' construct makes this
+ before the first repeat. Also used as an intermediary kind
+ of jump when compiling an alternative. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ dummy_failure_jump,
+
+ /* Push a dummy failure point and continue. Used at the end of
+ alternatives. */
+ push_dummy_failure,
+
+ /* Followed by two-byte relative address and two-byte number n.
+ After matching N times, jump to the address upon failure. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ succeed_n,
+
+ /* Followed by two-byte relative address, and two-byte number n.
+ Jump to the address N times, then fail. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ jump_n,
+
+ /* Set the following two-byte relative address to the
+ subsequent two-byte number. The address *includes* the two
+ bytes of number. */
+ /* ifdef MBS_SUPPORT, the size of address is 1. */
+ set_number_at,
+
+ wordchar, /* Matches any word-constituent character. */
+ notwordchar, /* Matches any char that is not a word-constituent. */
+
+ wordbeg, /* Succeeds if at word beginning. */
+ wordend, /* Succeeds if at word end. */
+
+ wordbound, /* Succeeds if at a word boundary. */
+ notwordbound /* Succeeds if not at a word boundary. */
+
+# ifdef emacs
+ ,before_dot, /* Succeeds if before point. */
+ at_dot, /* Succeeds if at point. */
+ after_dot, /* Succeeds if after point. */
+
+ /* Matches any character whose syntax is specified. Followed by
+ a byte which contains a syntax code, e.g., Sword. */
+ syntaxspec,
+
+ /* Matches any character whose syntax is not that specified. */
+ notsyntaxspec
+# endif /* emacs */
+} re_opcode_t;
+#endif /* not INSIDE_RECURSION */
+
+
+#ifdef BYTE
+# define CHAR_T char
+# define UCHAR_T unsigned char
+# define COMPILED_BUFFER_VAR bufp->buffer
+# define OFFSET_ADDRESS_SIZE 2
+# define PREFIX(name) byte_##name
+# define ARG_PREFIX(name) name
+# define PUT_CHAR(c) putchar (c)
+#else
+# ifdef WCHAR
+# define CHAR_T wchar_t
+# define UCHAR_T wchar_t
+# define COMPILED_BUFFER_VAR wc_buffer
+# define OFFSET_ADDRESS_SIZE 1 /* the size which STORE_NUMBER macro use */
+# define CHAR_CLASS_SIZE ((__alignof__(wctype_t)+sizeof(wctype_t))/sizeof(CHAR_T)+1)
+# define PREFIX(name) wcs_##name
+# define ARG_PREFIX(name) c##name
+/* Should we use wide stream?? */
+# define PUT_CHAR(c) printf ("%C", c);
+# define TRUE 1
+# define FALSE 0
+# else
+# ifdef MBS_SUPPORT
+# define WCHAR
+# define INSIDE_RECURSION
+# include "regex.c"
+# undef INSIDE_RECURSION
+# endif
+# define BYTE
+# define INSIDE_RECURSION
+# include "regex.c"
+# undef INSIDE_RECURSION
+# endif
+#endif
+#include "unlocked-io.h"
+
+#ifdef INSIDE_RECURSION
+/* Common operations on the compiled pattern. */
+
+/* Store NUMBER in two contiguous bytes starting at DESTINATION. */
+/* ifdef MBS_SUPPORT, we store NUMBER in 1 element. */
+
+# ifdef WCHAR
+# define STORE_NUMBER(destination, number) \
+ do { \
+ *(destination) = (UCHAR_T)(number); \
+ } while (0)
+# else /* BYTE */
+# define STORE_NUMBER(destination, number) \
+ do { \
+ (destination)[0] = (number) & 0377; \
+ (destination)[1] = (number) >> 8; \
+ } while (0)
+# endif /* WCHAR */
+
+/* Same as STORE_NUMBER, except increment DESTINATION to
+ the byte after where the number is stored. Therefore, DESTINATION
+ must be an lvalue. */
+/* ifdef MBS_SUPPORT, we store NUMBER in 1 element. */
+
+# define STORE_NUMBER_AND_INCR(destination, number) \
+ do { \
+ STORE_NUMBER (destination, number); \
+ (destination) += OFFSET_ADDRESS_SIZE; \
+ } while (0)
+
+/* Put into DESTINATION a number stored in two contiguous bytes starting
+ at SOURCE. */
+/* ifdef MBS_SUPPORT, we store NUMBER in 1 element. */
+
+# ifdef WCHAR
+# define EXTRACT_NUMBER(destination, source) \
+ do { \
+ (destination) = *(source); \
+ } while (0)
+# else /* BYTE */
+# define EXTRACT_NUMBER(destination, source) \
+ do { \
+ (destination) = *(source) & 0377; \
+ (destination) += (signed char) (*((source) + 1)) << 8; \
+ } while (0)
+# endif
+
+# ifdef DEBUG
+static void
+PREFIX(extract_number) (int *dest, UCHAR_T *source)
+{
+# ifdef WCHAR
+ *dest = *source;
+# else /* BYTE */
+ signed char temp = source[1];
+ *dest = *source & 0377;
+ *dest += temp << 8;
+# endif
+}
+
+# ifndef EXTRACT_MACROS /* To debug the macros. */
+# undef EXTRACT_NUMBER
+# define EXTRACT_NUMBER(dest, src) PREFIX(extract_number) (&dest, src)
+# endif /* not EXTRACT_MACROS */
+
+# endif /* DEBUG */
+
+/* Same as EXTRACT_NUMBER, except increment SOURCE to after the number.
+ SOURCE must be an lvalue. */
+
+# define EXTRACT_NUMBER_AND_INCR(destination, source) \
+ do { \
+ EXTRACT_NUMBER (destination, source); \
+ (source) += OFFSET_ADDRESS_SIZE; \
+ } while (0)
+
+# ifdef DEBUG
+static void
+PREFIX(extract_number_and_incr) (int *destination, UCHAR_T **source)
+{
+ PREFIX(extract_number) (destination, *source);
+ *source += OFFSET_ADDRESS_SIZE;
+}
+
+# ifndef EXTRACT_MACROS
+# undef EXTRACT_NUMBER_AND_INCR
+# define EXTRACT_NUMBER_AND_INCR(dest, src) \
+ PREFIX(extract_number_and_incr) (&dest, &src)
+# endif /* not EXTRACT_MACROS */
+
+# endif /* DEBUG */
+
+
+
+/* If DEBUG is defined, Regex prints many voluminous messages about what
+ it is doing (if the variable `debug' is nonzero). If linked with the
+ main program in `iregex.c', you can enter patterns and strings
+ interactively. And if linked with the main program in `main.c' and
+ the other test files, you can run the already-written tests. */
+
+# ifdef DEBUG
+
+# ifndef DEFINED_ONCE
+
+/* We use standard I/O for debugging. */
+# include <stdio.h>
+
+/* It is useful to test things that ``must'' be true when debugging. */
+# include <assert.h>
+
+static int debug;
+
+# define DEBUG_STATEMENT(e) e
+# define DEBUG_PRINT1(x) if (debug) printf (x)
+# define DEBUG_PRINT2(x1, x2) if (debug) printf (x1, x2)
+# define DEBUG_PRINT3(x1, x2, x3) if (debug) printf (x1, x2, x3)
+# define DEBUG_PRINT4(x1, x2, x3, x4) if (debug) printf (x1, x2, x3, x4)
+# endif /* not DEFINED_ONCE */
+
+# define DEBUG_PRINT_COMPILED_PATTERN(p, s, e) \
+ if (debug) PREFIX(print_partial_compiled_pattern) (s, e)
+# define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2) \
+ if (debug) PREFIX(print_double_string) (w, s1, sz1, s2, sz2)
+
+
+/* Print the fastmap in human-readable form. */
+
+# ifndef DEFINED_ONCE
+void
+print_fastmap (char *fastmap)
+{
+ unsigned was_a_range = 0;
+ unsigned i = 0;
+
+ while (i < (1 << BYTEWIDTH))
+ {
+ if (fastmap[i++])
+ {
+ was_a_range = 0;
+ putchar (i - 1);
+ while (i < (1 << BYTEWIDTH) && fastmap[i])
+ {
+ was_a_range = 1;
+ i++;
+ }
+ if (was_a_range)
+ {
+ printf ("-");
+ putchar (i - 1);
+ }
+ }
+ }
+ putchar ('\n');
+}
+# endif /* not DEFINED_ONCE */
+
+
+/* Print a compiled pattern string in human-readable form, starting at
+ the START pointer into it and ending just before the pointer END. */
+
+void
+PREFIX(print_partial_compiled_pattern) (UCHAR_T *start, UCHAR_T *end)
+{
+ int mcnt, mcnt2;
+ UCHAR_T *p1;
+ UCHAR_T *p = start;
+ UCHAR_T *pend = end;
+
+ if (start == NULL)
+ {
+ printf ("(null)\n");
+ return;
+ }
+
+ /* Loop over pattern commands. */
+ while (p < pend)
+ {
+# ifdef _LIBC
+ printf ("%td:\t", p - start);
+# else
+ printf ("%ld:\t", (long int) (p - start));
+# endif
+
+ switch ((re_opcode_t) *p++)
+ {
+ case no_op:
+ printf ("/no_op");
+ break;
+
+ case exactn:
+ mcnt = *p++;
+ printf ("/exactn/%d", mcnt);
+ do
+ {
+ putchar ('/');
+ PUT_CHAR (*p++);
+ }
+ while (--mcnt);
+ break;
+
+# ifdef MBS_SUPPORT
+ case exactn_bin:
+ mcnt = *p++;
+ printf ("/exactn_bin/%d", mcnt);
+ do
+ {
+ printf("/%lx", (long int) *p++);
+ }
+ while (--mcnt);
+ break;
+# endif /* MBS_SUPPORT */
+
+ case start_memory:
+ mcnt = *p++;
+ printf ("/start_memory/%d/%ld", mcnt, (long int) *p++);
+ break;
+
+ case stop_memory:
+ mcnt = *p++;
+ printf ("/stop_memory/%d/%ld", mcnt, (long int) *p++);
+ break;
+
+ case duplicate:
+ printf ("/duplicate/%ld", (long int) *p++);
+ break;
+
+ case anychar:
+ printf ("/anychar");
+ break;
+
+ case charset:
+ case charset_not:
+ {
+# ifdef WCHAR
+ int i, length;
+ wchar_t *workp = p;
+ printf ("/charset [%s",
+ (re_opcode_t) *(workp - 1) == charset_not ? "^" : "");
+ p += 5;
+ length = *workp++; /* the length of char_classes */
+ for (i=0 ; i<length ; i++)
+ printf("[:%lx:]", (long int) *p++);
+ length = *workp++; /* the length of collating_symbol */
+ for (i=0 ; i<length ;)
+ {
+ printf("[.");
+ while(*p != 0)
+ PUT_CHAR((i++,*p++));
+ i++,p++;
+ printf(".]");
+ }
+ length = *workp++; /* the length of equivalence_class */
+ for (i=0 ; i<length ;)
+ {
+ printf("[=");
+ while(*p != 0)
+ PUT_CHAR((i++,*p++));
+ i++,p++;
+ printf("=]");
+ }
+ length = *workp++; /* the length of char_range */
+ for (i=0 ; i<length ; i++)
+ {
+ wchar_t range_start = *p++;
+ wchar_t range_end = *p++;
+ printf("%C-%C", range_start, range_end);
+ }
+ length = *workp++; /* the length of char */
+ for (i=0 ; i<length ; i++)
+ printf("%C", *p++);
+ putchar (']');
+# else
+ register int c, last = -100;
+ register int in_range = 0;
+
+ printf ("/charset [%s",
+ (re_opcode_t) *(p - 1) == charset_not ? "^" : "");
+
+ assert (p + *p < pend);
+
+ for (c = 0; c < 256; c++)
+ if (c / 8 < *p
+ && (p[1 + (c/8)] & (1 << (c % 8))))
+ {
+ /* Are we starting a range? */
+ if (last + 1 == c && ! in_range)
+ {
+ putchar ('-');
+ in_range = 1;
+ }
+ /* Have we broken a range? */
+ else if (last + 1 != c && in_range)
+ {
+ putchar (last);
+ in_range = 0;
+ }
+
+ if (! in_range)
+ putchar (c);
+
+ last = c;
+ }
+
+ if (in_range)
+ putchar (last);
+
+ putchar (']');
+
+ p += 1 + *p;
+# endif /* WCHAR */
+ }
+ break;
+
+ case begline:
+ printf ("/begline");
+ break;
+
+ case endline:
+ printf ("/endline");
+ break;
+
+ case on_failure_jump:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/on_failure_jump to %td", p + mcnt - start);
+# else
+ printf ("/on_failure_jump to %ld", (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case on_failure_keep_string_jump:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/on_failure_keep_string_jump to %td", p + mcnt - start);
+# else
+ printf ("/on_failure_keep_string_jump to %ld",
+ (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case dummy_failure_jump:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/dummy_failure_jump to %td", p + mcnt - start);
+# else
+ printf ("/dummy_failure_jump to %ld", (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case push_dummy_failure:
+ printf ("/push_dummy_failure");
+ break;
+
+ case maybe_pop_jump:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/maybe_pop_jump to %td", p + mcnt - start);
+# else
+ printf ("/maybe_pop_jump to %ld", (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case pop_failure_jump:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/pop_failure_jump to %td", p + mcnt - start);
+# else
+ printf ("/pop_failure_jump to %ld", (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case jump_past_alt:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/jump_past_alt to %td", p + mcnt - start);
+# else
+ printf ("/jump_past_alt to %ld", (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case jump:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+# ifdef _LIBC
+ printf ("/jump to %td", p + mcnt - start);
+# else
+ printf ("/jump to %ld", (long int) (p + mcnt - start));
+# endif
+ break;
+
+ case succeed_n:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+ p1 = p + mcnt;
+ PREFIX(extract_number_and_incr) (&mcnt2, &p);
+# ifdef _LIBC
+ printf ("/succeed_n to %td, %d times", p1 - start, mcnt2);
+# else
+ printf ("/succeed_n to %ld, %d times",
+ (long int) (p1 - start), mcnt2);
+# endif
+ break;
+
+ case jump_n:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+ p1 = p + mcnt;
+ PREFIX(extract_number_and_incr) (&mcnt2, &p);
+ printf ("/jump_n to %d, %d times", p1 - start, mcnt2);
+ break;
+
+ case set_number_at:
+ PREFIX(extract_number_and_incr) (&mcnt, &p);
+ p1 = p + mcnt;
+ PREFIX(extract_number_and_incr) (&mcnt2, &p);
+# ifdef _LIBC
+ printf ("/set_number_at location %td to %d", p1 - start, mcnt2);
+# else
+ printf ("/set_number_at location %ld to %d",
+ (long int) (p1 - start), mcnt2);
+# endif
+ break;
+
+ case wordbound:
+ printf ("/wordbound");
+ break;
+
+ case notwordbound:
+ printf ("/notwordbound");
+ break;
+
+ case wordbeg:
+ printf ("/wordbeg");
+ break;
+
+ case wordend:
+ printf ("/wordend");
+ break;
+
+# ifdef emacs
+ case before_dot:
+ printf ("/before_dot");
+ break;
+
+ case at_dot:
+ printf ("/at_dot");
+ break;
+
+ case after_dot:
+ printf ("/after_dot");
+ break;
+
+ case syntaxspec:
+ printf ("/syntaxspec");
+ mcnt = *p++;
+ printf ("/%d", mcnt);
+ break;
+
+ case notsyntaxspec:
+ printf ("/notsyntaxspec");
+ mcnt = *p++;
+ printf ("/%d", mcnt);
+ break;
+# endif /* emacs */
+
+ case wordchar:
+ printf ("/wordchar");
+ break;
+
+ case notwordchar:
+ printf ("/notwordchar");
+ break;
+
+ case begbuf:
+ printf ("/begbuf");
+ break;
+
+ case endbuf:
+ printf ("/endbuf");
+ break;
+
+ default:
+ printf ("?%ld", (long int) *(p-1));
+ }
+
+ putchar ('\n');
+ }
+
+# ifdef _LIBC
+ printf ("%td:\tend of pattern.\n", p - start);
+# else
+ printf ("%ld:\tend of pattern.\n", (long int) (p - start));
+# endif
+}
+
+
+void
+PREFIX(print_compiled_pattern) (struct re_pattern_buffer *bufp)
+{
+ UCHAR_T *buffer = (UCHAR_T*) bufp->buffer;
+
+ PREFIX(print_partial_compiled_pattern) (buffer, buffer
+ + bufp->used / sizeof(UCHAR_T));
+ printf ("%ld bytes used/%ld bytes allocated.\n",
+ bufp->used, bufp->allocated);
+
+ if (bufp->fastmap_accurate && bufp->fastmap)
+ {
+ printf ("fastmap: ");
+ print_fastmap (bufp->fastmap);
+ }
+
+# ifdef _LIBC
+ printf ("re_nsub: %Zd\t", bufp->re_nsub);
+# else
+ printf ("re_nsub: %ld\t", (long int) bufp->re_nsub);
+# endif
+ printf ("regs_alloc: %d\t", bufp->regs_allocated);
+ printf ("can_be_null: %d\t", bufp->can_be_null);
+ printf ("newline_anchor: %d\n", bufp->newline_anchor);
+ printf ("no_sub: %d\t", bufp->no_sub);
+ printf ("not_bol: %d\t", bufp->not_bol);
+ printf ("not_eol: %d\t", bufp->not_eol);
+ printf ("syntax: %lx\n", bufp->syntax);
+ /* Perhaps we should print the translate table? */
+}
+
+
+void
+PREFIX(print_double_string) (const CHAR_T *where,
+ const CHAR_T *string1,
+ const CHAR_T *string2,
+ int size1,
+ int size2)
+{
+ int this_char;
+
+ if (where == NULL)
+ printf ("(null)");
+ else
+ {
+ int cnt;
+
+ if (FIRST_STRING_P (where))
+ {
+ for (this_char = where - string1; this_char < size1; this_char++)
+ PUT_CHAR (string1[this_char]);
+
+ where = string2;
+ }
+
+ cnt = 0;
+ for (this_char = where - string2; this_char < size2; this_char++)
+ {
+ PUT_CHAR (string2[this_char]);
+ if (++cnt > 100)
+ {
+ fputs ("...", stdout);
+ break;
+ }
+ }
+ }
+}
+
+# ifndef DEFINED_ONCE
+void
+printchar (c)
+ int c;
+{
+ putc (c, stderr);
+}
+# endif
+
+# else /* not DEBUG */
+
+# ifndef DEFINED_ONCE
+# undef assert
+# define assert(e)
+
+# define DEBUG_STATEMENT(e)
+# define DEBUG_PRINT1(x)
+# define DEBUG_PRINT2(x1, x2)
+# define DEBUG_PRINT3(x1, x2, x3)
+# define DEBUG_PRINT4(x1, x2, x3, x4)
+# endif /* not DEFINED_ONCE */
+# define DEBUG_PRINT_COMPILED_PATTERN(p, s, e)
+# define DEBUG_PRINT_DOUBLE_STRING(w, s1, sz1, s2, sz2)
+
+# endif /* not DEBUG */
+
+
+
+# ifdef WCHAR
+/* This convert a multibyte string to a wide character string.
+ And write their correspondances to offset_buffer(see below)
+ and write whether each wchar_t is binary data to is_binary.
+ This assume invalid multibyte sequences as binary data.
+ We assume offset_buffer and is_binary is already allocated
+ enough space. */
+
+static size_t
+convert_mbs_to_wcs (CHAR_T *dest,
+ const unsigned char* src,
+
+ /* The length of multibyte string. */
+ size_t len,
+
+ /* Correspondences between src(char string) and
+ dest(wchar_t string) for optimization. E.g.:
+ src = "xxxyzz"
+ dest = {'X', 'Y', 'Z'}
+ (each "xxx", "y" and "zz" represent one
+ multibyte character corresponding to 'X',
+ 'Y' and 'Z'.)
+ offset_buffer = {0, 0+3("xxx"), 0+3+1("y"),
+ 0+3+1+2("zz")}
+ = {0, 3, 4, 6} */
+ int *offset_buffer,
+
+ char *is_binary)
+{
+ wchar_t *pdest = dest;
+ const unsigned char *psrc = src;
+ size_t wc_count = 0;
+
+ mbstate_t mbs;
+ int i, consumed;
+ size_t mb_remain = len;
+ size_t mb_count = 0;
+
+ /* Initialize the conversion state. */
+ memset (&mbs, 0, sizeof (mbstate_t));
+
+ offset_buffer[0] = 0;
+ for( ; mb_remain > 0 ; ++wc_count, ++pdest, mb_remain -= consumed,
+ psrc += consumed)
+ {
+ consumed = mbrtowc (pdest, psrc, mb_remain, &mbs);
+
+ if (consumed <= 0)
+ /* failed to convert. maybe src contains binary data.
+ So we consume 1 byte manualy. */
+ {
+ *pdest = *psrc;
+ consumed = 1;
+ is_binary[wc_count] = TRUE;
+ }
+ else
+ is_binary[wc_count] = FALSE;
+ /* In sjis encoding, we use yen sign as escape character in
+ place of reverse solidus. So we convert 0x5c(yen sign in
+ sjis) to not 0xa5(yen sign in UCS2) but 0x5c(reverse
+ solidus in UCS2). */
+ if (consumed == 1 && (int) *psrc == 0x5c && (int) *pdest == 0xa5)
+ *pdest = (wchar_t) *psrc;
+
+ offset_buffer[wc_count + 1] = mb_count += consumed;
+ }
+
+ /* Fill remain of the buffer with sentinel. */
+ for (i = wc_count + 1 ; i <= len ; i++)
+ offset_buffer[i] = mb_count + 1;
+
+ return wc_count;
+}
+
+# endif /* WCHAR */
+
+#else /* not INSIDE_RECURSION */
+
+/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can
+ also be assigned to arbitrarily: each pattern buffer stores its own
+ syntax, so it can be changed between regex compilations. */
+/* This has no initializer because initialized variables in Emacs
+ become read-only after dumping. */
+reg_syntax_t re_syntax_options;
+
+
+/* Specify the precise syntax of regexps for compilation. This provides
+ for compatibility for various utilities which historically have
+ different, incompatible syntaxes.
+
+ The argument SYNTAX is a bit mask comprised of the various bits
+ defined in regex.h. We return the old syntax. */
+
+reg_syntax_t
+re_set_syntax (reg_syntax_t syntax)
+{
+ reg_syntax_t ret = re_syntax_options;
+
+ re_syntax_options = syntax;
+# ifdef DEBUG
+ if (syntax & RE_DEBUG)
+ debug = 1;
+ else if (debug) /* was on but now is not */
+ debug = 0;
+# endif /* DEBUG */
+ return ret;
+}
+# ifdef _LIBC
+weak_alias (__re_set_syntax, re_set_syntax)
+# endif
+
+/* This table gives an error message for each of the error codes listed
+ in regex.h. Obviously the order here has to be same as there.
+ POSIX doesn't require that we do anything for REG_NOERROR,
+ but why not be nice? */
+
+static const char re_error_msgid[] =
+ {
+# define REG_NOERROR_IDX 0
+ gettext_noop ("Success") /* REG_NOERROR */
+ "\0"
+# define REG_NOMATCH_IDX (REG_NOERROR_IDX + sizeof "Success")
+ gettext_noop ("No match") /* REG_NOMATCH */
+ "\0"
+# define REG_BADPAT_IDX (REG_NOMATCH_IDX + sizeof "No match")
+ gettext_noop ("Invalid regular expression") /* REG_BADPAT */
+ "\0"
+# define REG_ECOLLATE_IDX (REG_BADPAT_IDX + sizeof "Invalid regular expression")
+ gettext_noop ("Invalid collation character") /* REG_ECOLLATE */
+ "\0"
+# define REG_ECTYPE_IDX (REG_ECOLLATE_IDX + sizeof "Invalid collation character")
+ gettext_noop ("Invalid character class name") /* REG_ECTYPE */
+ "\0"
+# define REG_EESCAPE_IDX (REG_ECTYPE_IDX + sizeof "Invalid character class name")
+ gettext_noop ("Trailing backslash") /* REG_EESCAPE */
+ "\0"
+# define REG_ESUBREG_IDX (REG_EESCAPE_IDX + sizeof "Trailing backslash")
+ gettext_noop ("Invalid back reference") /* REG_ESUBREG */
+ "\0"
+# define REG_EBRACK_IDX (REG_ESUBREG_IDX + sizeof "Invalid back reference")
+ gettext_noop ("Unmatched [ or [^") /* REG_EBRACK */
+ "\0"
+# define REG_EPAREN_IDX (REG_EBRACK_IDX + sizeof "Unmatched [ or [^")
+ gettext_noop ("Unmatched ( or \\(") /* REG_EPAREN */
+ "\0"
+# define REG_EBRACE_IDX (REG_EPAREN_IDX + sizeof "Unmatched ( or \\(")
+ gettext_noop ("Unmatched \\{") /* REG_EBRACE */
+ "\0"
+# define REG_BADBR_IDX (REG_EBRACE_IDX + sizeof "Unmatched \\{")
+ gettext_noop ("Invalid content of \\{\\}") /* REG_BADBR */
+ "\0"
+# define REG_ERANGE_IDX (REG_BADBR_IDX + sizeof "Invalid content of \\{\\}")
+ gettext_noop ("Invalid range end") /* REG_ERANGE */
+ "\0"
+# define REG_ESPACE_IDX (REG_ERANGE_IDX + sizeof "Invalid range end")
+ gettext_noop ("Memory exhausted") /* REG_ESPACE */
+ "\0"
+# define REG_BADRPT_IDX (REG_ESPACE_IDX + sizeof "Memory exhausted")
+ gettext_noop ("Invalid preceding regular expression") /* REG_BADRPT */
+ "\0"
+# define REG_EEND_IDX (REG_BADRPT_IDX + sizeof "Invalid preceding regular expression")
+ gettext_noop ("Premature end of regular expression") /* REG_EEND */
+ "\0"
+# define REG_ESIZE_IDX (REG_EEND_IDX + sizeof "Premature end of regular expression")
+ gettext_noop ("Regular expression too big") /* REG_ESIZE */
+ "\0"
+# define REG_ERPAREN_IDX (REG_ESIZE_IDX + sizeof "Regular expression too big")
+ gettext_noop ("Unmatched ) or \\)") /* REG_ERPAREN */
+ };
+
+static const size_t re_error_msgid_idx[] =
+ {
+ REG_NOERROR_IDX,
+ REG_NOMATCH_IDX,
+ REG_BADPAT_IDX,
+ REG_ECOLLATE_IDX,
+ REG_ECTYPE_IDX,
+ REG_EESCAPE_IDX,
+ REG_ESUBREG_IDX,
+ REG_EBRACK_IDX,
+ REG_EPAREN_IDX,
+ REG_EBRACE_IDX,
+ REG_BADBR_IDX,
+ REG_ERANGE_IDX,
+ REG_ESPACE_IDX,
+ REG_BADRPT_IDX,
+ REG_EEND_IDX,
+ REG_ESIZE_IDX,
+ REG_ERPAREN_IDX
+ };
+
+#endif /* INSIDE_RECURSION */
+
+#ifndef DEFINED_ONCE
+/* Avoiding alloca during matching, to placate r_alloc. */
+
+/* Define MATCH_MAY_ALLOCATE unless we need to make sure that the
+ searching and matching functions should not call alloca. On some
+ systems, alloca is implemented in terms of malloc, and if we're
+ using the relocating allocator routines, then malloc could cause a
+ relocation, which might (if the strings being searched are in the
+ ralloc heap) shift the data out from underneath the regexp
+ routines.
+
+ Here's another reason to avoid allocation: Emacs
+ processes input from X in a signal handler; processing X input may
+ call malloc; if input arrives while a matching routine is calling
+ malloc, then we're scrod. But Emacs can't just block input while
+ calling matching routines; then we don't notice interrupts when
+ they come in. So, Emacs blocks input around all regexp calls
+ except the matching calls, which it leaves unprotected, in the
+ faith that they will not malloc. */
+
+/* Normally, this is fine. */
+# define MATCH_MAY_ALLOCATE
+
+/* When using GNU C, we are not REALLY using the C alloca, no matter
+ what config.h may say. So don't take precautions for it. */
+# ifdef __GNUC__
+# undef C_ALLOCA
+# endif
+
+/* The match routines may not allocate if (1) they would do it with malloc
+ and (2) it's not safe for them to use malloc.
+ Note that if REL_ALLOC is defined, matching would not use malloc for the
+ failure stack, but we would still use it for the register vectors;
+ so REL_ALLOC should not affect this. */
+# if (defined C_ALLOCA || defined REGEX_MALLOC) && defined emacs
+# undef MATCH_MAY_ALLOCATE
+# endif
+#endif /* not DEFINED_ONCE */
+
+#ifdef INSIDE_RECURSION
+/* Failure stack declarations and macros; both re_compile_fastmap and
+ re_match_2 use a failure stack. These have to be macros because of
+ REGEX_ALLOCATE_STACK. */
+
+
+/* Number of failure points for which to initially allocate space
+ when matching. If this number is exceeded, we allocate more
+ space, so it is not a hard limit. */
+# ifndef INIT_FAILURE_ALLOC
+# define INIT_FAILURE_ALLOC 5
+# endif
+
+/* Roughly the maximum number of failure points on the stack. Would be
+ exactly that if always used MAX_FAILURE_ITEMS items each time we failed.
+ This is a variable only so users of regex can assign to it; we never
+ change it ourselves. */
+
+# ifdef INT_IS_16BIT
+
+# ifndef DEFINED_ONCE
+# if defined MATCH_MAY_ALLOCATE
+/* 4400 was enough to cause a crash on Alpha OSF/1,
+ whose default stack limit is 2mb. */
+long int re_max_failures = 4000;
+# else
+long int re_max_failures = 2000;
+# endif
+# endif
+
+union PREFIX(fail_stack_elt)
+{
+ UCHAR_T *pointer;
+ long int integer;
+};
+
+typedef union PREFIX(fail_stack_elt) PREFIX(fail_stack_elt_t);
+
+typedef struct
+{
+ PREFIX(fail_stack_elt_t) *stack;
+ unsigned long int size;
+ unsigned long int avail; /* Offset of next open position. */
+} PREFIX(fail_stack_type);
+
+# else /* not INT_IS_16BIT */
+
+# ifndef DEFINED_ONCE
+# if defined MATCH_MAY_ALLOCATE
+/* 4400 was enough to cause a crash on Alpha OSF/1,
+ whose default stack limit is 2mb. */
+int re_max_failures = 4000;
+# else
+int re_max_failures = 2000;
+# endif
+# endif
+
+union PREFIX(fail_stack_elt)
+{
+ UCHAR_T *pointer;
+ int integer;
+};
+
+typedef union PREFIX(fail_stack_elt) PREFIX(fail_stack_elt_t);
+
+typedef struct
+{
+ PREFIX(fail_stack_elt_t) *stack;
+ unsigned size;
+ unsigned avail; /* Offset of next open position. */
+} PREFIX(fail_stack_type);
+
+# endif /* INT_IS_16BIT */
+
+# ifndef DEFINED_ONCE
+# define FAIL_STACK_EMPTY() (fail_stack.avail == 0)
+# define FAIL_STACK_PTR_EMPTY() (fail_stack_ptr->avail == 0)
+# define FAIL_STACK_FULL() (fail_stack.avail == fail_stack.size)
+# endif
+
+
+/* Define macros to initialize and free the failure stack.
+ Do `return -2' if the alloc fails. */
+
+# ifdef MATCH_MAY_ALLOCATE
+# define INIT_FAIL_STACK() \
+ do { \
+ fail_stack.stack = (PREFIX(fail_stack_elt_t) *) \
+ REGEX_ALLOCATE_STACK (INIT_FAILURE_ALLOC * sizeof (PREFIX(fail_stack_elt_t))); \
+ \
+ if (fail_stack.stack == NULL) \
+ return -2; \
+ \
+ fail_stack.size = INIT_FAILURE_ALLOC; \
+ fail_stack.avail = 0; \
+ } while (0)
+
+# define RESET_FAIL_STACK() REGEX_FREE_STACK (fail_stack.stack)
+# else
+# define INIT_FAIL_STACK() \
+ do { \
+ fail_stack.avail = 0; \
+ } while (0)
+
+# define RESET_FAIL_STACK()
+# endif
+
+
+/* Double the size of FAIL_STACK, up to approximately `re_max_failures' items.
+
+ Return 1 if succeeds, and 0 if either ran out of memory
+ allocating space for it or it was already too large.
+
+ REGEX_REALLOCATE_STACK requires `destination' be declared. */
+
+# define DOUBLE_FAIL_STACK(fail_stack) \
+ ((fail_stack).size > (unsigned) (re_max_failures * MAX_FAILURE_ITEMS) \
+ ? 0 \
+ : ((fail_stack).stack = (PREFIX(fail_stack_elt_t) *) \
+ REGEX_REALLOCATE_STACK ((fail_stack).stack, \
+ (fail_stack).size * sizeof (PREFIX(fail_stack_elt_t)), \
+ ((fail_stack).size << 1) * sizeof (PREFIX(fail_stack_elt_t))),\
+ \
+ (fail_stack).stack == NULL \
+ ? 0 \
+ : ((fail_stack).size <<= 1, \
+ 1)))
+
+
+/* Push pointer POINTER on FAIL_STACK.
+ Return 1 if was able to do so and 0 if ran out of memory allocating
+ space to do so. */
+# define PUSH_PATTERN_OP(POINTER, FAIL_STACK) \
+ ((FAIL_STACK_FULL () \
+ && !DOUBLE_FAIL_STACK (FAIL_STACK)) \
+ ? 0 \
+ : ((FAIL_STACK).stack[(FAIL_STACK).avail++].pointer = POINTER, \
+ 1))
+
+/* Push a pointer value onto the failure stack.
+ Assumes the variable `fail_stack'. Probably should only
+ be called from within `PUSH_FAILURE_POINT'. */
+# define PUSH_FAILURE_POINTER(item) \
+ fail_stack.stack[fail_stack.avail++].pointer = (UCHAR_T *) (item)
+
+/* This pushes an integer-valued item onto the failure stack.
+ Assumes the variable `fail_stack'. Probably should only
+ be called from within `PUSH_FAILURE_POINT'. */
+# define PUSH_FAILURE_INT(item) \
+ fail_stack.stack[fail_stack.avail++].integer = (item)
+
+/* Push a fail_stack_elt_t value onto the failure stack.
+ Assumes the variable `fail_stack'. Probably should only
+ be called from within `PUSH_FAILURE_POINT'. */
+# define PUSH_FAILURE_ELT(item) \
+ fail_stack.stack[fail_stack.avail++] = (item)
+
+/* These three POP... operations complement the three PUSH... operations.
+ All assume that `fail_stack' is nonempty. */
+# define POP_FAILURE_POINTER() fail_stack.stack[--fail_stack.avail].pointer
+# define POP_FAILURE_INT() fail_stack.stack[--fail_stack.avail].integer
+# define POP_FAILURE_ELT() fail_stack.stack[--fail_stack.avail]
+
+/* Used to omit pushing failure point id's when we're not debugging. */
+# ifdef DEBUG
+# define DEBUG_PUSH PUSH_FAILURE_INT
+# define DEBUG_POP(item_addr) *(item_addr) = POP_FAILURE_INT ()
+# else
+# define DEBUG_PUSH(item)
+# define DEBUG_POP(item_addr)
+# endif
+
+
+/* Push the information about the state we will need
+ if we ever fail back to it.
+
+ Requires variables fail_stack, regstart, regend, reg_info, and
+ num_regs_pushed be declared. DOUBLE_FAIL_STACK requires `destination'
+ be declared.
+
+ Does `return FAILURE_CODE' if runs out of memory. */
+
+# define PUSH_FAILURE_POINT(pattern_place, string_place, failure_code) \
+ do { \
+ char *destination; \
+ /* Must be int, so when we don't save any registers, the arithmetic \
+ of 0 + -1 isn't done as unsigned. */ \
+ /* Can't be int, since there is not a shred of a guarantee that int \
+ is wide enough to hold a value of something to which pointer can \
+ be assigned */ \
+ active_reg_t this_reg; \
+ \
+ DEBUG_STATEMENT (failure_id++); \
+ DEBUG_STATEMENT (nfailure_points_pushed++); \
+ DEBUG_PRINT2 ("\nPUSH_FAILURE_POINT #%u:\n", failure_id); \
+ DEBUG_PRINT2 (" Before push, next avail: %d\n", (fail_stack).avail);\
+ DEBUG_PRINT2 (" size: %d\n", (fail_stack).size);\
+ \
+ DEBUG_PRINT2 (" slots needed: %ld\n", NUM_FAILURE_ITEMS); \
+ DEBUG_PRINT2 (" available: %d\n", REMAINING_AVAIL_SLOTS); \
+ \
+ /* Ensure we have enough space allocated for what we will push. */ \
+ while (REMAINING_AVAIL_SLOTS < NUM_FAILURE_ITEMS) \
+ { \
+ if (!DOUBLE_FAIL_STACK (fail_stack)) \
+ return failure_code; \
+ \
+ DEBUG_PRINT2 ("\n Doubled stack; size now: %d\n", \
+ (fail_stack).size); \
+ DEBUG_PRINT2 (" slots available: %d\n", REMAINING_AVAIL_SLOTS);\
+ } \
+ \
+ /* Push the info, starting with the registers. */ \
+ DEBUG_PRINT1 ("\n"); \
+ \
+ if (1) \
+ for (this_reg = lowest_active_reg; this_reg <= highest_active_reg; \
+ this_reg++) \
+ { \
+ DEBUG_PRINT2 (" Pushing reg: %lu\n", this_reg); \
+ DEBUG_STATEMENT (num_regs_pushed++); \
+ \
+ DEBUG_PRINT2 (" start: %p\n", regstart[this_reg]); \
+ PUSH_FAILURE_POINTER (regstart[this_reg]); \
+ \
+ DEBUG_PRINT2 (" end: %p\n", regend[this_reg]); \
+ PUSH_FAILURE_POINTER (regend[this_reg]); \
+ \
+ DEBUG_PRINT2 (" info: %p\n ", \
+ reg_info[this_reg].word.pointer); \
+ DEBUG_PRINT2 (" match_null=%d", \
+ REG_MATCH_NULL_STRING_P (reg_info[this_reg])); \
+ DEBUG_PRINT2 (" active=%d", IS_ACTIVE (reg_info[this_reg])); \
+ DEBUG_PRINT2 (" matched_something=%d", \
+ MATCHED_SOMETHING (reg_info[this_reg])); \
+ DEBUG_PRINT2 (" ever_matched=%d", \
+ EVER_MATCHED_SOMETHING (reg_info[this_reg])); \
+ DEBUG_PRINT1 ("\n"); \
+ PUSH_FAILURE_ELT (reg_info[this_reg].word); \
+ } \
+ \
+ DEBUG_PRINT2 (" Pushing low active reg: %ld\n", lowest_active_reg);\
+ PUSH_FAILURE_INT (lowest_active_reg); \
+ \
+ DEBUG_PRINT2 (" Pushing high active reg: %ld\n", highest_active_reg);\
+ PUSH_FAILURE_INT (highest_active_reg); \
+ \
+ DEBUG_PRINT2 (" Pushing pattern %p:\n", pattern_place); \
+ DEBUG_PRINT_COMPILED_PATTERN (bufp, pattern_place, pend); \
+ PUSH_FAILURE_POINTER (pattern_place); \
+ \
+ DEBUG_PRINT2 (" Pushing string %p: `", string_place); \
+ DEBUG_PRINT_DOUBLE_STRING (string_place, string1, size1, string2, \
+ size2); \
+ DEBUG_PRINT1 ("'\n"); \
+ PUSH_FAILURE_POINTER (string_place); \
+ \
+ DEBUG_PRINT2 (" Pushing failure id: %u\n", failure_id); \
+ DEBUG_PUSH (failure_id); \
+ } while (0)
+
+# ifndef DEFINED_ONCE
+/* This is the number of items that are pushed and popped on the stack
+ for each register. */
+# define NUM_REG_ITEMS 3
+
+/* Individual items aside from the registers. */
+# ifdef DEBUG
+# define NUM_NONREG_ITEMS 5 /* Includes failure point id. */
+# else
+# define NUM_NONREG_ITEMS 4
+# endif
+
+/* We push at most this many items on the stack. */
+/* We used to use (num_regs - 1), which is the number of registers
+ this regexp will save; but that was changed to 5
+ to avoid stack overflow for a regexp with lots of parens. */
+# define MAX_FAILURE_ITEMS (5 * NUM_REG_ITEMS + NUM_NONREG_ITEMS)
+
+/* We actually push this many items. */
+# define NUM_FAILURE_ITEMS \
+ (((0 \
+ ? 0 : highest_active_reg - lowest_active_reg + 1) \
+ * NUM_REG_ITEMS) \
+ + NUM_NONREG_ITEMS)
+
+/* How many items can still be added to the stack without overflowing it. */
+# define REMAINING_AVAIL_SLOTS ((fail_stack).size - (fail_stack).avail)
+# endif /* not DEFINED_ONCE */
+
+
+/* Pops what PUSH_FAIL_STACK pushes.
+
+ We restore into the parameters, all of which should be lvalues:
+ STR -- the saved data position.
+ PAT -- the saved pattern position.
+ LOW_REG, HIGH_REG -- the highest and lowest active registers.
+ REGSTART, REGEND -- arrays of string positions.
+ REG_INFO -- array of information about each subexpression.
+
+ Also assumes the variables `fail_stack' and (if debugging), `bufp',
+ `pend', `string1', `size1', `string2', and `size2'. */
+# define POP_FAILURE_POINT(str, pat, low_reg, high_reg, regstart, regend, reg_info)\
+{ \
+ DEBUG_STATEMENT (unsigned failure_id;) \
+ active_reg_t this_reg; \
+ const UCHAR_T *string_temp; \
+ \
+ assert (!FAIL_STACK_EMPTY ()); \
+ \
+ /* Remove failure points and point to how many regs pushed. */ \
+ DEBUG_PRINT1 ("POP_FAILURE_POINT:\n"); \
+ DEBUG_PRINT2 (" Before pop, next avail: %d\n", fail_stack.avail); \
+ DEBUG_PRINT2 (" size: %d\n", fail_stack.size); \
+ \
+ assert (fail_stack.avail >= NUM_NONREG_ITEMS); \
+ \
+ DEBUG_POP (&failure_id); \
+ DEBUG_PRINT2 (" Popping failure id: %u\n", failure_id); \
+ \
+ /* If the saved string location is NULL, it came from an \
+ on_failure_keep_string_jump opcode, and we want to throw away the \
+ saved NULL, thus retaining our current position in the string. */ \
+ string_temp = POP_FAILURE_POINTER (); \
+ if (string_temp != NULL) \
+ str = (const CHAR_T *) string_temp; \
+ \
+ DEBUG_PRINT2 (" Popping string %p: `", str); \
+ DEBUG_PRINT_DOUBLE_STRING (str, string1, size1, string2, size2); \
+ DEBUG_PRINT1 ("'\n"); \
+ \
+ pat = (UCHAR_T *) POP_FAILURE_POINTER (); \
+ DEBUG_PRINT2 (" Popping pattern %p:\n", pat); \
+ DEBUG_PRINT_COMPILED_PATTERN (bufp, pat, pend); \
+ \
+ /* Restore register info. */ \
+ high_reg = (active_reg_t) POP_FAILURE_INT (); \
+ DEBUG_PRINT2 (" Popping high active reg: %ld\n", high_reg); \
+ \
+ low_reg = (active_reg_t) POP_FAILURE_INT (); \
+ DEBUG_PRINT2 (" Popping low active reg: %ld\n", low_reg); \
+ \
+ if (1) \
+ for (this_reg = high_reg; this_reg >= low_reg; this_reg--) \
+ { \
+ DEBUG_PRINT2 (" Popping reg: %ld\n", this_reg); \
+ \
+ reg_info[this_reg].word = POP_FAILURE_ELT (); \
+ DEBUG_PRINT2 (" info: %p\n", \
+ reg_info[this_reg].word.pointer); \
+ \
+ regend[this_reg] = (const CHAR_T *) POP_FAILURE_POINTER (); \
+ DEBUG_PRINT2 (" end: %p\n", regend[this_reg]); \
+ \
+ regstart[this_reg] = (const CHAR_T *) POP_FAILURE_POINTER (); \
+ DEBUG_PRINT2 (" start: %p\n", regstart[this_reg]); \
+ } \
+ else \
+ { \
+ for (this_reg = highest_active_reg; this_reg > high_reg; this_reg--) \
+ { \
+ reg_info[this_reg].word.integer = 0; \
+ regend[this_reg] = 0; \
+ regstart[this_reg] = 0; \
+ } \
+ highest_active_reg = high_reg; \
+ } \
+ \
+ set_regs_matched_done = 0; \
+ DEBUG_STATEMENT (nfailure_points_popped++); \
+} /* POP_FAILURE_POINT */
+
+/* Structure for per-register (a.k.a. per-group) information.
+ Other register information, such as the
+ starting and ending positions (which are addresses), and the list of
+ inner groups (which is a bits list) are maintained in separate
+ variables.
+
+ We are making a (strictly speaking) nonportable assumption here: that
+ the compiler will pack our bit fields into something that fits into
+ the type of `word', i.e., is something that fits into one item on the
+ failure stack. */
+
+
+/* Declarations and macros for re_match_2. */
+
+typedef union
+{
+ PREFIX(fail_stack_elt_t) word;
+ struct
+ {
+ /* This field is one if this group can match the empty string,
+ zero if not. If not yet determined, `MATCH_NULL_UNSET_VALUE'. */
+# define MATCH_NULL_UNSET_VALUE 3
+ unsigned match_null_string_p : 2;
+ unsigned is_active : 1;
+ unsigned matched_something : 1;
+ unsigned ever_matched_something : 1;
+ } bits;
+} PREFIX(register_info_type);
+
+# ifndef DEFINED_ONCE
+# define REG_MATCH_NULL_STRING_P(R) ((R).bits.match_null_string_p)
+# define IS_ACTIVE(R) ((R).bits.is_active)
+# define MATCHED_SOMETHING(R) ((R).bits.matched_something)
+# define EVER_MATCHED_SOMETHING(R) ((R).bits.ever_matched_something)
+
+
+/* Call this when have matched a real character; it sets `matched' flags
+ for the subexpressions which we are currently inside. Also records
+ that those subexprs have matched. */
+# define SET_REGS_MATCHED() \
+ do \
+ { \
+ if (!set_regs_matched_done) \
+ { \
+ active_reg_t r; \
+ set_regs_matched_done = 1; \
+ for (r = lowest_active_reg; r <= highest_active_reg; r++) \
+ { \
+ MATCHED_SOMETHING (reg_info[r]) \
+ = EVER_MATCHED_SOMETHING (reg_info[r]) \
+ = 1; \
+ } \
+ } \
+ } \
+ while (0)
+# endif /* not DEFINED_ONCE */
+
+/* Registers are set to a sentinel when they haven't yet matched. */
+static CHAR_T PREFIX(reg_unset_dummy);
+# define REG_UNSET_VALUE (&PREFIX(reg_unset_dummy))
+# define REG_UNSET(e) ((e) == REG_UNSET_VALUE)
+
+/* Subroutine declarations and macros for regex_compile. */
+static void PREFIX(store_op1) (re_opcode_t op, UCHAR_T *loc, int arg);
+static void PREFIX(store_op2) (re_opcode_t op, UCHAR_T *loc,
+ int arg1, int arg2);
+static void PREFIX(insert_op1) (re_opcode_t op, UCHAR_T *loc,
+ int arg, UCHAR_T *end);
+static void PREFIX(insert_op2) (re_opcode_t op, UCHAR_T *loc,
+ int arg1, int arg2, UCHAR_T *end);
+static boolean PREFIX(at_begline_loc_p) (const CHAR_T *pattern,
+ const CHAR_T *p,
+ reg_syntax_t syntax);
+static boolean PREFIX(at_endline_loc_p) (const CHAR_T *p,
+ const CHAR_T *pend,
+ reg_syntax_t syntax);
+# ifdef WCHAR
+static reg_errcode_t wcs_compile_range (CHAR_T range_start,
+ const CHAR_T **p_ptr,
+ const CHAR_T *pend,
+ char *translate,
+ reg_syntax_t syntax,
+ UCHAR_T *b,
+ CHAR_T *char_set);
+static void insert_space (int num, CHAR_T *loc, CHAR_T *end);
+# else /* BYTE */
+static reg_errcode_t byte_compile_range (unsigned int range_start,
+ const char **p_ptr,
+ const char *pend,
+ char *translate,
+ reg_syntax_t syntax,
+ unsigned char *b);
+# endif /* WCHAR */
+
+/* Fetch the next character in the uncompiled pattern---translating it
+ if necessary. Also cast from a signed character in the constant
+ string passed to us by the user to an unsigned char that we can use
+ as an array index (in, e.g., `translate'). */
+/* ifdef MBS_SUPPORT, we translate only if character <= 0xff,
+ because it is impossible to allocate 4GB array for some encodings
+ which have 4 byte character_set like UCS4. */
+# ifndef PATFETCH
+# ifdef WCHAR
+# define PATFETCH(c) \
+ do {if (p == pend) return REG_EEND; \
+ c = (UCHAR_T) *p++; \
+ if (translate && (c <= 0xff)) c = (UCHAR_T) translate[c]; \
+ } while (0)
+# else /* BYTE */
+# define PATFETCH(c) \
+ do {if (p == pend) return REG_EEND; \
+ c = (unsigned char) *p++; \
+ if (translate) c = (unsigned char) translate[c]; \
+ } while (0)
+# endif /* WCHAR */
+# endif
+
+/* Fetch the next character in the uncompiled pattern, with no
+ translation. */
+# define PATFETCH_RAW(c) \
+ do {if (p == pend) return REG_EEND; \
+ c = (UCHAR_T) *p++; \
+ } while (0)
+
+/* Go backwards one character in the pattern. */
+# define PATUNFETCH p--
+
+
+/* If `translate' is non-null, return translate[D], else just D. We
+ cast the subscript to translate because some data is declared as
+ `char *', to avoid warnings when a string constant is passed. But
+ when we use a character as a subscript we must make it unsigned. */
+/* ifdef MBS_SUPPORT, we translate only if character <= 0xff,
+ because it is impossible to allocate 4GB array for some encodings
+ which have 4 byte character_set like UCS4. */
+
+# ifndef TRANSLATE
+# ifdef WCHAR
+# define TRANSLATE(d) \
+ ((translate && ((UCHAR_T) (d)) <= 0xff) \
+ ? (char) translate[(unsigned char) (d)] : (d))
+# else /* BYTE */
+# define TRANSLATE(d) \
+ (translate ? (char) translate[(unsigned char) (d)] : (d))
+# endif /* WCHAR */
+# endif
+
+
+/* Macros for outputting the compiled pattern into `buffer'. */
+
+/* If the buffer isn't allocated when it comes in, use this. */
+# define INIT_BUF_SIZE (32 * sizeof(UCHAR_T))
+
+/* Make sure we have at least N more bytes of space in buffer. */
+# ifdef WCHAR
+# define GET_BUFFER_SPACE(n) \
+ while (((unsigned long)b - (unsigned long)COMPILED_BUFFER_VAR \
+ + (n)*sizeof(CHAR_T)) > bufp->allocated) \
+ EXTEND_BUFFER ()
+# else /* BYTE */
+# define GET_BUFFER_SPACE(n) \
+ while ((unsigned long) (b - bufp->buffer + (n)) > bufp->allocated) \
+ EXTEND_BUFFER ()
+# endif /* WCHAR */
+
+/* Make sure we have one more byte of buffer space and then add C to it. */
+# define BUF_PUSH(c) \
+ do { \
+ GET_BUFFER_SPACE (1); \
+ *b++ = (UCHAR_T) (c); \
+ } while (0)
+
+
+/* Ensure we have two more bytes of buffer space and then append C1 and C2. */
+# define BUF_PUSH_2(c1, c2) \
+ do { \
+ GET_BUFFER_SPACE (2); \
+ *b++ = (UCHAR_T) (c1); \
+ *b++ = (UCHAR_T) (c2); \
+ } while (0)
+
+
+/* As with BUF_PUSH_2, except for three bytes. */
+# define BUF_PUSH_3(c1, c2, c3) \
+ do { \
+ GET_BUFFER_SPACE (3); \
+ *b++ = (UCHAR_T) (c1); \
+ *b++ = (UCHAR_T) (c2); \
+ *b++ = (UCHAR_T) (c3); \
+ } while (0)
+
+/* Store a jump with opcode OP at LOC to location TO. We store a
+ relative address offset by the three bytes the jump itself occupies. */
+# define STORE_JUMP(op, loc, to) \
+ PREFIX(store_op1) (op, loc, (int) ((to) - (loc) - (1 + OFFSET_ADDRESS_SIZE)))
+
+/* Likewise, for a two-argument jump. */
+# define STORE_JUMP2(op, loc, to, arg) \
+ PREFIX(store_op2) (op, loc, (int) ((to) - (loc) - (1 + OFFSET_ADDRESS_SIZE)), arg)
+
+/* Like `STORE_JUMP', but for inserting. Assume `b' is the buffer end. */
+# define INSERT_JUMP(op, loc, to) \
+ PREFIX(insert_op1) (op, loc, (int) ((to) - (loc) - (1 + OFFSET_ADDRESS_SIZE)), b)
+
+/* Like `STORE_JUMP2', but for inserting. Assume `b' is the buffer end. */
+# define INSERT_JUMP2(op, loc, to, arg) \
+ PREFIX(insert_op2) (op, loc, (int) ((to) - (loc) - (1 + OFFSET_ADDRESS_SIZE)),\
+ arg, b)
+
+/* This is not an arbitrary limit: the arguments which represent offsets
+ into the pattern are two bytes long. So if 2^16 bytes turns out to
+ be too small, many things would have to change. */
+/* Any other compiler which, like MSC, has allocation limit below 2^16
+ bytes will have to use approach similar to what was done below for
+ MSC and drop MAX_BUF_SIZE a bit. Otherwise you may end up
+ reallocating to 0 bytes. Such thing is not going to work too well.
+ You have been warned!! */
+# ifndef DEFINED_ONCE
+# if defined _MSC_VER && !defined WIN32
+/* Microsoft C 16-bit versions limit malloc to approx 65512 bytes.
+ The REALLOC define eliminates a flurry of conversion warnings,
+ but is not required. */
+# define MAX_BUF_SIZE 65500L
+# define REALLOC(p,s) realloc ((p), (size_t) (s))
+# else
+# define MAX_BUF_SIZE (1L << 16)
+# define REALLOC(p,s) realloc ((p), (s))
+# endif
+
+/* Extend the buffer by twice its current size via realloc and
+ reset the pointers that pointed into the old block to point to the
+ correct places in the new one. If extending the buffer results in it
+ being larger than MAX_BUF_SIZE, then flag memory exhausted. */
+# if __BOUNDED_POINTERS__
+# define SET_HIGH_BOUND(P) (__ptrhigh (P) = __ptrlow (P) + bufp->allocated)
+# define MOVE_BUFFER_POINTER(P) \
+ (__ptrlow (P) += incr, SET_HIGH_BOUND (P), __ptrvalue (P) += incr)
+# define ELSE_EXTEND_BUFFER_HIGH_BOUND \
+ else \
+ { \
+ SET_HIGH_BOUND (b); \
+ SET_HIGH_BOUND (begalt); \
+ if (fixup_alt_jump) \
+ SET_HIGH_BOUND (fixup_alt_jump); \
+ if (laststart) \
+ SET_HIGH_BOUND (laststart); \
+ if (pending_exact) \
+ SET_HIGH_BOUND (pending_exact); \
+ }
+# else
+# define MOVE_BUFFER_POINTER(P) (P) += incr
+# define ELSE_EXTEND_BUFFER_HIGH_BOUND
+# endif
+# endif /* not DEFINED_ONCE */
+
+# ifdef WCHAR
+# define EXTEND_BUFFER() \
+ do { \
+ UCHAR_T *old_buffer = COMPILED_BUFFER_VAR; \
+ int wchar_count; \
+ if (bufp->allocated + sizeof(UCHAR_T) > MAX_BUF_SIZE) \
+ return REG_ESIZE; \
+ bufp->allocated <<= 1; \
+ if (bufp->allocated > MAX_BUF_SIZE) \
+ bufp->allocated = MAX_BUF_SIZE; \
+ /* How many characters the new buffer can have? */ \
+ wchar_count = bufp->allocated / sizeof(UCHAR_T); \
+ if (wchar_count == 0) wchar_count = 1; \
+ /* Truncate the buffer to CHAR_T align. */ \
+ bufp->allocated = wchar_count * sizeof(UCHAR_T); \
+ RETALLOC (COMPILED_BUFFER_VAR, wchar_count, UCHAR_T); \
+ bufp->buffer = (char*)COMPILED_BUFFER_VAR; \
+ if (COMPILED_BUFFER_VAR == NULL) \
+ return REG_ESPACE; \
+ /* If the buffer moved, move all the pointers into it. */ \
+ if (old_buffer != COMPILED_BUFFER_VAR) \
+ { \
+ int incr = COMPILED_BUFFER_VAR - old_buffer; \
+ MOVE_BUFFER_POINTER (b); \
+ MOVE_BUFFER_POINTER (begalt); \
+ if (fixup_alt_jump) \
+ MOVE_BUFFER_POINTER (fixup_alt_jump); \
+ if (laststart) \
+ MOVE_BUFFER_POINTER (laststart); \
+ if (pending_exact) \
+ MOVE_BUFFER_POINTER (pending_exact); \
+ } \
+ ELSE_EXTEND_BUFFER_HIGH_BOUND \
+ } while (0)
+# else /* BYTE */
+# define EXTEND_BUFFER() \
+ do { \
+ UCHAR_T *old_buffer = COMPILED_BUFFER_VAR; \
+ if (bufp->allocated == MAX_BUF_SIZE) \
+ return REG_ESIZE; \
+ bufp->allocated <<= 1; \
+ if (bufp->allocated > MAX_BUF_SIZE) \
+ bufp->allocated = MAX_BUF_SIZE; \
+ bufp->buffer = REALLOC (COMPILED_BUFFER_VAR, bufp->allocated); \
+ if (COMPILED_BUFFER_VAR == NULL) \
+ return REG_ESPACE; \
+ /* If the buffer moved, move all the pointers into it. */ \
+ if (old_buffer != COMPILED_BUFFER_VAR) \
+ { \
+ int incr = COMPILED_BUFFER_VAR - old_buffer; \
+ MOVE_BUFFER_POINTER (b); \
+ MOVE_BUFFER_POINTER (begalt); \
+ if (fixup_alt_jump) \
+ MOVE_BUFFER_POINTER (fixup_alt_jump); \
+ if (laststart) \
+ MOVE_BUFFER_POINTER (laststart); \
+ if (pending_exact) \
+ MOVE_BUFFER_POINTER (pending_exact); \
+ } \
+ ELSE_EXTEND_BUFFER_HIGH_BOUND \
+ } while (0)
+# endif /* WCHAR */
+
+# ifndef DEFINED_ONCE
+/* Since we have one byte reserved for the register number argument to
+ {start,stop}_memory, the maximum number of groups we can report
+ things about is what fits in that byte. */
+# define MAX_REGNUM 255
+
+/* But patterns can have more than `MAX_REGNUM' registers. We just
+ ignore the excess. */
+typedef unsigned regnum_t;
+
+
+/* Macros for the compile stack. */
+
+/* Since offsets can go either forwards or backwards, this type needs to
+ be able to hold values from -(MAX_BUF_SIZE - 1) to MAX_BUF_SIZE - 1. */
+/* int may be not enough when sizeof(int) == 2. */
+typedef long pattern_offset_t;
+
+typedef struct
+{
+ pattern_offset_t begalt_offset;
+ pattern_offset_t fixup_alt_jump;
+ pattern_offset_t inner_group_offset;
+ pattern_offset_t laststart_offset;
+ regnum_t regnum;
+} compile_stack_elt_t;
+
+
+typedef struct
+{
+ compile_stack_elt_t *stack;
+ unsigned size;
+ unsigned avail; /* Offset of next open position. */
+} compile_stack_type;
+
+
+# define INIT_COMPILE_STACK_SIZE 32
+
+# define COMPILE_STACK_EMPTY (compile_stack.avail == 0)
+# define COMPILE_STACK_FULL (compile_stack.avail == compile_stack.size)
+
+/* The next available element. */
+# define COMPILE_STACK_TOP (compile_stack.stack[compile_stack.avail])
+
+# endif /* not DEFINED_ONCE */
+
+/* Set the bit for character C in a list. */
+# ifndef DEFINED_ONCE
+# define SET_LIST_BIT(c) \
+ (b[((unsigned char) (c)) / BYTEWIDTH] \
+ |= 1 << (((unsigned char) c) % BYTEWIDTH))
+# endif /* DEFINED_ONCE */
+
+/* Get the next unsigned number in the uncompiled pattern. */
+# define GET_UNSIGNED_NUMBER(num) \
+ { \
+ while (p != pend) \
+ { \
+ PATFETCH (c); \
+ if (c < '0' || c > '9') \
+ break; \
+ if (num <= RE_DUP_MAX) \
+ { \
+ if (num < 0) \
+ num = 0; \
+ num = num * 10 + c - '0'; \
+ } \
+ } \
+ }
+
+# ifndef DEFINED_ONCE
+# if defined _LIBC || WIDE_CHAR_SUPPORT
+/* The GNU C library provides support for user-defined character classes
+ and the functions from ISO C amendement 1. */
+# ifdef CHARCLASS_NAME_MAX
+# define CHAR_CLASS_MAX_LENGTH CHARCLASS_NAME_MAX
+# else
+/* This shouldn't happen but some implementation might still have this
+ problem. Use a reasonable default value. */
+# define CHAR_CLASS_MAX_LENGTH 256
+# endif
+
+# ifdef _LIBC
+# define IS_CHAR_CLASS(string) __wctype (string)
+# else
+# define IS_CHAR_CLASS(string) wctype (string)
+# endif
+# else
+# define CHAR_CLASS_MAX_LENGTH 6 /* Namely, `xdigit'. */
+
+# define IS_CHAR_CLASS(string) \
+ (STREQ (string, "alpha") || STREQ (string, "upper") \
+ || STREQ (string, "lower") || STREQ (string, "digit") \
+ || STREQ (string, "alnum") || STREQ (string, "xdigit") \
+ || STREQ (string, "space") || STREQ (string, "print") \
+ || STREQ (string, "punct") || STREQ (string, "graph") \
+ || STREQ (string, "cntrl") || STREQ (string, "blank"))
+# endif
+# endif /* DEFINED_ONCE */
+
+# ifndef MATCH_MAY_ALLOCATE
+
+/* If we cannot allocate large objects within re_match_2_internal,
+ we make the fail stack and register vectors global.
+ The fail stack, we grow to the maximum size when a regexp
+ is compiled.
+ The register vectors, we adjust in size each time we
+ compile a regexp, according to the number of registers it needs. */
+
+static PREFIX(fail_stack_type) fail_stack;
+
+/* Size with which the following vectors are currently allocated.
+ That is so we can make them bigger as needed,
+ but never make them smaller. */
+# ifdef DEFINED_ONCE
+static int regs_allocated_size;
+
+static const char ** regstart, ** regend;
+static const char ** old_regstart, ** old_regend;
+static const char **best_regstart, **best_regend;
+static const char **reg_dummy;
+# endif /* DEFINED_ONCE */
+
+static PREFIX(register_info_type) *PREFIX(reg_info);
+static PREFIX(register_info_type) *PREFIX(reg_info_dummy);
+
+/* Make the register vectors big enough for NUM_REGS registers,
+ but don't make them smaller. */
+
+static void
+PREFIX(regex_grow_registers) (int num_regs)
+{
+ if (num_regs > regs_allocated_size)
+ {
+ RETALLOC_IF (regstart, num_regs, const char *);
+ RETALLOC_IF (regend, num_regs, const char *);
+ RETALLOC_IF (old_regstart, num_regs, const char *);
+ RETALLOC_IF (old_regend, num_regs, const char *);
+ RETALLOC_IF (best_regstart, num_regs, const char *);
+ RETALLOC_IF (best_regend, num_regs, const char *);
+ RETALLOC_IF (PREFIX(reg_info), num_regs, PREFIX(register_info_type));
+ RETALLOC_IF (reg_dummy, num_regs, const char *);
+ RETALLOC_IF (PREFIX(reg_info_dummy), num_regs, PREFIX(register_info_type));
+
+ regs_allocated_size = num_regs;
+ }
+}
+
+# endif /* not MATCH_MAY_ALLOCATE */
+
+# ifndef DEFINED_ONCE
+static boolean group_in_compile_stack (compile_stack_type
+ compile_stack,
+ regnum_t regnum);
+# endif /* not DEFINED_ONCE */
+
+/* `regex_compile' compiles PATTERN (of length SIZE) according to SYNTAX.
+ Returns one of error codes defined in `regex.h', or zero for success.
+
+ Assumes the `allocated' (and perhaps `buffer') and `translate'
+ fields are set in BUFP on entry.
+
+ If it succeeds, results are put in BUFP (if it returns an error, the
+ contents of BUFP are undefined):
+ `buffer' is the compiled pattern;
+ `syntax' is set to SYNTAX;
+ `used' is set to the length of the compiled pattern;
+ `fastmap_accurate' is zero;
+ `re_nsub' is the number of subexpressions in PATTERN;
+ `not_bol' and `not_eol' are zero;
+
+ The `fastmap' and `newline_anchor' fields are neither
+ examined nor set. */
+
+/* Return, freeing storage we allocated. */
+# ifdef WCHAR
+# define FREE_STACK_RETURN(value) \
+ return (free(pattern), free(mbs_offset), free(is_binary), free (compile_stack.stack), value)
+# else
+# define FREE_STACK_RETURN(value) \
+ return (free (compile_stack.stack), value)
+# endif /* WCHAR */
+
+static reg_errcode_t
+PREFIX(regex_compile) (const char *ARG_PREFIX(pattern),
+ size_t ARG_PREFIX(size),
+ reg_syntax_t syntax,
+ struct re_pattern_buffer *bufp)
+{
+ /* We fetch characters from PATTERN here. Even though PATTERN is
+ `char *' (i.e., signed), we declare these variables as unsigned, so
+ they can be reliably used as array indices. */
+ register UCHAR_T c, c1;
+
+#ifdef WCHAR
+ /* A temporary space to keep wchar_t pattern and compiled pattern. */
+ CHAR_T *pattern, *COMPILED_BUFFER_VAR;
+ size_t size;
+ /* offset buffer for optimization. See convert_mbs_to_wc. */
+ int *mbs_offset = NULL;
+ /* It hold whether each wchar_t is binary data or not. */
+ char *is_binary = NULL;
+ /* A flag whether exactn is handling binary data or not. */
+ char is_exactn_bin = FALSE;
+#endif /* WCHAR */
+
+ /* A random temporary spot in PATTERN. */
+ const CHAR_T *p1;
+
+ /* Points to the end of the buffer, where we should append. */
+ register UCHAR_T *b;
+
+ /* Keeps track of unclosed groups. */
+ compile_stack_type compile_stack;
+
+ /* Points to the current (ending) position in the pattern. */
+#ifdef WCHAR
+ const CHAR_T *p;
+ const CHAR_T *pend;
+#else /* BYTE */
+ const CHAR_T *p = pattern;
+ const CHAR_T *pend = pattern + size;
+#endif /* WCHAR */
+
+ /* How to translate the characters in the pattern. */
+ RE_TRANSLATE_TYPE translate = bufp->translate;
+
+ /* Address of the count-byte of the most recently inserted `exactn'
+ command. This makes it possible to tell if a new exact-match
+ character can be added to that command or if the character requires
+ a new `exactn' command. */
+ UCHAR_T *pending_exact = 0;
+
+ /* Address of start of the most recently finished expression.
+ This tells, e.g., postfix * where to find the start of its
+ operand. Reset at the beginning of groups and alternatives. */
+ UCHAR_T *laststart = 0;
+
+ /* Address of beginning of regexp, or inside of last group. */
+ UCHAR_T *begalt;
+
+ /* Address of the place where a forward jump should go to the end of
+ the containing expression. Each alternative of an `or' -- except the
+ last -- ends with a forward jump of this sort. */
+ UCHAR_T *fixup_alt_jump = 0;
+
+ /* Counts open-groups as they are encountered. Remembered for the
+ matching close-group on the compile stack, so the same register
+ number is put in the stop_memory as the start_memory. */
+ regnum_t regnum = 0;
+
+#ifdef WCHAR
+ /* Initialize the wchar_t PATTERN and offset_buffer. */
+ p = pend = pattern = TALLOC(csize + 1, CHAR_T);
+ mbs_offset = TALLOC(csize + 1, int);
+ is_binary = TALLOC(csize + 1, char);
+ if (pattern == NULL || mbs_offset == NULL || is_binary == NULL)
+ {
+ free(pattern);
+ free(mbs_offset);
+ free(is_binary);
+ return REG_ESPACE;
+ }
+ pattern[csize] = L'\0'; /* sentinel */
+ size = convert_mbs_to_wcs(pattern, cpattern, csize, mbs_offset, is_binary);
+ pend = p + size;
+ if (size < 0)
+ {
+ free(pattern);
+ free(mbs_offset);
+ free(is_binary);
+ return REG_BADPAT;
+ }
+#endif
+
+#ifdef DEBUG
+ DEBUG_PRINT1 ("\nCompiling pattern: ");
+ if (debug)
+ {
+ unsigned debug_count;
+
+ for (debug_count = 0; debug_count < size; debug_count++)
+ PUT_CHAR (pattern[debug_count]);
+ putchar ('\n');
+ }
+#endif /* DEBUG */
+
+ /* Initialize the compile stack. */
+ compile_stack.stack = TALLOC (INIT_COMPILE_STACK_SIZE, compile_stack_elt_t);
+ if (compile_stack.stack == NULL)
+ {
+#ifdef WCHAR
+ free(pattern);
+ free(mbs_offset);
+ free(is_binary);
+#endif
+ return REG_ESPACE;
+ }
+
+ compile_stack.size = INIT_COMPILE_STACK_SIZE;
+ compile_stack.avail = 0;
+
+ /* Initialize the pattern buffer. */
+ bufp->syntax = syntax;
+ bufp->fastmap_accurate = 0;
+ bufp->not_bol = bufp->not_eol = 0;
+
+ /* Set `used' to zero, so that if we return an error, the pattern
+ printer (for debugging) will think there's no pattern. We reset it
+ at the end. */
+ bufp->used = 0;
+
+ /* Always count groups, whether or not bufp->no_sub is set. */
+ bufp->re_nsub = 0;
+
+#if !defined emacs && !defined SYNTAX_TABLE
+ /* Initialize the syntax table. */
+ init_syntax_once ();
+#endif
+
+ if (bufp->allocated == 0)
+ {
+ if (bufp->buffer)
+ { /* If zero allocated, but buffer is non-null, try to realloc
+ enough space. This loses if buffer's address is bogus, but
+ that is the user's responsibility. */
+#ifdef WCHAR
+ /* Free bufp->buffer and allocate an array for wchar_t pattern
+ buffer. */
+ free(bufp->buffer);
+ COMPILED_BUFFER_VAR = TALLOC (INIT_BUF_SIZE/sizeof(UCHAR_T),
+ UCHAR_T);
+#else
+ RETALLOC (COMPILED_BUFFER_VAR, INIT_BUF_SIZE, UCHAR_T);
+#endif /* WCHAR */
+ }
+ else
+ { /* Caller did not allocate a buffer. Do it for them. */
+ COMPILED_BUFFER_VAR = TALLOC (INIT_BUF_SIZE / sizeof(UCHAR_T),
+ UCHAR_T);
+ }
+
+ if (!COMPILED_BUFFER_VAR) FREE_STACK_RETURN (REG_ESPACE);
+#ifdef WCHAR
+ bufp->buffer = (char*)COMPILED_BUFFER_VAR;
+#endif /* WCHAR */
+ bufp->allocated = INIT_BUF_SIZE;
+ }
+#ifdef WCHAR
+ else
+ COMPILED_BUFFER_VAR = (UCHAR_T*) bufp->buffer;
+#endif
+
+ begalt = b = COMPILED_BUFFER_VAR;
+
+ /* Loop through the uncompiled pattern until we're at the end. */
+ while (p != pend)
+ {
+ PATFETCH (c);
+
+ switch (c)
+ {
+ case '^':
+ {
+ if ( /* If at start of pattern, it's an operator. */
+ p == pattern + 1
+ /* If context independent, it's an operator. */
+ || syntax & RE_CONTEXT_INDEP_ANCHORS
+ /* Otherwise, depends on what's come before. */
+ || PREFIX(at_begline_loc_p) (pattern, p, syntax))
+ BUF_PUSH (begline);
+ else
+ goto normal_char;
+ }
+ break;
+
+
+ case '$':
+ {
+ if ( /* If at end of pattern, it's an operator. */
+ p == pend
+ /* If context independent, it's an operator. */
+ || syntax & RE_CONTEXT_INDEP_ANCHORS
+ /* Otherwise, depends on what's next. */
+ || PREFIX(at_endline_loc_p) (p, pend, syntax))
+ BUF_PUSH (endline);
+ else
+ goto normal_char;
+ }
+ break;
+
+
+ case '+':
+ case '?':
+ if ((syntax & RE_BK_PLUS_QM)
+ || (syntax & RE_LIMITED_OPS))
+ goto normal_char;
+ handle_plus:
+ case '*':
+ /* If there is no previous pattern... */
+ if (!laststart)
+ {
+ if (syntax & RE_CONTEXT_INVALID_OPS)
+ FREE_STACK_RETURN (REG_BADRPT);
+ else if (!(syntax & RE_CONTEXT_INDEP_OPS))
+ goto normal_char;
+ }
+
+ {
+ /* Are we optimizing this jump? */
+ boolean keep_string_p = false;
+
+ /* 1 means zero (many) matches is allowed. */
+ char zero_times_ok = 0, many_times_ok = 0;
+
+ /* If there is a sequence of repetition chars, collapse it
+ down to just one (the right one). We can't combine
+ interval operators with these because of, e.g., `a{2}*',
+ which should only match an even number of `a's. */
+
+ for (;;)
+ {
+ zero_times_ok |= c != '+';
+ many_times_ok |= c != '?';
+
+ if (p == pend)
+ break;
+
+ PATFETCH (c);
+
+ if (c == '*'
+ || (!(syntax & RE_BK_PLUS_QM) && (c == '+' || c == '?')))
+ ;
+
+ else if (syntax & RE_BK_PLUS_QM && c == '\\')
+ {
+ if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
+
+ PATFETCH (c1);
+ if (!(c1 == '+' || c1 == '?'))
+ {
+ PATUNFETCH;
+ PATUNFETCH;
+ break;
+ }
+
+ c = c1;
+ }
+ else
+ {
+ PATUNFETCH;
+ break;
+ }
+
+ /* If we get here, we found another repeat character. */
+ }
+
+ /* Star, etc. applied to an empty pattern is equivalent
+ to an empty pattern. */
+ if (!laststart)
+ break;
+
+ /* Now we know whether or not zero matches is allowed
+ and also whether or not two or more matches is allowed. */
+ if (many_times_ok)
+ { /* More than one repetition is allowed, so put in at the
+ end a backward relative jump from `b' to before the next
+ jump we're going to put in below (which jumps from
+ laststart to after this jump).
+
+ But if we are at the `*' in the exact sequence `.*\n',
+ insert an unconditional jump backwards to the .,
+ instead of the beginning of the loop. This way we only
+ push a failure point once, instead of every time
+ through the loop. */
+ assert (p - 1 > pattern);
+
+ /* Allocate the space for the jump. */
+ GET_BUFFER_SPACE (1 + OFFSET_ADDRESS_SIZE);
+
+ /* We know we are not at the first character of the pattern,
+ because laststart was nonzero. And we've already
+ incremented `p', by the way, to be the character after
+ the `*'. Do we have to do something analogous here
+ for null bytes, because of RE_DOT_NOT_NULL? */
+ if (TRANSLATE (*(p - 2)) == TRANSLATE ('.')
+ && zero_times_ok
+ && p < pend && TRANSLATE (*p) == TRANSLATE ('\n')
+ && !(syntax & RE_DOT_NEWLINE))
+ { /* We have .*\n. */
+ STORE_JUMP (jump, b, laststart);
+ keep_string_p = true;
+ }
+ else
+ /* Anything else. */
+ STORE_JUMP (maybe_pop_jump, b, laststart -
+ (1 + OFFSET_ADDRESS_SIZE));
+
+ /* We've added more stuff to the buffer. */
+ b += 1 + OFFSET_ADDRESS_SIZE;
+ }
+
+ /* On failure, jump from laststart to b + 3, which will be the
+ end of the buffer after this jump is inserted. */
+ /* ifdef WCHAR, 'b + 1 + OFFSET_ADDRESS_SIZE' instead of
+ 'b + 3'. */
+ GET_BUFFER_SPACE (1 + OFFSET_ADDRESS_SIZE);
+ INSERT_JUMP (keep_string_p ? on_failure_keep_string_jump
+ : on_failure_jump,
+ laststart, b + 1 + OFFSET_ADDRESS_SIZE);
+ pending_exact = 0;
+ b += 1 + OFFSET_ADDRESS_SIZE;
+
+ if (!zero_times_ok)
+ {
+ /* At least one repetition is required, so insert a
+ `dummy_failure_jump' before the initial
+ `on_failure_jump' instruction of the loop. This
+ effects a skip over that instruction the first time
+ we hit that loop. */
+ GET_BUFFER_SPACE (1 + OFFSET_ADDRESS_SIZE);
+ INSERT_JUMP (dummy_failure_jump, laststart, laststart +
+ 2 + 2 * OFFSET_ADDRESS_SIZE);
+ b += 1 + OFFSET_ADDRESS_SIZE;
+ }
+ }
+ break;
+
+
+ case '.':
+ laststart = b;
+ BUF_PUSH (anychar);
+ break;
+
+
+ case '[':
+ {
+ boolean had_char_class = false;
+#ifdef WCHAR
+ CHAR_T range_start = 0xffffffff;
+#else
+ unsigned int range_start = 0xffffffff;
+#endif
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+#ifdef WCHAR
+ /* We assume a charset(_not) structure as a wchar_t array.
+ charset[0] = (re_opcode_t) charset(_not)
+ charset[1] = l (= length of char_classes)
+ charset[2] = m (= length of collating_symbols)
+ charset[3] = n (= length of equivalence_classes)
+ charset[4] = o (= length of char_ranges)
+ charset[5] = p (= length of chars)
+
+ charset[6] = char_class (wctype_t)
+ charset[6+CHAR_CLASS_SIZE] = char_class (wctype_t)
+ ...
+ charset[l+5] = char_class (wctype_t)
+
+ charset[l+6] = collating_symbol (wchar_t)
+ ...
+ charset[l+m+5] = collating_symbol (wchar_t)
+ ifdef _LIBC we use the index if
+ _NL_COLLATE_SYMB_EXTRAMB instead of
+ wchar_t string.
+
+ charset[l+m+6] = equivalence_classes (wchar_t)
+ ...
+ charset[l+m+n+5] = equivalence_classes (wchar_t)
+ ifdef _LIBC we use the index in
+ _NL_COLLATE_WEIGHT instead of
+ wchar_t string.
+
+ charset[l+m+n+6] = range_start
+ charset[l+m+n+7] = range_end
+ ...
+ charset[l+m+n+2o+4] = range_start
+ charset[l+m+n+2o+5] = range_end
+ ifdef _LIBC we use the value looked up
+ in _NL_COLLATE_COLLSEQ instead of
+ wchar_t character.
+
+ charset[l+m+n+2o+6] = char
+ ...
+ charset[l+m+n+2o+p+5] = char
+
+ */
+
+ /* We need at least 6 spaces: the opcode, the length of
+ char_classes, the length of collating_symbols, the length of
+ equivalence_classes, the length of char_ranges, the length of
+ chars. */
+ GET_BUFFER_SPACE (6);
+
+ /* Save b as laststart. And We use laststart as the pointer
+ to the first element of the charset here.
+ In other words, laststart[i] indicates charset[i]. */
+ laststart = b;
+
+ /* We test `*p == '^' twice, instead of using an if
+ statement, so we only need one BUF_PUSH. */
+ BUF_PUSH (*p == '^' ? charset_not : charset);
+ if (*p == '^')
+ p++;
+
+ /* Push the length of char_classes, the length of
+ collating_symbols, the length of equivalence_classes, the
+ length of char_ranges and the length of chars. */
+ BUF_PUSH_3 (0, 0, 0);
+ BUF_PUSH_2 (0, 0);
+
+ /* Remember the first position in the bracket expression. */
+ p1 = p;
+
+ /* charset_not matches newline according to a syntax bit. */
+ if ((re_opcode_t) b[-6] == charset_not
+ && (syntax & RE_HAT_LISTS_NOT_NEWLINE))
+ {
+ BUF_PUSH('\n');
+ laststart[5]++; /* Update the length of characters */
+ }
+
+ /* Read in characters and ranges, setting map bits. */
+ for (;;)
+ {
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ PATFETCH (c);
+
+ /* \ might escape characters inside [...] and [^...]. */
+ if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\')
+ {
+ if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
+
+ PATFETCH (c1);
+ BUF_PUSH(c1);
+ laststart[5]++; /* Update the length of chars */
+ range_start = c1;
+ continue;
+ }
+
+ /* Could be the end of the bracket expression. If it's
+ not (i.e., when the bracket expression is `[]' so
+ far), the ']' character bit gets set way below. */
+ if (c == ']' && p != p1 + 1)
+ break;
+
+ /* Look ahead to see if it's a range when the last thing
+ was a character class. */
+ if (had_char_class && c == '-' && *p != ']')
+ FREE_STACK_RETURN (REG_ERANGE);
+
+ /* Look ahead to see if it's a range when the last thing
+ was a character: if this is a hyphen not at the
+ beginning or the end of a list, then it's the range
+ operator. */
+ if (c == '-'
+ && !(p - 2 >= pattern && p[-2] == '[')
+ && !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^')
+ && *p != ']')
+ {
+ reg_errcode_t ret;
+ /* Allocate the space for range_start and range_end. */
+ GET_BUFFER_SPACE (2);
+ /* Update the pointer to indicate end of buffer. */
+ b += 2;
+ ret = wcs_compile_range (range_start, &p, pend, translate,
+ syntax, b, laststart);
+ if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
+ range_start = 0xffffffff;
+ }
+ else if (p[0] == '-' && p[1] != ']')
+ { /* This handles ranges made up of characters only. */
+ reg_errcode_t ret;
+
+ /* Move past the `-'. */
+ PATFETCH (c1);
+ /* Allocate the space for range_start and range_end. */
+ GET_BUFFER_SPACE (2);
+ /* Update the pointer to indicate end of buffer. */
+ b += 2;
+ ret = wcs_compile_range (c, &p, pend, translate, syntax, b,
+ laststart);
+ if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
+ range_start = 0xffffffff;
+ }
+
+ /* See if we're at the beginning of a possible character
+ class. */
+ else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == ':')
+ { /* Leave room for the null. */
+ char str[CHAR_CLASS_MAX_LENGTH + 1];
+
+ PATFETCH (c);
+ c1 = 0;
+
+ /* If pattern is `[[:'. */
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (;;)
+ {
+ PATFETCH (c);
+ if ((c == ':' && *p == ']') || p == pend)
+ break;
+ if (c1 < CHAR_CLASS_MAX_LENGTH)
+ str[c1++] = c;
+ else
+ /* This is in any case an invalid class name. */
+ str[0] = '\0';
+ }
+ str[c1] = '\0';
+
+ /* If isn't a word bracketed by `[:' and `:]':
+ undo the ending character, the letters, and leave
+ the leading `:' and `[' (but store them as character). */
+ if (c == ':' && *p == ']')
+ {
+ wctype_t wt;
+ uintptr_t alignedp;
+
+ /* Query the character class as wctype_t. */
+ wt = IS_CHAR_CLASS (str);
+ if (wt == 0)
+ FREE_STACK_RETURN (REG_ECTYPE);
+
+ /* Throw away the ] at the end of the character
+ class. */
+ PATFETCH (c);
+
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ /* Allocate the space for character class. */
+ GET_BUFFER_SPACE(CHAR_CLASS_SIZE);
+ /* Update the pointer to indicate end of buffer. */
+ b += CHAR_CLASS_SIZE;
+ /* Move data which follow character classes
+ not to violate the data. */
+ insert_space(CHAR_CLASS_SIZE,
+ laststart + 6 + laststart[1],
+ b - 1);
+ alignedp = ((uintptr_t)(laststart + 6 + laststart[1])
+ + __alignof__(wctype_t) - 1)
+ & ~(uintptr_t)(__alignof__(wctype_t) - 1);
+ /* Store the character class. */
+ *((wctype_t*)alignedp) = wt;
+ /* Update length of char_classes */
+ laststart[1] += CHAR_CLASS_SIZE;
+
+ had_char_class = true;
+ }
+ else
+ {
+ c1++;
+ while (c1--)
+ PATUNFETCH;
+ BUF_PUSH ('[');
+ BUF_PUSH (':');
+ laststart[5] += 2; /* Update the length of characters */
+ range_start = ':';
+ had_char_class = false;
+ }
+ }
+ else if (syntax & RE_CHAR_CLASSES && c == '[' && (*p == '='
+ || *p == '.'))
+ {
+ CHAR_T str[128]; /* Should be large enough. */
+ CHAR_T delim = *p; /* '=' or '.' */
+# ifdef _LIBC
+ uint32_t nrules =
+ _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+# endif
+ PATFETCH (c);
+ c1 = 0;
+
+ /* If pattern is `[[=' or '[[.'. */
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (;;)
+ {
+ PATFETCH (c);
+ if ((c == delim && *p == ']') || p == pend)
+ break;
+ if (c1 < sizeof (str) - 1)
+ str[c1++] = c;
+ else
+ /* This is in any case an invalid class name. */
+ str[0] = '\0';
+ }
+ str[c1] = '\0';
+
+ if (c == delim && *p == ']' && str[0] != '\0')
+ {
+ unsigned int i, offset;
+ /* If we have no collation data we use the default
+ collation in which each character is in a class
+ by itself. It also means that ASCII is the
+ character set and therefore we cannot have character
+ with more than one byte in the multibyte
+ representation. */
+
+ /* If not defined _LIBC, we push the name and
+ `\0' for the sake of matching performance. */
+ int datasize = c1 + 1;
+
+# ifdef _LIBC
+ int32_t idx = 0;
+ if (nrules == 0)
+# endif
+ {
+ if (c1 != 1)
+ FREE_STACK_RETURN (REG_ECOLLATE);
+ }
+# ifdef _LIBC
+ else
+ {
+ const int32_t *table;
+ const int32_t *weights;
+ const int32_t *extra;
+ const int32_t *indirect;
+ wint_t *cp;
+
+ /* This #include defines a local function! */
+# include <locale/weightwc.h>
+
+ if(delim == '=')
+ {
+ /* We push the index for equivalence class. */
+ cp = (wint_t*)str;
+
+ table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_TABLEWC);
+ weights = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_WEIGHTWC);
+ extra = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_EXTRAWC);
+ indirect = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_INDIRECTWC);
+
+ idx = findidx ((const wint_t**)&cp);
+ if (idx == 0 || cp < (wint_t*) str + c1)
+ /* This is no valid character. */
+ FREE_STACK_RETURN (REG_ECOLLATE);
+
+ str[0] = (wchar_t)idx;
+ }
+ else /* delim == '.' */
+ {
+ /* We push collation sequence value
+ for collating symbol. */
+ int32_t table_size;
+ const int32_t *symb_table;
+ const unsigned char *extra;
+ int32_t idx;
+ int32_t elem;
+ int32_t second;
+ int32_t hash;
+ char char_str[c1];
+
+ /* We have to convert the name to a single-byte
+ string. This is possible since the names
+ consist of ASCII characters and the internal
+ representation is UCS4. */
+ for (i = 0; i < c1; ++i)
+ char_str[i] = str[i];
+
+ table_size =
+ _NL_CURRENT_WORD (LC_COLLATE,
+ _NL_COLLATE_SYMB_HASH_SIZEMB);
+ symb_table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_TABLEMB);
+ extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_EXTRAMB);
+
+ /* Locate the character in the hashing table. */
+ hash = elem_hash (char_str, c1);
+
+ idx = 0;
+ elem = hash % table_size;
+ second = hash % (table_size - 2);
+ while (symb_table[2 * elem] != 0)
+ {
+ /* First compare the hashing value. */
+ if (symb_table[2 * elem] == hash
+ && c1 == extra[symb_table[2 * elem + 1]]
+ && memcmp (char_str,
+ &extra[symb_table[2 * elem + 1]
+ + 1], c1) == 0)
+ {
+ /* Yep, this is the entry. */
+ idx = symb_table[2 * elem + 1];
+ idx += 1 + extra[idx];
+ break;
+ }
+
+ /* Next entry. */
+ elem += second;
+ }
+
+ if (symb_table[2 * elem] != 0)
+ {
+ /* Compute the index of the byte sequence
+ in the table. */
+ idx += 1 + extra[idx];
+ /* Adjust for the alignment. */
+ idx = (idx + 3) & ~3;
+
+ str[0] = (wchar_t) idx + 4;
+ }
+ else if (symb_table[2 * elem] == 0 && c1 == 1)
+ {
+ /* No valid character. Match it as a
+ single byte character. */
+ had_char_class = false;
+ BUF_PUSH(str[0]);
+ /* Update the length of characters */
+ laststart[5]++;
+ range_start = str[0];
+
+ /* Throw away the ] at the end of the
+ collating symbol. */
+ PATFETCH (c);
+ /* exit from the switch block. */
+ continue;
+ }
+ else
+ FREE_STACK_RETURN (REG_ECOLLATE);
+ }
+ datasize = 1;
+ }
+# endif
+ /* Throw away the ] at the end of the equivalence
+ class (or collating symbol). */
+ PATFETCH (c);
+
+ /* Allocate the space for the equivalence class
+ (or collating symbol) (and '\0' if needed). */
+ GET_BUFFER_SPACE(datasize);
+ /* Update the pointer to indicate end of buffer. */
+ b += datasize;
+
+ if (delim == '=')
+ { /* equivalence class */
+ /* Calculate the offset of char_ranges,
+ which is next to equivalence_classes. */
+ offset = laststart[1] + laststart[2]
+ + laststart[3] +6;
+ /* Insert space. */
+ insert_space(datasize, laststart + offset, b - 1);
+
+ /* Write the equivalence_class and \0. */
+ for (i = 0 ; i < datasize ; i++)
+ laststart[offset + i] = str[i];
+
+ /* Update the length of equivalence_classes. */
+ laststart[3] += datasize;
+ had_char_class = true;
+ }
+ else /* delim == '.' */
+ { /* collating symbol */
+ /* Calculate the offset of the equivalence_classes,
+ which is next to collating_symbols. */
+ offset = laststart[1] + laststart[2] + 6;
+ /* Insert space and write the collationg_symbol
+ and \0. */
+ insert_space(datasize, laststart + offset, b-1);
+ for (i = 0 ; i < datasize ; i++)
+ laststart[offset + i] = str[i];
+
+ /* In re_match_2_internal if range_start < -1, we
+ assume -range_start is the offset of the
+ collating symbol which is specified as
+ the character of the range start. So we assign
+ -(laststart[1] + laststart[2] + 6) to
+ range_start. */
+ range_start = -(laststart[1] + laststart[2] + 6);
+ /* Update the length of collating_symbol. */
+ laststart[2] += datasize;
+ had_char_class = false;
+ }
+ }
+ else
+ {
+ c1++;
+ while (c1--)
+ PATUNFETCH;
+ BUF_PUSH ('[');
+ BUF_PUSH (delim);
+ laststart[5] += 2; /* Update the length of characters */
+ range_start = delim;
+ had_char_class = false;
+ }
+ }
+ else
+ {
+ had_char_class = false;
+ BUF_PUSH(c);
+ laststart[5]++; /* Update the length of characters */
+ range_start = c;
+ }
+ }
+
+#else /* BYTE */
+ /* Ensure that we have enough space to push a charset: the
+ opcode, the length count, and the bitset; 34 bytes in all. */
+ GET_BUFFER_SPACE (34);
+
+ laststart = b;
+
+ /* We test `*p == '^' twice, instead of using an if
+ statement, so we only need one BUF_PUSH. */
+ BUF_PUSH (*p == '^' ? charset_not : charset);
+ if (*p == '^')
+ p++;
+
+ /* Remember the first position in the bracket expression. */
+ p1 = p;
+
+ /* Push the number of bytes in the bitmap. */
+ BUF_PUSH ((1 << BYTEWIDTH) / BYTEWIDTH);
+
+ /* Clear the whole map. */
+ bzero (b, (1 << BYTEWIDTH) / BYTEWIDTH);
+
+ /* charset_not matches newline according to a syntax bit. */
+ if ((re_opcode_t) b[-2] == charset_not
+ && (syntax & RE_HAT_LISTS_NOT_NEWLINE))
+ SET_LIST_BIT ('\n');
+
+ /* Read in characters and ranges, setting map bits. */
+ for (;;)
+ {
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ PATFETCH (c);
+
+ /* \ might escape characters inside [...] and [^...]. */
+ if ((syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) && c == '\\')
+ {
+ if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
+
+ PATFETCH (c1);
+ SET_LIST_BIT (c1);
+ range_start = c1;
+ continue;
+ }
+
+ /* Could be the end of the bracket expression. If it's
+ not (i.e., when the bracket expression is `[]' so
+ far), the ']' character bit gets set way below. */
+ if (c == ']' && p != p1 + 1)
+ break;
+
+ /* Look ahead to see if it's a range when the last thing
+ was a character class. */
+ if (had_char_class && c == '-' && *p != ']')
+ FREE_STACK_RETURN (REG_ERANGE);
+
+ /* Look ahead to see if it's a range when the last thing
+ was a character: if this is a hyphen not at the
+ beginning or the end of a list, then it's the range
+ operator. */
+ if (c == '-'
+ && !(p - 2 >= pattern && p[-2] == '[')
+ && !(p - 3 >= pattern && p[-3] == '[' && p[-2] == '^')
+ && *p != ']')
+ {
+ reg_errcode_t ret
+ = byte_compile_range (range_start, &p, pend, translate,
+ syntax, b);
+ if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
+ range_start = 0xffffffff;
+ }
+
+ else if (p[0] == '-' && p[1] != ']')
+ { /* This handles ranges made up of characters only. */
+ reg_errcode_t ret;
+
+ /* Move past the `-'. */
+ PATFETCH (c1);
+
+ ret = byte_compile_range (c, &p, pend, translate, syntax, b);
+ if (ret != REG_NOERROR) FREE_STACK_RETURN (ret);
+ range_start = 0xffffffff;
+ }
+
+ /* See if we're at the beginning of a possible character
+ class. */
+
+ else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == ':')
+ { /* Leave room for the null. */
+ char str[CHAR_CLASS_MAX_LENGTH + 1];
+
+ PATFETCH (c);
+ c1 = 0;
+
+ /* If pattern is `[[:'. */
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (;;)
+ {
+ PATFETCH (c);
+ if ((c == ':' && *p == ']') || p == pend)
+ break;
+ if (c1 < CHAR_CLASS_MAX_LENGTH)
+ str[c1++] = c;
+ else
+ /* This is in any case an invalid class name. */
+ str[0] = '\0';
+ }
+ str[c1] = '\0';
+
+ /* If isn't a word bracketed by `[:' and `:]':
+ undo the ending character, the letters, and leave
+ the leading `:' and `[' (but set bits for them). */
+ if (c == ':' && *p == ']')
+ {
+# if defined _LIBC || WIDE_CHAR_SUPPORT
+ boolean is_lower = STREQ (str, "lower");
+ boolean is_upper = STREQ (str, "upper");
+ wctype_t wt;
+ int ch;
+
+ wt = IS_CHAR_CLASS (str);
+ if (wt == 0)
+ FREE_STACK_RETURN (REG_ECTYPE);
+
+ /* Throw away the ] at the end of the character
+ class. */
+ PATFETCH (c);
+
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (ch = 0; ch < 1 << BYTEWIDTH; ++ch)
+ {
+ if (iswctype (btowc (ch), wt))
+ SET_LIST_BIT (ch);
+
+ if (translate && (is_upper || is_lower)
+ && (ISUPPER (ch) || ISLOWER (ch)))
+ SET_LIST_BIT (ch);
+ }
+
+ had_char_class = true;
+# else
+ int ch;
+ boolean is_alnum = STREQ (str, "alnum");
+ boolean is_alpha = STREQ (str, "alpha");
+ boolean is_blank = STREQ (str, "blank");
+ boolean is_cntrl = STREQ (str, "cntrl");
+ boolean is_digit = STREQ (str, "digit");
+ boolean is_graph = STREQ (str, "graph");
+ boolean is_lower = STREQ (str, "lower");
+ boolean is_print = STREQ (str, "print");
+ boolean is_punct = STREQ (str, "punct");
+ boolean is_space = STREQ (str, "space");
+ boolean is_upper = STREQ (str, "upper");
+ boolean is_xdigit = STREQ (str, "xdigit");
+
+ if (!IS_CHAR_CLASS (str))
+ FREE_STACK_RETURN (REG_ECTYPE);
+
+ /* Throw away the ] at the end of the character
+ class. */
+ PATFETCH (c);
+
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (ch = 0; ch < 1 << BYTEWIDTH; ch++)
+ {
+ /* This was split into 3 if's to
+ avoid an arbitrary limit in some compiler. */
+ if ( (is_alnum && ISALNUM (ch))
+ || (is_alpha && ISALPHA (ch))
+ || (is_blank && ISBLANK (ch))
+ || (is_cntrl && ISCNTRL (ch)))
+ SET_LIST_BIT (ch);
+ if ( (is_digit && ISDIGIT (ch))
+ || (is_graph && ISGRAPH (ch))
+ || (is_lower && ISLOWER (ch))
+ || (is_print && ISPRINT (ch)))
+ SET_LIST_BIT (ch);
+ if ( (is_punct && ISPUNCT (ch))
+ || (is_space && ISSPACE (ch))
+ || (is_upper && ISUPPER (ch))
+ || (is_xdigit && ISXDIGIT (ch)))
+ SET_LIST_BIT (ch);
+ if ( translate && (is_upper || is_lower)
+ && (ISUPPER (ch) || ISLOWER (ch)))
+ SET_LIST_BIT (ch);
+ }
+ had_char_class = true;
+# endif /* libc || wctype.h */
+ }
+ else
+ {
+ c1++;
+ while (c1--)
+ PATUNFETCH;
+ SET_LIST_BIT ('[');
+ SET_LIST_BIT (':');
+ range_start = ':';
+ had_char_class = false;
+ }
+ }
+ else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == '=')
+ {
+ unsigned char str[MB_LEN_MAX + 1];
+# ifdef _LIBC
+ uint32_t nrules =
+ _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+# endif
+
+ PATFETCH (c);
+ c1 = 0;
+
+ /* If pattern is `[[='. */
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (;;)
+ {
+ PATFETCH (c);
+ if ((c == '=' && *p == ']') || p == pend)
+ break;
+ if (c1 < MB_LEN_MAX)
+ str[c1++] = c;
+ else
+ /* This is in any case an invalid class name. */
+ str[0] = '\0';
+ }
+ str[c1] = '\0';
+
+ if (c == '=' && *p == ']' && str[0] != '\0')
+ {
+ /* If we have no collation data we use the default
+ collation in which each character is in a class
+ by itself. It also means that ASCII is the
+ character set and therefore we cannot have character
+ with more than one byte in the multibyte
+ representation. */
+# ifdef _LIBC
+ if (nrules == 0)
+# endif
+ {
+ if (c1 != 1)
+ FREE_STACK_RETURN (REG_ECOLLATE);
+
+ /* Throw away the ] at the end of the equivalence
+ class. */
+ PATFETCH (c);
+
+ /* Set the bit for the character. */
+ SET_LIST_BIT (str[0]);
+ }
+# ifdef _LIBC
+ else
+ {
+ /* Try to match the byte sequence in `str' against
+ those known to the collate implementation.
+ First find out whether the bytes in `str' are
+ actually from exactly one character. */
+ const int32_t *table;
+ const unsigned char *weights;
+ const unsigned char *extra;
+ const int32_t *indirect;
+ int32_t idx;
+ const unsigned char *cp = str;
+ int ch;
+
+ /* This #include defines a local function! */
+# include <locale/weight.h>
+
+ table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
+ weights = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
+ extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
+ indirect = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
+
+ idx = findidx (&cp);
+ if (idx == 0 || cp < str + c1)
+ /* This is no valid character. */
+ FREE_STACK_RETURN (REG_ECOLLATE);
+
+ /* Throw away the ] at the end of the equivalence
+ class. */
+ PATFETCH (c);
+
+ /* Now we have to go throught the whole table
+ and find all characters which have the same
+ first level weight.
+
+ XXX Note that this is not entirely correct.
+ we would have to match multibyte sequences
+ but this is not possible with the current
+ implementation. */
+ for (ch = 1; ch < 256; ++ch)
+ /* XXX This test would have to be changed if we
+ would allow matching multibyte sequences. */
+ if (table[ch] > 0)
+ {
+ int32_t idx2 = table[ch];
+ size_t len = weights[idx2];
+
+ /* Test whether the lenghts match. */
+ if (weights[idx] == len)
+ {
+ /* They do. New compare the bytes of
+ the weight. */
+ size_t cnt = 0;
+
+ while (cnt < len
+ && (weights[idx + 1 + cnt]
+ == weights[idx2 + 1 + cnt]))
+ ++cnt;
+
+ if (cnt == len)
+ /* They match. Mark the character as
+ acceptable. */
+ SET_LIST_BIT (ch);
+ }
+ }
+ }
+# endif
+ had_char_class = true;
+ }
+ else
+ {
+ c1++;
+ while (c1--)
+ PATUNFETCH;
+ SET_LIST_BIT ('[');
+ SET_LIST_BIT ('=');
+ range_start = '=';
+ had_char_class = false;
+ }
+ }
+ else if (syntax & RE_CHAR_CLASSES && c == '[' && *p == '.')
+ {
+ unsigned char str[128]; /* Should be large enough. */
+# ifdef _LIBC
+ uint32_t nrules =
+ _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+# endif
+
+ PATFETCH (c);
+ c1 = 0;
+
+ /* If pattern is `[[.'. */
+ if (p == pend) FREE_STACK_RETURN (REG_EBRACK);
+
+ for (;;)
+ {
+ PATFETCH (c);
+ if ((c == '.' && *p == ']') || p == pend)
+ break;
+ if (c1 < sizeof (str))
+ str[c1++] = c;
+ else
+ /* This is in any case an invalid class name. */
+ str[0] = '\0';
+ }
+ str[c1] = '\0';
+
+ if (c == '.' && *p == ']' && str[0] != '\0')
+ {
+ /* If we have no collation data we use the default
+ collation in which each character is the name
+ for its own class which contains only the one
+ character. It also means that ASCII is the
+ character set and therefore we cannot have character
+ with more than one byte in the multibyte
+ representation. */
+# ifdef _LIBC
+ if (nrules == 0)
+# endif
+ {
+ if (c1 != 1)
+ FREE_STACK_RETURN (REG_ECOLLATE);
+
+ /* Throw away the ] at the end of the equivalence
+ class. */
+ PATFETCH (c);
+
+ /* Set the bit for the character. */
+ SET_LIST_BIT (str[0]);
+ range_start = ((const unsigned char *) str)[0];
+ }
+# ifdef _LIBC
+ else
+ {
+ /* Try to match the byte sequence in `str' against
+ those known to the collate implementation.
+ First find out whether the bytes in `str' are
+ actually from exactly one character. */
+ int32_t table_size;
+ const int32_t *symb_table;
+ const unsigned char *extra;
+ int32_t idx;
+ int32_t elem;
+ int32_t second;
+ int32_t hash;
+
+ table_size =
+ _NL_CURRENT_WORD (LC_COLLATE,
+ _NL_COLLATE_SYMB_HASH_SIZEMB);
+ symb_table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_TABLEMB);
+ extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_SYMB_EXTRAMB);
+
+ /* Locate the character in the hashing table. */
+ hash = elem_hash (str, c1);
+
+ idx = 0;
+ elem = hash % table_size;
+ second = hash % (table_size - 2);
+ while (symb_table[2 * elem] != 0)
+ {
+ /* First compare the hashing value. */
+ if (symb_table[2 * elem] == hash
+ && c1 == extra[symb_table[2 * elem + 1]]
+ && memcmp (str,
+ &extra[symb_table[2 * elem + 1]
+ + 1],
+ c1) == 0)
+ {
+ /* Yep, this is the entry. */
+ idx = symb_table[2 * elem + 1];
+ idx += 1 + extra[idx];
+ break;
+ }
+
+ /* Next entry. */
+ elem += second;
+ }
+
+ if (symb_table[2 * elem] == 0)
+ /* This is no valid character. */
+ FREE_STACK_RETURN (REG_ECOLLATE);
+
+ /* Throw away the ] at the end of the equivalence
+ class. */
+ PATFETCH (c);
+
+ /* Now add the multibyte character(s) we found
+ to the accept list.
+
+ XXX Note that this is not entirely correct.
+ we would have to match multibyte sequences
+ but this is not possible with the current
+ implementation. Also, we have to match
+ collating symbols, which expand to more than
+ one file, as a whole and not allow the
+ individual bytes. */
+ c1 = extra[idx++];
+ if (c1 == 1)
+ range_start = extra[idx];
+ while (c1-- > 0)
+ {
+ SET_LIST_BIT (extra[idx]);
+ ++idx;
+ }
+ }
+# endif
+ had_char_class = false;
+ }
+ else
+ {
+ c1++;
+ while (c1--)
+ PATUNFETCH;
+ SET_LIST_BIT ('[');
+ SET_LIST_BIT ('.');
+ range_start = '.';
+ had_char_class = false;
+ }
+ }
+ else
+ {
+ had_char_class = false;
+ SET_LIST_BIT (c);
+ range_start = c;
+ }
+ }
+
+ /* Discard any (non)matching list bytes that are all 0 at the
+ end of the map. Decrease the map-length byte too. */
+ while ((int) b[-1] > 0 && b[b[-1] - 1] == 0)
+ b[-1]--;
+ b += b[-1];
+#endif /* WCHAR */
+ }
+ break;
+
+
+ case '(':
+ if (syntax & RE_NO_BK_PARENS)
+ goto handle_open;
+ else
+ goto normal_char;
+
+
+ case ')':
+ if (syntax & RE_NO_BK_PARENS)
+ goto handle_close;
+ else
+ goto normal_char;
+
+
+ case '\n':
+ if (syntax & RE_NEWLINE_ALT)
+ goto handle_alt;
+ else
+ goto normal_char;
+
+
+ case '|':
+ if (syntax & RE_NO_BK_VBAR)
+ goto handle_alt;
+ else
+ goto normal_char;
+
+
+ case '{':
+ if (syntax & RE_INTERVALS && syntax & RE_NO_BK_BRACES)
+ goto handle_interval;
+ else
+ goto normal_char;
+
+
+ case '\\':
+ if (p == pend) FREE_STACK_RETURN (REG_EESCAPE);
+
+ /* Do not translate the character after the \, so that we can
+ distinguish, e.g., \B from \b, even if we normally would
+ translate, e.g., B to b. */
+ PATFETCH_RAW (c);
+
+ switch (c)
+ {
+ case '(':
+ if (syntax & RE_NO_BK_PARENS)
+ goto normal_backslash;
+
+ handle_open:
+ bufp->re_nsub++;
+ regnum++;
+
+ if (COMPILE_STACK_FULL)
+ {
+ RETALLOC (compile_stack.stack, compile_stack.size << 1,
+ compile_stack_elt_t);
+ if (compile_stack.stack == NULL) return REG_ESPACE;
+
+ compile_stack.size <<= 1;
+ }
+
+ /* These are the values to restore when we hit end of this
+ group. They are all relative offsets, so that if the
+ whole pattern moves because of realloc, they will still
+ be valid. */
+ COMPILE_STACK_TOP.begalt_offset = begalt - COMPILED_BUFFER_VAR;
+ COMPILE_STACK_TOP.fixup_alt_jump
+ = fixup_alt_jump ? fixup_alt_jump - COMPILED_BUFFER_VAR + 1 : 0;
+ COMPILE_STACK_TOP.laststart_offset = b - COMPILED_BUFFER_VAR;
+ COMPILE_STACK_TOP.regnum = regnum;
+
+ /* We will eventually replace the 0 with the number of
+ groups inner to this one. But do not push a
+ start_memory for groups beyond the last one we can
+ represent in the compiled pattern. */
+ if (regnum <= MAX_REGNUM)
+ {
+ COMPILE_STACK_TOP.inner_group_offset = b
+ - COMPILED_BUFFER_VAR + 2;
+ BUF_PUSH_3 (start_memory, regnum, 0);
+ }
+
+ compile_stack.avail++;
+
+ fixup_alt_jump = 0;
+ laststart = 0;
+ begalt = b;
+ /* If we've reached MAX_REGNUM groups, then this open
+ won't actually generate any code, so we'll have to
+ clear pending_exact explicitly. */
+ pending_exact = 0;
+ break;
+
+
+ case ')':
+ if (syntax & RE_NO_BK_PARENS) goto normal_backslash;
+
+ if (COMPILE_STACK_EMPTY)
+ {
+ if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)
+ goto normal_backslash;
+ else
+ FREE_STACK_RETURN (REG_ERPAREN);
+ }
+
+ handle_close:
+ if (fixup_alt_jump)
+ { /* Push a dummy failure point at the end of the
+ alternative for a possible future
+ `pop_failure_jump' to pop. See comments at
+ `push_dummy_failure' in `re_match_2'. */
+ BUF_PUSH (push_dummy_failure);
+
+ /* We allocated space for this jump when we assigned
+ to `fixup_alt_jump', in the `handle_alt' case below. */
+ STORE_JUMP (jump_past_alt, fixup_alt_jump, b - 1);
+ }
+
+ /* See similar code for backslashed left paren above. */
+ if (COMPILE_STACK_EMPTY)
+ {
+ if (syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)
+ goto normal_char;
+ else
+ FREE_STACK_RETURN (REG_ERPAREN);
+ }
+
+ /* Since we just checked for an empty stack above, this
+ ``can't happen''. */
+ assert (compile_stack.avail != 0);
+ {
+ /* We don't just want to restore into `regnum', because
+ later groups should continue to be numbered higher,
+ as in `(ab)c(de)' -- the second group is #2. */
+ regnum_t this_group_regnum;
+
+ compile_stack.avail--;
+ begalt = COMPILED_BUFFER_VAR + COMPILE_STACK_TOP.begalt_offset;
+ fixup_alt_jump
+ = COMPILE_STACK_TOP.fixup_alt_jump
+ ? COMPILED_BUFFER_VAR + COMPILE_STACK_TOP.fixup_alt_jump - 1
+ : 0;
+ laststart = COMPILED_BUFFER_VAR + COMPILE_STACK_TOP.laststart_offset;
+ this_group_regnum = COMPILE_STACK_TOP.regnum;
+ /* If we've reached MAX_REGNUM groups, then this open
+ won't actually generate any code, so we'll have to
+ clear pending_exact explicitly. */
+ pending_exact = 0;
+
+ /* We're at the end of the group, so now we know how many
+ groups were inside this one. */
+ if (this_group_regnum <= MAX_REGNUM)
+ {
+ UCHAR_T *inner_group_loc
+ = COMPILED_BUFFER_VAR + COMPILE_STACK_TOP.inner_group_offset;
+
+ *inner_group_loc = regnum - this_group_regnum;
+ BUF_PUSH_3 (stop_memory, this_group_regnum,
+ regnum - this_group_regnum);
+ }
+ }
+ break;
+
+
+ case '|': /* `\|'. */
+ if (syntax & RE_LIMITED_OPS || syntax & RE_NO_BK_VBAR)
+ goto normal_backslash;
+ handle_alt:
+ if (syntax & RE_LIMITED_OPS)
+ goto normal_char;
+
+ /* Insert before the previous alternative a jump which
+ jumps to this alternative if the former fails. */
+ GET_BUFFER_SPACE (1 + OFFSET_ADDRESS_SIZE);
+ INSERT_JUMP (on_failure_jump, begalt,
+ b + 2 + 2 * OFFSET_ADDRESS_SIZE);
+ pending_exact = 0;
+ b += 1 + OFFSET_ADDRESS_SIZE;
+
+ /* The alternative before this one has a jump after it
+ which gets executed if it gets matched. Adjust that
+ jump so it will jump to this alternative's analogous
+ jump (put in below, which in turn will jump to the next
+ (if any) alternative's such jump, etc.). The last such
+ jump jumps to the correct final destination. A picture:
+ _____ _____
+ | | | |
+ | v | v
+ a | b | c
+
+ If we are at `b', then fixup_alt_jump right now points to a
+ three-byte space after `a'. We'll put in the jump, set
+ fixup_alt_jump to right after `b', and leave behind three
+ bytes which we'll fill in when we get to after `c'. */
+
+ if (fixup_alt_jump)
+ STORE_JUMP (jump_past_alt, fixup_alt_jump, b);
+
+ /* Mark and leave space for a jump after this alternative,
+ to be filled in later either by next alternative or
+ when know we're at the end of a series of alternatives. */
+ fixup_alt_jump = b;
+ GET_BUFFER_SPACE (1 + OFFSET_ADDRESS_SIZE);
+ b += 1 + OFFSET_ADDRESS_SIZE;
+
+ laststart = 0;
+ begalt = b;
+ break;
+
+
+ case '{':
+ /* If \{ is a literal. */
+ if (!(syntax & RE_INTERVALS)
+ /* If we're at `\{' and it's not the open-interval
+ operator. */
+ || (syntax & RE_NO_BK_BRACES))
+ goto normal_backslash;
+
+ handle_interval:
+ {
+ /* If got here, then the syntax allows intervals. */
+
+ /* At least (most) this many matches must be made. */
+ int lower_bound = -1, upper_bound = -1;
+
+ /* Place in the uncompiled pattern (i.e., just after
+ the '{') to go back to if the interval is invalid. */
+ const CHAR_T *beg_interval = p;
+
+ if (p == pend)
+ goto invalid_interval;
+
+ GET_UNSIGNED_NUMBER (lower_bound);
+
+ if (c == ',')
+ {
+ GET_UNSIGNED_NUMBER (upper_bound);
+ if (upper_bound < 0)
+ upper_bound = RE_DUP_MAX;
+ }
+ else
+ /* Interval such as `{1}' => match exactly once. */
+ upper_bound = lower_bound;
+
+ if (! (0 <= lower_bound && lower_bound <= upper_bound))
+ goto invalid_interval;
+
+ if (!(syntax & RE_NO_BK_BRACES))
+ {
+ if (c != '\\' || p == pend)
+ goto invalid_interval;
+ PATFETCH (c);
+ }
+
+ if (c != '}')
+ goto invalid_interval;
+
+ /* If it's invalid to have no preceding re. */
+ if (!laststart)
+ {
+ if (syntax & RE_CONTEXT_INVALID_OPS
+ && !(syntax & RE_INVALID_INTERVAL_ORD))
+ FREE_STACK_RETURN (REG_BADRPT);
+ else if (syntax & RE_CONTEXT_INDEP_OPS)
+ laststart = b;
+ else
+ goto unfetch_interval;
+ }
+
+ /* We just parsed a valid interval. */
+
+ if (RE_DUP_MAX < upper_bound)
+ FREE_STACK_RETURN (REG_BADBR);
+
+ /* If the upper bound is zero, don't want to succeed at
+ all; jump from `laststart' to `b + 3', which will be
+ the end of the buffer after we insert the jump. */
+ /* ifdef WCHAR, 'b + 1 + OFFSET_ADDRESS_SIZE'
+ instead of 'b + 3'. */
+ if (upper_bound == 0)
+ {
+ GET_BUFFER_SPACE (1 + OFFSET_ADDRESS_SIZE);
+ INSERT_JUMP (jump, laststart, b + 1
+ + OFFSET_ADDRESS_SIZE);
+ b += 1 + OFFSET_ADDRESS_SIZE;
+ }
+
+ /* Otherwise, we have a nontrivial interval. When
+ we're all done, the pattern will look like:
+ set_number_at <jump count> <upper bound>
+ set_number_at <succeed_n count> <lower bound>
+ succeed_n <after jump addr> <succeed_n count>
+ <body of loop>
+ jump_n <succeed_n addr> <jump count>
+ (The upper bound and `jump_n' are omitted if
+ `upper_bound' is 1, though.) */
+ else
+ { /* If the upper bound is > 1, we need to insert
+ more at the end of the loop. */
+ unsigned nbytes = 2 + 4 * OFFSET_ADDRESS_SIZE +
+ (upper_bound > 1) * (2 + 4 * OFFSET_ADDRESS_SIZE);
+
+ GET_BUFFER_SPACE (nbytes);
+
+ /* Initialize lower bound of the `succeed_n', even
+ though it will be set during matching by its
+ attendant `set_number_at' (inserted next),
+ because `re_compile_fastmap' needs to know.
+ Jump to the `jump_n' we might insert below. */
+ INSERT_JUMP2 (succeed_n, laststart,
+ b + 1 + 2 * OFFSET_ADDRESS_SIZE
+ + (upper_bound > 1) * (1 + 2 * OFFSET_ADDRESS_SIZE)
+ , lower_bound);
+ b += 1 + 2 * OFFSET_ADDRESS_SIZE;
+
+ /* Code to initialize the lower bound. Insert
+ before the `succeed_n'. The `5' is the last two
+ bytes of this `set_number_at', plus 3 bytes of
+ the following `succeed_n'. */
+ /* ifdef WCHAR, The '1+2*OFFSET_ADDRESS_SIZE'
+ is the 'set_number_at', plus '1+OFFSET_ADDRESS_SIZE'
+ of the following `succeed_n'. */
+ PREFIX(insert_op2) (set_number_at, laststart, 1
+ + 2 * OFFSET_ADDRESS_SIZE, lower_bound, b);
+ b += 1 + 2 * OFFSET_ADDRESS_SIZE;
+
+ if (upper_bound > 1)
+ { /* More than one repetition is allowed, so
+ append a backward jump to the `succeed_n'
+ that starts this interval.
+
+ When we've reached this during matching,
+ we'll have matched the interval once, so
+ jump back only `upper_bound - 1' times. */
+ STORE_JUMP2 (jump_n, b, laststart
+ + 2 * OFFSET_ADDRESS_SIZE + 1,
+ upper_bound - 1);
+ b += 1 + 2 * OFFSET_ADDRESS_SIZE;
+
+ /* The location we want to set is the second
+ parameter of the `jump_n'; that is `b-2' as
+ an absolute address. `laststart' will be
+ the `set_number_at' we're about to insert;
+ `laststart+3' the number to set, the source
+ for the relative address. But we are
+ inserting into the middle of the pattern --
+ so everything is getting moved up by 5.
+ Conclusion: (b - 2) - (laststart + 3) + 5,
+ i.e., b - laststart.
+
+ We insert this at the beginning of the loop
+ so that if we fail during matching, we'll
+ reinitialize the bounds. */
+ PREFIX(insert_op2) (set_number_at, laststart,
+ b - laststart,
+ upper_bound - 1, b);
+ b += 1 + 2 * OFFSET_ADDRESS_SIZE;
+ }
+ }
+ pending_exact = 0;
+ break;
+
+ invalid_interval:
+ if (!(syntax & RE_INVALID_INTERVAL_ORD))
+ FREE_STACK_RETURN (p == pend ? REG_EBRACE : REG_BADBR);
+ unfetch_interval:
+ /* Match the characters as literals. */
+ p = beg_interval;
+ c = '{';
+ if (syntax & RE_NO_BK_BRACES)
+ goto normal_char;
+ else
+ goto normal_backslash;
+ }
+
+#ifdef emacs
+ /* There is no way to specify the before_dot and after_dot
+ operators. rms says this is ok. --karl */
+ case '=':
+ BUF_PUSH (at_dot);
+ break;
+
+ case 's':
+ laststart = b;
+ PATFETCH (c);
+ BUF_PUSH_2 (syntaxspec, syntax_spec_code[c]);
+ break;
+
+ case 'S':
+ laststart = b;
+ PATFETCH (c);
+ BUF_PUSH_2 (notsyntaxspec, syntax_spec_code[c]);
+ break;
+#endif /* emacs */
+
+
+ case 'w':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ laststart = b;
+ BUF_PUSH (wordchar);
+ break;
+
+
+ case 'W':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ laststart = b;
+ BUF_PUSH (notwordchar);
+ break;
+
+
+ case '<':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ BUF_PUSH (wordbeg);
+ break;
+
+ case '>':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ BUF_PUSH (wordend);
+ break;
+
+ case 'b':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ BUF_PUSH (wordbound);
+ break;
+
+ case 'B':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ BUF_PUSH (notwordbound);
+ break;
+
+ case '`':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ BUF_PUSH (begbuf);
+ break;
+
+ case '\'':
+ if (syntax & RE_NO_GNU_OPS)
+ goto normal_char;
+ BUF_PUSH (endbuf);
+ break;
+
+ case '1': case '2': case '3': case '4': case '5':
+ case '6': case '7': case '8': case '9':
+ if (syntax & RE_NO_BK_REFS)
+ goto normal_char;
+
+ c1 = c - '0';
+
+ if (c1 > regnum)
+ FREE_STACK_RETURN (REG_ESUBREG);
+
+ /* Can't back reference to a subexpression if inside of it. */
+ if (group_in_compile_stack (compile_stack, (regnum_t) c1))
+ goto normal_char;
+
+ laststart = b;
+ BUF_PUSH_2 (duplicate, c1);
+ break;
+
+
+ case '+':
+ case '?':
+ if (syntax & RE_BK_PLUS_QM)
+ goto handle_plus;
+ else
+ goto normal_backslash;
+
+ default:
+ normal_backslash:
+ /* You might think it would be useful for \ to mean
+ not to translate; but if we don't translate it
+ it will never match anything. */
+ c = TRANSLATE (c);
+ goto normal_char;
+ }
+ break;
+
+
+ default:
+ /* Expects the character in `c'. */
+ normal_char:
+ /* If no exactn currently being built. */
+ if (!pending_exact
+#ifdef WCHAR
+ /* If last exactn handle binary(or character) and
+ new exactn handle character(or binary). */
+ || is_exactn_bin != is_binary[p - 1 - pattern]
+#endif /* WCHAR */
+
+ /* If last exactn not at current position. */
+ || pending_exact + *pending_exact + 1 != b
+
+ /* We have only one byte following the exactn for the count. */
+ || *pending_exact == (1 << BYTEWIDTH) - 1
+
+ /* If followed by a repetition operator. */
+ || *p == '*' || *p == '^'
+ || ((syntax & RE_BK_PLUS_QM)
+ ? *p == '\\' && (p[1] == '+' || p[1] == '?')
+ : (*p == '+' || *p == '?'))
+ || ((syntax & RE_INTERVALS)
+ && ((syntax & RE_NO_BK_BRACES)
+ ? *p == '{'
+ : (p[0] == '\\' && p[1] == '{'))))
+ {
+ /* Start building a new exactn. */
+
+ laststart = b;
+
+#ifdef WCHAR
+ /* Is this exactn binary data or character? */
+ is_exactn_bin = is_binary[p - 1 - pattern];
+ if (is_exactn_bin)
+ BUF_PUSH_2 (exactn_bin, 0);
+ else
+ BUF_PUSH_2 (exactn, 0);
+#else
+ BUF_PUSH_2 (exactn, 0);
+#endif /* WCHAR */
+ pending_exact = b - 1;
+ }
+
+ BUF_PUSH (c);
+ (*pending_exact)++;
+ break;
+ } /* switch (c) */
+ } /* while p != pend */
+
+
+ /* Through the pattern now. */
+
+ if (fixup_alt_jump)
+ STORE_JUMP (jump_past_alt, fixup_alt_jump, b);
+
+ if (!COMPILE_STACK_EMPTY)
+ FREE_STACK_RETURN (REG_EPAREN);
+
+ /* If we don't want backtracking, force success
+ the first time we reach the end of the compiled pattern. */
+ if (syntax & RE_NO_POSIX_BACKTRACKING)
+ BUF_PUSH (succeed);
+
+#ifdef WCHAR
+ free (pattern);
+ free (mbs_offset);
+ free (is_binary);
+#endif
+ free (compile_stack.stack);
+
+ /* We have succeeded; set the length of the buffer. */
+#ifdef WCHAR
+ bufp->used = (uintptr_t) b - (uintptr_t) COMPILED_BUFFER_VAR;
+#else
+ bufp->used = b - bufp->buffer;
+#endif
+
+#ifdef DEBUG
+ if (debug)
+ {
+ DEBUG_PRINT1 ("\nCompiled pattern: \n");
+ PREFIX(print_compiled_pattern) (bufp);
+ }
+#endif /* DEBUG */
+
+#ifndef MATCH_MAY_ALLOCATE
+ /* Initialize the failure stack to the largest possible stack. This
+ isn't necessary unless we're trying to avoid calling alloca in
+ the search and match routines. */
+ {
+ int num_regs = bufp->re_nsub + 1;
+
+ /* Since DOUBLE_FAIL_STACK refuses to double only if the current size
+ is strictly greater than re_max_failures, the largest possible stack
+ is 2 * re_max_failures failure points. */
+ if (fail_stack.size < (2 * re_max_failures * MAX_FAILURE_ITEMS))
+ {
+ fail_stack.size = (2 * re_max_failures * MAX_FAILURE_ITEMS);
+
+# ifdef emacs
+ if (! fail_stack.stack)
+ fail_stack.stack
+ = (PREFIX(fail_stack_elt_t) *) xmalloc (fail_stack.size
+ * sizeof (PREFIX(fail_stack_elt_t)));
+ else
+ fail_stack.stack
+ = (PREFIX(fail_stack_elt_t) *) xrealloc (fail_stack.stack,
+ (fail_stack.size
+ * sizeof (PREFIX(fail_stack_elt_t))));
+# else /* not emacs */
+ if (! fail_stack.stack)
+ fail_stack.stack
+ = malloc (fail_stack.size * sizeof (PREFIX(fail_stack_elt_t)));
+ else
+ fail_stack.stack
+ = realloc (fail_stack.stack,
+ fail_stack.size * sizeof (PREFIX(fail_stack_elt_t)));
+# endif /* not emacs */
+ }
+
+ PREFIX(regex_grow_registers) (num_regs);
+ }
+#endif /* not MATCH_MAY_ALLOCATE */
+
+ return REG_NOERROR;
+} /* regex_compile */
+
+/* Subroutines for `regex_compile'. */
+
+/* Store OP at LOC followed by two-byte integer parameter ARG. */
+/* ifdef WCHAR, integer parameter is 1 wchar_t. */
+
+static void
+PREFIX(store_op1) (re_opcode_t op, UCHAR_T *loc, int arg)
+{
+ *loc = (UCHAR_T) op;
+ STORE_NUMBER (loc + 1, arg);
+}
+
+
+/* Like `store_op1', but for two two-byte parameters ARG1 and ARG2. */
+/* ifdef WCHAR, integer parameter is 1 wchar_t. */
+
+static void
+PREFIX(store_op2) (re_opcode_t op, UCHAR_T *loc, int arg1, int arg2)
+{
+ *loc = (UCHAR_T) op;
+ STORE_NUMBER (loc + 1, arg1);
+ STORE_NUMBER (loc + 1 + OFFSET_ADDRESS_SIZE, arg2);
+}
+
+
+/* Copy the bytes from LOC to END to open up three bytes of space at LOC
+ for OP followed by two-byte integer parameter ARG. */
+/* ifdef WCHAR, integer parameter is 1 wchar_t. */
+
+static void
+PREFIX(insert_op1) (re_opcode_t op, UCHAR_T *loc, int arg, UCHAR_T *end)
+{
+ register UCHAR_T *pfrom = end;
+ register UCHAR_T *pto = end + 1 + OFFSET_ADDRESS_SIZE;
+
+ while (pfrom != loc)
+ *--pto = *--pfrom;
+
+ PREFIX(store_op1) (op, loc, arg);
+}
+
+
+/* Like `insert_op1', but for two two-byte parameters ARG1 and ARG2. */
+/* ifdef WCHAR, integer parameter is 1 wchar_t. */
+
+static void
+PREFIX(insert_op2) (re_opcode_t op, UCHAR_T *loc, int arg1, int arg2,
+ UCHAR_T *end)
+{
+ register UCHAR_T *pfrom = end;
+ register UCHAR_T *pto = end + 1 + 2 * OFFSET_ADDRESS_SIZE;
+
+ while (pfrom != loc)
+ *--pto = *--pfrom;
+
+ PREFIX(store_op2) (op, loc, arg1, arg2);
+}
+
+
+/* P points to just after a ^ in PATTERN. Return true if that ^ comes
+ after an alternative or a begin-subexpression. We assume there is at
+ least one character before the ^. */
+
+static boolean
+PREFIX(at_begline_loc_p) (const CHAR_T *pattern, const CHAR_T *p,
+ reg_syntax_t syntax)
+{
+ const CHAR_T *prev = p - 2;
+ boolean prev_prev_backslash = prev > pattern && prev[-1] == '\\';
+
+ return
+ /* After a subexpression? */
+ (*prev == '(' && (syntax & RE_NO_BK_PARENS || prev_prev_backslash))
+ /* After an alternative? */
+ || (*prev == '|' && (syntax & RE_NO_BK_VBAR || prev_prev_backslash));
+}
+
+
+/* The dual of at_begline_loc_p. This one is for $. We assume there is
+ at least one character after the $, i.e., `P < PEND'. */
+
+static boolean
+PREFIX(at_endline_loc_p) (const CHAR_T *p, const CHAR_T *pend,
+ reg_syntax_t syntax)
+{
+ const CHAR_T *next = p;
+ boolean next_backslash = *next == '\\';
+ const CHAR_T *next_next = p + 1 < pend ? p + 1 : 0;
+
+ return
+ /* Before a subexpression? */
+ (syntax & RE_NO_BK_PARENS ? *next == ')'
+ : next_backslash && next_next && *next_next == ')')
+ /* Before an alternative? */
+ || (syntax & RE_NO_BK_VBAR ? *next == '|'
+ : next_backslash && next_next && *next_next == '|');
+}
+
+#else /* not INSIDE_RECURSION */
+
+/* Returns true if REGNUM is in one of COMPILE_STACK's elements and
+ false if it's not. */
+
+static boolean
+group_in_compile_stack (compile_stack_type compile_stack,
+ regnum_t regnum)
+{
+ int this_element;
+
+ for (this_element = compile_stack.avail - 1;
+ this_element >= 0;
+ this_element--)
+ if (compile_stack.stack[this_element].regnum == regnum)
+ return true;
+
+ return false;
+}
+#endif /* not INSIDE_RECURSION */
+
+#ifdef INSIDE_RECURSION
+
+#ifdef WCHAR
+/* This insert space, which size is "num", into the pattern at "loc".
+ "end" must point the end of the allocated buffer. */
+static void
+insert_space (int num, CHAR_T *loc, CHAR_T *end)
+{
+ register CHAR_T *pto = end;
+ register CHAR_T *pfrom = end - num;
+
+ while (pfrom >= loc)
+ *pto-- = *pfrom--;
+}
+#endif /* WCHAR */
+
+#ifdef WCHAR
+static reg_errcode_t
+wcs_compile_range (CHAR_T range_start_char,
+ const CHAR_T **p_ptr, const CHAR_T *pend,
+ RE_TRANSLATE_TYPE translate, reg_syntax_t syntax,
+ CHAR_T *b, CHAR_T *char_set)
+{
+ const CHAR_T *p = *p_ptr;
+ CHAR_T range_start, range_end;
+ reg_errcode_t ret;
+# ifdef _LIBC
+ uint32_t nrules;
+ uint32_t start_val, end_val;
+# endif
+ if (p == pend)
+ return REG_ERANGE;
+
+# ifdef _LIBC
+ nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+ if (nrules != 0)
+ {
+ const char *collseq = (const char *) _NL_CURRENT(LC_COLLATE,
+ _NL_COLLATE_COLLSEQWC);
+ const unsigned char *extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB);
+
+ if (range_start_char < -1)
+ {
+ /* range_start is a collating symbol. */
+ int32_t *wextra;
+ /* Retreive the index and get collation sequence value. */
+ wextra = (int32_t*)(extra + char_set[-range_start_char]);
+ start_val = wextra[1 + *wextra];
+ }
+ else
+ start_val = collseq_table_lookup(collseq, TRANSLATE(range_start_char));
+
+ end_val = collseq_table_lookup (collseq, TRANSLATE (p[0]));
+
+ /* Report an error if the range is empty and the syntax prohibits
+ this. */
+ ret = ((syntax & RE_NO_EMPTY_RANGES)
+ && (start_val > end_val))? REG_ERANGE : REG_NOERROR;
+
+ /* Insert space to the end of the char_ranges. */
+ insert_space(2, b - char_set[5] - 2, b - 1);
+ *(b - char_set[5] - 2) = (wchar_t)start_val;
+ *(b - char_set[5] - 1) = (wchar_t)end_val;
+ char_set[4]++; /* ranges_index */
+ }
+ else
+# endif
+ {
+ range_start = (range_start_char >= 0)? TRANSLATE (range_start_char):
+ range_start_char;
+ range_end = TRANSLATE (p[0]);
+ /* Report an error if the range is empty and the syntax prohibits
+ this. */
+ ret = ((syntax & RE_NO_EMPTY_RANGES)
+ && (range_start > range_end))? REG_ERANGE : REG_NOERROR;
+
+ /* Insert space to the end of the char_ranges. */
+ insert_space(2, b - char_set[5] - 2, b - 1);
+ *(b - char_set[5] - 2) = range_start;
+ *(b - char_set[5] - 1) = range_end;
+ char_set[4]++; /* ranges_index */
+ }
+ /* Have to increment the pointer into the pattern string, so the
+ caller isn't still at the ending character. */
+ (*p_ptr)++;
+
+ return ret;
+}
+#else /* BYTE */
+/* Read the ending character of a range (in a bracket expression) from the
+ uncompiled pattern *P_PTR (which ends at PEND). We assume the
+ starting character is in `P[-2]'. (`P[-1]' is the character `-'.)
+ Then we set the translation of all bits between the starting and
+ ending characters (inclusive) in the compiled pattern B.
+
+ Return an error code.
+
+ We use these short variable names so we can use the same macros as
+ `regex_compile' itself. */
+
+static reg_errcode_t
+byte_compile_range (unsigned int range_start_char,
+ const char **p_ptr, const char *pend,
+ RE_TRANSLATE_TYPE translate, reg_syntax_t syntax,
+ unsigned char *b)
+{
+ unsigned this_char;
+ const char *p = *p_ptr;
+ reg_errcode_t ret;
+# if _LIBC
+ const unsigned char *collseq;
+ unsigned int start_colseq;
+ unsigned int end_colseq;
+# else
+ unsigned end_char;
+# endif
+
+ if (p == pend)
+ return REG_ERANGE;
+
+ /* Have to increment the pointer into the pattern string, so the
+ caller isn't still at the ending character. */
+ (*p_ptr)++;
+
+ /* Report an error if the range is empty and the syntax prohibits this. */
+ ret = syntax & RE_NO_EMPTY_RANGES ? REG_ERANGE : REG_NOERROR;
+
+# if _LIBC
+ collseq = (const unsigned char *) _NL_CURRENT (LC_COLLATE,
+ _NL_COLLATE_COLLSEQMB);
+
+ start_colseq = collseq[(unsigned char) TRANSLATE (range_start_char)];
+ end_colseq = collseq[(unsigned char) TRANSLATE (p[0])];
+ for (this_char = 0; this_char <= (unsigned char) -1; ++this_char)
+ {
+ unsigned int this_colseq = collseq[(unsigned char) TRANSLATE (this_char)];
+
+ if (start_colseq <= this_colseq && this_colseq <= end_colseq)
+ {
+ SET_LIST_BIT (TRANSLATE (this_char));
+ ret = REG_NOERROR;
+ }
+ }
+# else
+ /* Here we see why `this_char' has to be larger than an `unsigned
+ char' -- we would otherwise go into an infinite loop, since all
+ characters <= 0xff. */
+ range_start_char = TRANSLATE (range_start_char);
+ /* TRANSLATE(p[0]) is casted to char (not unsigned char) in TRANSLATE,
+ and some compilers cast it to int implicitly, so following for_loop
+ may fall to (almost) infinite loop.
+ e.g. If translate[p[0]] = 0xff, end_char may equals to 0xffffffff.
+ To avoid this, we cast p[0] to unsigned int and truncate it. */
+ end_char = ((unsigned)TRANSLATE(p[0]) & ((1 << BYTEWIDTH) - 1));
+
+ for (this_char = range_start_char; this_char <= end_char; ++this_char)
+ {
+ SET_LIST_BIT (TRANSLATE (this_char));
+ ret = REG_NOERROR;
+ }
+# endif
+
+ return ret;
+}
+#endif /* WCHAR */
+
+/* re_compile_fastmap computes a ``fastmap'' for the compiled pattern in
+ BUFP. A fastmap records which of the (1 << BYTEWIDTH) possible
+ characters can start a string that matches the pattern. This fastmap
+ is used by re_search to skip quickly over impossible starting points.
+
+ The caller must supply the address of a (1 << BYTEWIDTH)-byte data
+ area as BUFP->fastmap.
+
+ We set the `fastmap', `fastmap_accurate', and `can_be_null' fields in
+ the pattern buffer.
+
+ Returns 0 if we succeed, -2 if an internal error. */
+
+#ifdef WCHAR
+/* local function for re_compile_fastmap.
+ truncate wchar_t character to char. */
+
+static unsigned char
+truncate_wchar (CHAR_T c)
+{
+ unsigned char buf[MB_CUR_MAX];
+ mbstate_t state;
+ int retval;
+ memset (&state, '\0', sizeof (state));
+ retval = wcrtomb (buf, c, &state);
+ return retval > 0 ? buf[0] : (unsigned char) c;
+}
+#endif /* WCHAR */
+
+static int
+PREFIX(re_compile_fastmap) (struct re_pattern_buffer *bufp)
+{
+ int j, k;
+#ifdef MATCH_MAY_ALLOCATE
+ PREFIX(fail_stack_type) fail_stack;
+#endif
+#ifndef REGEX_MALLOC
+ char *destination;
+#endif
+
+ register char *fastmap = bufp->fastmap;
+
+#ifdef WCHAR
+ /* We need to cast pattern to (wchar_t*), because we casted this compiled
+ pattern to (char*) in regex_compile. */
+ UCHAR_T *pattern = (UCHAR_T*)bufp->buffer;
+ register UCHAR_T *pend = (UCHAR_T*) (bufp->buffer + bufp->used);
+#else /* BYTE */
+ UCHAR_T *pattern = bufp->buffer;
+ register UCHAR_T *pend = pattern + bufp->used;
+#endif /* WCHAR */
+ UCHAR_T *p = pattern;
+
+#ifdef REL_ALLOC
+ /* This holds the pointer to the failure stack, when
+ it is allocated relocatably. */
+ fail_stack_elt_t *failure_stack_ptr;
+#endif
+
+ /* Assume that each path through the pattern can be null until
+ proven otherwise. We set this false at the bottom of switch
+ statement, to which we get only if a particular path doesn't
+ match the empty string. */
+ boolean path_can_be_null = true;
+
+ /* We aren't doing a `succeed_n' to begin with. */
+ boolean succeed_n_p = false;
+
+ assert (fastmap != NULL && p != NULL);
+
+ INIT_FAIL_STACK ();
+ bzero (fastmap, 1 << BYTEWIDTH); /* Assume nothing's valid. */
+ bufp->fastmap_accurate = 1; /* It will be when we're done. */
+ bufp->can_be_null = 0;
+
+ while (1)
+ {
+ if (p == pend || *p == succeed)
+ {
+ /* We have reached the (effective) end of pattern. */
+ if (!FAIL_STACK_EMPTY ())
+ {
+ bufp->can_be_null |= path_can_be_null;
+
+ /* Reset for next path. */
+ path_can_be_null = true;
+
+ p = fail_stack.stack[--fail_stack.avail].pointer;
+
+ continue;
+ }
+ else
+ break;
+ }
+
+ /* We should never be about to go beyond the end of the pattern. */
+ assert (p < pend);
+
+ switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++))
+ {
+
+ /* I guess the idea here is to simply not bother with a fastmap
+ if a backreference is used, since it's too hard to figure out
+ the fastmap for the corresponding group. Setting
+ `can_be_null' stops `re_search_2' from using the fastmap, so
+ that is all we do. */
+ case duplicate:
+ bufp->can_be_null = 1;
+ goto done;
+
+
+ /* Following are the cases which match a character. These end
+ with `break'. */
+
+#ifdef WCHAR
+ case exactn:
+ fastmap[truncate_wchar(p[1])] = 1;
+ break;
+#else /* BYTE */
+ case exactn:
+ fastmap[p[1]] = 1;
+ break;
+#endif /* WCHAR */
+#ifdef MBS_SUPPORT
+ case exactn_bin:
+ fastmap[p[1]] = 1;
+ break;
+#endif
+
+#ifdef WCHAR
+ /* It is hard to distinguish fastmap from (multi byte) characters
+ which depends on current locale. */
+ case charset:
+ case charset_not:
+ case wordchar:
+ case notwordchar:
+ bufp->can_be_null = 1;
+ goto done;
+#else /* BYTE */
+ case charset:
+ for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--)
+ if (p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH)))
+ fastmap[j] = 1;
+ break;
+
+
+ case charset_not:
+ /* Chars beyond end of map must be allowed. */
+ for (j = *p * BYTEWIDTH; j < (1 << BYTEWIDTH); j++)
+ fastmap[j] = 1;
+
+ for (j = *p++ * BYTEWIDTH - 1; j >= 0; j--)
+ if (!(p[j / BYTEWIDTH] & (1 << (j % BYTEWIDTH))))
+ fastmap[j] = 1;
+ break;
+
+
+ case wordchar:
+ for (j = 0; j < (1 << BYTEWIDTH); j++)
+ if (SYNTAX (j) == Sword)
+ fastmap[j] = 1;
+ break;
+
+
+ case notwordchar:
+ for (j = 0; j < (1 << BYTEWIDTH); j++)
+ if (SYNTAX (j) != Sword)
+ fastmap[j] = 1;
+ break;
+#endif /* WCHAR */
+
+ case anychar:
+ {
+ int fastmap_newline = fastmap['\n'];
+
+ /* `.' matches anything ... */
+ for (j = 0; j < (1 << BYTEWIDTH); j++)
+ fastmap[j] = 1;
+
+ /* ... except perhaps newline. */
+ if (!(bufp->syntax & RE_DOT_NEWLINE))
+ fastmap['\n'] = fastmap_newline;
+
+ /* Return if we have already set `can_be_null'; if we have,
+ then the fastmap is irrelevant. Something's wrong here. */
+ else if (bufp->can_be_null)
+ goto done;
+
+ /* Otherwise, have to check alternative paths. */
+ break;
+ }
+
+#ifdef emacs
+ case syntaxspec:
+ k = *p++;
+ for (j = 0; j < (1 << BYTEWIDTH); j++)
+ if (SYNTAX (j) == (enum syntaxcode) k)
+ fastmap[j] = 1;
+ break;
+
+
+ case notsyntaxspec:
+ k = *p++;
+ for (j = 0; j < (1 << BYTEWIDTH); j++)
+ if (SYNTAX (j) != (enum syntaxcode) k)
+ fastmap[j] = 1;
+ break;
+
+
+ /* All cases after this match the empty string. These end with
+ `continue'. */
+
+
+ case before_dot:
+ case at_dot:
+ case after_dot:
+ continue;
+#endif /* emacs */
+
+
+ case no_op:
+ case begline:
+ case endline:
+ case begbuf:
+ case endbuf:
+ case wordbound:
+ case notwordbound:
+ case wordbeg:
+ case wordend:
+ case push_dummy_failure:
+ continue;
+
+
+ case jump_n:
+ case pop_failure_jump:
+ case maybe_pop_jump:
+ case jump:
+ case jump_past_alt:
+ case dummy_failure_jump:
+ EXTRACT_NUMBER_AND_INCR (j, p);
+ p += j;
+ if (j > 0)
+ continue;
+
+ /* Jump backward implies we just went through the body of a
+ loop and matched nothing. Opcode jumped to should be
+ `on_failure_jump' or `succeed_n'. Just treat it like an
+ ordinary jump. For a * loop, it has pushed its failure
+ point already; if so, discard that as redundant. */
+ if ((re_opcode_t) *p != on_failure_jump
+ && (re_opcode_t) *p != succeed_n)
+ continue;
+
+ p++;
+ EXTRACT_NUMBER_AND_INCR (j, p);
+ p += j;
+
+ /* If what's on the stack is where we are now, pop it. */
+ if (!FAIL_STACK_EMPTY ()
+ && fail_stack.stack[fail_stack.avail - 1].pointer == p)
+ fail_stack.avail--;
+
+ continue;
+
+
+ case on_failure_jump:
+ case on_failure_keep_string_jump:
+ handle_on_failure_jump:
+ EXTRACT_NUMBER_AND_INCR (j, p);
+
+ /* For some patterns, e.g., `(a?)?', `p+j' here points to the
+ end of the pattern. We don't want to push such a point,
+ since when we restore it above, entering the switch will
+ increment `p' past the end of the pattern. We don't need
+ to push such a point since we obviously won't find any more
+ fastmap entries beyond `pend'. Such a pattern can match
+ the null string, though. */
+ if (p + j < pend)
+ {
+ if (!PUSH_PATTERN_OP (p + j, fail_stack))
+ {
+ RESET_FAIL_STACK ();
+ return -2;
+ }
+ }
+ else
+ bufp->can_be_null = 1;
+
+ if (succeed_n_p)
+ {
+ EXTRACT_NUMBER_AND_INCR (k, p); /* Skip the n. */
+ succeed_n_p = false;
+ }
+
+ continue;
+
+
+ case succeed_n:
+ /* Get to the number of times to succeed. */
+ p += OFFSET_ADDRESS_SIZE;
+
+ /* Increment p past the n for when k != 0. */
+ EXTRACT_NUMBER_AND_INCR (k, p);
+ if (k == 0)
+ {
+ p -= 2 * OFFSET_ADDRESS_SIZE;
+ succeed_n_p = true; /* Spaghetti code alert. */
+ goto handle_on_failure_jump;
+ }
+ continue;
+
+
+ case set_number_at:
+ p += 2 * OFFSET_ADDRESS_SIZE;
+ continue;
+
+
+ case start_memory:
+ case stop_memory:
+ p += 2;
+ continue;
+
+
+ default:
+ abort (); /* We have listed all the cases. */
+ } /* switch *p++ */
+
+ /* Getting here means we have found the possible starting
+ characters for one path of the pattern -- and that the empty
+ string does not match. We need not follow this path further.
+ Instead, look at the next alternative (remembered on the
+ stack), or quit if no more. The test at the top of the loop
+ does these things. */
+ path_can_be_null = false;
+ p = pend;
+ } /* while p */
+
+ /* Set `can_be_null' for the last path (also the first path, if the
+ pattern is empty). */
+ bufp->can_be_null |= path_can_be_null;
+
+ done:
+ RESET_FAIL_STACK ();
+ return 0;
+}
+
+#else /* not INSIDE_RECURSION */
+
+int
+re_compile_fastmap (struct re_pattern_buffer *bufp)
+{
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ return wcs_re_compile_fastmap(bufp);
+ else
+# endif
+ return byte_re_compile_fastmap(bufp);
+} /* re_compile_fastmap */
+#ifdef _LIBC
+weak_alias (__re_compile_fastmap, re_compile_fastmap)
+#endif
+
+
+/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
+ ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use
+ this memory for recording register information. STARTS and ENDS
+ must be allocated using the malloc library routine, and must each
+ be at least NUM_REGS * sizeof (regoff_t) bytes long.
+
+ If NUM_REGS == 0, then subsequent matches should allocate their own
+ register data.
+
+ Unless this function is called, the first search or match using
+ PATTERN_BUFFER will allocate its own register data, without
+ freeing the old data. */
+
+void
+re_set_registers (struct re_pattern_buffer *bufp,
+ struct re_registers *regs,
+ unsigned int num_regs,
+ regoff_t *starts, regoff_t *ends)
+{
+ if (num_regs)
+ {
+ bufp->regs_allocated = REGS_REALLOCATE;
+ regs->num_regs = num_regs;
+ regs->start = starts;
+ regs->end = ends;
+ }
+ else
+ {
+ bufp->regs_allocated = REGS_UNALLOCATED;
+ regs->num_regs = 0;
+ regs->start = regs->end = (regoff_t *) 0;
+ }
+}
+#ifdef _LIBC
+weak_alias (__re_set_registers, re_set_registers)
+#endif
+
+/* Searching routines. */
+
+/* Like re_search_2, below, but only one string is specified, and
+ doesn't let you say where to stop matching. */
+
+int
+re_search (struct re_pattern_buffer *bufp,
+ const char *string,
+ int size, int startpos, int range,
+ struct re_registers *regs)
+{
+ return re_search_2 (bufp, NULL, 0, string, size, startpos, range,
+ regs, size);
+}
+#ifdef _LIBC
+weak_alias (__re_search, re_search)
+#endif
+
+
+/* Using the compiled pattern in BUFP->buffer, first tries to match the
+ virtual concatenation of STRING1 and STRING2, starting first at index
+ STARTPOS, then at STARTPOS + 1, and so on.
+
+ STRING1 and STRING2 have length SIZE1 and SIZE2, respectively.
+
+ RANGE is how far to scan while trying to match. RANGE = 0 means try
+ only at STARTPOS; in general, the last start tried is STARTPOS +
+ RANGE.
+
+ In REGS, return the indices of the virtual concatenation of STRING1
+ and STRING2 that matched the entire BUFP->buffer and its contained
+ subexpressions.
+
+ Do not consider matching one past the index STOP in the virtual
+ concatenation of STRING1 and STRING2.
+
+ We return either the position in the strings at which the match was
+ found, -1 if no match, or -2 if error (such as failure
+ stack overflow). */
+
+int
+re_search_2 (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int startpos, int range,
+ struct re_registers *regs,
+ int stop)
+{
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ return wcs_re_search_2 (bufp, string1, size1, string2, size2, startpos,
+ range, regs, stop);
+ else
+# endif
+ return byte_re_search_2 (bufp, string1, size1, string2, size2, startpos,
+ range, regs, stop);
+} /* re_search_2 */
+#ifdef _LIBC
+weak_alias (__re_search_2, re_search_2)
+#endif
+
+#endif /* not INSIDE_RECURSION */
+
+#ifdef INSIDE_RECURSION
+
+#ifdef MATCH_MAY_ALLOCATE
+# define FREE_VAR(var) if (var) REGEX_FREE (var); var = NULL
+#else
+# define FREE_VAR(var) if (var) free (var); var = NULL
+#endif
+
+#ifdef WCHAR
+# define MAX_ALLOCA_SIZE 2000
+
+# define FREE_WCS_BUFFERS() \
+ do { \
+ if (size1 > MAX_ALLOCA_SIZE) \
+ { \
+ free (wcs_string1); \
+ free (mbs_offset1); \
+ } \
+ else \
+ { \
+ FREE_VAR (wcs_string1); \
+ FREE_VAR (mbs_offset1); \
+ } \
+ if (size2 > MAX_ALLOCA_SIZE) \
+ { \
+ free (wcs_string2); \
+ free (mbs_offset2); \
+ } \
+ else \
+ { \
+ FREE_VAR (wcs_string2); \
+ FREE_VAR (mbs_offset2); \
+ } \
+ } while (0)
+
+#endif
+
+
+static int
+PREFIX(re_search_2) (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int startpos, int range,
+ struct re_registers *regs,
+ int stop)
+{
+ int val;
+ register char *fastmap = bufp->fastmap;
+ register RE_TRANSLATE_TYPE translate = bufp->translate;
+ int total_size = size1 + size2;
+ int endpos = startpos + range;
+#ifdef WCHAR
+ /* We need wchar_t* buffers correspond to cstring1, cstring2. */
+ wchar_t *wcs_string1 = NULL, *wcs_string2 = NULL;
+ /* We need the size of wchar_t buffers correspond to csize1, csize2. */
+ int wcs_size1 = 0, wcs_size2 = 0;
+ /* offset buffer for optimizatoin. See convert_mbs_to_wc. */
+ int *mbs_offset1 = NULL, *mbs_offset2 = NULL;
+ /* They hold whether each wchar_t is binary data or not. */
+ char *is_binary = NULL;
+#endif /* WCHAR */
+
+ /* Check for out-of-range STARTPOS. */
+ if (startpos < 0 || startpos > total_size)
+ return -1;
+
+ /* Fix up RANGE if it might eventually take us outside
+ the virtual concatenation of STRING1 and STRING2.
+ Make sure we won't move STARTPOS below 0 or above TOTAL_SIZE. */
+ if (endpos < 0)
+ range = 0 - startpos;
+ else if (endpos > total_size)
+ range = total_size - startpos;
+
+ /* If the search isn't to be a backwards one, don't waste time in a
+ search for a pattern that must be anchored. */
+ if (bufp->used > 0 && range > 0
+ && ((re_opcode_t) bufp->buffer[0] == begbuf
+ /* `begline' is like `begbuf' if it cannot match at newlines. */
+ || ((re_opcode_t) bufp->buffer[0] == begline
+ && !bufp->newline_anchor)))
+ {
+ if (startpos > 0)
+ return -1;
+ else
+ range = 1;
+ }
+
+#ifdef emacs
+ /* In a forward search for something that starts with \=.
+ don't keep searching past point. */
+ if (bufp->used > 0 && (re_opcode_t) bufp->buffer[0] == at_dot && range > 0)
+ {
+ range = PT - startpos;
+ if (range <= 0)
+ return -1;
+ }
+#endif /* emacs */
+
+ /* Update the fastmap now if not correct already. */
+ if (fastmap && !bufp->fastmap_accurate)
+ if (re_compile_fastmap (bufp) == -2)
+ return -2;
+
+#ifdef WCHAR
+ /* Allocate wchar_t array for wcs_string1 and wcs_string2 and
+ fill them with converted string. */
+ if (size1 != 0)
+ {
+ if (size1 > MAX_ALLOCA_SIZE)
+ {
+ wcs_string1 = TALLOC (size1 + 1, CHAR_T);
+ mbs_offset1 = TALLOC (size1 + 1, int);
+ is_binary = TALLOC (size1 + 1, char);
+ }
+ else
+ {
+ wcs_string1 = REGEX_TALLOC (size1 + 1, CHAR_T);
+ mbs_offset1 = REGEX_TALLOC (size1 + 1, int);
+ is_binary = REGEX_TALLOC (size1 + 1, char);
+ }
+ if (!wcs_string1 || !mbs_offset1 || !is_binary)
+ {
+ if (size1 > MAX_ALLOCA_SIZE)
+ {
+ free (wcs_string1);
+ free (mbs_offset1);
+ free (is_binary);
+ }
+ else
+ {
+ FREE_VAR (wcs_string1);
+ FREE_VAR (mbs_offset1);
+ FREE_VAR (is_binary);
+ }
+ return -2;
+ }
+ wcs_size1 = convert_mbs_to_wcs(wcs_string1, string1, size1,
+ mbs_offset1, is_binary);
+ wcs_string1[wcs_size1] = L'\0'; /* for a sentinel */
+ if (size1 > MAX_ALLOCA_SIZE)
+ free (is_binary);
+ else
+ FREE_VAR (is_binary);
+ }
+ if (size2 != 0)
+ {
+ if (size2 > MAX_ALLOCA_SIZE)
+ {
+ wcs_string2 = TALLOC (size2 + 1, CHAR_T);
+ mbs_offset2 = TALLOC (size2 + 1, int);
+ is_binary = TALLOC (size2 + 1, char);
+ }
+ else
+ {
+ wcs_string2 = REGEX_TALLOC (size2 + 1, CHAR_T);
+ mbs_offset2 = REGEX_TALLOC (size2 + 1, int);
+ is_binary = REGEX_TALLOC (size2 + 1, char);
+ }
+ if (!wcs_string2 || !mbs_offset2 || !is_binary)
+ {
+ FREE_WCS_BUFFERS ();
+ if (size2 > MAX_ALLOCA_SIZE)
+ free (is_binary);
+ else
+ FREE_VAR (is_binary);
+ return -2;
+ }
+ wcs_size2 = convert_mbs_to_wcs(wcs_string2, string2, size2,
+ mbs_offset2, is_binary);
+ wcs_string2[wcs_size2] = L'\0'; /* for a sentinel */
+ if (size2 > MAX_ALLOCA_SIZE)
+ free (is_binary);
+ else
+ FREE_VAR (is_binary);
+ }
+#endif /* WCHAR */
+
+
+ /* Loop through the string, looking for a place to start matching. */
+ for (;;)
+ {
+ /* If a fastmap is supplied, skip quickly over characters that
+ cannot be the start of a match. If the pattern can match the
+ null string, however, we don't need to skip characters; we want
+ the first null string. */
+ if (fastmap && startpos < total_size && !bufp->can_be_null)
+ {
+ if (range > 0) /* Searching forwards. */
+ {
+ register const char *d;
+ register int lim = 0;
+ int irange = range;
+
+ if (startpos < size1 && startpos + range >= size1)
+ lim = range - (size1 - startpos);
+
+ d = (startpos >= size1 ? string2 - size1 : string1) + startpos;
+
+ /* Written out as an if-else to avoid testing `translate'
+ inside the loop. */
+ if (translate)
+ while (range > lim
+ && !fastmap[(unsigned char)
+ translate[(unsigned char) *d++]])
+ range--;
+ else
+ while (range > lim && !fastmap[(unsigned char) *d++])
+ range--;
+
+ startpos += irange - range;
+ }
+ else /* Searching backwards. */
+ {
+ register CHAR_T c = (size1 == 0 || startpos >= size1
+ ? string2[startpos - size1]
+ : string1[startpos]);
+
+ if (!fastmap[(unsigned char) TRANSLATE (c)])
+ goto advance;
+ }
+ }
+
+ /* If can't match the null string, and that's all we have left, fail. */
+ if (range >= 0 && startpos == total_size && fastmap
+ && !bufp->can_be_null)
+ {
+#ifdef WCHAR
+ FREE_WCS_BUFFERS ();
+#endif
+ return -1;
+ }
+
+#ifdef WCHAR
+ val = wcs_re_match_2_internal (bufp, string1, size1, string2,
+ size2, startpos, regs, stop,
+ wcs_string1, wcs_size1,
+ wcs_string2, wcs_size2,
+ mbs_offset1, mbs_offset2);
+#else /* BYTE */
+ val = byte_re_match_2_internal (bufp, string1, size1, string2,
+ size2, startpos, regs, stop);
+#endif /* BYTE */
+
+#ifndef REGEX_MALLOC
+# ifdef C_ALLOCA
+ alloca (0);
+# endif
+#endif
+
+ if (val >= 0)
+ {
+#ifdef WCHAR
+ FREE_WCS_BUFFERS ();
+#endif
+ return startpos;
+ }
+
+ if (val == -2)
+ {
+#ifdef WCHAR
+ FREE_WCS_BUFFERS ();
+#endif
+ return -2;
+ }
+
+ advance:
+ if (!range)
+ break;
+ else if (range > 0)
+ {
+ range--;
+ startpos++;
+ }
+ else
+ {
+ range++;
+ startpos--;
+ }
+ }
+#ifdef WCHAR
+ FREE_WCS_BUFFERS ();
+#endif
+ return -1;
+}
+
+#ifdef WCHAR
+/* This converts PTR, a pointer into one of the search wchar_t strings
+ `string1' and `string2' into an multibyte string offset from the
+ beginning of that string. We use mbs_offset to optimize.
+ See convert_mbs_to_wcs. */
+# define POINTER_TO_OFFSET(ptr) \
+ (FIRST_STRING_P (ptr) \
+ ? ((regoff_t)(mbs_offset1 != NULL? mbs_offset1[(ptr)-string1] : 0)) \
+ : ((regoff_t)((mbs_offset2 != NULL? mbs_offset2[(ptr)-string2] : 0) \
+ + csize1)))
+#else /* BYTE */
+/* This converts PTR, a pointer into one of the search strings `string1'
+ and `string2' into an offset from the beginning of that string. */
+# define POINTER_TO_OFFSET(ptr) \
+ (FIRST_STRING_P (ptr) \
+ ? ((regoff_t) ((ptr) - string1)) \
+ : ((regoff_t) ((ptr) - string2 + size1)))
+#endif /* WCHAR */
+
+/* Macros for dealing with the split strings in re_match_2. */
+
+#define MATCHING_IN_FIRST_STRING (dend == end_match_1)
+
+/* Call before fetching a character with *d. This switches over to
+ string2 if necessary. */
+#define PREFETCH() \
+ while (d == dend) \
+ { \
+ /* End of string2 => fail. */ \
+ if (dend == end_match_2) \
+ goto fail; \
+ /* End of string1 => advance to string2. */ \
+ d = string2; \
+ dend = end_match_2; \
+ }
+
+/* Test if at very beginning or at very end of the virtual concatenation
+ of `string1' and `string2'. If only one string, it's `string2'. */
+#define AT_STRINGS_BEG(d) ((d) == (size1 ? string1 : string2) || !size2)
+#define AT_STRINGS_END(d) ((d) == end2)
+
+
+/* Test if D points to a character which is word-constituent. We have
+ two special cases to check for: if past the end of string1, look at
+ the first character in string2; and if before the beginning of
+ string2, look at the last character in string1. */
+#ifdef WCHAR
+/* Use internationalized API instead of SYNTAX. */
+# define WORDCHAR_P(d) \
+ (iswalnum ((wint_t)((d) == end1 ? *string2 \
+ : (d) == string2 - 1 ? *(end1 - 1) : *(d))) != 0 \
+ || ((d) == end1 ? *string2 \
+ : (d) == string2 - 1 ? *(end1 - 1) : *(d)) == L'_')
+#else /* BYTE */
+# define WORDCHAR_P(d) \
+ (SYNTAX ((d) == end1 ? *string2 \
+ : (d) == string2 - 1 ? *(end1 - 1) : *(d)) \
+ == Sword)
+#endif /* WCHAR */
+
+/* Disabled due to a compiler bug -- see comment at case wordbound */
+#if 0
+/* Test if the character before D and the one at D differ with respect
+ to being word-constituent. */
+#define AT_WORD_BOUNDARY(d) \
+ (AT_STRINGS_BEG (d) || AT_STRINGS_END (d) \
+ || WORDCHAR_P (d - 1) != WORDCHAR_P (d))
+#endif
+
+/* Free everything we malloc. */
+#ifdef MATCH_MAY_ALLOCATE
+# ifdef WCHAR
+# define FREE_VARIABLES() \
+ do { \
+ REGEX_FREE_STACK (fail_stack.stack); \
+ FREE_VAR (regstart); \
+ FREE_VAR (regend); \
+ FREE_VAR (old_regstart); \
+ FREE_VAR (old_regend); \
+ FREE_VAR (best_regstart); \
+ FREE_VAR (best_regend); \
+ FREE_VAR (reg_info); \
+ FREE_VAR (reg_dummy); \
+ FREE_VAR (reg_info_dummy); \
+ if (!cant_free_wcs_buf) \
+ { \
+ FREE_VAR (string1); \
+ FREE_VAR (string2); \
+ FREE_VAR (mbs_offset1); \
+ FREE_VAR (mbs_offset2); \
+ } \
+ } while (0)
+# else /* BYTE */
+# define FREE_VARIABLES() \
+ do { \
+ REGEX_FREE_STACK (fail_stack.stack); \
+ FREE_VAR (regstart); \
+ FREE_VAR (regend); \
+ FREE_VAR (old_regstart); \
+ FREE_VAR (old_regend); \
+ FREE_VAR (best_regstart); \
+ FREE_VAR (best_regend); \
+ FREE_VAR (reg_info); \
+ FREE_VAR (reg_dummy); \
+ FREE_VAR (reg_info_dummy); \
+ } while (0)
+# endif /* WCHAR */
+#else
+# ifdef WCHAR
+# define FREE_VARIABLES() \
+ do { \
+ if (!cant_free_wcs_buf) \
+ { \
+ FREE_VAR (string1); \
+ FREE_VAR (string2); \
+ FREE_VAR (mbs_offset1); \
+ FREE_VAR (mbs_offset2); \
+ } \
+ } while (0)
+# else /* BYTE */
+# define FREE_VARIABLES() ((void)0) /* Do nothing! But inhibit gcc warning. */
+# endif /* WCHAR */
+#endif /* not MATCH_MAY_ALLOCATE */
+
+/* These values must meet several constraints. They must not be valid
+ register values; since we have a limit of 255 registers (because
+ we use only one byte in the pattern for the register number), we can
+ use numbers larger than 255. They must differ by 1, because of
+ NUM_FAILURE_ITEMS above. And the value for the lowest register must
+ be larger than the value for the highest register, so we do not try
+ to actually save any registers when none are active. */
+#define NO_HIGHEST_ACTIVE_REG (1 << BYTEWIDTH)
+#define NO_LOWEST_ACTIVE_REG (NO_HIGHEST_ACTIVE_REG + 1)
+
+#else /* not INSIDE_RECURSION */
+/* Matching routines. */
+
+#ifndef emacs /* Emacs never uses this. */
+/* re_match is like re_match_2 except it takes only a single string. */
+
+int
+re_match (struct re_pattern_buffer *bufp,
+ const char *string,
+ int size, int pos,
+ struct re_registers *regs)
+{
+ int result;
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ result = wcs_re_match_2_internal (bufp, NULL, 0, string, size,
+ pos, regs, size,
+ NULL, 0, NULL, 0, NULL, NULL);
+ else
+# endif
+ result = byte_re_match_2_internal (bufp, NULL, 0, string, size,
+ pos, regs, size);
+# ifndef REGEX_MALLOC
+# ifdef C_ALLOCA
+ alloca (0);
+# endif
+# endif
+ return result;
+}
+# ifdef _LIBC
+weak_alias (__re_match, re_match)
+# endif
+#endif /* not emacs */
+
+#endif /* not INSIDE_RECURSION */
+
+#ifdef INSIDE_RECURSION
+static boolean PREFIX(group_match_null_string_p) (UCHAR_T **p,
+ UCHAR_T *end,
+ PREFIX(register_info_type) *reg_info);
+static boolean PREFIX(alt_match_null_string_p) (UCHAR_T *p,
+ UCHAR_T *end,
+ PREFIX(register_info_type) *reg_info);
+static boolean PREFIX(common_op_match_null_string_p) (UCHAR_T **p,
+ UCHAR_T *end,
+ PREFIX(register_info_type) *reg_info);
+static int PREFIX(bcmp_translate) (const CHAR_T *s1, const CHAR_T *s2,
+ int len, char *translate);
+#else /* not INSIDE_RECURSION */
+
+/* re_match_2 matches the compiled pattern in BUFP against the
+ the (virtual) concatenation of STRING1 and STRING2 (of length SIZE1
+ and SIZE2, respectively). We start matching at POS, and stop
+ matching at STOP.
+
+ If REGS is non-null and the `no_sub' field of BUFP is nonzero, we
+ store offsets for the substring each group matched in REGS. See the
+ documentation for exactly how many groups we fill.
+
+ We return -1 if no match, -2 if an internal error (such as the
+ failure stack overflowing). Otherwise, we return the length of the
+ matched substring. */
+
+int
+re_match_2 (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int pos, struct re_registers *regs,
+ int stop)
+{
+ int result;
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ result = wcs_re_match_2_internal (bufp, string1, size1, string2, size2,
+ pos, regs, stop,
+ NULL, 0, NULL, 0, NULL, NULL);
+ else
+# endif
+ result = byte_re_match_2_internal (bufp, string1, size1, string2, size2,
+ pos, regs, stop);
+
+#ifndef REGEX_MALLOC
+# ifdef C_ALLOCA
+ alloca (0);
+# endif
+#endif
+ return result;
+}
+#ifdef _LIBC
+weak_alias (__re_match_2, re_match_2)
+#endif
+
+#endif /* not INSIDE_RECURSION */
+
+#ifdef INSIDE_RECURSION
+
+#ifdef WCHAR
+
+/* This check the substring (from 0, to length) of the multibyte string,
+ to which offset_buffer correspond. And count how many wchar_t_characters
+ the substring occupy. We use offset_buffer to optimization.
+ See convert_mbs_to_wcs. */
+
+static int
+count_mbs_length (int *offset_buffer, int length)
+{
+ int upper, lower;
+
+ /* Check whether the size is valid. */
+ if (length < 0)
+ return -1;
+
+ if (offset_buffer == NULL)
+ return 0;
+
+ /* If there are no multibyte character, offset_buffer[i] == i.
+ Optmize for this case. */
+ if (offset_buffer[length] == length)
+ return length;
+
+ /* Set up upper with length. (because for all i, offset_buffer[i] >= i) */
+ upper = length;
+ lower = 0;
+
+ while (true)
+ {
+ int middle = (lower + upper) / 2;
+ if (middle == lower || middle == upper)
+ break;
+ if (offset_buffer[middle] > length)
+ upper = middle;
+ else if (offset_buffer[middle] < length)
+ lower = middle;
+ else
+ return middle;
+ }
+
+ return -1;
+}
+#endif /* WCHAR */
+
+/* This is a separate function so that we can force an alloca cleanup
+ afterwards. */
+#ifdef WCHAR
+static int
+wcs_re_match_2_internal (struct re_pattern_buffer *bufp,
+ const char *cstring1, int csize1,
+ const char *cstring2, int csize2,
+ int pos,
+ struct re_registers *regs,
+ int stop,
+ /* string1 == string2 == NULL means
+ string1/2, size1/2 and mbs_offset1/2 need
+ setting up in this function. */
+ /* We need wchar_t * buffers corresponding to
+ cstring1, cstring2. */
+ wchar_t *string1, int size1,
+ wchar_t *string2, int size2,
+ /* Offset buffer for optimization. See
+ convert_mbs_to_wc. */
+ int *mbs_offset1,
+ int *mbs_offset2)
+#else /* BYTE */
+static int
+byte_re_match_2_internal (struct re_pattern_buffer *bufp,
+ const char *string1, int size1,
+ const char *string2, int size2,
+ int pos,
+ struct re_registers *regs,
+ int stop)
+#endif /* BYTE */
+{
+ /* General temporaries. */
+ int mcnt;
+ UCHAR_T *p1;
+#ifdef WCHAR
+ /* They hold whether each wchar_t is binary data or not. */
+ char *is_binary = NULL;
+ /* If true, we can't free string1/2, mbs_offset1/2. */
+ int cant_free_wcs_buf = 1;
+#endif /* WCHAR */
+
+ /* Just past the end of the corresponding string. */
+ const CHAR_T *end1, *end2;
+
+ /* Pointers into string1 and string2, just past the last characters in
+ each to consider matching. */
+ const CHAR_T *end_match_1, *end_match_2;
+
+ /* Where we are in the data, and the end of the current string. */
+ const CHAR_T *d, *dend;
+
+ /* Where we are in the pattern, and the end of the pattern. */
+#ifdef WCHAR
+ UCHAR_T *pattern, *p;
+ register UCHAR_T *pend;
+#else /* BYTE */
+ UCHAR_T *p = bufp->buffer;
+ register UCHAR_T *pend = p + bufp->used;
+#endif /* WCHAR */
+
+ /* Mark the opcode just after a start_memory, so we can test for an
+ empty subpattern when we get to the stop_memory. */
+ UCHAR_T *just_past_start_mem = 0;
+
+ /* We use this to map every character in the string. */
+ RE_TRANSLATE_TYPE translate = bufp->translate;
+
+ /* Failure point stack. Each place that can handle a failure further
+ down the line pushes a failure point on this stack. It consists of
+ restart, regend, and reg_info for all registers corresponding to
+ the subexpressions we're currently inside, plus the number of such
+ registers, and, finally, two char *'s. The first char * is where
+ to resume scanning the pattern; the second one is where to resume
+ scanning the strings. If the latter is zero, the failure point is
+ a ``dummy''; if a failure happens and the failure point is a dummy,
+ it gets discarded and the next next one is tried. */
+#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */
+ PREFIX(fail_stack_type) fail_stack;
+#endif
+#ifdef DEBUG
+ static unsigned failure_id;
+ unsigned nfailure_points_pushed = 0, nfailure_points_popped = 0;
+#endif
+
+#ifdef REL_ALLOC
+ /* This holds the pointer to the failure stack, when
+ it is allocated relocatably. */
+ fail_stack_elt_t *failure_stack_ptr;
+#endif
+
+ /* We fill all the registers internally, independent of what we
+ return, for use in backreferences. The number here includes
+ an element for register zero. */
+ size_t num_regs = bufp->re_nsub + 1;
+
+ /* The currently active registers. */
+ active_reg_t lowest_active_reg = NO_LOWEST_ACTIVE_REG;
+ active_reg_t highest_active_reg = NO_HIGHEST_ACTIVE_REG;
+
+ /* Information on the contents of registers. These are pointers into
+ the input strings; they record just what was matched (on this
+ attempt) by a subexpression part of the pattern, that is, the
+ regnum-th regstart pointer points to where in the pattern we began
+ matching and the regnum-th regend points to right after where we
+ stopped matching the regnum-th subexpression. (The zeroth register
+ keeps track of what the whole pattern matches.) */
+#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
+ const CHAR_T **regstart, **regend;
+#endif
+
+ /* If a group that's operated upon by a repetition operator fails to
+ match anything, then the register for its start will need to be
+ restored because it will have been set to wherever in the string we
+ are when we last see its open-group operator. Similarly for a
+ register's end. */
+#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
+ const CHAR_T **old_regstart, **old_regend;
+#endif
+
+ /* The is_active field of reg_info helps us keep track of which (possibly
+ nested) subexpressions we are currently in. The matched_something
+ field of reg_info[reg_num] helps us tell whether or not we have
+ matched any of the pattern so far this time through the reg_num-th
+ subexpression. These two fields get reset each time through any
+ loop their register is in. */
+#ifdef MATCH_MAY_ALLOCATE /* otherwise, this is global. */
+ PREFIX(register_info_type) *reg_info;
+#endif
+
+ /* The following record the register info as found in the above
+ variables when we find a match better than any we've seen before.
+ This happens as we backtrack through the failure points, which in
+ turn happens only if we have not yet matched the entire string. */
+ unsigned best_regs_set = false;
+#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
+ const CHAR_T **best_regstart, **best_regend;
+#endif
+
+ /* Logically, this is `best_regend[0]'. But we don't want to have to
+ allocate space for that if we're not allocating space for anything
+ else (see below). Also, we never need info about register 0 for
+ any of the other register vectors, and it seems rather a kludge to
+ treat `best_regend' differently than the rest. So we keep track of
+ the end of the best match so far in a separate variable. We
+ initialize this to NULL so that when we backtrack the first time
+ and need to test it, it's not garbage. */
+ const CHAR_T *match_end = NULL;
+
+ /* This helps SET_REGS_MATCHED avoid doing redundant work. */
+ int set_regs_matched_done = 0;
+
+ /* Used when we pop values we don't care about. */
+#ifdef MATCH_MAY_ALLOCATE /* otherwise, these are global. */
+ const CHAR_T **reg_dummy;
+ PREFIX(register_info_type) *reg_info_dummy;
+#endif
+
+#ifdef DEBUG
+ /* Counts the total number of registers pushed. */
+ unsigned num_regs_pushed = 0;
+#endif
+
+ /* Definitions for state transitions. More efficiently for gcc. */
+#ifdef __GNUC__
+# if defined HAVE_SUBTRACT_LOCAL_LABELS && defined SHARED
+# define NEXT \
+ do \
+ { \
+ int offset; \
+ const void *__unbounded ptr; \
+ offset = (p == pend \
+ ? 0 : jmptable[SWITCH_ENUM_CAST ((re_opcode_t) *p++)]); \
+ ptr = &&end_of_pattern + offset; \
+ goto *ptr; \
+ } \
+ while (0)
+# define REF(x) \
+ &&label_##x - &&end_of_pattern
+# define JUMP_TABLE_TYPE const int
+# else
+# define NEXT \
+ do \
+ { \
+ const void *__unbounded ptr; \
+ ptr = (p == pend ? &&end_of_pattern \
+ : jmptable[SWITCH_ENUM_CAST ((re_opcode_t) *p++)]); \
+ goto *ptr; \
+ } \
+ while (0)
+# define REF(x) \
+ &&label_##x
+# define JUMP_TABLE_TYPE const void *const
+# endif
+# define CASE(x) label_##x
+ static JUMP_TABLE_TYPE jmptable[] =
+ {
+ REF (no_op),
+ REF (succeed),
+ REF (exactn),
+# ifdef MBS_SUPPORT
+ REF (exactn_bin),
+# endif
+ REF (anychar),
+ REF (charset),
+ REF (charset_not),
+ REF (start_memory),
+ REF (stop_memory),
+ REF (duplicate),
+ REF (begline),
+ REF (endline),
+ REF (begbuf),
+ REF (endbuf),
+ REF (jump),
+ REF (jump_past_alt),
+ REF (on_failure_jump),
+ REF (on_failure_keep_string_jump),
+ REF (pop_failure_jump),
+ REF (maybe_pop_jump),
+ REF (dummy_failure_jump),
+ REF (push_dummy_failure),
+ REF (succeed_n),
+ REF (jump_n),
+ REF (set_number_at),
+ REF (wordchar),
+ REF (notwordchar),
+ REF (wordbeg),
+ REF (wordend),
+ REF (wordbound),
+ REF (notwordbound)
+# ifdef emacs
+ ,REF (before_dot),
+ REF (at_dot),
+ REF (after_dot),
+ REF (syntaxspec),
+ REF (notsyntaxspec)
+# endif
+ };
+#else
+# define NEXT \
+ break
+# define CASE(x) \
+ case x
+#endif
+
+ DEBUG_PRINT1 ("\n\nEntering re_match_2.\n");
+
+ INIT_FAIL_STACK ();
+
+#ifdef MATCH_MAY_ALLOCATE
+ /* Do not bother to initialize all the register variables if there are
+ no groups in the pattern, as it takes a fair amount of time. If
+ there are groups, we include space for register 0 (the whole
+ pattern), even though we never use it, since it simplifies the
+ array indexing. We should fix this. */
+ if (bufp->re_nsub)
+ {
+ regstart = REGEX_TALLOC (num_regs, const CHAR_T *);
+ regend = REGEX_TALLOC (num_regs, const CHAR_T *);
+ old_regstart = REGEX_TALLOC (num_regs, const CHAR_T *);
+ old_regend = REGEX_TALLOC (num_regs, const CHAR_T *);
+ best_regstart = REGEX_TALLOC (num_regs, const CHAR_T *);
+ best_regend = REGEX_TALLOC (num_regs, const CHAR_T *);
+ reg_info = REGEX_TALLOC (num_regs, PREFIX(register_info_type));
+ reg_dummy = REGEX_TALLOC (num_regs, const CHAR_T *);
+ reg_info_dummy = REGEX_TALLOC (num_regs, PREFIX(register_info_type));
+
+ if (!(regstart && regend && old_regstart && old_regend && reg_info
+ && best_regstart && best_regend && reg_dummy && reg_info_dummy))
+ {
+ FREE_VARIABLES ();
+ return -2;
+ }
+ }
+ else
+ {
+ /* We must initialize all our variables to NULL, so that
+ `FREE_VARIABLES' doesn't try to free them. */
+ regstart = regend = old_regstart = old_regend = best_regstart
+ = best_regend = reg_dummy = NULL;
+ reg_info = reg_info_dummy = (PREFIX(register_info_type) *) NULL;
+ }
+#endif /* MATCH_MAY_ALLOCATE */
+
+ /* The starting position is bogus. */
+#ifdef WCHAR
+ if (pos < 0 || pos > csize1 + csize2)
+#else /* BYTE */
+ if (pos < 0 || pos > size1 + size2)
+#endif
+ {
+ FREE_VARIABLES ();
+ return -1;
+ }
+
+#ifdef WCHAR
+ /* Allocate wchar_t array for string1 and string2 and
+ fill them with converted string. */
+ if (string1 == NULL && string2 == NULL)
+ {
+ /* We need seting up buffers here. */
+
+ /* We must free wcs buffers in this function. */
+ cant_free_wcs_buf = 0;
+
+ if (csize1 != 0)
+ {
+ string1 = REGEX_TALLOC (csize1 + 1, CHAR_T);
+ mbs_offset1 = REGEX_TALLOC (csize1 + 1, int);
+ is_binary = REGEX_TALLOC (csize1 + 1, char);
+ if (!string1 || !mbs_offset1 || !is_binary)
+ {
+ FREE_VAR (string1);
+ FREE_VAR (mbs_offset1);
+ FREE_VAR (is_binary);
+ return -2;
+ }
+ }
+ if (csize2 != 0)
+ {
+ string2 = REGEX_TALLOC (csize2 + 1, CHAR_T);
+ mbs_offset2 = REGEX_TALLOC (csize2 + 1, int);
+ is_binary = REGEX_TALLOC (csize2 + 1, char);
+ if (!string2 || !mbs_offset2 || !is_binary)
+ {
+ FREE_VAR (string1);
+ FREE_VAR (mbs_offset1);
+ FREE_VAR (string2);
+ FREE_VAR (mbs_offset2);
+ FREE_VAR (is_binary);
+ return -2;
+ }
+ size2 = convert_mbs_to_wcs(string2, cstring2, csize2,
+ mbs_offset2, is_binary);
+ string2[size2] = L'\0'; /* for a sentinel */
+ FREE_VAR (is_binary);
+ }
+ }
+
+ /* We need to cast pattern to (wchar_t*), because we casted this compiled
+ pattern to (char*) in regex_compile. */
+ p = pattern = (CHAR_T*)bufp->buffer;
+ pend = (CHAR_T*)(bufp->buffer + bufp->used);
+
+#endif /* WCHAR */
+
+ /* Initialize subexpression text positions to -1 to mark ones that no
+ start_memory/stop_memory has been seen for. Also initialize the
+ register information struct. */
+ for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
+ {
+ regstart[mcnt] = regend[mcnt]
+ = old_regstart[mcnt] = old_regend[mcnt] = REG_UNSET_VALUE;
+
+ REG_MATCH_NULL_STRING_P (reg_info[mcnt]) = MATCH_NULL_UNSET_VALUE;
+ IS_ACTIVE (reg_info[mcnt]) = 0;
+ MATCHED_SOMETHING (reg_info[mcnt]) = 0;
+ EVER_MATCHED_SOMETHING (reg_info[mcnt]) = 0;
+ }
+
+ /* We move `string1' into `string2' if the latter's empty -- but not if
+ `string1' is null. */
+ if (size2 == 0 && string1 != NULL)
+ {
+ string2 = string1;
+ size2 = size1;
+ string1 = 0;
+ size1 = 0;
+#ifdef WCHAR
+ mbs_offset2 = mbs_offset1;
+ csize2 = csize1;
+ mbs_offset1 = NULL;
+ csize1 = 0;
+#endif
+ }
+ end1 = string1 + size1;
+ end2 = string2 + size2;
+
+ /* Compute where to stop matching, within the two strings. */
+#ifdef WCHAR
+ if (stop <= csize1)
+ {
+ mcnt = count_mbs_length(mbs_offset1, stop);
+ end_match_1 = string1 + mcnt;
+ end_match_2 = string2;
+ }
+ else
+ {
+ if (stop > csize1 + csize2)
+ stop = csize1 + csize2;
+ end_match_1 = end1;
+ mcnt = count_mbs_length(mbs_offset2, stop-csize1);
+ end_match_2 = string2 + mcnt;
+ }
+ if (mcnt < 0)
+ { /* count_mbs_length return error. */
+ FREE_VARIABLES ();
+ return -1;
+ }
+#else
+ if (stop <= size1)
+ {
+ end_match_1 = string1 + stop;
+ end_match_2 = string2;
+ }
+ else
+ {
+ end_match_1 = end1;
+ end_match_2 = string2 + stop - size1;
+ }
+#endif /* WCHAR */
+
+ /* `p' scans through the pattern as `d' scans through the data.
+ `dend' is the end of the input string that `d' points within. `d'
+ is advanced into the following input string whenever necessary, but
+ this happens before fetching; therefore, at the beginning of the
+ loop, `d' can be pointing at the end of a string, but it cannot
+ equal `string2'. */
+#ifdef WCHAR
+ if (size1 > 0 && pos <= csize1)
+ {
+ mcnt = count_mbs_length(mbs_offset1, pos);
+ d = string1 + mcnt;
+ dend = end_match_1;
+ }
+ else
+ {
+ mcnt = count_mbs_length(mbs_offset2, pos-csize1);
+ d = string2 + mcnt;
+ dend = end_match_2;
+ }
+
+ if (mcnt < 0)
+ { /* count_mbs_length return error. */
+ FREE_VARIABLES ();
+ return -1;
+ }
+#else
+ if (size1 > 0 && pos <= size1)
+ {
+ d = string1 + pos;
+ dend = end_match_1;
+ }
+ else
+ {
+ d = string2 + pos - size1;
+ dend = end_match_2;
+ }
+#endif /* WCHAR */
+
+ DEBUG_PRINT1 ("The compiled pattern is:\n");
+ DEBUG_PRINT_COMPILED_PATTERN (bufp, p, pend);
+ DEBUG_PRINT1 ("The string to match is: `");
+ DEBUG_PRINT_DOUBLE_STRING (d, string1, size1, string2, size2);
+ DEBUG_PRINT1 ("'\n");
+
+ /* This loops over pattern commands. It exits by returning from the
+ function if the match is complete, or it drops through if the match
+ fails at this starting point in the input data. */
+ for (;;)
+ {
+#ifdef _LIBC
+ DEBUG_PRINT2 ("\n%p: ", p);
+#else
+ DEBUG_PRINT2 ("\n0x%x: ", p);
+#endif
+
+#ifdef __GNUC__
+ NEXT;
+#else
+ if (p == pend)
+#endif
+ {
+#ifdef __GNUC__
+ end_of_pattern:
+#endif
+ /* End of pattern means we might have succeeded. */
+ DEBUG_PRINT1 ("end of pattern ... ");
+
+ /* If we haven't matched the entire string, and we want the
+ longest match, try backtracking. */
+ if (d != end_match_2)
+ {
+ /* 1 if this match ends in the same string (string1 or string2)
+ as the best previous match. */
+ boolean same_str_p = (FIRST_STRING_P (match_end)
+ == MATCHING_IN_FIRST_STRING);
+ /* 1 if this match is the best seen so far. */
+ boolean best_match_p;
+
+ /* AIX compiler got confused when this was combined
+ with the previous declaration. */
+ if (same_str_p)
+ best_match_p = d > match_end;
+ else
+ best_match_p = !MATCHING_IN_FIRST_STRING;
+
+ DEBUG_PRINT1 ("backtracking.\n");
+
+ if (!FAIL_STACK_EMPTY ())
+ { /* More failure points to try. */
+
+ /* If exceeds best match so far, save it. */
+ if (!best_regs_set || best_match_p)
+ {
+ best_regs_set = true;
+ match_end = d;
+
+ DEBUG_PRINT1 ("\nSAVING match as best so far.\n");
+
+ for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
+ {
+ best_regstart[mcnt] = regstart[mcnt];
+ best_regend[mcnt] = regend[mcnt];
+ }
+ }
+ goto fail;
+ }
+
+ /* If no failure points, don't restore garbage. And if
+ last match is real best match, don't restore second
+ best one. */
+ else if (best_regs_set && !best_match_p)
+ {
+ restore_best_regs:
+ /* Restore best match. It may happen that `dend ==
+ end_match_1' while the restored d is in string2.
+ For example, the pattern `x.*y.*z' against the
+ strings `x-' and `y-z-', if the two strings are
+ not consecutive in memory. */
+ DEBUG_PRINT1 ("Restoring best registers.\n");
+
+ d = match_end;
+ dend = ((d >= string1 && d <= end1)
+ ? end_match_1 : end_match_2);
+
+ for (mcnt = 1; (unsigned) mcnt < num_regs; mcnt++)
+ {
+ regstart[mcnt] = best_regstart[mcnt];
+ regend[mcnt] = best_regend[mcnt];
+ }
+ }
+ } /* d != end_match_2 */
+
+ succeed_label:
+ DEBUG_PRINT1 ("Accepting match.\n");
+ /* If caller wants register contents data back, do it. */
+ if (regs && !bufp->no_sub)
+ {
+ /* Have the register data arrays been allocated? */
+ if (bufp->regs_allocated == REGS_UNALLOCATED)
+ { /* No. So allocate them with malloc. We need one
+ extra element beyond `num_regs' for the `-1' marker
+ GNU code uses. */
+ regs->num_regs = MAX (RE_NREGS, num_regs + 1);
+ regs->start = TALLOC (regs->num_regs, regoff_t);
+ regs->end = TALLOC (regs->num_regs, regoff_t);
+ if (regs->start == NULL || regs->end == NULL)
+ {
+ FREE_VARIABLES ();
+ return -2;
+ }
+ bufp->regs_allocated = REGS_REALLOCATE;
+ }
+ else if (bufp->regs_allocated == REGS_REALLOCATE)
+ { /* Yes. If we need more elements than were already
+ allocated, reallocate them. If we need fewer, just
+ leave it alone. */
+ if (regs->num_regs < num_regs + 1)
+ {
+ regs->num_regs = num_regs + 1;
+ RETALLOC (regs->start, regs->num_regs, regoff_t);
+ RETALLOC (regs->end, regs->num_regs, regoff_t);
+ if (regs->start == NULL || regs->end == NULL)
+ {
+ FREE_VARIABLES ();
+ return -2;
+ }
+ }
+ }
+ else
+ {
+ /* These braces fend off a "empty body in an else-statement"
+ warning under GCC when assert expands to nothing. */
+ assert (bufp->regs_allocated == REGS_FIXED);
+ }
+
+ /* Convert the pointer data in `regstart' and `regend' to
+ indices. Register zero has to be set differently,
+ since we haven't kept track of any info for it. */
+ if (regs->num_regs > 0)
+ {
+ regs->start[0] = pos;
+#ifdef WCHAR
+ if (MATCHING_IN_FIRST_STRING)
+ regs->end[0] = (mbs_offset1 != NULL ?
+ mbs_offset1[d-string1] : 0);
+ else
+ regs->end[0] = csize1 + (mbs_offset2 != NULL
+ ? mbs_offset2[d-string2] : 0);
+#else
+ regs->end[0] = (MATCHING_IN_FIRST_STRING
+ ? ((regoff_t) (d - string1))
+ : ((regoff_t) (d - string2 + size1)));
+#endif /* WCHAR */
+ }
+
+ /* Go through the first `min (num_regs, regs->num_regs)'
+ registers, since that is all we initialized. */
+ for (mcnt = 1; (unsigned) mcnt < MIN (num_regs, regs->num_regs);
+ mcnt++)
+ {
+ if (REG_UNSET (regstart[mcnt]) || REG_UNSET (regend[mcnt]))
+ regs->start[mcnt] = regs->end[mcnt] = -1;
+ else
+ {
+ regs->start[mcnt]
+ = (regoff_t) POINTER_TO_OFFSET (regstart[mcnt]);
+ regs->end[mcnt]
+ = (regoff_t) POINTER_TO_OFFSET (regend[mcnt]);
+ }
+ }
+
+ /* If the regs structure we return has more elements than
+ were in the pattern, set the extra elements to -1. If
+ we (re)allocated the registers, this is the case,
+ because we always allocate enough to have at least one
+ -1 at the end. */
+ for (mcnt = num_regs; (unsigned) mcnt < regs->num_regs; mcnt++)
+ regs->start[mcnt] = regs->end[mcnt] = -1;
+ } /* regs && !bufp->no_sub */
+
+ DEBUG_PRINT4 ("%u failure points pushed, %u popped (%u remain).\n",
+ nfailure_points_pushed, nfailure_points_popped,
+ nfailure_points_pushed - nfailure_points_popped);
+ DEBUG_PRINT2 ("%u registers pushed.\n", num_regs_pushed);
+
+#ifdef WCHAR
+ if (MATCHING_IN_FIRST_STRING)
+ mcnt = mbs_offset1 != NULL ? mbs_offset1[d-string1] : 0;
+ else
+ mcnt = (mbs_offset2 != NULL ? mbs_offset2[d-string2] : 0) +
+ csize1;
+ mcnt -= pos;
+#else
+ mcnt = d - pos - (MATCHING_IN_FIRST_STRING
+ ? string1 : string2 - size1);
+#endif /* WCHAR */
+
+ DEBUG_PRINT2 ("Returning %d from re_match_2.\n", mcnt);
+
+ FREE_VARIABLES ();
+ return mcnt;
+ }
+
+#ifndef __GNUC__
+ /* Otherwise match next pattern command. */
+ switch (SWITCH_ENUM_CAST ((re_opcode_t) *p++))
+ {
+#endif
+ /* Ignore these. Used to ignore the n of succeed_n's which
+ currently have n == 0. */
+ CASE (no_op):
+ DEBUG_PRINT1 ("EXECUTING no_op.\n");
+ NEXT;
+
+ CASE (succeed):
+ DEBUG_PRINT1 ("EXECUTING succeed.\n");
+ goto succeed_label;
+
+ /* Match the next n pattern characters exactly. The following
+ byte in the pattern defines n, and the n bytes after that
+ are the characters to match. */
+ CASE (exactn):
+#ifdef MBS_SUPPORT
+ CASE (exactn_bin):
+#endif
+ mcnt = *p++;
+ DEBUG_PRINT2 ("EXECUTING exactn %d.\n", mcnt);
+
+ /* This is written out as an if-else so we don't waste time
+ testing `translate' inside the loop. */
+ if (translate)
+ {
+ do
+ {
+ PREFETCH ();
+#ifdef WCHAR
+ if (*d <= 0xff)
+ {
+ if ((UCHAR_T) translate[(unsigned char) *d++]
+ != (UCHAR_T) *p++)
+ goto fail;
+ }
+ else
+ {
+ if (*d++ != (CHAR_T) *p++)
+ goto fail;
+ }
+#else
+ if ((UCHAR_T) translate[(unsigned char) *d++]
+ != (UCHAR_T) *p++)
+ goto fail;
+#endif /* WCHAR */
+ }
+ while (--mcnt);
+ }
+ else
+ {
+ do
+ {
+ PREFETCH ();
+ if (*d++ != (CHAR_T) *p++) goto fail;
+ }
+ while (--mcnt);
+ }
+ SET_REGS_MATCHED ();
+ NEXT;
+
+
+ /* Match any character except possibly a newline or a null. */
+ CASE (anychar):
+ DEBUG_PRINT1 ("EXECUTING anychar.\n");
+
+ PREFETCH ();
+
+ if ((!(bufp->syntax & RE_DOT_NEWLINE) && TRANSLATE (*d) == '\n')
+ || (bufp->syntax & RE_DOT_NOT_NULL && TRANSLATE (*d) == '\000'))
+ goto fail;
+
+ SET_REGS_MATCHED ();
+ DEBUG_PRINT2 (" Matched `%ld'.\n", (long int) *d);
+ d++;
+ NEXT;
+
+
+ CASE (charset):
+ CASE (charset_not):
+ {
+ register UCHAR_T c;
+#ifdef WCHAR
+ unsigned int i, char_class_length, coll_symbol_length,
+ equiv_class_length, ranges_length, chars_length, length;
+ CHAR_T *workp, *workp2, *charset_top;
+#define WORK_BUFFER_SIZE 128
+ CHAR_T str_buf[WORK_BUFFER_SIZE];
+# ifdef _LIBC
+ uint32_t nrules;
+# endif /* _LIBC */
+#endif /* WCHAR */
+ boolean not = (re_opcode_t) *(p - 1) == charset_not;
+
+ DEBUG_PRINT2 ("EXECUTING charset%s.\n", not ? "_not" : "");
+ PREFETCH ();
+ c = TRANSLATE (*d); /* The character to match. */
+#ifdef WCHAR
+# ifdef _LIBC
+ nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
+# endif /* _LIBC */
+ charset_top = p - 1;
+ char_class_length = *p++;
+ coll_symbol_length = *p++;
+ equiv_class_length = *p++;
+ ranges_length = *p++;
+ chars_length = *p++;
+ /* p points charset[6], so the address of the next instruction
+ (charset[l+m+n+2o+k+p']) equals p[l+m+n+2*o+p'],
+ where l=length of char_classes, m=length of collating_symbol,
+ n=equivalence_class, o=length of char_range,
+ p'=length of character. */
+ workp = p;
+ /* Update p to indicate the next instruction. */
+ p += char_class_length + coll_symbol_length+ equiv_class_length +
+ 2*ranges_length + chars_length;
+
+ /* match with char_class? */
+ for (i = 0; i < char_class_length ; i += CHAR_CLASS_SIZE)
+ {
+ wctype_t wctype;
+ uintptr_t alignedp = ((uintptr_t)workp
+ + __alignof__(wctype_t) - 1)
+ & ~(uintptr_t)(__alignof__(wctype_t) - 1);
+ wctype = *((wctype_t*)alignedp);
+ workp += CHAR_CLASS_SIZE;
+ if (iswctype((wint_t)c, wctype))
+ goto char_set_matched;
+ }
+
+ /* match with collating_symbol? */
+# ifdef _LIBC
+ if (nrules != 0)
+ {
+ const unsigned char *extra = (const unsigned char *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB);
+
+ for (workp2 = workp + coll_symbol_length ; workp < workp2 ;
+ workp++)
+ {
+ int32_t *wextra;
+ wextra = (int32_t*)(extra + *workp++);
+ for (i = 0; i < *wextra; ++i)
+ if (TRANSLATE(d[i]) != wextra[1 + i])
+ break;
+
+ if (i == *wextra)
+ {
+ /* Update d, however d will be incremented at
+ char_set_matched:, we decrement d here. */
+ d += i - 1;
+ goto char_set_matched;
+ }
+ }
+ }
+ else /* (nrules == 0) */
+# endif
+ /* If we can't look up collation data, we use wcscoll
+ instead. */
+ {
+ for (workp2 = workp + coll_symbol_length ; workp < workp2 ;)
+ {
+ const CHAR_T *backup_d = d, *backup_dend = dend;
+ length = wcslen (workp);
+
+ /* If wcscoll(the collating symbol, whole string) > 0,
+ any substring of the string never match with the
+ collating symbol. */
+ if (wcscoll (workp, d) > 0)
+ {
+ workp += length + 1;
+ continue;
+ }
+
+ /* First, we compare the collating symbol with
+ the first character of the string.
+ If it don't match, we add the next character to
+ the compare buffer in turn. */
+ for (i = 0 ; i < WORK_BUFFER_SIZE-1 ; i++, d++)
+ {
+ int match;
+ if (d == dend)
+ {
+ if (dend == end_match_2)
+ break;
+ d = string2;
+ dend = end_match_2;
+ }
+
+ /* add next character to the compare buffer. */
+ str_buf[i] = TRANSLATE(*d);
+ str_buf[i+1] = '\0';
+
+ match = wcscoll (workp, str_buf);
+ if (match == 0)
+ goto char_set_matched;
+
+ if (match < 0)
+ /* (str_buf > workp) indicate (str_buf + X > workp),
+ because for all X (str_buf + X > str_buf).
+ So we don't need continue this loop. */
+ break;
+
+ /* Otherwise(str_buf < workp),
+ (str_buf+next_character) may equals (workp).
+ So we continue this loop. */
+ }
+ /* not matched */
+ d = backup_d;
+ dend = backup_dend;
+ workp += length + 1;
+ }
+ }
+ /* match with equivalence_class? */
+# ifdef _LIBC
+ if (nrules != 0)
+ {
+ const CHAR_T *backup_d = d, *backup_dend = dend;
+ /* Try to match the equivalence class against
+ those known to the collate implementation. */
+ const int32_t *table;
+ const int32_t *weights;
+ const int32_t *extra;
+ const int32_t *indirect;
+ int32_t idx, idx2;
+ wint_t *cp;
+ size_t len;
+
+ /* This #include defines a local function! */
+# include <locale/weightwc.h>
+
+ table = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEWC);
+ weights = (const wint_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTWC);
+ extra = (const wint_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAWC);
+ indirect = (const int32_t *)
+ _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTWC);
+
+ /* Write 1 collating element to str_buf, and
+ get its index. */
+ idx2 = 0;
+
+ for (i = 0 ; idx2 == 0 && i < WORK_BUFFER_SIZE - 1; i++)
+ {
+ cp = (wint_t*)str_buf;
+ if (d == dend)
+ {
+ if (dend == end_match_2)
+ break;
+ d = string2;
+ dend = end_match_2;
+ }
+ str_buf[i] = TRANSLATE(*(d+i));
+ str_buf[i+1] = '\0'; /* sentinel */
+ idx2 = findidx ((const wint_t**)&cp);
+ }
+
+ /* Update d, however d will be incremented at
+ char_set_matched:, we decrement d here. */
+ d = backup_d + ((wchar_t*)cp - (wchar_t*)str_buf - 1);
+ if (d >= dend)
+ {
+ if (dend == end_match_2)
+ d = dend;
+ else
+ {
+ d = string2;
+ dend = end_match_2;
+ }
+ }
+
+ len = weights[idx2];
+
+ for (workp2 = workp + equiv_class_length ; workp < workp2 ;
+ workp++)
+ {
+ idx = (int32_t)*workp;
+ /* We already checked idx != 0 in regex_compile. */
+
+ if (idx2 != 0 && len == weights[idx])
+ {
+ int cnt = 0;
+ while (cnt < len && (weights[idx + 1 + cnt]
+ == weights[idx2 + 1 + cnt]))
+ ++cnt;
+
+ if (cnt == len)
+ goto char_set_matched;
+ }
+ }
+ /* not matched */
+ d = backup_d;
+ dend = backup_dend;
+ }
+ else /* (nrules == 0) */
+# endif
+ /* If we can't look up collation data, we use wcscoll
+ instead. */
+ {
+ for (workp2 = workp + equiv_class_length ; workp < workp2 ;)
+ {
+ const CHAR_T *backup_d = d, *backup_dend = dend;
+ length = wcslen (workp);
+
+ /* If wcscoll(the collating symbol, whole string) > 0,
+ any substring of the string never match with the
+ collating symbol. */
+ if (wcscoll (workp, d) > 0)
+ {
+ workp += length + 1;
+ break;
+ }
+
+ /* First, we compare the equivalence class with
+ the first character of the string.
+ If it don't match, we add the next character to
+ the compare buffer in turn. */
+ for (i = 0 ; i < WORK_BUFFER_SIZE - 1 ; i++, d++)
+ {
+ int match;
+ if (d == dend)
+ {
+ if (dend == end_match_2)
+ break;
+ d = string2;
+ dend = end_match_2;
+ }
+
+ /* add next character to the compare buffer. */
+ str_buf[i] = TRANSLATE(*d);
+ str_buf[i+1] = '\0';
+
+ match = wcscoll (workp, str_buf);
+
+ if (match == 0)
+ goto char_set_matched;
+
+ if (match < 0)
+ /* (str_buf > workp) indicate (str_buf + X > workp),
+ because for all X (str_buf + X > str_buf).
+ So we don't need continue this loop. */
+ break;
+
+ /* Otherwise(str_buf < workp),
+ (str_buf+next_character) may equals (workp).
+ So we continue this loop. */
+ }
+ /* not matched */
+ d = backup_d;
+ dend = backup_dend;
+ workp += length + 1;
+ }
+ }
+
+ /* match with char_range? */
+# ifdef _LIBC
+ if (nrules != 0)
+ {
+ uint32_t collseqval;
+ const char *collseq = (const char *)
+ _NL_CURRENT(LC_COLLATE, _NL_COLLATE_COLLSEQWC);
+
+ collseqval = collseq_table_lookup (collseq, c);
+
+ for (; workp < p - chars_length ;)
+ {
+ uint32_t start_val, end_val;
+
+ /* We already compute the collation sequence value
+ of the characters (or collating symbols). */
+ start_val = (uint32_t) *workp++; /* range_start */
+ end_val = (uint32_t) *workp++; /* range_end */
+
+ if (start_val <= collseqval && collseqval <= end_val)
+ goto char_set_matched;
+ }
+ }
+ else
+# endif
+ {
+ /* We set range_start_char at str_buf[0], range_end_char
+ at str_buf[4], and compared char at str_buf[2]. */
+ str_buf[1] = 0;
+ str_buf[2] = c;
+ str_buf[3] = 0;
+ str_buf[5] = 0;
+ for (; workp < p - chars_length ;)
+ {
+ wchar_t *range_start_char, *range_end_char;
+
+ /* match if (range_start_char <= c <= range_end_char). */
+
+ /* If range_start(or end) < 0, we assume -range_start(end)
+ is the offset of the collating symbol which is specified
+ as the character of the range start(end). */
+
+ /* range_start */
+ if (*workp < 0)
+ range_start_char = charset_top - (*workp++);
+ else
+ {
+ str_buf[0] = *workp++;
+ range_start_char = str_buf;
+ }
+
+ /* range_end */
+ if (*workp < 0)
+ range_end_char = charset_top - (*workp++);
+ else
+ {
+ str_buf[4] = *workp++;
+ range_end_char = str_buf + 4;
+ }
+
+ if (wcscoll (range_start_char, str_buf+2) <= 0
+ && wcscoll (str_buf+2, range_end_char) <= 0)
+ goto char_set_matched;
+ }
+ }
+
+ /* match with char? */
+ for (; workp < p ; workp++)
+ if (c == *workp)
+ goto char_set_matched;
+
+ not = !not;
+
+ char_set_matched:
+ if (not) goto fail;
+#else
+ /* Cast to `unsigned' instead of `unsigned char' in case the
+ bit list is a full 32 bytes long. */
+ if (c < (unsigned) (*p * BYTEWIDTH)
+ && p[1 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH)))
+ not = !not;
+
+ p += 1 + *p;
+
+ if (!not) goto fail;
+#undef WORK_BUFFER_SIZE
+#endif /* WCHAR */
+ SET_REGS_MATCHED ();
+ d++;
+ NEXT;
+ }
+
+
+ /* The beginning of a group is represented by start_memory.
+ The arguments are the register number in the next byte, and the
+ number of groups inner to this one in the next. The text
+ matched within the group is recorded (in the internal
+ registers data structure) under the register number. */
+ CASE (start_memory):
+ DEBUG_PRINT3 ("EXECUTING start_memory %ld (%ld):\n",
+ (long int) *p, (long int) p[1]);
+
+ /* Find out if this group can match the empty string. */
+ p1 = p; /* To send to group_match_null_string_p. */
+
+ if (REG_MATCH_NULL_STRING_P (reg_info[*p]) == MATCH_NULL_UNSET_VALUE)
+ REG_MATCH_NULL_STRING_P (reg_info[*p])
+ = PREFIX(group_match_null_string_p) (&p1, pend, reg_info);
+
+ /* Save the position in the string where we were the last time
+ we were at this open-group operator in case the group is
+ operated upon by a repetition operator, e.g., with `(a*)*b'
+ against `ab'; then we want to ignore where we are now in
+ the string in case this attempt to match fails. */
+ old_regstart[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p])
+ ? REG_UNSET (regstart[*p]) ? d : regstart[*p]
+ : regstart[*p];
+ DEBUG_PRINT2 (" old_regstart: %d\n",
+ POINTER_TO_OFFSET (old_regstart[*p]));
+
+ regstart[*p] = d;
+ DEBUG_PRINT2 (" regstart: %d\n", POINTER_TO_OFFSET (regstart[*p]));
+
+ IS_ACTIVE (reg_info[*p]) = 1;
+ MATCHED_SOMETHING (reg_info[*p]) = 0;
+
+ /* Clear this whenever we change the register activity status. */
+ set_regs_matched_done = 0;
+
+ /* This is the new highest active register. */
+ highest_active_reg = *p;
+
+ /* If nothing was active before, this is the new lowest active
+ register. */
+ if (lowest_active_reg == NO_LOWEST_ACTIVE_REG)
+ lowest_active_reg = *p;
+
+ /* Move past the register number and inner group count. */
+ p += 2;
+ just_past_start_mem = p;
+
+ NEXT;
+
+
+ /* The stop_memory opcode represents the end of a group. Its
+ arguments are the same as start_memory's: the register
+ number, and the number of inner groups. */
+ CASE (stop_memory):
+ DEBUG_PRINT3 ("EXECUTING stop_memory %ld (%ld):\n",
+ (long int) *p, (long int) p[1]);
+
+ /* We need to save the string position the last time we were at
+ this close-group operator in case the group is operated
+ upon by a repetition operator, e.g., with `((a*)*(b*)*)*'
+ against `aba'; then we want to ignore where we are now in
+ the string in case this attempt to match fails. */
+ old_regend[*p] = REG_MATCH_NULL_STRING_P (reg_info[*p])
+ ? REG_UNSET (regend[*p]) ? d : regend[*p]
+ : regend[*p];
+ DEBUG_PRINT2 (" old_regend: %d\n",
+ POINTER_TO_OFFSET (old_regend[*p]));
+
+ regend[*p] = d;
+ DEBUG_PRINT2 (" regend: %d\n", POINTER_TO_OFFSET (regend[*p]));
+
+ /* This register isn't active anymore. */
+ IS_ACTIVE (reg_info[*p]) = 0;
+
+ /* Clear this whenever we change the register activity status. */
+ set_regs_matched_done = 0;
+
+ /* If this was the only register active, nothing is active
+ anymore. */
+ if (lowest_active_reg == highest_active_reg)
+ {
+ lowest_active_reg = NO_LOWEST_ACTIVE_REG;
+ highest_active_reg = NO_HIGHEST_ACTIVE_REG;
+ }
+ else
+ { /* We must scan for the new highest active register, since
+ it isn't necessarily one less than now: consider
+ (a(b)c(d(e)f)g). When group 3 ends, after the f), the
+ new highest active register is 1. */
+ UCHAR_T r = *p - 1;
+ while (r > 0 && !IS_ACTIVE (reg_info[r]))
+ r--;
+
+ /* If we end up at register zero, that means that we saved
+ the registers as the result of an `on_failure_jump', not
+ a `start_memory', and we jumped to past the innermost
+ `stop_memory'. For example, in ((.)*) we save
+ registers 1 and 2 as a result of the *, but when we pop
+ back to the second ), we are at the stop_memory 1.
+ Thus, nothing is active. */
+ if (r == 0)
+ {
+ lowest_active_reg = NO_LOWEST_ACTIVE_REG;
+ highest_active_reg = NO_HIGHEST_ACTIVE_REG;
+ }
+ else
+ highest_active_reg = r;
+ }
+
+ /* If just failed to match something this time around with a
+ group that's operated on by a repetition operator, try to
+ force exit from the ``loop'', and restore the register
+ information for this group that we had before trying this
+ last match. */
+ if ((!MATCHED_SOMETHING (reg_info[*p])
+ || just_past_start_mem == p - 1)
+ && (p + 2) < pend)
+ {
+ boolean is_a_jump_n = false;
+
+ p1 = p + 2;
+ mcnt = 0;
+ switch ((re_opcode_t) *p1++)
+ {
+ case jump_n:
+ is_a_jump_n = true;
+ case pop_failure_jump:
+ case maybe_pop_jump:
+ case jump:
+ case dummy_failure_jump:
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ if (is_a_jump_n)
+ p1 += OFFSET_ADDRESS_SIZE;
+ break;
+
+ default:
+ /* do nothing */ ;
+ }
+ p1 += mcnt;
+
+ /* If the next operation is a jump backwards in the pattern
+ to an on_failure_jump right before the start_memory
+ corresponding to this stop_memory, exit from the loop
+ by forcing a failure after pushing on the stack the
+ on_failure_jump's jump in the pattern, and d. */
+ if (mcnt < 0 && (re_opcode_t) *p1 == on_failure_jump
+ && (re_opcode_t) p1[1+OFFSET_ADDRESS_SIZE] == start_memory
+ && p1[2+OFFSET_ADDRESS_SIZE] == *p)
+ {
+ /* If this group ever matched anything, then restore
+ what its registers were before trying this last
+ failed match, e.g., with `(a*)*b' against `ab' for
+ regstart[1], and, e.g., with `((a*)*(b*)*)*'
+ against `aba' for regend[3].
+
+ Also restore the registers for inner groups for,
+ e.g., `((a*)(b*))*' against `aba' (register 3 would
+ otherwise get trashed). */
+
+ if (EVER_MATCHED_SOMETHING (reg_info[*p]))
+ {
+ unsigned r;
+
+ EVER_MATCHED_SOMETHING (reg_info[*p]) = 0;
+
+ /* Restore this and inner groups' (if any) registers. */
+ for (r = *p; r < (unsigned) *p + (unsigned) *(p + 1);
+ r++)
+ {
+ regstart[r] = old_regstart[r];
+
+ /* xx why this test? */
+ if (old_regend[r] >= regstart[r])
+ regend[r] = old_regend[r];
+ }
+ }
+ p1++;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ PUSH_FAILURE_POINT (p1 + mcnt, d, -2);
+
+ goto fail;
+ }
+ }
+
+ /* Move past the register number and the inner group count. */
+ p += 2;
+ NEXT;
+
+
+ /* \<digit> has been turned into a `duplicate' command which is
+ followed by the numeric value of <digit> as the register number. */
+ CASE (duplicate):
+ {
+ register const CHAR_T *d2, *dend2;
+ int regno = *p++; /* Get which register to match against. */
+ DEBUG_PRINT2 ("EXECUTING duplicate %d.\n", regno);
+
+ /* Can't back reference a group which we've never matched. */
+ if (REG_UNSET (regstart[regno]) || REG_UNSET (regend[regno]))
+ goto fail;
+
+ /* Where in input to try to start matching. */
+ d2 = regstart[regno];
+
+ /* Where to stop matching; if both the place to start and
+ the place to stop matching are in the same string, then
+ set to the place to stop, otherwise, for now have to use
+ the end of the first string. */
+
+ dend2 = ((FIRST_STRING_P (regstart[regno])
+ == FIRST_STRING_P (regend[regno]))
+ ? regend[regno] : end_match_1);
+ for (;;)
+ {
+ /* If necessary, advance to next segment in register
+ contents. */
+ while (d2 == dend2)
+ {
+ if (dend2 == end_match_2) break;
+ if (dend2 == regend[regno]) break;
+
+ /* End of string1 => advance to string2. */
+ d2 = string2;
+ dend2 = regend[regno];
+ }
+ /* At end of register contents => success */
+ if (d2 == dend2) break;
+
+ /* If necessary, advance to next segment in data. */
+ PREFETCH ();
+
+ /* How many characters left in this segment to match. */
+ mcnt = dend - d;
+
+ /* Want how many consecutive characters we can match in
+ one shot, so, if necessary, adjust the count. */
+ if (mcnt > dend2 - d2)
+ mcnt = dend2 - d2;
+
+ /* Compare that many; failure if mismatch, else move
+ past them. */
+ if (translate
+ ? PREFIX(bcmp_translate) (d, d2, mcnt, translate)
+ : memcmp (d, d2, mcnt*sizeof(UCHAR_T)))
+ goto fail;
+ d += mcnt, d2 += mcnt;
+
+ /* Do this because we've match some characters. */
+ SET_REGS_MATCHED ();
+ }
+ }
+ NEXT;
+
+
+ /* begline matches the empty string at the beginning of the string
+ (unless `not_bol' is set in `bufp'), and, if
+ `newline_anchor' is set, after newlines. */
+ CASE (begline):
+ DEBUG_PRINT1 ("EXECUTING begline.\n");
+
+ if (AT_STRINGS_BEG (d))
+ {
+ if (!bufp->not_bol)
+ {
+ NEXT;
+ }
+ }
+ else if (d[-1] == '\n' && bufp->newline_anchor)
+ {
+ NEXT;
+ }
+ /* In all other cases, we fail. */
+ goto fail;
+
+
+ /* endline is the dual of begline. */
+ CASE (endline):
+ DEBUG_PRINT1 ("EXECUTING endline.\n");
+
+ if (AT_STRINGS_END (d))
+ {
+ if (!bufp->not_eol)
+ {
+ NEXT;
+ }
+ }
+
+ /* We have to ``prefetch'' the next character. */
+ else if ((d == end1 ? *string2 : *d) == '\n'
+ && bufp->newline_anchor)
+ {
+ NEXT;
+ }
+ goto fail;
+
+
+ /* Match at the very beginning of the data. */
+ CASE (begbuf):
+ DEBUG_PRINT1 ("EXECUTING begbuf.\n");
+ if (AT_STRINGS_BEG (d))
+ {
+ NEXT;
+ }
+ goto fail;
+
+
+ /* Match at the very end of the data. */
+ CASE (endbuf):
+ DEBUG_PRINT1 ("EXECUTING endbuf.\n");
+ if (AT_STRINGS_END (d))
+ {
+ NEXT;
+ }
+ goto fail;
+
+
+ /* on_failure_keep_string_jump is used to optimize `.*\n'. It
+ pushes NULL as the value for the string on the stack. Then
+ `pop_failure_point' will keep the current value for the
+ string, instead of restoring it. To see why, consider
+ matching `foo\nbar' against `.*\n'. The .* matches the foo;
+ then the . fails against the \n. But the next thing we want
+ to do is match the \n against the \n; if we restored the
+ string value, we would be back at the foo.
+
+ Because this is used only in specific cases, we don't need to
+ check all the things that `on_failure_jump' does, to make
+ sure the right things get saved on the stack. Hence we don't
+ share its code. The only reason to push anything on the
+ stack at all is that otherwise we would have to change
+ `anychar's code to do something besides goto fail in this
+ case; that seems worse than this. */
+ CASE (on_failure_keep_string_jump):
+ DEBUG_PRINT1 ("EXECUTING on_failure_keep_string_jump");
+
+ EXTRACT_NUMBER_AND_INCR (mcnt, p);
+#ifdef _LIBC
+ DEBUG_PRINT3 (" %d (to %p):\n", mcnt, p + mcnt);
+#else
+ DEBUG_PRINT3 (" %d (to 0x%x):\n", mcnt, p + mcnt);
+#endif
+
+ PUSH_FAILURE_POINT (p + mcnt, NULL, -2);
+ NEXT;
+
+
+ /* Uses of on_failure_jump:
+
+ Each alternative starts with an on_failure_jump that points
+ to the beginning of the next alternative. Each alternative
+ except the last ends with a jump that in effect jumps past
+ the rest of the alternatives. (They really jump to the
+ ending jump of the following alternative, because tensioning
+ these jumps is a hassle.)
+
+ Repeats start with an on_failure_jump that points past both
+ the repetition text and either the following jump or
+ pop_failure_jump back to this on_failure_jump. */
+ CASE (on_failure_jump):
+ on_failure:
+ DEBUG_PRINT1 ("EXECUTING on_failure_jump");
+
+ EXTRACT_NUMBER_AND_INCR (mcnt, p);
+#ifdef _LIBC
+ DEBUG_PRINT3 (" %d (to %p)", mcnt, p + mcnt);
+#else
+ DEBUG_PRINT3 (" %d (to 0x%x)", mcnt, p + mcnt);
+#endif
+
+ /* If this on_failure_jump comes right before a group (i.e.,
+ the original * applied to a group), save the information
+ for that group and all inner ones, so that if we fail back
+ to this point, the group's information will be correct.
+ For example, in \(a*\)*\1, we need the preceding group,
+ and in \(zz\(a*\)b*\)\2, we need the inner group. */
+
+ /* We can't use `p' to check ahead because we push
+ a failure point to `p + mcnt' after we do this. */
+ p1 = p;
+
+ /* We need to skip no_op's before we look for the
+ start_memory in case this on_failure_jump is happening as
+ the result of a completed succeed_n, as in \(a\)\{1,3\}b\1
+ against aba. */
+ while (p1 < pend && (re_opcode_t) *p1 == no_op)
+ p1++;
+
+ if (p1 < pend && (re_opcode_t) *p1 == start_memory)
+ {
+ /* We have a new highest active register now. This will
+ get reset at the start_memory we are about to get to,
+ but we will have saved all the registers relevant to
+ this repetition op, as described above. */
+ highest_active_reg = *(p1 + 1) + *(p1 + 2);
+ if (lowest_active_reg == NO_LOWEST_ACTIVE_REG)
+ lowest_active_reg = *(p1 + 1);
+ }
+
+ DEBUG_PRINT1 (":\n");
+ PUSH_FAILURE_POINT (p + mcnt, d, -2);
+ NEXT;
+
+
+ /* A smart repeat ends with `maybe_pop_jump'.
+ We change it to either `pop_failure_jump' or `jump'. */
+ CASE (maybe_pop_jump):
+ EXTRACT_NUMBER_AND_INCR (mcnt, p);
+ DEBUG_PRINT2 ("EXECUTING maybe_pop_jump %d.\n", mcnt);
+ {
+ register UCHAR_T *p2 = p;
+
+ /* Compare the beginning of the repeat with what in the
+ pattern follows its end. If we can establish that there
+ is nothing that they would both match, i.e., that we
+ would have to backtrack because of (as in, e.g., `a*a')
+ then we can change to pop_failure_jump, because we'll
+ never have to backtrack.
+
+ This is not true in the case of alternatives: in
+ `(a|ab)*' we do need to backtrack to the `ab' alternative
+ (e.g., if the string was `ab'). But instead of trying to
+ detect that here, the alternative has put on a dummy
+ failure point which is what we will end up popping. */
+
+ /* Skip over open/close-group commands.
+ If what follows this loop is a ...+ construct,
+ look at what begins its body, since we will have to
+ match at least one of that. */
+ while (1)
+ {
+ if (p2 + 2 < pend
+ && ((re_opcode_t) *p2 == stop_memory
+ || (re_opcode_t) *p2 == start_memory))
+ p2 += 3;
+ else if (p2 + 2 + 2 * OFFSET_ADDRESS_SIZE < pend
+ && (re_opcode_t) *p2 == dummy_failure_jump)
+ p2 += 2 + 2 * OFFSET_ADDRESS_SIZE;
+ else
+ break;
+ }
+
+ p1 = p + mcnt;
+ /* p1[0] ... p1[2] are the `on_failure_jump' corresponding
+ to the `maybe_finalize_jump' of this case. Examine what
+ follows. */
+
+ /* If we're at the end of the pattern, we can change. */
+ if (p2 == pend)
+ {
+ /* Consider what happens when matching ":\(.*\)"
+ against ":/". I don't really understand this code
+ yet. */
+ p[-(1+OFFSET_ADDRESS_SIZE)] = (UCHAR_T)
+ pop_failure_jump;
+ DEBUG_PRINT1
+ (" End of pattern: change to `pop_failure_jump'.\n");
+ }
+
+ else if ((re_opcode_t) *p2 == exactn
+#ifdef MBS_SUPPORT
+ || (re_opcode_t) *p2 == exactn_bin
+#endif
+ || (bufp->newline_anchor && (re_opcode_t) *p2 == endline))
+ {
+ register UCHAR_T c
+ = *p2 == (UCHAR_T) endline ? '\n' : p2[2];
+
+ if (((re_opcode_t) p1[1+OFFSET_ADDRESS_SIZE] == exactn
+#ifdef MBS_SUPPORT
+ || (re_opcode_t) p1[1+OFFSET_ADDRESS_SIZE] == exactn_bin
+#endif
+ ) && p1[3+OFFSET_ADDRESS_SIZE] != c)
+ {
+ p[-(1+OFFSET_ADDRESS_SIZE)] = (UCHAR_T)
+ pop_failure_jump;
+#ifdef WCHAR
+ DEBUG_PRINT3 (" %C != %C => pop_failure_jump.\n",
+ (wint_t) c,
+ (wint_t) p1[3+OFFSET_ADDRESS_SIZE]);
+#else
+ DEBUG_PRINT3 (" %c != %c => pop_failure_jump.\n",
+ (char) c,
+ (char) p1[3+OFFSET_ADDRESS_SIZE]);
+#endif
+ }
+
+#ifndef WCHAR
+ else if ((re_opcode_t) p1[3] == charset
+ || (re_opcode_t) p1[3] == charset_not)
+ {
+ int not = (re_opcode_t) p1[3] == charset_not;
+
+ if (c < (unsigned) (p1[4] * BYTEWIDTH)
+ && p1[5 + c / BYTEWIDTH] & (1 << (c % BYTEWIDTH)))
+ not = !not;
+
+ /* `not' is equal to 1 if c would match, which means
+ that we can't change to pop_failure_jump. */
+ if (!not)
+ {
+ p[-3] = (unsigned char) pop_failure_jump;
+ DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
+ }
+ }
+#endif /* not WCHAR */
+ }
+#ifndef WCHAR
+ else if ((re_opcode_t) *p2 == charset)
+ {
+ /* We win if the first character of the loop is not part
+ of the charset. */
+ if ((re_opcode_t) p1[3] == exactn
+ && ! ((int) p2[1] * BYTEWIDTH > (int) p1[5]
+ && (p2[2 + p1[5] / BYTEWIDTH]
+ & (1 << (p1[5] % BYTEWIDTH)))))
+ {
+ p[-3] = (unsigned char) pop_failure_jump;
+ DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
+ }
+
+ else if ((re_opcode_t) p1[3] == charset_not)
+ {
+ int idx;
+ /* We win if the charset_not inside the loop
+ lists every character listed in the charset after. */
+ for (idx = 0; idx < (int) p2[1]; idx++)
+ if (! (p2[2 + idx] == 0
+ || (idx < (int) p1[4]
+ && ((p2[2 + idx] & ~ p1[5 + idx]) == 0))))
+ break;
+
+ if (idx == p2[1])
+ {
+ p[-3] = (unsigned char) pop_failure_jump;
+ DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
+ }
+ }
+ else if ((re_opcode_t) p1[3] == charset)
+ {
+ int idx;
+ /* We win if the charset inside the loop
+ has no overlap with the one after the loop. */
+ for (idx = 0;
+ idx < (int) p2[1] && idx < (int) p1[4];
+ idx++)
+ if ((p2[2 + idx] & p1[5 + idx]) != 0)
+ break;
+
+ if (idx == p2[1] || idx == p1[4])
+ {
+ p[-3] = (unsigned char) pop_failure_jump;
+ DEBUG_PRINT1 (" No match => pop_failure_jump.\n");
+ }
+ }
+ }
+#endif /* not WCHAR */
+ }
+ p -= OFFSET_ADDRESS_SIZE; /* Point at relative address again. */
+ if ((re_opcode_t) p[-1] != pop_failure_jump)
+ {
+ p[-1] = (UCHAR_T) jump;
+ DEBUG_PRINT1 (" Match => jump.\n");
+ goto unconditional_jump;
+ }
+ /* Note fall through. */
+
+
+ /* The end of a simple repeat has a pop_failure_jump back to
+ its matching on_failure_jump, where the latter will push a
+ failure point. The pop_failure_jump takes off failure
+ points put on by this pop_failure_jump's matching
+ on_failure_jump; we got through the pattern to here from the
+ matching on_failure_jump, so didn't fail. */
+ CASE (pop_failure_jump):
+ {
+ /* We need to pass separate storage for the lowest and
+ highest registers, even though we don't care about the
+ actual values. Otherwise, we will restore only one
+ register from the stack, since lowest will == highest in
+ `pop_failure_point'. */
+ active_reg_t dummy_low_reg, dummy_high_reg;
+ UCHAR_T *pdummy = NULL;
+ const CHAR_T *sdummy = NULL;
+
+ DEBUG_PRINT1 ("EXECUTING pop_failure_jump.\n");
+ POP_FAILURE_POINT (sdummy, pdummy,
+ dummy_low_reg, dummy_high_reg,
+ reg_dummy, reg_dummy, reg_info_dummy);
+ }
+ /* Note fall through. */
+
+ unconditional_jump:
+#ifdef _LIBC
+ DEBUG_PRINT2 ("\n%p: ", p);
+#else
+ DEBUG_PRINT2 ("\n0x%x: ", p);
+#endif
+ /* Note fall through. */
+
+ /* Unconditionally jump (without popping any failure points). */
+ CASE (jump):
+ EXTRACT_NUMBER_AND_INCR (mcnt, p); /* Get the amount to jump. */
+ DEBUG_PRINT2 ("EXECUTING jump %d ", mcnt);
+ p += mcnt; /* Do the jump. */
+#ifdef _LIBC
+ DEBUG_PRINT2 ("(to %p).\n", p);
+#else
+ DEBUG_PRINT2 ("(to 0x%x).\n", p);
+#endif
+ NEXT;
+
+
+ /* We need this opcode so we can detect where alternatives end
+ in `group_match_null_string_p' et al. */
+ CASE (jump_past_alt):
+ DEBUG_PRINT1 ("EXECUTING jump_past_alt.\n");
+ goto unconditional_jump;
+
+
+ /* Normally, the on_failure_jump pushes a failure point, which
+ then gets popped at pop_failure_jump. We will end up at
+ pop_failure_jump, also, and with a pattern of, say, `a+', we
+ are skipping over the on_failure_jump, so we have to push
+ something meaningless for pop_failure_jump to pop. */
+ CASE (dummy_failure_jump):
+ DEBUG_PRINT1 ("EXECUTING dummy_failure_jump.\n");
+ /* It doesn't matter what we push for the string here. What
+ the code at `fail' tests is the value for the pattern. */
+ PUSH_FAILURE_POINT (NULL, NULL, -2);
+ goto unconditional_jump;
+
+
+ /* At the end of an alternative, we need to push a dummy failure
+ point in case we are followed by a `pop_failure_jump', because
+ we don't want the failure point for the alternative to be
+ popped. For example, matching `(a|ab)*' against `aab'
+ requires that we match the `ab' alternative. */
+ CASE (push_dummy_failure):
+ DEBUG_PRINT1 ("EXECUTING push_dummy_failure.\n");
+ /* See comments just above at `dummy_failure_jump' about the
+ two zeroes. */
+ PUSH_FAILURE_POINT (NULL, NULL, -2);
+ NEXT;
+
+ /* Have to succeed matching what follows at least n times.
+ After that, handle like `on_failure_jump'. */
+ CASE (succeed_n):
+ EXTRACT_NUMBER (mcnt, p + OFFSET_ADDRESS_SIZE);
+ DEBUG_PRINT2 ("EXECUTING succeed_n %d.\n", mcnt);
+
+ assert (mcnt >= 0);
+ /* Originally, this is how many times we HAVE to succeed. */
+ if (mcnt > 0)
+ {
+ mcnt--;
+ p += OFFSET_ADDRESS_SIZE;
+ STORE_NUMBER_AND_INCR (p, mcnt);
+#ifdef _LIBC
+ DEBUG_PRINT3 (" Setting %p to %d.\n", p - OFFSET_ADDRESS_SIZE
+ , mcnt);
+#else
+ DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p - OFFSET_ADDRESS_SIZE
+ , mcnt);
+#endif
+ }
+ else if (mcnt == 0)
+ {
+#ifdef _LIBC
+ DEBUG_PRINT2 (" Setting two bytes from %p to no_op.\n",
+ p + OFFSET_ADDRESS_SIZE);
+#else
+ DEBUG_PRINT2 (" Setting two bytes from 0x%x to no_op.\n",
+ p + OFFSET_ADDRESS_SIZE);
+#endif /* _LIBC */
+
+#ifdef WCHAR
+ p[1] = (UCHAR_T) no_op;
+#else
+ p[2] = (UCHAR_T) no_op;
+ p[3] = (UCHAR_T) no_op;
+#endif /* WCHAR */
+ goto on_failure;
+ }
+ NEXT;
+
+ CASE (jump_n):
+ EXTRACT_NUMBER (mcnt, p + OFFSET_ADDRESS_SIZE);
+ DEBUG_PRINT2 ("EXECUTING jump_n %d.\n", mcnt);
+
+ /* Originally, this is how many times we CAN jump. */
+ if (mcnt)
+ {
+ mcnt--;
+ STORE_NUMBER (p + OFFSET_ADDRESS_SIZE, mcnt);
+
+#ifdef _LIBC
+ DEBUG_PRINT3 (" Setting %p to %d.\n", p + OFFSET_ADDRESS_SIZE,
+ mcnt);
+#else
+ DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p + OFFSET_ADDRESS_SIZE,
+ mcnt);
+#endif /* _LIBC */
+ goto unconditional_jump;
+ }
+ /* If don't have to jump any more, skip over the rest of command. */
+ else
+ p += 2 * OFFSET_ADDRESS_SIZE;
+ NEXT;
+
+ CASE (set_number_at):
+ {
+ DEBUG_PRINT1 ("EXECUTING set_number_at.\n");
+
+ EXTRACT_NUMBER_AND_INCR (mcnt, p);
+ p1 = p + mcnt;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p);
+#ifdef _LIBC
+ DEBUG_PRINT3 (" Setting %p to %d.\n", p1, mcnt);
+#else
+ DEBUG_PRINT3 (" Setting 0x%x to %d.\n", p1, mcnt);
+#endif
+ STORE_NUMBER (p1, mcnt);
+ NEXT;
+ }
+
+#if 0
+ /* The DEC Alpha C compiler 3.x generates incorrect code for the
+ test WORDCHAR_P (d - 1) != WORDCHAR_P (d) in the expansion of
+ AT_WORD_BOUNDARY, so this code is disabled. Expanding the
+ macro and introducing temporary variables works around the bug. */
+
+ CASE (wordbound):
+ DEBUG_PRINT1 ("EXECUTING wordbound.\n");
+ if (AT_WORD_BOUNDARY (d))
+ {
+ NEXT;
+ }
+ goto fail;
+
+ CASE (notwordbound):
+ DEBUG_PRINT1 ("EXECUTING notwordbound.\n");
+ if (AT_WORD_BOUNDARY (d))
+ goto fail;
+ NEXT;
+#else
+ CASE (wordbound):
+ {
+ boolean prevchar, thischar;
+
+ DEBUG_PRINT1 ("EXECUTING wordbound.\n");
+ if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d))
+ {
+ NEXT;
+ }
+
+ prevchar = WORDCHAR_P (d - 1);
+ thischar = WORDCHAR_P (d);
+ if (prevchar != thischar)
+ {
+ NEXT;
+ }
+ goto fail;
+ }
+
+ CASE (notwordbound):
+ {
+ boolean prevchar, thischar;
+
+ DEBUG_PRINT1 ("EXECUTING notwordbound.\n");
+ if (AT_STRINGS_BEG (d) || AT_STRINGS_END (d))
+ goto fail;
+
+ prevchar = WORDCHAR_P (d - 1);
+ thischar = WORDCHAR_P (d);
+ if (prevchar != thischar)
+ goto fail;
+ NEXT;
+ }
+#endif
+
+ CASE (wordbeg):
+ DEBUG_PRINT1 ("EXECUTING wordbeg.\n");
+ if (!AT_STRINGS_END (d) && WORDCHAR_P (d)
+ && (AT_STRINGS_BEG (d) || !WORDCHAR_P (d - 1)))
+ {
+ NEXT;
+ }
+ goto fail;
+
+ CASE (wordend):
+ DEBUG_PRINT1 ("EXECUTING wordend.\n");
+ if (!AT_STRINGS_BEG (d) && WORDCHAR_P (d - 1)
+ && (AT_STRINGS_END (d) || !WORDCHAR_P (d)))
+ {
+ NEXT;
+ }
+ goto fail;
+
+#ifdef emacs
+ CASE (before_dot):
+ DEBUG_PRINT1 ("EXECUTING before_dot.\n");
+ if (PTR_CHAR_POS ((unsigned char *) d) >= point)
+ goto fail;
+ NEXT;
+
+ CASE (at_dot):
+ DEBUG_PRINT1 ("EXECUTING at_dot.\n");
+ if (PTR_CHAR_POS ((unsigned char *) d) != point)
+ goto fail;
+ NEXT;
+
+ CASE (after_dot):
+ DEBUG_PRINT1 ("EXECUTING after_dot.\n");
+ if (PTR_CHAR_POS ((unsigned char *) d) <= point)
+ goto fail;
+ NEXT;
+
+ CASE (syntaxspec):
+ DEBUG_PRINT2 ("EXECUTING syntaxspec %d.\n", mcnt);
+ mcnt = *p++;
+ goto matchsyntax;
+
+ CASE (wordchar):
+ DEBUG_PRINT1 ("EXECUTING Emacs wordchar.\n");
+ mcnt = (int) Sword;
+ matchsyntax:
+ PREFETCH ();
+ /* Can't use *d++ here; SYNTAX may be an unsafe macro. */
+ d++;
+ if (SYNTAX (d[-1]) != (enum syntaxcode) mcnt)
+ goto fail;
+ SET_REGS_MATCHED ();
+ NEXT;
+
+ CASE (notsyntaxspec):
+ DEBUG_PRINT2 ("EXECUTING notsyntaxspec %d.\n", mcnt);
+ mcnt = *p++;
+ goto matchnotsyntax;
+
+ CASE (notwordchar):
+ DEBUG_PRINT1 ("EXECUTING Emacs notwordchar.\n");
+ mcnt = (int) Sword;
+ matchnotsyntax:
+ PREFETCH ();
+ /* Can't use *d++ here; SYNTAX may be an unsafe macro. */
+ d++;
+ if (SYNTAX (d[-1]) == (enum syntaxcode) mcnt)
+ goto fail;
+ SET_REGS_MATCHED ();
+ NEXT;
+
+#else /* not emacs */
+ CASE (wordchar):
+ DEBUG_PRINT1 ("EXECUTING non-Emacs wordchar.\n");
+ PREFETCH ();
+ if (!WORDCHAR_P (d))
+ goto fail;
+ SET_REGS_MATCHED ();
+ d++;
+ NEXT;
+
+ CASE (notwordchar):
+ DEBUG_PRINT1 ("EXECUTING non-Emacs notwordchar.\n");
+ PREFETCH ();
+ if (WORDCHAR_P (d))
+ goto fail;
+ SET_REGS_MATCHED ();
+ d++;
+ NEXT;
+#endif /* not emacs */
+
+#ifndef __GNUC__
+ default:
+ abort ();
+ }
+ continue; /* Successfully executed one pattern command; keep going. */
+#endif
+
+
+ /* We goto here if a matching operation fails. */
+ fail:
+ if (!FAIL_STACK_EMPTY ())
+ { /* A restart point is known. Restore to that state. */
+ DEBUG_PRINT1 ("\nFAIL:\n");
+ POP_FAILURE_POINT (d, p,
+ lowest_active_reg, highest_active_reg,
+ regstart, regend, reg_info);
+
+ /* If this failure point is a dummy, try the next one. */
+ if (!p)
+ goto fail;
+
+ /* If we failed to the end of the pattern, don't examine *p. */
+ assert (p <= pend);
+ if (p < pend)
+ {
+ boolean is_a_jump_n = false;
+
+ /* If failed to a backwards jump that's part of a repetition
+ loop, need to pop this failure point and use the next one. */
+ switch ((re_opcode_t) *p)
+ {
+ case jump_n:
+ is_a_jump_n = true;
+ case maybe_pop_jump:
+ case pop_failure_jump:
+ case jump:
+ p1 = p + 1;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ p1 += mcnt;
+
+ if ((is_a_jump_n && (re_opcode_t) *p1 == succeed_n)
+ || (!is_a_jump_n
+ && (re_opcode_t) *p1 == on_failure_jump))
+ goto fail;
+ break;
+ default:
+ /* do nothing */ ;
+ }
+ }
+
+ if (d >= string1 && d <= end1)
+ dend = end_match_1;
+ }
+ else
+ break; /* Matching at this starting point really fails. */
+ } /* for (;;) */
+
+ if (best_regs_set)
+ goto restore_best_regs;
+
+ FREE_VARIABLES ();
+
+ return -1; /* Failure to match. */
+} /* re_match_2 */
+
+/* Subroutine definitions for re_match_2. */
+
+
+/* We are passed P pointing to a register number after a start_memory.
+
+ Return true if the pattern up to the corresponding stop_memory can
+ match the empty string, and false otherwise.
+
+ If we find the matching stop_memory, sets P to point to one past its number.
+ Otherwise, sets P to an undefined byte less than or equal to END.
+
+ We don't handle duplicates properly (yet). */
+
+static boolean
+PREFIX(group_match_null_string_p) (UCHAR_T **p, UCHAR_T *end,
+ PREFIX(register_info_type) *reg_info)
+{
+ int mcnt;
+ /* Point to after the args to the start_memory. */
+ UCHAR_T *p1 = *p + 2;
+
+ while (p1 < end)
+ {
+ /* Skip over opcodes that can match nothing, and return true or
+ false, as appropriate, when we get to one that can't, or to the
+ matching stop_memory. */
+
+ switch ((re_opcode_t) *p1)
+ {
+ /* Could be either a loop or a series of alternatives. */
+ case on_failure_jump:
+ p1++;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+
+ /* If the next operation is not a jump backwards in the
+ pattern. */
+
+ if (mcnt >= 0)
+ {
+ /* Go through the on_failure_jumps of the alternatives,
+ seeing if any of the alternatives cannot match nothing.
+ The last alternative starts with only a jump,
+ whereas the rest start with on_failure_jump and end
+ with a jump, e.g., here is the pattern for `a|b|c':
+
+ /on_failure_jump/0/6/exactn/1/a/jump_past_alt/0/6
+ /on_failure_jump/0/6/exactn/1/b/jump_past_alt/0/3
+ /exactn/1/c
+
+ So, we have to first go through the first (n-1)
+ alternatives and then deal with the last one separately. */
+
+
+ /* Deal with the first (n-1) alternatives, which start
+ with an on_failure_jump (see above) that jumps to right
+ past a jump_past_alt. */
+
+ while ((re_opcode_t) p1[mcnt-(1+OFFSET_ADDRESS_SIZE)] ==
+ jump_past_alt)
+ {
+ /* `mcnt' holds how many bytes long the alternative
+ is, including the ending `jump_past_alt' and
+ its number. */
+
+ if (!PREFIX(alt_match_null_string_p) (p1, p1 + mcnt -
+ (1 + OFFSET_ADDRESS_SIZE),
+ reg_info))
+ return false;
+
+ /* Move to right after this alternative, including the
+ jump_past_alt. */
+ p1 += mcnt;
+
+ /* Break if it's the beginning of an n-th alternative
+ that doesn't begin with an on_failure_jump. */
+ if ((re_opcode_t) *p1 != on_failure_jump)
+ break;
+
+ /* Still have to check that it's not an n-th
+ alternative that starts with an on_failure_jump. */
+ p1++;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ if ((re_opcode_t) p1[mcnt-(1+OFFSET_ADDRESS_SIZE)] !=
+ jump_past_alt)
+ {
+ /* Get to the beginning of the n-th alternative. */
+ p1 -= 1 + OFFSET_ADDRESS_SIZE;
+ break;
+ }
+ }
+
+ /* Deal with the last alternative: go back and get number
+ of the `jump_past_alt' just before it. `mcnt' contains
+ the length of the alternative. */
+ EXTRACT_NUMBER (mcnt, p1 - OFFSET_ADDRESS_SIZE);
+
+ if (!PREFIX(alt_match_null_string_p) (p1, p1 + mcnt, reg_info))
+ return false;
+
+ p1 += mcnt; /* Get past the n-th alternative. */
+ } /* if mcnt > 0 */
+ break;
+
+
+ case stop_memory:
+ assert (p1[1] == **p);
+ *p = p1 + 2;
+ return true;
+
+
+ default:
+ if (!PREFIX(common_op_match_null_string_p) (&p1, end, reg_info))
+ return false;
+ }
+ } /* while p1 < end */
+
+ return false;
+} /* group_match_null_string_p */
+
+
+/* Similar to group_match_null_string_p, but doesn't deal with alternatives:
+ It expects P to be the first byte of a single alternative and END one
+ byte past the last. The alternative can contain groups. */
+
+static boolean
+PREFIX(alt_match_null_string_p) (UCHAR_T *p, UCHAR_T *end,
+ PREFIX(register_info_type) *reg_info)
+{
+ int mcnt;
+ UCHAR_T *p1 = p;
+
+ while (p1 < end)
+ {
+ /* Skip over opcodes that can match nothing, and break when we get
+ to one that can't. */
+
+ switch ((re_opcode_t) *p1)
+ {
+ /* It's a loop. */
+ case on_failure_jump:
+ p1++;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ p1 += mcnt;
+ break;
+
+ default:
+ if (!PREFIX(common_op_match_null_string_p) (&p1, end, reg_info))
+ return false;
+ }
+ } /* while p1 < end */
+
+ return true;
+} /* alt_match_null_string_p */
+
+
+/* Deals with the ops common to group_match_null_string_p and
+ alt_match_null_string_p.
+
+ Sets P to one after the op and its arguments, if any. */
+
+static boolean
+PREFIX(common_op_match_null_string_p) (UCHAR_T **p, UCHAR_T *end,
+ PREFIX(register_info_type) *reg_info)
+{
+ int mcnt;
+ boolean ret;
+ int reg_no;
+ UCHAR_T *p1 = *p;
+
+ switch ((re_opcode_t) *p1++)
+ {
+ case no_op:
+ case begline:
+ case endline:
+ case begbuf:
+ case endbuf:
+ case wordbeg:
+ case wordend:
+ case wordbound:
+ case notwordbound:
+#ifdef emacs
+ case before_dot:
+ case at_dot:
+ case after_dot:
+#endif
+ break;
+
+ case start_memory:
+ reg_no = *p1;
+ assert (reg_no > 0 && reg_no <= MAX_REGNUM);
+ ret = PREFIX(group_match_null_string_p) (&p1, end, reg_info);
+
+ /* Have to set this here in case we're checking a group which
+ contains a group and a back reference to it. */
+
+ if (REG_MATCH_NULL_STRING_P (reg_info[reg_no]) == MATCH_NULL_UNSET_VALUE)
+ REG_MATCH_NULL_STRING_P (reg_info[reg_no]) = ret;
+
+ if (!ret)
+ return false;
+ break;
+
+ /* If this is an optimized succeed_n for zero times, make the jump. */
+ case jump:
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ if (mcnt >= 0)
+ p1 += mcnt;
+ else
+ return false;
+ break;
+
+ case succeed_n:
+ /* Get to the number of times to succeed. */
+ p1 += OFFSET_ADDRESS_SIZE;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+
+ if (mcnt == 0)
+ {
+ p1 -= 2 * OFFSET_ADDRESS_SIZE;
+ EXTRACT_NUMBER_AND_INCR (mcnt, p1);
+ p1 += mcnt;
+ }
+ else
+ return false;
+ break;
+
+ case duplicate:
+ if (!REG_MATCH_NULL_STRING_P (reg_info[*p1]))
+ return false;
+ break;
+
+ case set_number_at:
+ p1 += 2 * OFFSET_ADDRESS_SIZE;
+
+ default:
+ /* All other opcodes mean we cannot match the empty string. */
+ return false;
+ }
+
+ *p = p1;
+ return true;
+} /* common_op_match_null_string_p */
+
+
+/* Return zero if TRANSLATE[S1] and TRANSLATE[S2] are identical for LEN
+ bytes; nonzero otherwise. */
+
+static int
+PREFIX(bcmp_translate) (const CHAR_T *s1, const CHAR_T *s2,
+ register int len,
+ RE_TRANSLATE_TYPE translate)
+{
+ register const UCHAR_T *p1 = (const UCHAR_T *) s1;
+ register const UCHAR_T *p2 = (const UCHAR_T *) s2;
+ while (len)
+ {
+#ifdef WCHAR
+ if (((*p1<=0xff)?translate[*p1++]:*p1++)
+ != ((*p2<=0xff)?translate[*p2++]:*p2++))
+ return 1;
+#else /* BYTE */
+ if (translate[*p1++] != translate[*p2++]) return 1;
+#endif /* WCHAR */
+ len--;
+ }
+ return 0;
+}
+
+
+#else /* not INSIDE_RECURSION */
+
+/* Entry points for GNU code. */
+
+/* re_compile_pattern is the GNU regular expression compiler: it
+ compiles PATTERN (of length SIZE) and puts the result in BUFP.
+ Returns 0 if the pattern was valid, otherwise an error string.
+
+ Assumes the `allocated' (and perhaps `buffer') and `translate' fields
+ are set in BUFP on entry.
+
+ We call regex_compile to do the actual compilation. */
+
+const char *
+re_compile_pattern (const char *pattern,
+ size_t length,
+ struct re_pattern_buffer *bufp)
+{
+ reg_errcode_t ret;
+
+ /* GNU code is written to assume at least RE_NREGS registers will be set
+ (and at least one extra will be -1). */
+ bufp->regs_allocated = REGS_UNALLOCATED;
+
+ /* And GNU code determines whether or not to get register information
+ by passing null for the REGS argument to re_match, etc., not by
+ setting no_sub. */
+ bufp->no_sub = 0;
+
+ /* Match anchors at newline. */
+ bufp->newline_anchor = 1;
+
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ ret = wcs_regex_compile (pattern, length, re_syntax_options, bufp);
+ else
+# endif
+ ret = byte_regex_compile (pattern, length, re_syntax_options, bufp);
+
+ if (!ret)
+ return NULL;
+ return gettext (re_error_msgid + re_error_msgid_idx[(int) ret]);
+}
+#ifdef _LIBC
+weak_alias (__re_compile_pattern, re_compile_pattern)
+#endif
+
+/* Entry points compatible with 4.2 BSD regex library. We don't define
+ them unless specifically requested. */
+
+#if defined _REGEX_RE_COMP || defined _LIBC
+
+/* BSD has one and only one pattern buffer. */
+static struct re_pattern_buffer re_comp_buf;
+
+char *
+#ifdef _LIBC
+/* Make these definitions weak in libc, so POSIX programs can redefine
+ these names if they don't use our functions, and still use
+ regcomp/regexec below without link errors. */
+weak_function
+#endif
+re_comp (const char *s)
+{
+ reg_errcode_t ret;
+
+ if (!s)
+ {
+ if (!re_comp_buf.buffer)
+ return gettext ("No previous regular expression");
+ return 0;
+ }
+
+ if (!re_comp_buf.buffer)
+ {
+ re_comp_buf.buffer = malloc (200);
+ if (re_comp_buf.buffer == NULL)
+ return (char *) gettext (re_error_msgid
+ + re_error_msgid_idx[(int) REG_ESPACE]);
+ re_comp_buf.allocated = 200;
+
+ re_comp_buf.fastmap = malloc (1 << BYTEWIDTH);
+ if (re_comp_buf.fastmap == NULL)
+ return (char *) gettext (re_error_msgid
+ + re_error_msgid_idx[(int) REG_ESPACE]);
+ }
+
+ /* Since `re_exec' always passes NULL for the `regs' argument, we
+ don't need to initialize the pattern buffer fields which affect it. */
+
+ /* Match anchors at newlines. */
+ re_comp_buf.newline_anchor = 1;
+
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ ret = wcs_regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf);
+ else
+# endif
+ ret = byte_regex_compile (s, strlen (s), re_syntax_options, &re_comp_buf);
+
+ if (!ret)
+ return NULL;
+
+ /* Yes, we're discarding `const' here if !HAVE_LIBINTL. */
+ return (char *) gettext (re_error_msgid + re_error_msgid_idx[(int) ret]);
+}
+
+
+int
+#ifdef _LIBC
+weak_function
+#endif
+re_exec (const char *s)
+{
+ const int len = strlen (s);
+ return
+ 0 <= re_search (&re_comp_buf, s, len, 0, len, 0);
+}
+
+#endif /* _REGEX_RE_COMP */
+
+/* POSIX.2 functions. Don't define these for Emacs. */
+
+#ifndef emacs
+
+/* regcomp takes a regular expression as a string and compiles it.
+
+ PREG is a regex_t *. We do not expect any fields to be initialized,
+ since POSIX says we shouldn't. Thus, we set
+
+ `buffer' to the compiled pattern;
+ `used' to the length of the compiled pattern;
+ `syntax' to RE_SYNTAX_POSIX_EXTENDED if the
+ REG_EXTENDED bit in CFLAGS is set; otherwise, to
+ RE_SYNTAX_POSIX_BASIC;
+ `newline_anchor' to REG_NEWLINE being set in CFLAGS;
+ `fastmap' to an allocated space for the fastmap;
+ `fastmap_accurate' to zero;
+ `re_nsub' to the number of subexpressions in PATTERN.
+
+ PATTERN is the address of the pattern string.
+
+ CFLAGS is a series of bits which affect compilation.
+
+ If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we
+ use POSIX basic syntax.
+
+ If REG_NEWLINE is set, then . and [^...] don't match newline.
+ Also, regexec will try a match beginning after every newline.
+
+ If REG_ICASE is set, then we considers upper- and lowercase
+ versions of letters to be equivalent when matching.
+
+ If REG_NOSUB is set, then when PREG is passed to regexec, that
+ routine will report only success or failure, and nothing about the
+ registers.
+
+ It returns 0 if it succeeds, nonzero if it doesn't. (See regex.h for
+ the return codes and their meanings.) */
+
+int
+regcomp (regex_t *preg, const char *pattern, int cflags)
+{
+ reg_errcode_t ret;
+ reg_syntax_t syntax
+ = (cflags & REG_EXTENDED) ?
+ RE_SYNTAX_POSIX_EXTENDED : RE_SYNTAX_POSIX_BASIC;
+
+ /* regex_compile will allocate the space for the compiled pattern. */
+ preg->buffer = 0;
+ preg->allocated = 0;
+ preg->used = 0;
+
+ /* Try to allocate space for the fastmap. */
+ preg->fastmap = malloc (1 << BYTEWIDTH);
+
+ if (cflags & REG_ICASE)
+ {
+ unsigned i;
+
+ preg->translate = malloc (CHAR_SET_SIZE
+ * sizeof (*(RE_TRANSLATE_TYPE)0));
+ if (preg->translate == NULL)
+ return (int) REG_ESPACE;
+
+ /* Map uppercase characters to corresponding lowercase ones. */
+ for (i = 0; i < CHAR_SET_SIZE; i++)
+ preg->translate[i] = ISUPPER (i) ? TOLOWER (i) : i;
+ }
+ else
+ preg->translate = NULL;
+
+ /* If REG_NEWLINE is set, newlines are treated differently. */
+ if (cflags & REG_NEWLINE)
+ { /* REG_NEWLINE implies neither . nor [^...] match newline. */
+ syntax &= ~RE_DOT_NEWLINE;
+ syntax |= RE_HAT_LISTS_NOT_NEWLINE;
+ /* It also changes the matching behavior. */
+ preg->newline_anchor = 1;
+ }
+ else
+ preg->newline_anchor = 0;
+
+ preg->no_sub = !!(cflags & REG_NOSUB);
+
+ /* POSIX says a null character in the pattern terminates it, so we
+ can use strlen here in compiling the pattern. */
+# ifdef MBS_SUPPORT
+ if (MB_CUR_MAX != 1)
+ ret = wcs_regex_compile (pattern, strlen (pattern), syntax, preg);
+ else
+# endif
+ ret = byte_regex_compile (pattern, strlen (pattern), syntax, preg);
+
+ /* POSIX doesn't distinguish between an unmatched open-group and an
+ unmatched close-group: both are REG_EPAREN. */
+ if (ret == REG_ERPAREN) ret = REG_EPAREN;
+
+ if (ret == REG_NOERROR && preg->fastmap)
+ {
+ /* Compute the fastmap now, since regexec cannot modify the pattern
+ buffer. */
+ if (re_compile_fastmap (preg) == -2)
+ {
+ /* Some error occurred while computing the fastmap, just forget
+ about it. */
+ free (preg->fastmap);
+ preg->fastmap = NULL;
+ }
+ }
+
+ return (int) ret;
+}
+#ifdef _LIBC
+weak_alias (__regcomp, regcomp)
+#endif
+
+
+/* regexec searches for a given pattern, specified by PREG, in the
+ string STRING.
+
+ If NMATCH is zero or REG_NOSUB was set in the cflags argument to
+ `regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at
+ least NMATCH elements, and we set them to the offsets of the
+ corresponding matched substrings.
+
+ EFLAGS specifies `execution flags' which affect matching: if
+ REG_NOTBOL is set, then ^ does not match at the beginning of the
+ string; if REG_NOTEOL is set, then $ does not match at the end.
+
+ We return 0 if we find a match and REG_NOMATCH if not. */
+
+int
+regexec (const regex_t *preg, const char *string,
+ size_t nmatch, regmatch_t pmatch[], int eflags)
+{
+ int ret;
+ struct re_registers regs;
+ regex_t private_preg;
+ int len = strlen (string);
+ boolean want_reg_info = !preg->no_sub && nmatch > 0;
+
+ private_preg = *preg;
+
+ private_preg.not_bol = !!(eflags & REG_NOTBOL);
+ private_preg.not_eol = !!(eflags & REG_NOTEOL);
+
+ /* The user has told us exactly how many registers to return
+ information about, via `nmatch'. We have to pass that on to the
+ matching routines. */
+ private_preg.regs_allocated = REGS_FIXED;
+
+ if (want_reg_info)
+ {
+ regs.num_regs = nmatch;
+ regs.start = TALLOC (nmatch * 2, regoff_t);
+ if (regs.start == NULL)
+ return (int) REG_NOMATCH;
+ regs.end = regs.start + nmatch;
+ }
+
+ /* Perform the searching operation. */
+ ret = re_search (&private_preg, string, len,
+ /* start: */ 0, /* range: */ len,
+ want_reg_info ? &regs : 0);
+
+ /* Copy the register information to the POSIX structure. */
+ if (want_reg_info)
+ {
+ if (ret >= 0)
+ {
+ unsigned r;
+
+ for (r = 0; r < nmatch; r++)
+ {
+ pmatch[r].rm_so = regs.start[r];
+ pmatch[r].rm_eo = regs.end[r];
+ }
+ }
+
+ /* If we needed the temporary register info, free the space now. */
+ free (regs.start);
+ }
+
+ /* We want zero return to mean success, unlike `re_search'. */
+ return ret >= 0 ? (int) REG_NOERROR : (int) REG_NOMATCH;
+}
+#ifdef _LIBC
+weak_alias (__regexec, regexec)
+#endif
+
+
+/* Returns a message corresponding to an error code, ERRCODE, returned
+ from either regcomp or regexec. We don't use PREG here. */
+
+size_t
+regerror (int errcode, const regex_t *preg, char *errbuf, size_t errbuf_size)
+{
+ const char *msg;
+ size_t msg_size;
+
+ if (errcode < 0
+ || errcode >= (int) (sizeof (re_error_msgid_idx)
+ / sizeof (re_error_msgid_idx[0])))
+ /* Only error codes returned by the rest of the code should be passed
+ to this routine. If we are given anything else, or if other regex
+ code generates an invalid error code, then the program has a bug.
+ Dump core so we can fix it. */
+ abort ();
+
+ msg = gettext (re_error_msgid + re_error_msgid_idx[errcode]);
+
+ msg_size = strlen (msg) + 1; /* Includes the null. */
+
+ if (errbuf_size != 0)
+ {
+ if (msg_size > errbuf_size)
+ {
+#if defined HAVE_MEMPCPY || defined _LIBC
+ *((char *) __mempcpy (errbuf, msg, errbuf_size - 1)) = '\0';
+#else
+ memcpy (errbuf, msg, errbuf_size - 1);
+ errbuf[errbuf_size - 1] = 0;
+#endif
+ }
+ else
+ memcpy (errbuf, msg, msg_size);
+ }
+
+ return msg_size;
+}
+#ifdef _LIBC
+weak_alias (__regerror, regerror)
+#endif
+
+
+/* Free dynamically allocated space used by PREG. */
+
+void
+regfree (regex_t *preg)
+{
+ if (preg->buffer != NULL)
+ free (preg->buffer);
+ preg->buffer = NULL;
+
+ preg->allocated = 0;
+ preg->used = 0;
+
+ if (preg->fastmap != NULL)
+ free (preg->fastmap);
+ preg->fastmap = NULL;
+ preg->fastmap_accurate = 0;
+
+ if (preg->translate != NULL)
+ free (preg->translate);
+ preg->translate = NULL;
+}
+#ifdef _LIBC
+weak_alias (__regfree, regfree)
+#endif
+
+#endif /* not emacs */
+
+#endif /* not INSIDE_RECURSION */
+
+
+#undef STORE_NUMBER
+#undef STORE_NUMBER_AND_INCR
+#undef EXTRACT_NUMBER
+#undef EXTRACT_NUMBER_AND_INCR
+
+#undef DEBUG_PRINT_COMPILED_PATTERN
+#undef DEBUG_PRINT_DOUBLE_STRING
+
+#undef INIT_FAIL_STACK
+#undef RESET_FAIL_STACK
+#undef DOUBLE_FAIL_STACK
+#undef PUSH_PATTERN_OP
+#undef PUSH_FAILURE_POINTER
+#undef PUSH_FAILURE_INT
+#undef PUSH_FAILURE_ELT
+#undef POP_FAILURE_POINTER
+#undef POP_FAILURE_INT
+#undef POP_FAILURE_ELT
+#undef DEBUG_PUSH
+#undef DEBUG_POP
+#undef PUSH_FAILURE_POINT
+#undef POP_FAILURE_POINT
+
+#undef REG_UNSET_VALUE
+#undef REG_UNSET
+
+#undef PATFETCH
+#undef PATFETCH_RAW
+#undef PATUNFETCH
+#undef TRANSLATE
+
+#undef INIT_BUF_SIZE
+#undef GET_BUFFER_SPACE
+#undef BUF_PUSH
+#undef BUF_PUSH_2
+#undef BUF_PUSH_3
+#undef STORE_JUMP
+#undef STORE_JUMP2
+#undef INSERT_JUMP
+#undef INSERT_JUMP2
+#undef EXTEND_BUFFER
+#undef GET_UNSIGNED_NUMBER
+#undef FREE_STACK_RETURN
+
+# undef POINTER_TO_OFFSET
+# undef MATCHING_IN_FRST_STRING
+# undef PREFETCH
+# undef AT_STRINGS_BEG
+# undef AT_STRINGS_END
+# undef WORDCHAR_P
+# undef FREE_VAR
+# undef FREE_VARIABLES
+# undef NO_HIGHEST_ACTIVE_REG
+# undef NO_LOWEST_ACTIVE_REG
+
+# undef CHAR_T
+# undef UCHAR_T
+# undef COMPILED_BUFFER_VAR
+# undef OFFSET_ADDRESS_SIZE
+# undef CHAR_CLASS_SIZE
+# undef PREFIX
+# undef ARG_PREFIX
+# undef PUT_CHAR
+# undef BYTE
+# undef WCHAR
+
+# define DEFINED_ONCE
diff --git a/contrib/diff/lib/regex.h b/contrib/diff/lib/regex.h
new file mode 100644
index 0000000..b4bbc01
--- /dev/null
+++ b/contrib/diff/lib/regex.h
@@ -0,0 +1,556 @@
+/* Definitions for data structures and routines for the regular
+ expression library.
+ Copyright (C) 1985,1989-93,1995-98,2000,2001,2002,2003
+ Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _REGEX_H
+#define _REGEX_H 1
+
+#include <sys/types.h>
+
+/* Allow the use in C++ code. */
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* POSIX says that <sys/types.h> must be included (by the caller) before
+ <regex.h>. */
+
+#if !defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE && defined VMS
+/* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it
+ should be there. */
+# include <stddef.h>
+#endif
+
+/* The following two types have to be signed and unsigned integer type
+ wide enough to hold a value of a pointer. For most ANSI compilers
+ ptrdiff_t and size_t should be likely OK. Still size of these two
+ types is 2 for Microsoft C. Ugh... */
+typedef long int s_reg_t;
+typedef unsigned long int active_reg_t;
+
+/* The following bits are used to determine the regexp syntax we
+ recognize. The set/not-set meanings are chosen so that Emacs syntax
+ remains the value 0. The bits are given in alphabetical order, and
+ the definitions shifted by one from the previous bit; thus, when we
+ add or remove a bit, only one other definition need change. */
+typedef unsigned long int reg_syntax_t;
+
+/* If this bit is not set, then \ inside a bracket expression is literal.
+ If set, then such a \ quotes the following character. */
+#define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1)
+
+/* If this bit is not set, then + and ? are operators, and \+ and \? are
+ literals.
+ If set, then \+ and \? are operators and + and ? are literals. */
+#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1)
+
+/* If this bit is set, then character classes are supported. They are:
+ [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:],
+ [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:].
+ If not set, then character classes are not supported. */
+#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1)
+
+/* If this bit is set, then ^ and $ are always anchors (outside bracket
+ expressions, of course).
+ If this bit is not set, then it depends:
+ ^ is an anchor if it is at the beginning of a regular
+ expression or after an open-group or an alternation operator;
+ $ is an anchor if it is at the end of a regular expression, or
+ before a close-group or an alternation operator.
+
+ This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because
+ POSIX draft 11.2 says that * etc. in leading positions is undefined.
+ We already implemented a previous draft which made those constructs
+ invalid, though, so we haven't changed the code back. */
+#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1)
+
+/* If this bit is set, then special characters are always special
+ regardless of where they are in the pattern.
+ If this bit is not set, then special characters are special only in
+ some contexts; otherwise they are ordinary. Specifically,
+ * + ? and intervals are only special when not after the beginning,
+ open-group, or alternation operator. */
+#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1)
+
+/* If this bit is set, then *, +, ?, and { cannot be first in an re or
+ immediately after an alternation or begin-group operator. */
+#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1)
+
+/* If this bit is set, then . matches newline.
+ If not set, then it doesn't. */
+#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1)
+
+/* If this bit is set, then . doesn't match NUL.
+ If not set, then it does. */
+#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1)
+
+/* If this bit is set, nonmatching lists [^...] do not match newline.
+ If not set, they do. */
+#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1)
+
+/* If this bit is set, either \{...\} or {...} defines an
+ interval, depending on RE_NO_BK_BRACES.
+ If not set, \{, \}, {, and } are literals. */
+#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1)
+
+/* If this bit is set, +, ? and | aren't recognized as operators.
+ If not set, they are. */
+#define RE_LIMITED_OPS (RE_INTERVALS << 1)
+
+/* If this bit is set, newline is an alternation operator.
+ If not set, newline is literal. */
+#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1)
+
+/* If this bit is set, then `{...}' defines an interval, and \{ and \}
+ are literals.
+ If not set, then `\{...\}' defines an interval. */
+#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1)
+
+/* If this bit is set, (...) defines a group, and \( and \) are literals.
+ If not set, \(...\) defines a group, and ( and ) are literals. */
+#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1)
+
+/* If this bit is set, then \<digit> matches <digit>.
+ If not set, then \<digit> is a back-reference. */
+#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1)
+
+/* If this bit is set, then | is an alternation operator, and \| is literal.
+ If not set, then \| is an alternation operator, and | is literal. */
+#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1)
+
+/* If this bit is set, then an ending range point collating higher
+ than the starting range point, as in [z-a], is invalid.
+ If not set, then when ending range point collates higher than the
+ starting range point, the range is ignored. */
+#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1)
+
+/* If this bit is set, then an unmatched ) is ordinary.
+ If not set, then an unmatched ) is invalid. */
+#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1)
+
+/* If this bit is set, succeed as soon as we match the whole pattern,
+ without further backtracking. */
+#define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1)
+
+/* If this bit is set, do not process the GNU regex operators.
+ If not set, then the GNU regex operators are recognized. */
+#define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1)
+
+/* If this bit is set, turn on internal regex debugging.
+ If not set, and debugging was on, turn it off.
+ This only works if regex.c is compiled -DDEBUG.
+ We define this bit always, so that all that's needed to turn on
+ debugging is to recompile regex.c; the calling code can always have
+ this bit set, and it won't affect anything in the normal case. */
+#define RE_DEBUG (RE_NO_GNU_OPS << 1)
+
+/* If this bit is set, a syntactically invalid interval is treated as
+ a string of ordinary characters. For example, the ERE 'a{1' is
+ treated as 'a\{1'. */
+#define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1)
+
+/* If this bit is set, then ignore case when matching.
+ If not set, then case is significant. */
+#define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1)
+
+/* This global variable defines the particular regexp syntax to use (for
+ some interfaces). When a regexp is compiled, the syntax used is
+ stored in the pattern buffer, so changing this does not affect
+ already-compiled regexps. */
+extern reg_syntax_t re_syntax_options;
+
+/* Define combinations of the above bits for the standard possibilities.
+ (The [[[ comments delimit what gets put into the Texinfo file, so
+ don't delete them!) */
+/* [[[begin syntaxes]]] */
+#define RE_SYNTAX_EMACS 0
+
+#define RE_SYNTAX_AWK \
+ (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \
+ | RE_NO_BK_PARENS | RE_NO_BK_REFS \
+ | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \
+ | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \
+ | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS)
+
+#define RE_SYNTAX_GNU_AWK \
+ ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \
+ & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS \
+ | RE_CONTEXT_INVALID_OPS ))
+
+#define RE_SYNTAX_POSIX_AWK \
+ (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \
+ | RE_INTERVALS | RE_NO_GNU_OPS)
+
+#define RE_SYNTAX_GREP \
+ (RE_BK_PLUS_QM | RE_CHAR_CLASSES \
+ | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \
+ | RE_NEWLINE_ALT)
+
+#define RE_SYNTAX_EGREP \
+ (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \
+ | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \
+ | RE_NEWLINE_ALT | RE_NO_BK_PARENS \
+ | RE_NO_BK_VBAR)
+
+#define RE_SYNTAX_POSIX_EGREP \
+ (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \
+ | RE_INVALID_INTERVAL_ORD)
+
+/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */
+#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC
+
+#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC
+
+/* Syntax bits common to both basic and extended POSIX regex syntax. */
+#define _RE_SYNTAX_POSIX_COMMON \
+ (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \
+ | RE_INTERVALS | RE_NO_EMPTY_RANGES)
+
+#define RE_SYNTAX_POSIX_BASIC \
+ (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM)
+
+/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes
+ RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this
+ isn't minimal, since other operators, such as \`, aren't disabled. */
+#define RE_SYNTAX_POSIX_MINIMAL_BASIC \
+ (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS)
+
+#define RE_SYNTAX_POSIX_EXTENDED \
+ (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
+ | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \
+ | RE_NO_BK_PARENS | RE_NO_BK_VBAR \
+ | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD)
+
+/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is
+ removed and RE_NO_BK_REFS is added. */
+#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \
+ (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \
+ | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \
+ | RE_NO_BK_PARENS | RE_NO_BK_REFS \
+ | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD)
+/* [[[end syntaxes]]] */
+
+/* Maximum number of duplicates an interval can allow. Some systems
+ (erroneously) define this in other header files, but we want our
+ value, so remove any previous define. */
+#ifdef RE_DUP_MAX
+# undef RE_DUP_MAX
+#endif
+/* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */
+#define RE_DUP_MAX (0x7fff)
+
+
+/* POSIX `cflags' bits (i.e., information for `regcomp'). */
+
+/* If this bit is set, then use extended regular expression syntax.
+ If not set, then use basic regular expression syntax. */
+#define REG_EXTENDED 1
+
+/* If this bit is set, then ignore case when matching.
+ If not set, then case is significant. */
+#define REG_ICASE (REG_EXTENDED << 1)
+
+/* If this bit is set, then anchors do not match at newline
+ characters in the string.
+ If not set, then anchors do match at newlines. */
+#define REG_NEWLINE (REG_ICASE << 1)
+
+/* If this bit is set, then report only success or fail in regexec.
+ If not set, then returns differ between not matching and errors. */
+#define REG_NOSUB (REG_NEWLINE << 1)
+
+
+/* POSIX `eflags' bits (i.e., information for regexec). */
+
+/* If this bit is set, then the beginning-of-line operator doesn't match
+ the beginning of the string (presumably because it's not the
+ beginning of a line).
+ If not set, then the beginning-of-line operator does match the
+ beginning of the string. */
+#define REG_NOTBOL 1
+
+/* Like REG_NOTBOL, except for the end-of-line. */
+#define REG_NOTEOL (1 << 1)
+
+
+/* If any error codes are removed, changed, or added, update the
+ `re_error_msg' table in regex.c. */
+typedef enum
+{
+#ifdef _XOPEN_SOURCE
+ REG_ENOSYS = -1, /* This will never happen for this implementation. */
+#endif
+
+ REG_NOERROR = 0, /* Success. */
+ REG_NOMATCH, /* Didn't find a match (for regexec). */
+
+ /* POSIX regcomp return error codes. (In the order listed in the
+ standard.) */
+ REG_BADPAT, /* Invalid pattern. */
+ REG_ECOLLATE, /* Not implemented. */
+ REG_ECTYPE, /* Invalid character class name. */
+ REG_EESCAPE, /* Trailing backslash. */
+ REG_ESUBREG, /* Invalid back reference. */
+ REG_EBRACK, /* Unmatched left bracket. */
+ REG_EPAREN, /* Parenthesis imbalance. */
+ REG_EBRACE, /* Unmatched \{. */
+ REG_BADBR, /* Invalid contents of \{\}. */
+ REG_ERANGE, /* Invalid range end. */
+ REG_ESPACE, /* Ran out of memory. */
+ REG_BADRPT, /* No preceding re for repetition op. */
+
+ /* Error codes we've added. */
+ REG_EEND, /* Premature end. */
+ REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */
+ REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */
+} reg_errcode_t;
+
+/* This data structure represents a compiled pattern. Before calling
+ the pattern compiler, the fields `buffer', `allocated', `fastmap',
+ `translate', and `no_sub' can be set. After the pattern has been
+ compiled, the `re_nsub' field is available. All other fields are
+ private to the regex routines. */
+
+#ifndef RE_TRANSLATE_TYPE
+# define RE_TRANSLATE_TYPE char *
+#endif
+
+struct re_pattern_buffer
+{
+/* [[[begin pattern_buffer]]] */
+ /* Space that holds the compiled pattern. It is declared as
+ `unsigned char *' because its elements are
+ sometimes used as array indexes. */
+ unsigned char *buffer;
+
+ /* Number of bytes to which `buffer' points. */
+ unsigned long int allocated;
+
+ /* Number of bytes actually used in `buffer'. */
+ unsigned long int used;
+
+ /* Syntax setting with which the pattern was compiled. */
+ reg_syntax_t syntax;
+
+ /* Pointer to a fastmap, if any, otherwise zero. re_search uses
+ the fastmap, if there is one, to skip over impossible
+ starting points for matches. */
+ char *fastmap;
+
+ /* Either a translate table to apply to all characters before
+ comparing them, or zero for no translation. The translation
+ is applied to a pattern when it is compiled and to a string
+ when it is matched. */
+ RE_TRANSLATE_TYPE translate;
+
+ /* Number of subexpressions found by the compiler. */
+ size_t re_nsub;
+
+ /* Zero if this pattern cannot match the empty string, one else.
+ Well, in truth it's used only in `re_search_2', to see
+ whether or not we should use the fastmap, so we don't set
+ this absolutely perfectly; see `re_compile_fastmap' (the
+ `duplicate' case). */
+ unsigned can_be_null : 1;
+
+ /* If REGS_UNALLOCATED, allocate space in the `regs' structure
+ for `max (RE_NREGS, re_nsub + 1)' groups.
+ If REGS_REALLOCATE, reallocate space if necessary.
+ If REGS_FIXED, use what's there. */
+#define REGS_UNALLOCATED 0
+#define REGS_REALLOCATE 1
+#define REGS_FIXED 2
+ unsigned regs_allocated : 2;
+
+ /* Set to zero when `regex_compile' compiles a pattern; set to one
+ by `re_compile_fastmap' if it updates the fastmap. */
+ unsigned fastmap_accurate : 1;
+
+ /* If set, `re_match_2' does not return information about
+ subexpressions. */
+ unsigned no_sub : 1;
+
+ /* If set, a beginning-of-line anchor doesn't match at the
+ beginning of the string. */
+ unsigned not_bol : 1;
+
+ /* Similarly for an end-of-line anchor. */
+ unsigned not_eol : 1;
+
+ /* If true, an anchor at a newline matches. */
+ unsigned newline_anchor : 1;
+
+/* [[[end pattern_buffer]]] */
+};
+
+typedef struct re_pattern_buffer regex_t;
+
+/* Type for byte offsets within the string. POSIX mandates this. */
+typedef int regoff_t;
+
+
+/* This is the structure we store register match data in. See
+ regex.texinfo for a full description of what registers match. */
+struct re_registers
+{
+ unsigned num_regs;
+ regoff_t *start;
+ regoff_t *end;
+};
+
+
+/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer,
+ `re_match_2' returns information about at least this many registers
+ the first time a `regs' structure is passed. */
+#ifndef RE_NREGS
+# define RE_NREGS 30
+#endif
+
+
+/* POSIX specification for registers. Aside from the different names than
+ `re_registers', POSIX uses an array of structures, instead of a
+ structure of arrays. */
+typedef struct
+{
+ regoff_t rm_so; /* Byte offset from string's start to substring's start. */
+ regoff_t rm_eo; /* Byte offset from string's start to substring's end. */
+} regmatch_t;
+
+/* Declarations for routines. */
+
+/* Sets the current default syntax to SYNTAX, and return the old syntax.
+ You can also simply assign to the `re_syntax_options' variable. */
+extern reg_syntax_t re_set_syntax (reg_syntax_t syntax);
+
+/* Compile the regular expression PATTERN, with length LENGTH
+ and syntax given by the global `re_syntax_options', into the buffer
+ BUFFER. Return NULL if successful, and an error string if not. */
+extern const char *re_compile_pattern (const char *pattern, size_t length,
+ struct re_pattern_buffer *buffer);
+
+
+/* Compile a fastmap for the compiled pattern in BUFFER; used to
+ accelerate searches. Return 0 if successful and -2 if was an
+ internal error. */
+extern int re_compile_fastmap (struct re_pattern_buffer *buffer);
+
+
+/* Search in the string STRING (with length LENGTH) for the pattern
+ compiled into BUFFER. Start searching at position START, for RANGE
+ characters. Return the starting position of the match, -1 for no
+ match, or -2 for an internal error. Also return register
+ information in REGS (if REGS and BUFFER->no_sub are nonzero). */
+extern int re_search (struct re_pattern_buffer *buffer, const char *string,
+ int length, int start, int range,
+ struct re_registers *regs);
+
+
+/* Like `re_search', but search in the concatenation of STRING1 and
+ STRING2. Also, stop searching at index START + STOP. */
+extern int re_search_2 (struct re_pattern_buffer *buffer, const char *string1,
+ int length1, const char *string2, int length2,
+ int start, int range, struct re_registers *regs,
+ int stop);
+
+
+/* Like `re_search', but return how many characters in STRING the regexp
+ in BUFFER matched, starting at position START. */
+extern int re_match (struct re_pattern_buffer *buffer, const char *string,
+ int length, int start, struct re_registers *regs);
+
+
+/* Relates to `re_match' as `re_search_2' relates to `re_search'. */
+extern int re_match_2 (struct re_pattern_buffer *buffer, const char *string1,
+ int length1, const char *string2, int length2,
+ int start, struct re_registers *regs, int stop);
+
+
+/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
+ ENDS. Subsequent matches using BUFFER and REGS will use this memory
+ for recording register information. STARTS and ENDS must be
+ allocated with malloc, and must each be at least `NUM_REGS * sizeof
+ (regoff_t)' bytes long.
+
+ If NUM_REGS == 0, then subsequent matches should allocate their own
+ register data.
+
+ Unless this function is called, the first search or match using
+ PATTERN_BUFFER will allocate its own register data, without
+ freeing the old data. */
+extern void re_set_registers (struct re_pattern_buffer *buffer,
+ struct re_registers *regs, unsigned num_regs,
+ regoff_t *starts, regoff_t *ends);
+
+#if defined _REGEX_RE_COMP || defined _LIBC
+# ifndef _CRAY
+/* 4.2 bsd compatibility. */
+extern char *re_comp (const char *);
+extern int re_exec (const char *);
+# endif
+#endif
+
+/* GCC 2.95 and later have "__restrict"; C99 compilers have
+ "restrict", and "configure" may have defined "restrict". */
+#ifndef __restrict
+# if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__))
+# if defined restrict || 199901L <= __STDC_VERSION__
+# define __restrict restrict
+# else
+# define __restrict
+# endif
+# endif
+#endif
+/* gcc 3.1 and up support the [restrict] syntax. */
+#ifndef __restrict_arr
+# if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
+# define __restrict_arr __restrict
+# else
+# define __restrict_arr
+# endif
+#endif
+
+/* POSIX compatibility. */
+extern int regcomp (regex_t *__restrict __preg,
+ const char *__restrict __pattern,
+ int __cflags);
+
+extern int regexec (const regex_t *__restrict __preg,
+ const char *__restrict __string, size_t __nmatch,
+ regmatch_t __pmatch[__restrict_arr],
+ int __eflags);
+
+extern size_t regerror (int __errcode, const regex_t *__preg,
+ char *__errbuf, size_t __errbuf_size);
+
+extern void regfree (regex_t *__preg);
+
+
+#ifdef __cplusplus
+}
+#endif /* C++ */
+
+#endif /* regex.h */
+
+/*
+Local variables:
+make-backup-files: t
+version-control: t
+trim-versions-without-asking: nil
+End:
+*/
diff --git a/contrib/diff/lib/setmode.c b/contrib/diff/lib/setmode.c
new file mode 100644
index 0000000..b2b93a6
--- /dev/null
+++ b/contrib/diff/lib/setmode.c
@@ -0,0 +1,67 @@
+/* Set a file descriptor's mode to binary or to text.
+
+ Copyright (C) 2001, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert <eggert@twinsun.com> */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#if HAVE_STDBOOL_H
+# include <stdbool.h>
+#else
+typedef enum {false = 0, true = 1} bool;
+#endif
+
+#if HAVE_SETMODE_DOS
+# include <io.h>
+# if HAVE_FCNTL_H
+# include <fcntl.h>
+# endif
+# if HAVE_UNISTD_H
+# include <unistd.h>
+# endif
+#endif
+
+#include "setmode.h"
+#undef set_binary_mode
+
+#ifndef __attribute__
+# if __GNUC__ < 3 || __STRICT_ANSI__
+# define __attribute__(x)
+# endif
+#endif
+
+/* Set the binary mode of FD to MODE, returning its previous mode.
+ MODE is 1 for binary and 0 for text. If setting the mode might
+ cause problems, ignore the request and return MODE. Always return
+ 1 on POSIX platforms, which do not distinguish between text and
+ binary. */
+
+#if HAVE_SETMODE_DOS
+bool
+set_binary_mode (int fd, bool mode)
+{
+ if (isatty (fd))
+ return mode;
+ return setmode (fd, mode ? O_BINARY : O_TEXT) != O_TEXT;
+}
+#else
+static char dummy;
+#endif
diff --git a/contrib/diff/lib/setmode.h b/contrib/diff/lib/setmode.h
new file mode 100644
index 0000000..8381d12
--- /dev/null
+++ b/contrib/diff/lib/setmode.h
@@ -0,0 +1,27 @@
+/* Set a file descriptor's mode to binary or to text.
+
+ Copyright (C) 2001, 2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert <eggert@twinsun.com> */
+
+#ifndef set_binary_mode
+bool set_binary_mode (int, bool);
+# if ! HAVE_SETMODE_DOS
+# define set_binary_mode(fd, mode) true
+# endif
+#endif
diff --git a/contrib/diff/lib/stdbool_.h b/contrib/diff/lib/stdbool_.h
new file mode 100644
index 0000000..e33715a
--- /dev/null
+++ b/contrib/diff/lib/stdbool_.h
@@ -0,0 +1,93 @@
+/* Copyright (C) 2001, 2002, 2003 Free Software Foundation, Inc.
+ Written by Bruno Haible <haible@clisp.cons.org>, 2001.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _STDBOOL_H
+#define _STDBOOL_H
+
+/* ISO C 99 <stdbool.h> for platforms that lack it. */
+
+/* Usage suggestions:
+
+ Programs that use <stdbool.h> should be aware of some limitations
+ and standards compliance issues.
+
+ Standards compliance:
+
+ - <stdbool.h> must be #included before 'bool', 'false', 'true'
+ can be used.
+
+ - You cannot assume that sizeof (bool) == 1.
+
+ - Programs should not undefine the macros bool, true, and false,
+ as C99 lists that as an "obsolescent feature".
+
+ Limitations of this substitute, when used in a C89 environment:
+
+ - <stdbool.h> must be #included before the '_Bool' type can be used.
+
+ - You cannot assume that _Bool is a typedef; it might be a macro.
+
+ - In C99, casts and automatic conversions to '_Bool' or 'bool' are
+ performed in such a way that every nonzero value gets converted
+ to 'true', and zero gets converted to 'false'. This doesn't work
+ with this substitute. With this substitute, only the values 0 and 1
+ give the expected result when converted to _Bool' or 'bool'.
+
+ Also, it is suggested that programs use 'bool' rather than '_Bool';
+ this isn't required, but 'bool' is more common. */
+
+
+/* 7.16. Boolean type and values */
+
+/* BeOS <sys/socket.h> already #defines false 0, true 1. We use the same
+ definitions below, but temporarily we have to #undef them. */
+#ifdef __BEOS__
+# include <OS.h> /* defines bool but not _Bool */
+# undef false
+# undef true
+#endif
+
+/* For the sake of symbolic names in gdb, we define true and false as
+ enum constants, not only as macros.
+ It is tempting to write
+ typedef enum { false = 0, true = 1 } _Bool;
+ so that gdb prints values of type 'bool' symbolically. But if we do
+ this, values of type '_Bool' may promote to 'int' or 'unsigned int'
+ (see ISO C 99 6.7.2.2.(4)); however, '_Bool' must promote to 'int'
+ (see ISO C 99 6.3.1.1.(2)). So we add a negative value to the
+ enum; this ensures that '_Bool' promotes to 'int'. */
+#if !(defined __cplusplus || defined __BEOS__)
+# if !@HAVE__BOOL@
+# if defined __SUNPRO_C && (__SUNPRO_C < 0x550 || __STDC__ == 1)
+ /* Avoid stupid "warning: _Bool is a keyword in ISO C99". */
+# define _Bool signed char
+enum { false = 0, true = 1 };
+# else
+typedef enum { _Bool_must_promote_to_int = -1, false = 0, true = 1 } _Bool;
+# endif
+# endif
+#else
+typedef bool _Bool;
+#endif
+#define bool _Bool
+
+/* The other macros must be usable in preprocessor directives. */
+#define false 0
+#define true 1
+#define __bool_true_false_are_defined 1
+
+#endif /* _STDBOOL_H */
diff --git a/contrib/diff/lib/strcase.h b/contrib/diff/lib/strcase.h
new file mode 100644
index 0000000..f17e648
--- /dev/null
+++ b/contrib/diff/lib/strcase.h
@@ -0,0 +1,35 @@
+/* Case-insensitive string comparison functions.
+ Copyright (C) 1995-1996, 2001, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef _STRCASE_H
+#define _STRCASE_H
+
+#include <stddef.h>
+
+/* Compare strings S1 and S2, ignoring case, returning less than, equal to or
+ greater than zero if S1 is lexicographically less than, equal to or greater
+ than S2.
+ Note: This function does not work correctly in multibyte locales. */
+extern int strcasecmp (const char *s1, const char *s2);
+
+/* Compare no more than N characters of strings S1 and S2, ignoring case,
+ returning less than, equal to or greater than zero if S1 is
+ lexicographically less than, equal to or greater than S2.
+ Note: This function can not work correctly in multibyte locales. */
+extern int strncasecmp (const char *s1, const char *s2, size_t n);
+
+#endif /* _STRCASE_H */
diff --git a/contrib/diff/lib/strcasecmp.c b/contrib/diff/lib/strcasecmp.c
new file mode 100644
index 0000000..cf4ab88
--- /dev/null
+++ b/contrib/diff/lib/strcasecmp.c
@@ -0,0 +1,66 @@
+/* strcasecmp.c -- case insensitive string comparator
+ Copyright (C) 1998, 1999 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#ifdef LENGTH_LIMIT
+# define STRXCASECMP_FUNCTION strncasecmp
+# define STRXCASECMP_DECLARE_N , size_t n
+# define LENGTH_LIMIT_EXPR(Expr) Expr
+#else
+# define STRXCASECMP_FUNCTION strcasecmp
+# define STRXCASECMP_DECLARE_N /* empty */
+# define LENGTH_LIMIT_EXPR(Expr) 0
+#endif
+
+#include <stddef.h>
+#include <ctype.h>
+
+#define TOLOWER(Ch) (isupper (Ch) ? tolower (Ch) : (Ch))
+
+/* Compare {{no more than N characters of }}strings S1 and S2,
+ ignoring case, returning less than, equal to or
+ greater than zero if S1 is lexicographically less
+ than, equal to or greater than S2. */
+
+int
+STRXCASECMP_FUNCTION (const char *s1, const char *s2 STRXCASECMP_DECLARE_N)
+{
+ register const unsigned char *p1 = (const unsigned char *) s1;
+ register const unsigned char *p2 = (const unsigned char *) s2;
+ unsigned char c1, c2;
+
+ if (p1 == p2 || LENGTH_LIMIT_EXPR (n == 0))
+ return 0;
+
+ do
+ {
+ c1 = TOLOWER (*p1);
+ c2 = TOLOWER (*p2);
+
+ if (LENGTH_LIMIT_EXPR (--n == 0) || c1 == '\0')
+ break;
+
+ ++p1;
+ ++p2;
+ }
+ while (c1 == c2);
+
+ return c1 - c2;
+}
diff --git a/contrib/diff/lib/strftime.c b/contrib/diff/lib/strftime.c
new file mode 100644
index 0000000..dccbaf7
--- /dev/null
+++ b/contrib/diff/lib/strftime.c
@@ -0,0 +1,1330 @@
+/* Copyright (C) 1991-1999, 2000, 2001, 2003 Free Software Foundation, Inc.
+
+ NOTE: The canonical source of this file is maintained with the GNU C Library.
+ Bugs can be reported to bug-glibc@prep.ai.mit.edu.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifdef HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#ifdef _LIBC
+# define HAVE_MBLEN 1
+# define HAVE_MBRLEN 1
+# define HAVE_STRUCT_ERA_ENTRY 1
+# define HAVE_TM_GMTOFF 1
+# define HAVE_TM_ZONE 1
+# define HAVE_TZNAME 1
+# define HAVE_TZSET 1
+# define MULTIBYTE_IS_FORMAT_SAFE 1
+# include "../locale/localeinfo.h"
+#endif
+
+#include <ctype.h>
+#include <sys/types.h> /* Some systems define `time_t' here. */
+
+#ifdef TIME_WITH_SYS_TIME
+# include <sys/time.h>
+# include <time.h>
+#else
+# ifdef HAVE_SYS_TIME_H
+# include <sys/time.h>
+# else
+# include <time.h>
+# endif
+#endif
+#if HAVE_TZNAME
+extern char *tzname[];
+#endif
+
+/* Do multibyte processing if multibytes are supported, unless
+ multibyte sequences are safe in formats. Multibyte sequences are
+ safe if they cannot contain byte sequences that look like format
+ conversion specifications. The GNU C Library uses UTF8 multibyte
+ encoding, which is safe for formats, but strftime.c can be used
+ with other C libraries that use unsafe encodings. */
+#define DO_MULTIBYTE (HAVE_MBLEN && ! MULTIBYTE_IS_FORMAT_SAFE)
+
+#if DO_MULTIBYTE
+# if HAVE_MBRLEN
+# include <wchar.h>
+# else
+ /* Simulate mbrlen with mblen as best we can. */
+# define mbstate_t int
+# define mbrlen(s, n, ps) mblen (s, n)
+# define mbsinit(ps) (*(ps) == 0)
+# endif
+ static const mbstate_t mbstate_zero;
+#endif
+
+#include <limits.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef COMPILE_WIDE
+# include <endian.h>
+# define CHAR_T wchar_t
+# define UCHAR_T unsigned int
+# define L_(Str) L##Str
+# define NLW(Sym) _NL_W##Sym
+
+# define MEMCPY(d, s, n) __wmemcpy (d, s, n)
+# define STRLEN(s) __wcslen (s)
+
+#else
+# define CHAR_T char
+# define UCHAR_T unsigned char
+# define L_(Str) Str
+# define NLW(Sym) Sym
+
+# define MEMCPY(d, s, n) memcpy (d, s, n)
+# define STRLEN(s) strlen (s)
+
+# ifdef _LIBC
+# define MEMPCPY(d, s, n) __mempcpy (d, s, n)
+# else
+# ifndef HAVE_MEMPCPY
+# define MEMPCPY(d, s, n) ((void *) ((char *) memcpy (d, s, n) + (n)))
+# endif
+# endif
+#endif
+
+#define TYPE_SIGNED(t) ((t) -1 < 0)
+
+/* Bound on length of the string representing an integer value of type t.
+ Subtract one for the sign bit if t is signed;
+ 302 / 1000 is log10 (2) rounded up;
+ add one for integer division truncation;
+ add one more for a minus sign if t is signed. */
+#define INT_STRLEN_BOUND(t) \
+ ((sizeof (t) * CHAR_BIT - TYPE_SIGNED (t)) * 302 / 1000 + 1 + TYPE_SIGNED (t))
+
+#define TM_YEAR_BASE 1900
+
+#ifndef __isleap
+/* Nonzero if YEAR is a leap year (every 4 years,
+ except every 100th isn't, and every 400th is). */
+# define __isleap(year) \
+ ((year) % 4 == 0 && ((year) % 100 != 0 || (year) % 400 == 0))
+#endif
+
+
+#ifdef _LIBC
+# define tzname __tzname
+# define tzset __tzset
+#endif
+
+#if !HAVE_TM_GMTOFF
+/* Portable standalone applications should supply a "time_r.h" that
+ declares a POSIX-compliant localtime_r, for the benefit of older
+ implementations that lack localtime_r or have a nonstandard one.
+ See the gnulib time_r module for one way to implement this. */
+# include "time_r.h"
+# undef __gmtime_r
+# undef __localtime_r
+# define __gmtime_r gmtime_r
+# define __localtime_r localtime_r
+#endif
+
+
+#ifdef COMPILE_WIDE
+# define memset_space(P, Len) (wmemset (P, L' ', Len), (P) += (Len))
+# define memset_zero(P, Len) (wmemset (P, L'0', Len), (P) += (Len))
+#else
+# define memset_space(P, Len) (memset (P, ' ', Len), (P) += (Len))
+# define memset_zero(P, Len) (memset (P, '0', Len), (P) += (Len))
+#endif
+
+#define add(n, f) \
+ do \
+ { \
+ int _n = (n); \
+ int _delta = width - _n; \
+ int _incr = _n + (_delta > 0 ? _delta : 0); \
+ if ((size_t) _incr >= maxsize - i) \
+ return 0; \
+ if (p) \
+ { \
+ if (_delta > 0) \
+ { \
+ if (pad == L_('0')) \
+ memset_zero (p, _delta); \
+ else \
+ memset_space (p, _delta); \
+ } \
+ f; \
+ p += _n; \
+ } \
+ i += _incr; \
+ } while (0)
+
+#define cpy(n, s) \
+ add ((n), \
+ if (to_lowcase) \
+ memcpy_lowcase (p, (s), _n LOCALE_ARG); \
+ else if (to_uppcase) \
+ memcpy_uppcase (p, (s), _n LOCALE_ARG); \
+ else \
+ MEMCPY ((void *) p, (void const *) (s), _n))
+
+#ifdef COMPILE_WIDE
+# ifndef USE_IN_EXTENDED_LOCALE_MODEL
+# undef __mbsrtowcs_l
+# define __mbsrtowcs_l(d, s, l, st, loc) __mbsrtowcs (d, s, l, st)
+# endif
+# define widen(os, ws, l) \
+ { \
+ mbstate_t __st; \
+ const char *__s = os; \
+ memset (&__st, '\0', sizeof (__st)); \
+ l = __mbsrtowcs_l (NULL, &__s, 0, &__st, loc); \
+ ws = (wchar_t *) alloca ((l + 1) * sizeof (wchar_t)); \
+ (void) __mbsrtowcs_l (ws, &__s, l, &__st, loc); \
+ }
+#endif
+
+
+#if defined _LIBC && defined USE_IN_EXTENDED_LOCALE_MODEL
+/* We use this code also for the extended locale handling where the
+ function gets as an additional argument the locale which has to be
+ used. To access the values we have to redefine the _NL_CURRENT
+ macro. */
+# define strftime __strftime_l
+# define wcsftime __wcsftime_l
+# undef _NL_CURRENT
+# define _NL_CURRENT(category, item) \
+ (current->values[_NL_ITEM_INDEX (item)].string)
+# define LOCALE_ARG , loc
+# define LOCALE_PARAM_PROTO , __locale_t loc
+# define HELPER_LOCALE_ARG , current
+#else
+# define LOCALE_PARAM_PROTO
+# define LOCALE_ARG
+# ifdef _LIBC
+# define HELPER_LOCALE_ARG , _NL_CURRENT_DATA (LC_TIME)
+# else
+# define HELPER_LOCALE_ARG
+# endif
+#endif
+
+#ifdef COMPILE_WIDE
+# ifdef USE_IN_EXTENDED_LOCALE_MODEL
+# define TOUPPER(Ch, L) __towupper_l (Ch, L)
+# define TOLOWER(Ch, L) __towlower_l (Ch, L)
+# else
+# define TOUPPER(Ch, L) towupper (Ch)
+# define TOLOWER(Ch, L) towlower (Ch)
+# endif
+#else
+# ifdef _LIBC
+# ifdef USE_IN_EXTENDED_LOCALE_MODEL
+# define TOUPPER(Ch, L) __toupper_l (Ch, L)
+# define TOLOWER(Ch, L) __tolower_l (Ch, L)
+# else
+# define TOUPPER(Ch, L) toupper (Ch)
+# define TOLOWER(Ch, L) tolower (Ch)
+# endif
+# else
+# define TOUPPER(Ch, L) (islower (Ch) ? toupper (Ch) : (Ch))
+# define TOLOWER(Ch, L) (isupper (Ch) ? tolower (Ch) : (Ch))
+# endif
+#endif
+/* We don't use `isdigit' here since the locale dependent
+ interpretation is not what we want here. We only need to accept
+ the arabic digits in the ASCII range. One day there is perhaps a
+ more reliable way to accept other sets of digits. */
+#define ISDIGIT(Ch) ((unsigned int) (Ch) - L_('0') <= 9)
+
+static CHAR_T *
+memcpy_lowcase (CHAR_T *dest, const CHAR_T *src,
+ size_t len LOCALE_PARAM_PROTO)
+{
+ while (len-- > 0)
+ dest[len] = TOLOWER ((UCHAR_T) src[len], loc);
+ return dest;
+}
+
+static CHAR_T *
+memcpy_uppcase (CHAR_T *dest, const CHAR_T *src,
+ size_t len LOCALE_PARAM_PROTO)
+{
+ while (len-- > 0)
+ dest[len] = TOUPPER ((UCHAR_T) src[len], loc);
+ return dest;
+}
+
+
+#if ! HAVE_TM_GMTOFF
+/* Yield the difference between *A and *B,
+ measured in seconds, ignoring leap seconds. */
+# define tm_diff ftime_tm_diff
+static int
+tm_diff (const struct tm *a, const struct tm *b)
+{
+ /* Compute intervening leap days correctly even if year is negative.
+ Take care to avoid int overflow in leap day calculations,
+ but it's OK to assume that A and B are close to each other. */
+ int a4 = (a->tm_year >> 2) + (TM_YEAR_BASE >> 2) - ! (a->tm_year & 3);
+ int b4 = (b->tm_year >> 2) + (TM_YEAR_BASE >> 2) - ! (b->tm_year & 3);
+ int a100 = a4 / 25 - (a4 % 25 < 0);
+ int b100 = b4 / 25 - (b4 % 25 < 0);
+ int a400 = a100 >> 2;
+ int b400 = b100 >> 2;
+ int intervening_leap_days = (a4 - b4) - (a100 - b100) + (a400 - b400);
+ int years = a->tm_year - b->tm_year;
+ int days = (365 * years + intervening_leap_days
+ + (a->tm_yday - b->tm_yday));
+ return (60 * (60 * (24 * days + (a->tm_hour - b->tm_hour))
+ + (a->tm_min - b->tm_min))
+ + (a->tm_sec - b->tm_sec));
+}
+#endif /* ! HAVE_TM_GMTOFF */
+
+
+
+/* The number of days from the first day of the first ISO week of this
+ year to the year day YDAY with week day WDAY. ISO weeks start on
+ Monday; the first ISO week has the year's first Thursday. YDAY may
+ be as small as YDAY_MINIMUM. */
+#define ISO_WEEK_START_WDAY 1 /* Monday */
+#define ISO_WEEK1_WDAY 4 /* Thursday */
+#define YDAY_MINIMUM (-366)
+#ifdef __GNUC__
+__inline__
+#endif
+static int
+iso_week_days (int yday, int wday)
+{
+ /* Add enough to the first operand of % to make it nonnegative. */
+ int big_enough_multiple_of_7 = (-YDAY_MINIMUM / 7 + 2) * 7;
+ return (yday
+ - (yday - wday + ISO_WEEK1_WDAY + big_enough_multiple_of_7) % 7
+ + ISO_WEEK1_WDAY - ISO_WEEK_START_WDAY);
+}
+
+
+#if !(defined _NL_CURRENT || HAVE_STRFTIME)
+static CHAR_T const weekday_name[][10] =
+ {
+ L_("Sunday"), L_("Monday"), L_("Tuesday"), L_("Wednesday"),
+ L_("Thursday"), L_("Friday"), L_("Saturday")
+ };
+static CHAR_T const month_name[][10] =
+ {
+ L_("January"), L_("February"), L_("March"), L_("April"), L_("May"),
+ L_("June"), L_("July"), L_("August"), L_("September"), L_("October"),
+ L_("November"), L_("December")
+ };
+#endif
+
+
+/* When compiling this file, GNU applications can #define my_strftime
+ to a symbol (typically nstrftime) to get an extended strftime with
+ extra arguments UT and NS. Emacs is a special case for now, but
+ this Emacs-specific code can be removed once Emacs's config.h
+ defines my_strftime. */
+#if defined emacs && !defined my_strftime
+# define my_strftime nstrftime
+#endif
+
+#ifdef my_strftime
+# define extra_args , ut, ns
+# define extra_args_spec , int ut, int ns
+#else
+# ifdef COMPILE_WIDE
+# define my_strftime wcsftime
+# define nl_get_alt_digit _nl_get_walt_digit
+# else
+# define my_strftime strftime
+# define nl_get_alt_digit _nl_get_alt_digit
+# endif
+# define extra_args
+# define extra_args_spec
+/* We don't have this information in general. */
+# define ut 0
+# define ns 0
+#endif
+
+#if ! defined _LIBC && ! HAVE_RUN_TZSET_TEST
+/* Solaris 2.5.x and 2.6 tzset sometimes modify the storage returned
+ by localtime. On such systems, we must use the tzset and localtime
+ wrappers to work around the bug. */
+"you must run the autoconf test for a working tzset function"
+#endif
+
+
+/* Write information from TP into S according to the format
+ string FORMAT, writing no more that MAXSIZE characters
+ (including the terminating '\0') and returning number of
+ characters written. If S is NULL, nothing will be written
+ anywhere, so to determine how many characters would be
+ written, use NULL for S and (size_t) UINT_MAX for MAXSIZE. */
+size_t
+my_strftime (CHAR_T *s, size_t maxsize, const CHAR_T *format,
+ const struct tm *tp extra_args_spec LOCALE_PARAM_PROTO)
+{
+#if defined _LIBC && defined USE_IN_EXTENDED_LOCALE_MODEL
+ struct locale_data *const current = loc->__locales[LC_TIME];
+#endif
+
+ int hour12 = tp->tm_hour;
+#ifdef _NL_CURRENT
+ /* We cannot make the following values variables since we must delay
+ the evaluation of these values until really needed since some
+ expressions might not be valid in every situation. The `struct tm'
+ might be generated by a strptime() call that initialized
+ only a few elements. Dereference the pointers only if the format
+ requires this. Then it is ok to fail if the pointers are invalid. */
+# define a_wkday \
+ ((const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(ABDAY_1) + tp->tm_wday))
+# define f_wkday \
+ ((const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(DAY_1) + tp->tm_wday))
+# define a_month \
+ ((const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(ABMON_1) + tp->tm_mon))
+# define f_month \
+ ((const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(MON_1) + tp->tm_mon))
+# define ampm \
+ ((const CHAR_T *) _NL_CURRENT (LC_TIME, tp->tm_hour > 11 \
+ ? NLW(PM_STR) : NLW(AM_STR)))
+
+# define aw_len STRLEN (a_wkday)
+# define am_len STRLEN (a_month)
+# define ap_len STRLEN (ampm)
+#else
+# if !HAVE_STRFTIME
+# define f_wkday (weekday_name[tp->tm_wday])
+# define f_month (month_name[tp->tm_mon])
+# define a_wkday f_wkday
+# define a_month f_month
+# define ampm (L_("AMPM") + 2 * (tp->tm_hour > 11))
+
+ size_t aw_len = 3;
+ size_t am_len = 3;
+ size_t ap_len = 2;
+# endif
+#endif
+ const char *zone;
+ size_t i = 0;
+ CHAR_T *p = s;
+ const CHAR_T *f;
+#if DO_MULTIBYTE && !defined COMPILE_WIDE
+ const char *format_end = NULL;
+#endif
+
+ zone = NULL;
+#if HAVE_TM_ZONE
+ /* The POSIX test suite assumes that setting
+ the environment variable TZ to a new value before calling strftime()
+ will influence the result (the %Z format) even if the information in
+ TP is computed with a totally different time zone.
+ This is bogus: though POSIX allows bad behavior like this,
+ POSIX does not require it. Do the right thing instead. */
+ zone = (const char *) tp->tm_zone;
+#endif
+#if HAVE_TZNAME
+ if (ut)
+ {
+ if (! (zone && *zone))
+ zone = "GMT";
+ }
+ else
+ {
+ /* POSIX.1 requires that local time zone information be used as
+ though strftime called tzset. */
+# if HAVE_TZSET
+ tzset ();
+# endif
+ }
+#endif
+
+ if (hour12 > 12)
+ hour12 -= 12;
+ else
+ if (hour12 == 0)
+ hour12 = 12;
+
+ for (f = format; *f != '\0'; ++f)
+ {
+ int pad = 0; /* Padding for number ('-', '_', or 0). */
+ int modifier; /* Field modifier ('E', 'O', or 0). */
+ int digits; /* Max digits for numeric format. */
+ int number_value; /* Numeric value to be printed. */
+ int negative_number; /* 1 if the number is negative. */
+ const CHAR_T *subfmt;
+ CHAR_T *bufp;
+ CHAR_T buf[1 + (sizeof (int) < sizeof (time_t)
+ ? INT_STRLEN_BOUND (time_t)
+ : INT_STRLEN_BOUND (int))];
+ int width = -1;
+ int to_lowcase = 0;
+ int to_uppcase = 0;
+ int change_case = 0;
+ int format_char;
+
+#if DO_MULTIBYTE && !defined COMPILE_WIDE
+ switch (*f)
+ {
+ case L_('%'):
+ break;
+
+ case L_('\b'): case L_('\t'): case L_('\n'):
+ case L_('\v'): case L_('\f'): case L_('\r'):
+ case L_(' '): case L_('!'): case L_('"'): case L_('#'): case L_('&'):
+ case L_('\''): case L_('('): case L_(')'): case L_('*'): case L_('+'):
+ case L_(','): case L_('-'): case L_('.'): case L_('/'): case L_('0'):
+ case L_('1'): case L_('2'): case L_('3'): case L_('4'): case L_('5'):
+ case L_('6'): case L_('7'): case L_('8'): case L_('9'): case L_(':'):
+ case L_(';'): case L_('<'): case L_('='): case L_('>'): case L_('?'):
+ case L_('A'): case L_('B'): case L_('C'): case L_('D'): case L_('E'):
+ case L_('F'): case L_('G'): case L_('H'): case L_('I'): case L_('J'):
+ case L_('K'): case L_('L'): case L_('M'): case L_('N'): case L_('O'):
+ case L_('P'): case L_('Q'): case L_('R'): case L_('S'): case L_('T'):
+ case L_('U'): case L_('V'): case L_('W'): case L_('X'): case L_('Y'):
+ case L_('Z'): case L_('['): case L_('\\'): case L_(']'): case L_('^'):
+ case L_('_'): case L_('a'): case L_('b'): case L_('c'): case L_('d'):
+ case L_('e'): case L_('f'): case L_('g'): case L_('h'): case L_('i'):
+ case L_('j'): case L_('k'): case L_('l'): case L_('m'): case L_('n'):
+ case L_('o'): case L_('p'): case L_('q'): case L_('r'): case L_('s'):
+ case L_('t'): case L_('u'): case L_('v'): case L_('w'): case L_('x'):
+ case L_('y'): case L_('z'): case L_('{'): case L_('|'): case L_('}'):
+ case L_('~'):
+ /* The C Standard requires these 98 characters (plus '%') to
+ be in the basic execution character set. None of these
+ characters can start a multibyte sequence, so they need
+ not be analyzed further. */
+ add (1, *p = *f);
+ continue;
+
+ default:
+ /* Copy this multibyte sequence until we reach its end, find
+ an error, or come back to the initial shift state. */
+ {
+ mbstate_t mbstate = mbstate_zero;
+ size_t len = 0;
+ size_t fsize;
+
+ if (! format_end)
+ format_end = f + strlen (f) + 1;
+ fsize = format_end - f;
+
+ do
+ {
+ size_t bytes = mbrlen (f + len, fsize - len, &mbstate);
+
+ if (bytes == 0)
+ break;
+
+ if (bytes == (size_t) -2)
+ {
+ len += strlen (f + len);
+ break;
+ }
+
+ if (bytes == (size_t) -1)
+ {
+ len++;
+ break;
+ }
+
+ len += bytes;
+ }
+ while (! mbsinit (&mbstate));
+
+ cpy (len, f);
+ f += len - 1;
+ continue;
+ }
+ }
+
+#else /* ! DO_MULTIBYTE */
+
+ /* Either multibyte encodings are not supported, they are
+ safe for formats, so any non-'%' byte can be copied through,
+ or this is the wide character version. */
+ if (*f != L_('%'))
+ {
+ add (1, *p = *f);
+ continue;
+ }
+
+#endif /* ! DO_MULTIBYTE */
+
+ /* Check for flags that can modify a format. */
+ while (1)
+ {
+ switch (*++f)
+ {
+ /* This influences the number formats. */
+ case L_('_'):
+ case L_('-'):
+ case L_('0'):
+ pad = *f;
+ continue;
+
+ /* This changes textual output. */
+ case L_('^'):
+ to_uppcase = 1;
+ continue;
+ case L_('#'):
+ change_case = 1;
+ continue;
+
+ default:
+ break;
+ }
+ break;
+ }
+
+ /* As a GNU extension we allow to specify the field width. */
+ if (ISDIGIT (*f))
+ {
+ width = 0;
+ do
+ {
+ if (width > INT_MAX / 10
+ || (width == INT_MAX / 10 && *f - L_('0') > INT_MAX % 10))
+ /* Avoid overflow. */
+ width = INT_MAX;
+ else
+ {
+ width *= 10;
+ width += *f - L_('0');
+ }
+ ++f;
+ }
+ while (ISDIGIT (*f));
+ }
+
+ /* Check for modifiers. */
+ switch (*f)
+ {
+ case L_('E'):
+ case L_('O'):
+ modifier = *f++;
+ break;
+
+ default:
+ modifier = 0;
+ break;
+ }
+
+ /* Now do the specified format. */
+ format_char = *f;
+ switch (format_char)
+ {
+#define DO_NUMBER(d, v) \
+ digits = d > width ? d : width; \
+ number_value = v; goto do_number
+#define DO_NUMBER_SPACEPAD(d, v) \
+ digits = d > width ? d : width; \
+ number_value = v; goto do_number_spacepad
+
+ case L_('%'):
+ if (modifier != 0)
+ goto bad_format;
+ add (1, *p = *f);
+ break;
+
+ case L_('a'):
+ if (modifier != 0)
+ goto bad_format;
+ if (change_case)
+ {
+ to_uppcase = 1;
+ to_lowcase = 0;
+ }
+#if defined _NL_CURRENT || !HAVE_STRFTIME
+ cpy (aw_len, a_wkday);
+ break;
+#else
+ goto underlying_strftime;
+#endif
+
+ case 'A':
+ if (modifier != 0)
+ goto bad_format;
+ if (change_case)
+ {
+ to_uppcase = 1;
+ to_lowcase = 0;
+ }
+#if defined _NL_CURRENT || !HAVE_STRFTIME
+ cpy (STRLEN (f_wkday), f_wkday);
+ break;
+#else
+ goto underlying_strftime;
+#endif
+
+ case L_('b'):
+ case L_('h'):
+ if (change_case)
+ {
+ to_uppcase = 1;
+ to_lowcase = 0;
+ }
+ if (modifier != 0)
+ goto bad_format;
+#if defined _NL_CURRENT || !HAVE_STRFTIME
+ cpy (am_len, a_month);
+ break;
+#else
+ goto underlying_strftime;
+#endif
+
+ case L_('B'):
+ if (modifier != 0)
+ goto bad_format;
+ if (change_case)
+ {
+ to_uppcase = 1;
+ to_lowcase = 0;
+ }
+#if defined _NL_CURRENT || !HAVE_STRFTIME
+ cpy (STRLEN (f_month), f_month);
+ break;
+#else
+ goto underlying_strftime;
+#endif
+
+ case L_('c'):
+ if (modifier == L_('O'))
+ goto bad_format;
+#ifdef _NL_CURRENT
+ if (! (modifier == 'E'
+ && (*(subfmt =
+ (const CHAR_T *) _NL_CURRENT (LC_TIME,
+ NLW(ERA_D_T_FMT)))
+ != '\0')))
+ subfmt = (const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(D_T_FMT));
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# else
+ subfmt = L_("%a %b %e %H:%M:%S %Y");
+# endif
+#endif
+
+ subformat:
+ {
+ CHAR_T *old_start = p;
+ size_t len = my_strftime (NULL, (size_t) -1, subfmt,
+ tp extra_args LOCALE_ARG);
+ add (len, my_strftime (p, maxsize - i, subfmt,
+ tp extra_args LOCALE_ARG));
+
+ if (to_uppcase)
+ while (old_start < p)
+ {
+ *old_start = TOUPPER ((UCHAR_T) *old_start, loc);
+ ++old_start;
+ }
+ }
+ break;
+
+#if HAVE_STRFTIME && ! (defined _NL_CURRENT && HAVE_STRUCT_ERA_ENTRY)
+ underlying_strftime:
+ {
+ /* The relevant information is available only via the
+ underlying strftime implementation, so use that. */
+ char ufmt[4];
+ char *u = ufmt;
+ char ubuf[1024]; /* enough for any single format in practice */
+ size_t len;
+ /* Make sure we're calling the actual underlying strftime.
+ In some cases, config.h contains something like
+ "#define strftime rpl_strftime". */
+# ifdef strftime
+# undef strftime
+ size_t strftime ();
+# endif
+
+ *u++ = '%';
+ if (modifier != 0)
+ *u++ = modifier;
+ *u++ = format_char;
+ *u = '\0';
+ len = strftime (ubuf, sizeof ubuf, ufmt, tp);
+ if (len == 0 && ubuf[0] != '\0')
+ return 0;
+ cpy (len, ubuf);
+ }
+ break;
+#endif
+
+ case L_('C'):
+ if (modifier == L_('O'))
+ goto bad_format;
+ if (modifier == L_('E'))
+ {
+#if HAVE_STRUCT_ERA_ENTRY
+ struct era_entry *era = _nl_get_era_entry (tp HELPER_LOCALE_ARG);
+ if (era)
+ {
+# ifdef COMPILE_WIDE
+ size_t len = __wcslen (era->era_wname);
+ cpy (len, era->era_wname);
+# else
+ size_t len = strlen (era->era_name);
+ cpy (len, era->era_name);
+# endif
+ break;
+ }
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# endif
+#endif
+ }
+
+ {
+ int year = tp->tm_year + TM_YEAR_BASE;
+ DO_NUMBER (1, year / 100 - (year % 100 < 0));
+ }
+
+ case L_('x'):
+ if (modifier == L_('O'))
+ goto bad_format;
+#ifdef _NL_CURRENT
+ if (! (modifier == L_('E')
+ && (*(subfmt =
+ (const CHAR_T *)_NL_CURRENT (LC_TIME, NLW(ERA_D_FMT)))
+ != L_('\0'))))
+ subfmt = (const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(D_FMT));
+ goto subformat;
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# else
+ /* Fall through. */
+# endif
+#endif
+ case L_('D'):
+ if (modifier != 0)
+ goto bad_format;
+ subfmt = L_("%m/%d/%y");
+ goto subformat;
+
+ case L_('d'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, tp->tm_mday);
+
+ case L_('e'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER_SPACEPAD (2, tp->tm_mday);
+
+ /* All numeric formats set DIGITS and NUMBER_VALUE and then
+ jump to one of these two labels. */
+
+ do_number_spacepad:
+ /* Force `_' flag unless overridden by `0' or `-' flag. */
+ if (pad != L_('0') && pad != L_('-'))
+ pad = L_('_');
+
+ do_number:
+ /* Format the number according to the MODIFIER flag. */
+
+ if (modifier == L_('O') && 0 <= number_value)
+ {
+#ifdef _NL_CURRENT
+ /* Get the locale specific alternate representation of
+ the number NUMBER_VALUE. If none exist NULL is returned. */
+ const CHAR_T *cp = nl_get_alt_digit (number_value
+ HELPER_LOCALE_ARG);
+
+ if (cp != NULL)
+ {
+ size_t digitlen = STRLEN (cp);
+ if (digitlen != 0)
+ {
+ cpy (digitlen, cp);
+ break;
+ }
+ }
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# endif
+#endif
+ }
+ {
+ unsigned int u = number_value;
+
+ bufp = buf + sizeof (buf) / sizeof (buf[0]);
+ negative_number = number_value < 0;
+
+ if (negative_number)
+ u = -u;
+
+ do
+ *--bufp = u % 10 + L_('0');
+ while ((u /= 10) != 0);
+ }
+
+ do_number_sign_and_padding:
+ if (negative_number)
+ *--bufp = L_('-');
+
+ if (pad != L_('-'))
+ {
+ int padding = digits - (buf + (sizeof (buf) / sizeof (buf[0]))
+ - bufp);
+
+ if (padding > 0)
+ {
+ if (pad == L_('_'))
+ {
+ if ((size_t) padding >= maxsize - i)
+ return 0;
+
+ if (p)
+ memset_space (p, padding);
+ i += padding;
+ width = width > padding ? width - padding : 0;
+ }
+ else
+ {
+ if ((size_t) digits >= maxsize - i)
+ return 0;
+
+ if (negative_number)
+ {
+ ++bufp;
+
+ if (p)
+ *p++ = L_('-');
+ ++i;
+ }
+
+ if (p)
+ memset_zero (p, padding);
+ i += padding;
+ width = 0;
+ }
+ }
+ }
+
+ cpy (buf + sizeof (buf) / sizeof (buf[0]) - bufp, bufp);
+ break;
+
+ case L_('F'):
+ if (modifier != 0)
+ goto bad_format;
+ subfmt = L_("%Y-%m-%d");
+ goto subformat;
+
+ case L_('H'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, tp->tm_hour);
+
+ case L_('I'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, hour12);
+
+ case L_('k'): /* GNU extension. */
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER_SPACEPAD (2, tp->tm_hour);
+
+ case L_('l'): /* GNU extension. */
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER_SPACEPAD (2, hour12);
+
+ case L_('j'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (3, 1 + tp->tm_yday);
+
+ case L_('M'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, tp->tm_min);
+
+ case L_('m'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, tp->tm_mon + 1);
+
+#ifndef _LIBC
+ case L_('N'): /* GNU extension. */
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ number_value = ns;
+ if (width != -1)
+ {
+ /* Take an explicit width less than 9 as a precision. */
+ int j;
+ for (j = width; j < 9; j++)
+ number_value /= 10;
+ }
+
+ DO_NUMBER (9, number_value);
+#endif
+
+ case L_('n'):
+ add (1, *p = L_('\n'));
+ break;
+
+ case L_('P'):
+ to_lowcase = 1;
+#if !defined _NL_CURRENT && HAVE_STRFTIME
+ format_char = L_('p');
+#endif
+ /* FALLTHROUGH */
+
+ case L_('p'):
+ if (change_case)
+ {
+ to_uppcase = 0;
+ to_lowcase = 1;
+ }
+#if defined _NL_CURRENT || !HAVE_STRFTIME
+ cpy (ap_len, ampm);
+ break;
+#else
+ goto underlying_strftime;
+#endif
+
+ case L_('R'):
+ subfmt = L_("%H:%M");
+ goto subformat;
+
+ case L_('r'):
+#if !defined _NL_CURRENT && HAVE_STRFTIME
+ goto underlying_strftime;
+#else
+# ifdef _NL_CURRENT
+ if (*(subfmt = (const CHAR_T *) _NL_CURRENT (LC_TIME,
+ NLW(T_FMT_AMPM)))
+ == L_('\0'))
+# endif
+ subfmt = L_("%I:%M:%S %p");
+ goto subformat;
+#endif
+
+ case L_('S'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, tp->tm_sec);
+
+ case L_('s'): /* GNU extension. */
+ {
+ struct tm ltm;
+ time_t t;
+
+ ltm = *tp;
+ t = mktime (&ltm);
+
+ /* Generate string value for T using time_t arithmetic;
+ this works even if sizeof (long) < sizeof (time_t). */
+
+ bufp = buf + sizeof (buf) / sizeof (buf[0]);
+ negative_number = t < 0;
+
+ do
+ {
+ int d = t % 10;
+ t /= 10;
+
+ if (negative_number)
+ {
+ d = -d;
+
+ /* Adjust if division truncates to minus infinity. */
+ if (0 < -1 % 10 && d < 0)
+ {
+ t++;
+ d += 10;
+ }
+ }
+
+ *--bufp = d + L_('0');
+ }
+ while (t != 0);
+
+ digits = 1;
+ goto do_number_sign_and_padding;
+ }
+
+ case L_('X'):
+ if (modifier == L_('O'))
+ goto bad_format;
+#ifdef _NL_CURRENT
+ if (! (modifier == L_('E')
+ && (*(subfmt =
+ (const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(ERA_T_FMT)))
+ != L_('\0'))))
+ subfmt = (const CHAR_T *) _NL_CURRENT (LC_TIME, NLW(T_FMT));
+ goto subformat;
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# else
+ /* Fall through. */
+# endif
+#endif
+ case L_('T'):
+ subfmt = L_("%H:%M:%S");
+ goto subformat;
+
+ case L_('t'):
+ add (1, *p = L_('\t'));
+ break;
+
+ case L_('u'):
+ DO_NUMBER (1, (tp->tm_wday - 1 + 7) % 7 + 1);
+
+ case L_('U'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, (tp->tm_yday - tp->tm_wday + 7) / 7);
+
+ case L_('V'):
+ case L_('g'):
+ case L_('G'):
+ if (modifier == L_('E'))
+ goto bad_format;
+ {
+ int year = tp->tm_year + TM_YEAR_BASE;
+ int days = iso_week_days (tp->tm_yday, tp->tm_wday);
+
+ if (days < 0)
+ {
+ /* This ISO week belongs to the previous year. */
+ year--;
+ days = iso_week_days (tp->tm_yday + (365 + __isleap (year)),
+ tp->tm_wday);
+ }
+ else
+ {
+ int d = iso_week_days (tp->tm_yday - (365 + __isleap (year)),
+ tp->tm_wday);
+ if (0 <= d)
+ {
+ /* This ISO week belongs to the next year. */
+ year++;
+ days = d;
+ }
+ }
+
+ switch (*f)
+ {
+ case L_('g'):
+ DO_NUMBER (2, (year % 100 + 100) % 100);
+
+ case L_('G'):
+ DO_NUMBER (1, year);
+
+ default:
+ DO_NUMBER (2, days / 7 + 1);
+ }
+ }
+
+ case L_('W'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (2, (tp->tm_yday - (tp->tm_wday - 1 + 7) % 7 + 7) / 7);
+
+ case L_('w'):
+ if (modifier == L_('E'))
+ goto bad_format;
+
+ DO_NUMBER (1, tp->tm_wday);
+
+ case L_('Y'):
+ if (modifier == 'E')
+ {
+#if HAVE_STRUCT_ERA_ENTRY
+ struct era_entry *era = _nl_get_era_entry (tp HELPER_LOCALE_ARG);
+ if (era)
+ {
+# ifdef COMPILE_WIDE
+ subfmt = era->era_wformat;
+# else
+ subfmt = era->era_format;
+# endif
+ goto subformat;
+ }
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# endif
+#endif
+ }
+ if (modifier == L_('O'))
+ goto bad_format;
+ else
+ DO_NUMBER (1, tp->tm_year + TM_YEAR_BASE);
+
+ case L_('y'):
+ if (modifier == L_('E'))
+ {
+#if HAVE_STRUCT_ERA_ENTRY
+ struct era_entry *era = _nl_get_era_entry (tp HELPER_LOCALE_ARG);
+ if (era)
+ {
+ int delta = tp->tm_year - era->start_date[0];
+ DO_NUMBER (1, (era->offset
+ + delta * era->absolute_direction));
+ }
+#else
+# if HAVE_STRFTIME
+ goto underlying_strftime;
+# endif
+#endif
+ }
+ DO_NUMBER (2, (tp->tm_year % 100 + 100) % 100);
+
+ case L_('Z'):
+ if (change_case)
+ {
+ to_uppcase = 0;
+ to_lowcase = 1;
+ }
+
+#if HAVE_TZNAME
+ /* The tzset() call might have changed the value. */
+ if (!(zone && *zone) && tp->tm_isdst >= 0)
+ zone = tzname[tp->tm_isdst];
+#endif
+ if (! zone)
+ zone = "";
+
+#ifdef COMPILE_WIDE
+ {
+ /* The zone string is always given in multibyte form. We have
+ to transform it first. */
+ wchar_t *wczone;
+ size_t len;
+ widen (zone, wczone, len);
+ cpy (len, wczone);
+ }
+#else
+ cpy (strlen (zone), zone);
+#endif
+ break;
+
+ case L_('z'):
+ if (tp->tm_isdst < 0)
+ break;
+
+ {
+ int diff;
+#if HAVE_TM_GMTOFF
+ diff = tp->tm_gmtoff;
+#else
+ if (ut)
+ diff = 0;
+ else
+ {
+ struct tm gtm;
+ struct tm ltm;
+ time_t lt;
+
+ ltm = *tp;
+ lt = mktime (&ltm);
+
+ if (lt == (time_t) -1)
+ {
+ /* mktime returns -1 for errors, but -1 is also a
+ valid time_t value. Check whether an error really
+ occurred. */
+ struct tm tm;
+
+ if (! __localtime_r (&lt, &tm)
+ || ((ltm.tm_sec ^ tm.tm_sec)
+ | (ltm.tm_min ^ tm.tm_min)
+ | (ltm.tm_hour ^ tm.tm_hour)
+ | (ltm.tm_mday ^ tm.tm_mday)
+ | (ltm.tm_mon ^ tm.tm_mon)
+ | (ltm.tm_year ^ tm.tm_year)))
+ break;
+ }
+
+ if (! __gmtime_r (&lt, &gtm))
+ break;
+
+ diff = tm_diff (&ltm, &gtm);
+ }
+#endif
+
+ if (diff < 0)
+ {
+ add (1, *p = L_('-'));
+ diff = -diff;
+ }
+ else
+ add (1, *p = L_('+'));
+
+ diff /= 60;
+ DO_NUMBER (4, (diff / 60) * 100 + diff % 60);
+ }
+
+ case L_('\0'): /* GNU extension: % at end of format. */
+ --f;
+ /* Fall through. */
+ default:
+ /* Unknown format; output the format, including the '%',
+ since this is most likely the right thing to do if a
+ multibyte string has been misparsed. */
+ bad_format:
+ {
+ int flen;
+ for (flen = 1; f[1 - flen] != L_('%'); flen++)
+ continue;
+ cpy (flen, &f[1 - flen]);
+ }
+ break;
+ }
+ }
+
+ if (p && maxsize != 0)
+ *p = L_('\0');
+ return i;
+}
+#ifdef _LIBC
+libc_hidden_def (my_strftime)
+#endif
+
+
+#ifdef emacs
+/* For Emacs we have a separate interface which corresponds to the normal
+ strftime function plus the ut argument, but without the ns argument. */
+size_t
+emacs_strftimeu (char *s, size_t maxsize, const char *format,
+ const struct tm *tp, int ut)
+{
+ return my_strftime (s, maxsize, format, tp, ut, 0);
+}
+#endif
diff --git a/contrib/diff/lib/stripslash.c b/contrib/diff/lib/stripslash.c
new file mode 100644
index 0000000..c6b319e
--- /dev/null
+++ b/contrib/diff/lib/stripslash.c
@@ -0,0 +1,39 @@
+/* stripslash.c -- remove redundant trailing slashes from a file name
+ Copyright (C) 1990, 2001, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "dirname.h"
+
+/* Remove trailing slashes from PATH.
+ Return nonzero if a trailing slash was removed.
+ This is useful when using filename completion from a shell that
+ adds a "/" after directory names (such as tcsh and bash), because
+ the Unix rename and rmdir system calls return an "Invalid argument" error
+ when given a path that ends in "/" (except for the root directory). */
+
+int
+strip_trailing_slashes (char *path)
+{
+ char *base = base_name (path);
+ char *base_lim = base + base_len (base);
+ int had_slash = *base_lim;
+ *base_lim = '\0';
+ return had_slash;
+}
diff --git a/contrib/diff/lib/strncasecmp.c b/contrib/diff/lib/strncasecmp.c
new file mode 100644
index 0000000..68d95aa
--- /dev/null
+++ b/contrib/diff/lib/strncasecmp.c
@@ -0,0 +1,2 @@
+#define LENGTH_LIMIT
+#include "strcasecmp.c"
diff --git a/contrib/diff/lib/strtoimax.c b/contrib/diff/lib/strtoimax.c
new file mode 100644
index 0000000..4ce741c
--- /dev/null
+++ b/contrib/diff/lib/strtoimax.c
@@ -0,0 +1,80 @@
+/* Convert string representation of a number into an intmax_t value.
+ Copyright (C) 1999, 2001, 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#if HAVE_INTTYPES_H
+# include <inttypes.h>
+#elif HAVE_STDINT_H
+# include <stdint.h>
+#endif
+
+#include <stdlib.h>
+
+/* Verify a requirement at compile-time (unlike assert, which is runtime). */
+#define verify(name, assertion) struct name { char a[(assertion) ? 1 : -1]; }
+
+#ifdef UNSIGNED
+# ifndef HAVE_DECL_STRTOULL
+"this configure-time declaration test was not run"
+# endif
+# if !HAVE_DECL_STRTOULL && HAVE_UNSIGNED_LONG_LONG
+unsigned long long strtoull (char const *, char **, int);
+# endif
+
+#else
+
+# ifndef HAVE_DECL_STRTOLL
+"this configure-time declaration test was not run"
+# endif
+# if !HAVE_DECL_STRTOLL && HAVE_UNSIGNED_LONG_LONG
+long long strtoll (char const *, char **, int);
+# endif
+#endif
+
+#ifdef UNSIGNED
+# undef HAVE_LONG_LONG
+# define HAVE_LONG_LONG HAVE_UNSIGNED_LONG_LONG
+# define INT uintmax_t
+# define strtoimax strtoumax
+# define strtol strtoul
+# define strtoll strtoull
+#else
+# define INT intmax_t
+#endif
+
+INT
+strtoimax (char const *ptr, char **endptr, int base)
+{
+#if HAVE_LONG_LONG
+ verify (size_is_that_of_long_or_long_long,
+ (sizeof (INT) == sizeof (long)
+ || sizeof (INT) == sizeof (long long)));
+
+ if (sizeof (INT) != sizeof (long))
+ return strtoll (ptr, endptr, base);
+#else
+ verify (size_is_that_of_long,
+ sizeof (INT) == sizeof (long));
+#endif
+
+ return strtol (ptr, endptr, base);
+}
diff --git a/contrib/diff/lib/strtol.c b/contrib/diff/lib/strtol.c
new file mode 100644
index 0000000..e2b1589
--- /dev/null
+++ b/contrib/diff/lib/strtol.c
@@ -0,0 +1,432 @@
+/* Convert string representation of a number into an integer value.
+
+ Copyright (C) 1991, 1992, 1994, 1995, 1996, 1997, 1998, 1999, 2003
+ Free Software Foundation, Inc.
+
+ NOTE: The canonical source of this file is maintained with the GNU C
+ Library. Bugs can be reported to bug-glibc@gnu.org.
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2, or (at your option) any
+ later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#ifdef _LIBC
+# define USE_NUMBER_GROUPING
+#endif
+
+#include <ctype.h>
+#include <errno.h>
+#ifndef errno
+extern int errno;
+#endif
+#ifndef __set_errno
+# define __set_errno(Val) errno = (Val)
+#endif
+
+#include <limits.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+
+#ifdef USE_NUMBER_GROUPING
+# include "../locale/localeinfo.h"
+#endif
+
+/* Nonzero if we are defining `strtoul' or `strtoull', operating on
+ unsigned integers. */
+#ifndef UNSIGNED
+# define UNSIGNED 0
+# define INT LONG int
+#else
+# define INT unsigned LONG int
+#endif
+
+/* Determine the name. */
+#ifdef USE_IN_EXTENDED_LOCALE_MODEL
+# if UNSIGNED
+# ifdef USE_WIDE_CHAR
+# ifdef QUAD
+# define strtol __wcstoull_l
+# else
+# define strtol __wcstoul_l
+# endif
+# else
+# ifdef QUAD
+# define strtol __strtoull_l
+# else
+# define strtol __strtoul_l
+# endif
+# endif
+# else
+# ifdef USE_WIDE_CHAR
+# ifdef QUAD
+# define strtol __wcstoll_l
+# else
+# define strtol __wcstol_l
+# endif
+# else
+# ifdef QUAD
+# define strtol __strtoll_l
+# else
+# define strtol __strtol_l
+# endif
+# endif
+# endif
+#else
+# if UNSIGNED
+# ifdef USE_WIDE_CHAR
+# ifdef QUAD
+# define strtol wcstoull
+# else
+# define strtol wcstoul
+# endif
+# else
+# ifdef QUAD
+# define strtol strtoull
+# else
+# define strtol strtoul
+# endif
+# endif
+# else
+# ifdef USE_WIDE_CHAR
+# ifdef QUAD
+# define strtol wcstoll
+# else
+# define strtol wcstol
+# endif
+# else
+# ifdef QUAD
+# define strtol strtoll
+# endif
+# endif
+# endif
+#endif
+
+/* If QUAD is defined, we are defining `strtoll' or `strtoull',
+ operating on `long long int's. */
+#ifdef QUAD
+# define LONG long long
+# define STRTOL_LONG_MIN LONG_LONG_MIN
+# define STRTOL_LONG_MAX LONG_LONG_MAX
+# define STRTOL_ULONG_MAX ULONG_LONG_MAX
+
+/* The extra casts work around common compiler bugs,
+ e.g. Cray C 5.0.3.0 when t == time_t. */
+# ifndef TYPE_SIGNED
+# define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
+# endif
+# ifndef TYPE_MINIMUM
+# define TYPE_MINIMUM(t) ((t) (TYPE_SIGNED (t) \
+ ? ~ (t) 0 << (sizeof (t) * CHAR_BIT - 1) \
+ : (t) 0))
+# endif
+# ifndef TYPE_MAXIMUM
+# define TYPE_MAXIMUM(t) ((t) (~ (t) 0 - TYPE_MINIMUM (t)))
+# endif
+
+# ifndef ULONG_LONG_MAX
+# define ULONG_LONG_MAX TYPE_MAXIMUM (unsigned long long)
+# endif
+# ifndef LONG_LONG_MAX
+# define LONG_LONG_MAX TYPE_MAXIMUM (long long int)
+# endif
+# ifndef LONG_LONG_MIN
+# define LONG_LONG_MIN TYPE_MINIMUM (long long int)
+# endif
+
+# if __GNUC__ == 2 && __GNUC_MINOR__ < 7
+ /* Work around gcc bug with using this constant. */
+ static const unsigned long long int maxquad = ULONG_LONG_MAX;
+# undef STRTOL_ULONG_MAX
+# define STRTOL_ULONG_MAX maxquad
+# endif
+#else
+# define LONG long
+# define STRTOL_LONG_MIN LONG_MIN
+# define STRTOL_LONG_MAX LONG_MAX
+# define STRTOL_ULONG_MAX ULONG_MAX
+#endif
+
+
+/* We use this code also for the extended locale handling where the
+ function gets as an additional argument the locale which has to be
+ used. To access the values we have to redefine the _NL_CURRENT
+ macro. */
+#ifdef USE_IN_EXTENDED_LOCALE_MODEL
+# undef _NL_CURRENT
+# define _NL_CURRENT(category, item) \
+ (current->values[_NL_ITEM_INDEX (item)].string)
+# define LOCALE_PARAM , loc
+# define LOCALE_PARAM_PROTO , __locale_t loc
+#else
+# define LOCALE_PARAM
+# define LOCALE_PARAM_PROTO
+#endif
+
+#if defined _LIBC || defined HAVE_WCHAR_H
+# include <wchar.h>
+#endif
+
+#ifdef USE_WIDE_CHAR
+# include <wctype.h>
+# define L_(Ch) L##Ch
+# define UCHAR_TYPE wint_t
+# define STRING_TYPE wchar_t
+# ifdef USE_IN_EXTENDED_LOCALE_MODEL
+# define ISSPACE(Ch) __iswspace_l ((Ch), loc)
+# define ISALPHA(Ch) __iswalpha_l ((Ch), loc)
+# define TOUPPER(Ch) __towupper_l ((Ch), loc)
+# else
+# define ISSPACE(Ch) iswspace (Ch)
+# define ISALPHA(Ch) iswalpha (Ch)
+# define TOUPPER(Ch) towupper (Ch)
+# endif
+#else
+# if defined STDC_HEADERS || (!defined isascii && !defined HAVE_ISASCII)
+# define IN_CTYPE_DOMAIN(c) 1
+# else
+# define IN_CTYPE_DOMAIN(c) isascii(c)
+# endif
+# define L_(Ch) Ch
+# define UCHAR_TYPE unsigned char
+# define STRING_TYPE char
+# ifdef USE_IN_EXTENDED_LOCALE_MODEL
+# define ISSPACE(Ch) __isspace_l ((Ch), loc)
+# define ISALPHA(Ch) __isalpha_l ((Ch), loc)
+# define TOUPPER(Ch) __toupper_l ((Ch), loc)
+# else
+# define ISSPACE(Ch) (IN_CTYPE_DOMAIN (Ch) && isspace (Ch))
+# define ISALPHA(Ch) (IN_CTYPE_DOMAIN (Ch) && isalpha (Ch))
+# define TOUPPER(Ch) (IN_CTYPE_DOMAIN (Ch) ? toupper (Ch) : (Ch))
+# endif
+#endif
+
+#define INTERNAL(X) INTERNAL1(X)
+#define INTERNAL1(X) __##X##_internal
+#define WEAKNAME(X) WEAKNAME1(X)
+
+#ifdef USE_NUMBER_GROUPING
+/* This file defines a function to check for correct grouping. */
+# include "grouping.h"
+#endif
+
+
+
+/* Convert NPTR to an `unsigned long int' or `long int' in base BASE.
+ If BASE is 0 the base is determined by the presence of a leading
+ zero, indicating octal or a leading "0x" or "0X", indicating hexadecimal.
+ If BASE is < 2 or > 36, it is reset to 10.
+ If ENDPTR is not NULL, a pointer to the character after the last
+ one converted is stored in *ENDPTR. */
+
+INT
+INTERNAL (strtol) (const STRING_TYPE *nptr, STRING_TYPE **endptr,
+ int base, int group LOCALE_PARAM_PROTO)
+{
+ int negative;
+ register unsigned LONG int cutoff;
+ register unsigned int cutlim;
+ register unsigned LONG int i;
+ register const STRING_TYPE *s;
+ register UCHAR_TYPE c;
+ const STRING_TYPE *save, *end;
+ int overflow;
+
+#ifdef USE_NUMBER_GROUPING
+# ifdef USE_IN_EXTENDED_LOCALE_MODEL
+ struct locale_data *current = loc->__locales[LC_NUMERIC];
+# endif
+ /* The thousands character of the current locale. */
+ wchar_t thousands = L'\0';
+ /* The numeric grouping specification of the current locale,
+ in the format described in <locale.h>. */
+ const char *grouping;
+
+ if (group)
+ {
+ grouping = _NL_CURRENT (LC_NUMERIC, GROUPING);
+ if (*grouping <= 0 || *grouping == CHAR_MAX)
+ grouping = NULL;
+ else
+ {
+ /* Figure out the thousands separator character. */
+# if defined _LIBC || defined _HAVE_BTOWC
+ thousands = __btowc (*_NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP));
+ if (thousands == WEOF)
+ thousands = L'\0';
+# endif
+ if (thousands == L'\0')
+ grouping = NULL;
+ }
+ }
+ else
+ grouping = NULL;
+#endif
+
+ if (base < 0 || base == 1 || base > 36)
+ {
+ __set_errno (EINVAL);
+ return 0;
+ }
+
+ save = s = nptr;
+
+ /* Skip white space. */
+ while (ISSPACE (*s))
+ ++s;
+ if (*s == L_('\0'))
+ goto noconv;
+
+ /* Check for a sign. */
+ if (*s == L_('-'))
+ {
+ negative = 1;
+ ++s;
+ }
+ else if (*s == L_('+'))
+ {
+ negative = 0;
+ ++s;
+ }
+ else
+ negative = 0;
+
+ /* Recognize number prefix and if BASE is zero, figure it out ourselves. */
+ if (*s == L_('0'))
+ {
+ if ((base == 0 || base == 16) && TOUPPER (s[1]) == L_('X'))
+ {
+ s += 2;
+ base = 16;
+ }
+ else if (base == 0)
+ base = 8;
+ }
+ else if (base == 0)
+ base = 10;
+
+ /* Save the pointer so we can check later if anything happened. */
+ save = s;
+
+#ifdef USE_NUMBER_GROUPING
+ if (group)
+ {
+ /* Find the end of the digit string and check its grouping. */
+ end = s;
+ for (c = *end; c != L_('\0'); c = *++end)
+ if ((wchar_t) c != thousands
+ && ((wchar_t) c < L_('0') || (wchar_t) c > L_('9'))
+ && (!ISALPHA (c) || (int) (TOUPPER (c) - L_('A') + 10) >= base))
+ break;
+ if (*s == thousands)
+ end = s;
+ else
+ end = correctly_grouped_prefix (s, end, thousands, grouping);
+ }
+ else
+#endif
+ end = NULL;
+
+ cutoff = STRTOL_ULONG_MAX / (unsigned LONG int) base;
+ cutlim = STRTOL_ULONG_MAX % (unsigned LONG int) base;
+
+ overflow = 0;
+ i = 0;
+ for (c = *s; c != L_('\0'); c = *++s)
+ {
+ if (s == end)
+ break;
+ if (c >= L_('0') && c <= L_('9'))
+ c -= L_('0');
+ else if (ISALPHA (c))
+ c = TOUPPER (c) - L_('A') + 10;
+ else
+ break;
+ if ((int) c >= base)
+ break;
+ /* Check for overflow. */
+ if (i > cutoff || (i == cutoff && c > cutlim))
+ overflow = 1;
+ else
+ {
+ i *= (unsigned LONG int) base;
+ i += c;
+ }
+ }
+
+ /* Check if anything actually happened. */
+ if (s == save)
+ goto noconv;
+
+ /* Store in ENDPTR the address of one character
+ past the last character we converted. */
+ if (endptr != NULL)
+ *endptr = (STRING_TYPE *) s;
+
+#if !UNSIGNED
+ /* Check for a value that is within the range of
+ `unsigned LONG int', but outside the range of `LONG int'. */
+ if (overflow == 0
+ && i > (negative
+ ? -((unsigned LONG int) (STRTOL_LONG_MIN + 1)) + 1
+ : (unsigned LONG int) STRTOL_LONG_MAX))
+ overflow = 1;
+#endif
+
+ if (overflow)
+ {
+ __set_errno (ERANGE);
+#if UNSIGNED
+ return STRTOL_ULONG_MAX;
+#else
+ return negative ? STRTOL_LONG_MIN : STRTOL_LONG_MAX;
+#endif
+ }
+
+ /* Return the result of the appropriate sign. */
+ return negative ? -i : i;
+
+noconv:
+ /* We must handle a special case here: the base is 0 or 16 and the
+ first two characters are '0' and 'x', but the rest are no
+ hexadecimal digits. This is no error case. We return 0 and
+ ENDPTR points to the `x`. */
+ if (endptr != NULL)
+ {
+ if (save - nptr >= 2 && TOUPPER (save[-1]) == L_('X')
+ && save[-2] == L_('0'))
+ *endptr = (STRING_TYPE *) &save[-1];
+ else
+ /* There was no number to convert. */
+ *endptr = (STRING_TYPE *) nptr;
+ }
+
+ return 0L;
+}
+
+/* External user entry point. */
+
+
+INT
+#ifdef weak_function
+weak_function
+#endif
+strtol (const STRING_TYPE *nptr, STRING_TYPE **endptr,
+ int base LOCALE_PARAM_PROTO)
+{
+ return INTERNAL (strtol) (nptr, endptr, base, 0 LOCALE_PARAM);
+}
diff --git a/contrib/diff/lib/strtoll.c b/contrib/diff/lib/strtoll.c
new file mode 100644
index 0000000..8e9b261
--- /dev/null
+++ b/contrib/diff/lib/strtoll.c
@@ -0,0 +1,33 @@
+/* Function to parse a `long long int' from text.
+ Copyright (C) 1995, 1996, 1997, 1999, 2001 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#define QUAD 1
+
+#include <strtol.c>
+
+#ifdef _LIBC
+# ifdef SHARED
+# include <shlib-compat.h>
+
+# if SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_2)
+compat_symbol (libc, __strtoll_internal, __strtoq_internal, GLIBC_2_0);
+# endif
+
+# endif
+weak_alias (strtoll, strtoq)
+#endif
diff --git a/contrib/diff/lib/strtoul.c b/contrib/diff/lib/strtoul.c
new file mode 100644
index 0000000..bb62f9b
--- /dev/null
+++ b/contrib/diff/lib/strtoul.c
@@ -0,0 +1,20 @@
+/* Copyright (C) 1991, 1997 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#define UNSIGNED 1
+
+#include "strtol.c"
diff --git a/contrib/diff/lib/strtoull.c b/contrib/diff/lib/strtoull.c
new file mode 100644
index 0000000..d6aa1f8
--- /dev/null
+++ b/contrib/diff/lib/strtoull.c
@@ -0,0 +1,27 @@
+/* Function to parse an `unsigned long long int' from text.
+ Copyright (C) 1995, 1996, 1997, 1999 Free Software Foundation, Inc.
+ NOTE: The canonical source of this file is maintained with the GNU C
+ Library. Bugs can be reported to bug-glibc@gnu.org.
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2, or (at your option) any
+ later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#define QUAD 1
+
+#include "strtoul.c"
+
+#ifdef _LIBC
+strong_alias (__strtoull_internal, __strtouq_internal)
+weak_alias (strtoull, strtouq)
+#endif
diff --git a/contrib/diff/lib/strtoumax.c b/contrib/diff/lib/strtoumax.c
new file mode 100644
index 0000000..dc395d6
--- /dev/null
+++ b/contrib/diff/lib/strtoumax.c
@@ -0,0 +1,2 @@
+#define UNSIGNED 1
+#include "strtoimax.c"
diff --git a/contrib/diff/lib/tempname.c b/contrib/diff/lib/tempname.c
new file mode 100644
index 0000000..b3d0874
--- /dev/null
+++ b/contrib/diff/lib/tempname.c
@@ -0,0 +1,344 @@
+/* tempname.c - generate the name of a temporary file.
+
+ Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
+ 2000, 2001, 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <sys/types.h>
+#include <assert.h>
+
+#include <errno.h>
+#ifndef __set_errno
+# define __set_errno(Val) errno = (Val)
+#endif
+
+#include <stdio.h>
+#ifndef P_tmpdir
+# define P_tmpdir "/tmp"
+#endif
+#ifndef TMP_MAX
+# define TMP_MAX 238328
+#endif
+#ifndef __GT_FILE
+# define __GT_FILE 0
+# define __GT_BIGFILE 1
+# define __GT_DIR 2
+# define __GT_NOCREATE 3
+#endif
+
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+
+#if HAVE_FCNTL_H || _LIBC
+# include <fcntl.h>
+#endif
+
+#if HAVE_SYS_TIME_H || _LIBC
+# include <sys/time.h>
+#endif
+
+#if HAVE_STDINT_H || _LIBC
+# include <stdint.h>
+#endif
+#if HAVE_INTTYPES_H
+# include <inttypes.h>
+#endif
+
+#if HAVE_UNISTD_H || _LIBC
+# include <unistd.h>
+#endif
+
+#include <sys/stat.h>
+#if STAT_MACROS_BROKEN
+# undef S_ISDIR
+#endif
+#if !defined S_ISDIR && defined S_IFDIR
+# define S_ISDIR(mode) (((mode) & S_IFMT) == S_IFDIR)
+#endif
+#if !S_IRUSR && S_IREAD
+# define S_IRUSR S_IREAD
+#endif
+#if !S_IRUSR
+# define S_IRUSR 00400
+#endif
+#if !S_IWUSR && S_IWRITE
+# define S_IWUSR S_IWRITE
+#endif
+#if !S_IWUSR
+# define S_IWUSR 00200
+#endif
+#if !S_IXUSR && S_IEXEC
+# define S_IXUSR S_IEXEC
+#endif
+#if !S_IXUSR
+# define S_IXUSR 00100
+#endif
+
+#if _LIBC
+# define struct_stat64 struct stat64
+#else
+# define struct_stat64 struct stat
+# define __getpid getpid
+# define __gettimeofday gettimeofday
+# define __mkdir mkdir
+# define __open open
+# define __open64 open
+# define __lxstat64(version, path, buf) lstat (path, buf)
+# define __xstat64(version, path, buf) stat (path, buf)
+#endif
+
+#if ! (HAVE___SECURE_GETENV || _LIBC)
+# define __secure_getenv getenv
+#endif
+
+#ifdef _LIBC
+# include <hp-timing.h>
+# if HP_TIMING_AVAIL
+# define RANDOM_BITS(Var) \
+ if (__builtin_expect (value == UINT64_C (0), 0)) \
+ { \
+ /* If this is the first time this function is used initialize \
+ the variable we accumulate the value in to some somewhat \
+ random value. If we'd not do this programs at startup time \
+ might have a reduced set of possible names, at least on slow \
+ machines. */ \
+ struct timeval tv; \
+ __gettimeofday (&tv, NULL); \
+ value = ((uint64_t) tv.tv_usec << 16) ^ tv.tv_sec; \
+ } \
+ HP_TIMING_NOW (Var)
+# endif
+#endif
+
+/* Use the widest available unsigned type if uint64_t is not
+ available. The algorithm below extracts a number less than 62**6
+ (approximately 2**35.725) from uint64_t, so ancient hosts where
+ uintmax_t is only 32 bits lose about 3.725 bits of randomness,
+ which is better than not having mkstemp at all. */
+#if !defined UINT64_MAX && !defined uint64_t
+# define uint64_t uintmax_t
+#endif
+
+/* Return nonzero if DIR is an existent directory. */
+static int
+direxists (const char *dir)
+{
+ struct_stat64 buf;
+ return __xstat64 (_STAT_VER, dir, &buf) == 0 && S_ISDIR (buf.st_mode);
+}
+
+/* Path search algorithm, for tmpnam, tmpfile, etc. If DIR is
+ non-null and exists, uses it; otherwise uses the first of $TMPDIR,
+ P_tmpdir, /tmp that exists. Copies into TMPL a template suitable
+ for use with mk[s]temp. Will fail (-1) if DIR is non-null and
+ doesn't exist, none of the searched dirs exists, or there's not
+ enough space in TMPL. */
+int
+__path_search (char *tmpl, size_t tmpl_len, const char *dir, const char *pfx,
+ int try_tmpdir)
+{
+ const char *d;
+ size_t dlen, plen;
+
+ if (!pfx || !pfx[0])
+ {
+ pfx = "file";
+ plen = 4;
+ }
+ else
+ {
+ plen = strlen (pfx);
+ if (plen > 5)
+ plen = 5;
+ }
+
+ if (try_tmpdir)
+ {
+ d = __secure_getenv ("TMPDIR");
+ if (d != NULL && direxists (d))
+ dir = d;
+ else if (dir != NULL && direxists (dir))
+ /* nothing */ ;
+ else
+ dir = NULL;
+ }
+ if (dir == NULL)
+ {
+ if (direxists (P_tmpdir))
+ dir = P_tmpdir;
+ else if (strcmp (P_tmpdir, "/tmp") != 0 && direxists ("/tmp"))
+ dir = "/tmp";
+ else
+ {
+ __set_errno (ENOENT);
+ return -1;
+ }
+ }
+
+ dlen = strlen (dir);
+ while (dlen > 1 && dir[dlen - 1] == '/')
+ dlen--; /* remove trailing slashes */
+
+ /* check we have room for "${dir}/${pfx}XXXXXX\0" */
+ if (tmpl_len < dlen + 1 + plen + 6 + 1)
+ {
+ __set_errno (EINVAL);
+ return -1;
+ }
+
+ sprintf (tmpl, "%.*s/%.*sXXXXXX", (int) dlen, dir, (int) plen, pfx);
+ return 0;
+}
+
+/* These are the characters used in temporary filenames. */
+static const char letters[] =
+"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789";
+
+/* Generate a temporary file name based on TMPL. TMPL must match the
+ rules for mk[s]temp (i.e. end in "XXXXXX"). The name constructed
+ does not exist at the time of the call to __gen_tempname. TMPL is
+ overwritten with the result.
+
+ KIND may be one of:
+ __GT_NOCREATE: simply verify that the name does not exist
+ at the time of the call.
+ __GT_FILE: create the file using open(O_CREAT|O_EXCL)
+ and return a read-write fd. The file is mode 0600.
+ __GT_BIGFILE: same as __GT_FILE but use open64().
+ __GT_DIR: create a directory, which will be mode 0700.
+
+ We use a clever algorithm to get hard-to-predict names. */
+int
+__gen_tempname (char *tmpl, int kind)
+{
+ int len;
+ char *XXXXXX;
+ static uint64_t value;
+ uint64_t random_time_bits;
+ unsigned int count;
+ int fd = -1;
+ int save_errno = errno;
+ struct_stat64 st;
+
+ /* A lower bound on the number of temporary files to attempt to
+ generate. The maximum total number of temporary file names that
+ can exist for a given template is 62**6. It should never be
+ necessary to try all these combinations. Instead if a reasonable
+ number of names is tried (we define reasonable as 62**3) fail to
+ give the system administrator the chance to remove the problems. */
+ unsigned int attempts_min = 62 * 62 * 62;
+
+ /* The number of times to attempt to generate a temporary file. To
+ conform to POSIX, this must be no smaller than TMP_MAX. */
+ unsigned int attempts = attempts_min < TMP_MAX ? TMP_MAX : attempts_min;
+
+ len = strlen (tmpl);
+ if (len < 6 || strcmp (&tmpl[len - 6], "XXXXXX"))
+ {
+ __set_errno (EINVAL);
+ return -1;
+ }
+
+ /* This is where the Xs start. */
+ XXXXXX = &tmpl[len - 6];
+
+ /* Get some more or less random data. */
+#ifdef RANDOM_BITS
+ RANDOM_BITS (random_time_bits);
+#else
+# if HAVE_GETTIMEOFDAY || _LIBC
+ {
+ struct timeval tv;
+ __gettimeofday (&tv, NULL);
+ random_time_bits = ((uint64_t) tv.tv_usec << 16) ^ tv.tv_sec;
+ }
+# else
+ random_time_bits = time (NULL);
+# endif
+#endif
+ value += random_time_bits ^ __getpid ();
+
+ for (count = 0; count < attempts; value += 7777, ++count)
+ {
+ uint64_t v = value;
+
+ /* Fill in the random bits. */
+ XXXXXX[0] = letters[v % 62];
+ v /= 62;
+ XXXXXX[1] = letters[v % 62];
+ v /= 62;
+ XXXXXX[2] = letters[v % 62];
+ v /= 62;
+ XXXXXX[3] = letters[v % 62];
+ v /= 62;
+ XXXXXX[4] = letters[v % 62];
+ v /= 62;
+ XXXXXX[5] = letters[v % 62];
+
+ switch (kind)
+ {
+ case __GT_FILE:
+ fd = __open (tmpl, O_RDWR | O_CREAT | O_EXCL, S_IRUSR | S_IWUSR);
+ break;
+
+ case __GT_BIGFILE:
+ fd = __open64 (tmpl, O_RDWR | O_CREAT | O_EXCL, S_IRUSR | S_IWUSR);
+ break;
+
+ case __GT_DIR:
+ fd = __mkdir (tmpl, S_IRUSR | S_IWUSR | S_IXUSR);
+ break;
+
+ case __GT_NOCREATE:
+ /* This case is backward from the other three. __gen_tempname
+ succeeds if __xstat fails because the name does not exist.
+ Note the continue to bypass the common logic at the bottom
+ of the loop. */
+ if (__lxstat64 (_STAT_VER, tmpl, &st) < 0)
+ {
+ if (errno == ENOENT)
+ {
+ __set_errno (save_errno);
+ return 0;
+ }
+ else
+ /* Give up now. */
+ return -1;
+ }
+ continue;
+
+ default:
+ assert (! "invalid KIND in __gen_tempname");
+ }
+
+ if (fd >= 0)
+ {
+ __set_errno (save_errno);
+ return fd;
+ }
+ else if (errno != EEXIST)
+ return -1;
+ }
+
+ /* We got out of the loop because we ran out of combinations to try. */
+ __set_errno (EEXIST);
+ return -1;
+}
diff --git a/contrib/diff/lib/time_r.c b/contrib/diff/lib/time_r.c
new file mode 100644
index 0000000..6c42504
--- /dev/null
+++ b/contrib/diff/lib/time_r.c
@@ -0,0 +1,69 @@
+/* Reentrant time functions like localtime_r.
+
+ Copyright (C) 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "time_r.h"
+
+#include <string.h>
+
+static char *
+copy_string_result (char *dest, char const *src)
+{
+ if (! src)
+ return 0;
+ return strcpy (dest, src);
+}
+
+static struct tm *
+copy_tm_result (struct tm *dest, struct tm const *src)
+{
+ if (! src)
+ return 0;
+ *dest = *src;
+ return dest;
+}
+
+
+char *
+asctime_r (struct tm const * restrict tm, char * restrict buf)
+{
+ return copy_string_result (buf, asctime (tm));
+}
+
+char *
+ctime_r (time_t const *t, char *buf)
+{
+ return copy_string_result (buf, ctime (t));
+}
+
+struct tm *
+gmtime_r (time_t const * restrict t, struct tm * restrict tp)
+{
+ return copy_tm_result (tp, gmtime (t));
+}
+
+struct tm *
+localtime_r (time_t const * restrict t, struct tm * restrict tp)
+{
+ return copy_tm_result (tp, localtime (t));
+}
diff --git a/contrib/diff/lib/time_r.h b/contrib/diff/lib/time_r.h
new file mode 100644
index 0000000..fbd3375
--- /dev/null
+++ b/contrib/diff/lib/time_r.h
@@ -0,0 +1,46 @@
+/* Reentrant time functions like localtime_r.
+
+ Copyright (C) 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+#ifndef _TIME_R_H
+#define _TIME_R_H
+
+/* Include <time.h> first, since it may declare these functions with
+ signatures that disagree with POSIX, and we don't want to rename
+ those declarations. */
+#include <time.h>
+
+#if !HAVE_TIME_R_POSIX
+# undef asctime_r
+# undef ctime_r
+# undef gmtime_r
+# undef localtime_r
+
+# define asctime_r rpl_asctime_r
+# define ctime_r rpl_ctime_r
+# define gmtime_r rpl_gmtime_r
+# define localtime_r rpl_localtime_r
+
+char *asctime_r (struct tm const * restrict, char * restrict);
+char *ctime_r (time_t const *, char *);
+struct tm *gmtime_r (time_t const * restrict, struct tm * restrict);
+struct tm *localtime_r (time_t const * restrict, struct tm * restrict);
+#endif
+
+#endif
diff --git a/contrib/diff/lib/umaxtostr.c b/contrib/diff/lib/umaxtostr.c
new file mode 100644
index 0000000..4f49a7f
--- /dev/null
+++ b/contrib/diff/lib/umaxtostr.c
@@ -0,0 +1,3 @@
+#define inttostr umaxtostr
+#define inttype uintmax_t
+#include "inttostr.c"
diff --git a/contrib/diff/lib/unlocked-io.h b/contrib/diff/lib/unlocked-io.h
new file mode 100644
index 0000000..36a7a48
--- /dev/null
+++ b/contrib/diff/lib/unlocked-io.h
@@ -0,0 +1,132 @@
+/* Prefer faster, non-thread-safe stdio functions if available.
+
+ Copyright (C) 2001, 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Jim Meyering. */
+
+#ifndef UNLOCKED_IO_H
+# define UNLOCKED_IO_H 1
+
+# ifndef USE_UNLOCKED_IO
+# define USE_UNLOCKED_IO 1
+# endif
+
+# if USE_UNLOCKED_IO
+
+/* These are wrappers for functions/macros from the GNU C library, and
+ from other C libraries supporting POSIX's optional thread-safe functions.
+
+ The standard I/O functions are thread-safe. These *_unlocked ones are
+ more efficient but not thread-safe. That they're not thread-safe is
+ fine since all of the applications in this package are single threaded.
+
+ Also, some code that is shared with the GNU C library may invoke
+ the *_unlocked functions directly. On hosts that lack those
+ functions, invoke the non-thread-safe versions instead. */
+
+# include <stdio.h>
+
+# if HAVE_DECL_CLEARERR_UNLOCKED
+# undef clearerr
+# define clearerr(x) clearerr_unlocked (x)
+# else
+# define clearerr_unlocked(x) clearerr (x)
+# endif
+# if HAVE_DECL_FEOF_UNLOCKED
+# undef feof
+# define feof(x) feof_unlocked (x)
+# else
+# define feof_unlocked(x) feof (x)
+# endif
+# if HAVE_DECL_FERROR_UNLOCKED
+# undef ferror
+# define ferror(x) ferror_unlocked (x)
+# else
+# define ferror_unlocked(x) ferror (x)
+# endif
+# if HAVE_DECL_FFLUSH_UNLOCKED
+# undef fflush
+# define fflush(x) fflush_unlocked (x)
+# else
+# define fflush_unlocked(x) fflush (x)
+# endif
+# if HAVE_DECL_FGETS_UNLOCKED
+# undef fgets
+# define fgets(x,y,z) fgets_unlocked (x,y,z)
+# else
+# define fgets_unlocked(x,y,z) fgets (x,y,z)
+# endif
+# if HAVE_DECL_FPUTC_UNLOCKED
+# undef fputc
+# define fputc(x,y) fputc_unlocked (x,y)
+# else
+# define fputc_unlocked(x,y) fputc (x,y)
+# endif
+# if HAVE_DECL_FPUTS_UNLOCKED
+# undef fputs
+# define fputs(x,y) fputs_unlocked (x,y)
+# else
+# define fputs_unlocked(x,y) fputs (x,y)
+# endif
+# if HAVE_DECL_FREAD_UNLOCKED
+# undef fread
+# define fread(w,x,y,z) fread_unlocked (w,x,y,z)
+# else
+# define fread_unlocked(w,x,y,z) fread (w,x,y,z)
+# endif
+# if HAVE_DECL_FWRITE_UNLOCKED
+# undef fwrite
+# define fwrite(w,x,y,z) fwrite_unlocked (w,x,y,z)
+# else
+# define fwrite_unlocked(w,x,y,z) fwrite (w,x,y,z)
+# endif
+# if HAVE_DECL_GETC_UNLOCKED
+# undef getc
+# define getc(x) getc_unlocked (x)
+# else
+# define getc_unlocked(x) getc (x)
+# endif
+# if HAVE_DECL_GETCHAR_UNLOCKED
+# undef getchar
+# define getchar() getchar_unlocked ()
+# else
+# define getchar_unlocked() getchar ()
+# endif
+# if HAVE_DECL_PUTC_UNLOCKED
+# undef putc
+# define putc(x,y) putc_unlocked (x,y)
+# else
+# define putc_unlocked(x,y) putc (x,y)
+# endif
+# if HAVE_DECL_PUTCHAR_UNLOCKED
+# undef putchar
+# define putchar(x) putchar_unlocked (x)
+# else
+# define putchar_unlocked(x) putchar (x)
+# endif
+
+# undef flockfile
+# define flockfile(x) ((void) 0)
+
+# undef ftrylockfile
+# define ftrylockfile(x) 0
+
+# undef funlockfile
+# define funlockfile(x) ((void) 0)
+
+# endif /* USE_UNLOCKED_IO */
+#endif /* UNLOCKED_IO_H */
diff --git a/contrib/diff/lib/version-etc.c b/contrib/diff/lib/version-etc.c
new file mode 100644
index 0000000..7523b08
--- /dev/null
+++ b/contrib/diff/lib/version-etc.c
@@ -0,0 +1,176 @@
+/* Utility to help print --version output in a consistent format.
+ Copyright (C) 1999-2004 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Jim Meyering. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+/* Specification. */
+#include "version-etc.h"
+
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include "unlocked-io.h"
+
+#include "gettext.h"
+#define _(msgid) gettext (msgid)
+
+/* Default copyright goes to the FSF. */
+
+const char* version_etc_copyright =
+ /* Do *not* mark this string for translation. */
+ "Copyright (C) 2004 Free Software Foundation, Inc.";
+
+
+/* Like version_etc, below, but with the NULL-terminated author list
+ provided via a variable of type va_list. */
+void
+version_etc_va (FILE *stream,
+ const char *command_name, const char *package,
+ const char *version, va_list authors)
+{
+ unsigned int n_authors;
+
+ /* Count the number of authors. */
+ {
+ va_list tmp_authors;
+
+#ifdef __va_copy
+ __va_copy (tmp_authors, authors);
+#else
+ tmp_authors = authors;
+#endif
+
+ n_authors = 0;
+ while (va_arg (tmp_authors, const char *) != NULL)
+ ++n_authors;
+ }
+
+ if (command_name)
+ fprintf (stream, "%s (%s) %s\n", command_name, package, version);
+ else
+ fprintf (stream, "%s %s\n", package, version);
+
+ switch (n_authors)
+ {
+ case 0:
+ /* The caller must provide at least one author name. */
+ abort ();
+ case 1:
+ /* TRANSLATORS: %s denotes an author name. */
+ vfprintf (stream, _("Written by %s.\n"), authors);
+ break;
+ case 2:
+ /* TRANSLATORS: Each %s denotes an author name. */
+ vfprintf (stream, _("Written by %s and %s.\n"), authors);
+ break;
+ case 3:
+ /* TRANSLATORS: Each %s denotes an author name. */
+ vfprintf (stream, _("Written by %s, %s, and %s.\n"), authors);
+ break;
+ case 4:
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("Written by %s, %s, %s,\nand %s.\n"), authors);
+ break;
+ case 5:
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("Written by %s, %s, %s,\n%s, and %s.\n"), authors);
+ break;
+ case 6:
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("Written by %s, %s, %s,\n%s, %s, and %s.\n"),
+ authors);
+ break;
+ case 7:
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("Written by %s, %s, %s,\n%s, %s, %s, and %s.\n"),
+ authors);
+ break;
+ case 8:
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("\
+Written by %s, %s, %s,\n%s, %s, %s, %s,\nand %s.\n"),
+ authors);
+ break;
+ case 9:
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("\
+Written by %s, %s, %s,\n%s, %s, %s, %s,\n%s, and %s.\n"),
+ authors);
+ break;
+ default:
+ /* 10 or more authors. Use an abbreviation, since the human reader
+ will probably not want to read the entire list anyway. */
+ /* TRANSLATORS: Each %s denotes an author name.
+ You can use line breaks, estimating that each author name occupies
+ ca. 16 screen columns and that a screen line has ca. 80 columns. */
+ vfprintf (stream, _("\
+Written by %s, %s, %s,\n%s, %s, %s, %s,\n%s, %s, and others.\n"),
+ authors);
+ break;
+ }
+ va_end (authors);
+ putc ('\n', stream);
+
+ fputs (version_etc_copyright, stream);
+ putc ('\n', stream);
+
+ fputs (_("\
+This is free software; see the source for copying conditions. There is NO\n\
+warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n"),
+ stream);
+}
+
+
+/* Display the --version information the standard way.
+
+ If COMMAND_NAME is NULL, the PACKAGE is asumed to be the name of
+ the program. The formats are therefore:
+
+ PACKAGE VERSION
+
+ or
+
+ COMMAND_NAME (PACKAGE) VERSION.
+
+ The author names are passed as separate arguments, with an additional
+ NULL argument at the end. */
+void
+version_etc (FILE *stream,
+ const char *command_name, const char *package,
+ const char *version, /* const char *author1, ...*/ ...)
+{
+ va_list authors;
+
+ va_start (authors, version);
+ version_etc_va (stream, command_name, package, version, authors);
+}
diff --git a/contrib/diff/lib/version-etc.h b/contrib/diff/lib/version-etc.h
new file mode 100644
index 0000000..d505e75
--- /dev/null
+++ b/contrib/diff/lib/version-etc.h
@@ -0,0 +1,37 @@
+/* Utility to help print --version output in a consistent format.
+ Copyright (C) 1999, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Jim Meyering. */
+
+#ifndef VERSION_ETC_H
+# define VERSION_ETC_H 1
+
+# include <stdarg.h>
+# include <stdio.h>
+
+extern const char *version_etc_copyright;
+
+extern void version_etc_va (FILE *stream,
+ const char *command_name, const char *package,
+ const char *version, va_list authors);
+
+extern void version_etc (FILE *stream,
+ const char *command_name, const char *package,
+ const char *version,
+ /* const char *author1, ...*/ ...);
+
+#endif /* VERSION_ETC_H */
diff --git a/contrib/diff/lib/waitpid.c b/contrib/diff/lib/waitpid.c
new file mode 100644
index 0000000..bc07898
--- /dev/null
+++ b/contrib/diff/lib/waitpid.c
@@ -0,0 +1,72 @@
+/* Emulate waitpid on systems that just have wait.
+ Copyright (C) 1994, 1995, 1998, 1999 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING.
+ If not, write to the Free Software Foundation,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include <errno.h>
+#ifndef errno
+extern int errno;
+#endif
+
+#define WAITPID_CHILDREN 8
+static pid_t waited_pid[WAITPID_CHILDREN];
+static int waited_status[WAITPID_CHILDREN];
+
+pid_t
+waitpid (pid_t pid, int *stat_loc, int options)
+{
+ int i;
+ pid_t p;
+
+ if (!options && (pid == -1 || 0 < pid))
+ {
+ /* If we have already waited for this child, return it immediately. */
+ for (i = 0; i < WAITPID_CHILDREN; i++)
+ {
+ p = waited_pid[i];
+ if (p && (p == pid || pid == -1))
+ {
+ waited_pid[i] = 0;
+ goto success;
+ }
+ }
+
+ /* The child has not returned yet; wait for it, accumulating status. */
+ for (i = 0; i < WAITPID_CHILDREN; i++)
+ if (! waited_pid[i])
+ {
+ p = wait (&waited_status[i]);
+ if (p < 0)
+ return p;
+ if (p == pid || pid == -1)
+ goto success;
+ waited_pid[i] = p;
+ }
+ }
+
+ /* We cannot emulate this wait call, e.g. because of too many children. */
+ errno = EINVAL;
+ return -1;
+
+success:
+ if (stat_loc)
+ *stat_loc = waited_status[i];
+ return p;
+}
diff --git a/contrib/diff/lib/xalloc.h b/contrib/diff/lib/xalloc.h
new file mode 100644
index 0000000..4b65858
--- /dev/null
+++ b/contrib/diff/lib/xalloc.h
@@ -0,0 +1,87 @@
+/* xalloc.h -- malloc with out-of-memory checking
+
+ Copyright (C) 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
+ 1999, 2000, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef XALLOC_H_
+# define XALLOC_H_
+
+# include <stddef.h>
+
+# ifndef __attribute__
+# if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 8) || __STRICT_ANSI__
+# define __attribute__(x)
+# endif
+# endif
+
+# ifndef ATTRIBUTE_NORETURN
+# define ATTRIBUTE_NORETURN __attribute__ ((__noreturn__))
+# endif
+
+/* If this pointer is non-zero, run the specified function upon each
+ allocation failure. It is initialized to zero. */
+extern void (*xalloc_fail_func) (void);
+
+/* If XALLOC_FAIL_FUNC is undefined or a function that returns, this
+ message is output. It is translated via gettext.
+ Its value is "memory exhausted". */
+extern char const xalloc_msg_memory_exhausted[];
+
+/* This function is always triggered when memory is exhausted. It is
+ in charge of honoring the two previous items. It exits with status
+ exit_failure (defined in exitfail.h). This is the
+ function to call when one wants the program to die because of a
+ memory allocation failure. */
+extern void xalloc_die (void) ATTRIBUTE_NORETURN;
+
+void *xmalloc (size_t s);
+void *xnmalloc (size_t n, size_t s);
+void *xzalloc (size_t s);
+void *xcalloc (size_t n, size_t s);
+void *xrealloc (void *p, size_t s);
+void *xnrealloc (void *p, size_t n, size_t s);
+void *x2realloc (void *p, size_t *pn);
+void *x2nrealloc (void *p, size_t *pn, size_t s);
+void *xclone (void const *p, size_t s);
+char *xstrdup (const char *str);
+
+/* Return 1 if an array of N objects, each of size S, cannot exist due
+ to size arithmetic overflow. S must be positive and N must be
+ nonnegative. This is a macro, not an inline function, so that it
+ works correctly even when SIZE_MAX < N.
+
+ By gnulib convention, SIZE_MAX represents overflow in size
+ calculations, so the conservative dividend to use here is
+ SIZE_MAX - 1, since SIZE_MAX might represent an overflowed value.
+ However, malloc (SIZE_MAX) fails on all known hosts where
+ sizeof (ptrdiff_t) <= sizeof (size_t), so do not bother to test for
+ exactly-SIZE_MAX allocations on such hosts; this avoids a test and
+ branch when S is known to be 1. */
+# define xalloc_oversized(n, s) \
+ ((size_t) (sizeof (ptrdiff_t) <= sizeof (size_t) ? -1 : -2) / (s) < (n))
+
+/* These macros are deprecated; they will go away soon, and are retained
+ temporarily only to ease conversion to the functions described above. */
+# define CCLONE(p, n) xclone (p, (n) * sizeof *(p))
+# define CLONE(p) xclone (p, sizeof *(p))
+# define NEW(type, var) type *var = xmalloc (sizeof (type))
+# define XCALLOC(type, n) xcalloc (n, sizeof (type))
+# define XMALLOC(type, n) xnmalloc (n, sizeof (type))
+# define XREALLOC(p, type, n) xnrealloc (p, n, sizeof (type))
+# define XFREE(p) free (p)
+
+#endif /* !XALLOC_H_ */
diff --git a/contrib/diff/lib/xmalloc.c b/contrib/diff/lib/xmalloc.c
new file mode 100644
index 0000000..181006b
--- /dev/null
+++ b/contrib/diff/lib/xmalloc.c
@@ -0,0 +1,255 @@
+/* xmalloc.c -- malloc with out of memory checking
+
+ Copyright (C) 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 2003,
+ 1999, 2000, 2002, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#include "xalloc.h"
+
+#include <stdlib.h>
+#include <string.h>
+
+#include "gettext.h"
+#define _(msgid) gettext (msgid)
+#define N_(msgid) msgid
+
+#include "error.h"
+#include "exitfail.h"
+
+#ifndef SIZE_MAX
+# define SIZE_MAX ((size_t) -1)
+#endif
+
+#ifndef HAVE_MALLOC
+"you must run the autoconf test for a GNU libc compatible malloc"
+#endif
+
+#ifndef HAVE_REALLOC
+"you must run the autoconf test for a GNU libc compatible realloc"
+#endif
+
+/* If non NULL, call this function when memory is exhausted. */
+void (*xalloc_fail_func) (void) = 0;
+
+/* If XALLOC_FAIL_FUNC is NULL, or does return, display this message
+ before exiting when memory is exhausted. Goes through gettext. */
+char const xalloc_msg_memory_exhausted[] = N_("memory exhausted");
+
+void
+xalloc_die (void)
+{
+ if (xalloc_fail_func)
+ (*xalloc_fail_func) ();
+ error (exit_failure, 0, "%s", _(xalloc_msg_memory_exhausted));
+ /* The `noreturn' cannot be given to error, since it may return if
+ its first argument is 0. To help compilers understand the
+ xalloc_die does terminate, call abort. */
+ abort ();
+}
+
+/* Allocate an array of N objects, each with S bytes of memory,
+ dynamically, with error checking. S must be nonzero. */
+
+static inline void *
+xnmalloc_inline (size_t n, size_t s)
+{
+ void *p;
+ if (xalloc_oversized (n, s) || ! (p = malloc (n * s)))
+ xalloc_die ();
+ return p;
+}
+
+void *
+xnmalloc (size_t n, size_t s)
+{
+ return xnmalloc_inline (n, s);
+}
+
+/* Allocate N bytes of memory dynamically, with error checking. */
+
+void *
+xmalloc (size_t n)
+{
+ return xnmalloc_inline (n, 1);
+}
+
+/* Change the size of an allocated block of memory P to an array of N
+ objects each of S bytes, with error checking. S must be nonzero. */
+
+static inline void *
+xnrealloc_inline (void *p, size_t n, size_t s)
+{
+ if (xalloc_oversized (n, s) || ! (p = realloc (p, n * s)))
+ xalloc_die ();
+ return p;
+}
+
+void *
+xnrealloc (void *p, size_t n, size_t s)
+{
+ return xnrealloc_inline (p, n, s);
+}
+
+/* Change the size of an allocated block of memory P to N bytes,
+ with error checking. */
+
+void *
+xrealloc (void *p, size_t n)
+{
+ return xnrealloc_inline (p, n, 1);
+}
+
+
+/* If P is null, allocate a block of at least *PN such objects;
+ otherwise, reallocate P so that it contains more than *PN objects
+ each of S bytes. *PN must be nonzero unless P is null, and S must
+ be nonzero. Set *PN to the new number of objects, and return the
+ pointer to the new block. *PN is never set to zero, and the
+ returned pointer is never null.
+
+ Repeated reallocations are guaranteed to make progress, either by
+ allocating an initial block with a nonzero size, or by allocating a
+ larger block.
+
+ In the following implementation, nonzero sizes are doubled so that
+ repeated reallocations have O(N log N) overall cost rather than
+ O(N**2) cost, but the specification for this function does not
+ guarantee that sizes are doubled.
+
+ Here is an example of use:
+
+ int *p = NULL;
+ size_t used = 0;
+ size_t allocated = 0;
+
+ void
+ append_int (int value)
+ {
+ if (used == allocated)
+ p = x2nrealloc (p, &allocated, sizeof *p);
+ p[used++] = value;
+ }
+
+ This causes x2nrealloc to allocate a block of some nonzero size the
+ first time it is called.
+
+ To have finer-grained control over the initial size, set *PN to a
+ nonzero value before calling this function with P == NULL. For
+ example:
+
+ int *p = NULL;
+ size_t used = 0;
+ size_t allocated = 0;
+ size_t allocated1 = 1000;
+
+ void
+ append_int (int value)
+ {
+ if (used == allocated)
+ {
+ p = x2nrealloc (p, &allocated1, sizeof *p);
+ allocated = allocated1;
+ }
+ p[used++] = value;
+ }
+
+ */
+
+static inline void *
+x2nrealloc_inline (void *p, size_t *pn, size_t s)
+{
+ size_t n = *pn;
+
+ if (! p)
+ {
+ if (! n)
+ {
+ /* The approximate size to use for initial small allocation
+ requests, when the invoking code specifies an old size of
+ zero. 64 bytes is the largest "small" request for the
+ GNU C library malloc. */
+ enum { DEFAULT_MXFAST = 64 };
+
+ n = DEFAULT_MXFAST / s;
+ n += !n;
+ }
+ }
+ else
+ {
+ if (SIZE_MAX / 2 / s < n)
+ xalloc_die ();
+ n *= 2;
+ }
+
+ *pn = n;
+ return xrealloc (p, n * s);
+}
+
+void *
+x2nrealloc (void *p, size_t *pn, size_t s)
+{
+ return x2nrealloc_inline (p, pn, s);
+}
+
+/* If P is null, allocate a block of at least *PN bytes; otherwise,
+ reallocate P so that it contains more than *PN bytes. *PN must be
+ nonzero unless P is null. Set *PN to the new block's size, and
+ return the pointer to the new block. *PN is never set to zero, and
+ the returned pointer is never null. */
+
+void *
+x2realloc (void *p, size_t *pn)
+{
+ return x2nrealloc_inline (p, pn, 1);
+}
+
+/* Allocate S bytes of zeroed memory dynamically, with error checking.
+ There's no need for xnzalloc (N, S), since it would be equivalent
+ to xcalloc (N, S). */
+
+void *
+xzalloc (size_t s)
+{
+ return memset (xmalloc (s), 0, s);
+}
+
+/* Allocate zeroed memory for N elements of S bytes, with error
+ checking. S must be nonzero. */
+
+void *
+xcalloc (size_t n, size_t s)
+{
+ void *p;
+ /* Test for overflow, since some calloc implementations don't have
+ proper overflow checks. */
+ if (xalloc_oversized (n, s) || ! (p = calloc (n, s)))
+ xalloc_die ();
+ return p;
+}
+
+/* Clone an object P of size S, with error checking. There's no need
+ for xnclone (P, N, S), since xclone (P, N * S) works without any
+ need for an arithmetic overflow check. */
+
+void *
+xclone (void const *p, size_t s)
+{
+ return memcpy (xmalloc (s), p, s);
+}
diff --git a/contrib/diff/lib/xstrdup.c b/contrib/diff/lib/xstrdup.c
new file mode 100644
index 0000000..58f18be
--- /dev/null
+++ b/contrib/diff/lib/xstrdup.c
@@ -0,0 +1,33 @@
+/* xstrdup.c -- copy a string with out of memory checking
+ Copyright (C) 1990, 1996, 1998, 2001, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+/* Specification. */
+#include "xalloc.h"
+
+#include <string.h>
+
+/* Return a newly allocated copy of STRING. */
+
+char *
+xstrdup (const char *string)
+{
+ return xclone (string, strlen (string) + 1);
+}
diff --git a/contrib/diff/lib/xstrtol.c b/contrib/diff/lib/xstrtol.c
new file mode 100644
index 0000000..d0aa0a9
--- /dev/null
+++ b/contrib/diff/lib/xstrtol.c
@@ -0,0 +1,295 @@
+/* A more useful interface to strtol.
+
+ Copyright (C) 1995, 1996, 1998, 1999, 2000, 2001, 2003 Free
+ Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Jim Meyering. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#ifndef __strtol
+# define __strtol strtol
+# define __strtol_t long int
+# define __xstrtol xstrtol
+# define STRTOL_T_MINIMUM LONG_MIN
+# define STRTOL_T_MAXIMUM LONG_MAX
+#endif
+
+/* Some pre-ANSI implementations (e.g. SunOS 4)
+ need stderr defined if assertion checking is enabled. */
+#include <stdio.h>
+
+#include <assert.h>
+#include <ctype.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include <errno.h>
+#ifndef errno
+extern int errno;
+#endif
+
+#include <limits.h>
+
+/* The extra casts work around common compiler bugs. */
+#define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
+#define TYPE_MINIMUM(t) ((t) (TYPE_SIGNED (t) \
+ ? ~ (t) 0 << (sizeof (t) * CHAR_BIT - 1) \
+ : (t) 0))
+#define TYPE_MAXIMUM(t) ((t) (~ (t) 0 - TYPE_MINIMUM (t)))
+
+#ifndef STRTOL_T_MINIMUM
+# define STRTOL_T_MINIMUM TYPE_MINIMUM (__strtol_t)
+# define STRTOL_T_MAXIMUM TYPE_MAXIMUM (__strtol_t)
+#endif
+
+#if defined (STDC_HEADERS) || (!defined (isascii) && !defined (HAVE_ISASCII))
+# define IN_CTYPE_DOMAIN(c) 1
+#else
+# define IN_CTYPE_DOMAIN(c) isascii(c)
+#endif
+
+#define ISSPACE(c) (IN_CTYPE_DOMAIN (c) && isspace (c))
+
+#include "xstrtol.h"
+
+#if !HAVE_DECL_STRTOIMAX && !defined strtoimax
+intmax_t strtoimax ();
+#endif
+
+#if !HAVE_DECL_STRTOUMAX && !defined strtoumax
+uintmax_t strtoumax ();
+#endif
+
+static strtol_error
+bkm_scale (__strtol_t *x, int scale_factor)
+{
+ if (TYPE_SIGNED (__strtol_t) && *x < STRTOL_T_MINIMUM / scale_factor)
+ {
+ *x = STRTOL_T_MINIMUM;
+ return LONGINT_OVERFLOW;
+ }
+ if (STRTOL_T_MAXIMUM / scale_factor < *x)
+ {
+ *x = STRTOL_T_MAXIMUM;
+ return LONGINT_OVERFLOW;
+ }
+ *x *= scale_factor;
+ return LONGINT_OK;
+}
+
+static strtol_error
+bkm_scale_by_power (__strtol_t *x, int base, int power)
+{
+ strtol_error err = LONGINT_OK;
+ while (power--)
+ err |= bkm_scale (x, base);
+ return err;
+}
+
+/* FIXME: comment. */
+
+strtol_error
+__xstrtol (const char *s, char **ptr, int strtol_base,
+ __strtol_t *val, const char *valid_suffixes)
+{
+ char *t_ptr;
+ char **p;
+ __strtol_t tmp;
+ strtol_error err = LONGINT_OK;
+
+ assert (0 <= strtol_base && strtol_base <= 36);
+
+ p = (ptr ? ptr : &t_ptr);
+
+ if (! TYPE_SIGNED (__strtol_t))
+ {
+ const char *q = s;
+ while (ISSPACE ((unsigned char) *q))
+ ++q;
+ if (*q == '-')
+ return LONGINT_INVALID;
+ }
+
+ errno = 0;
+ tmp = __strtol (s, p, strtol_base);
+
+ if (*p == s)
+ {
+ /* If there is no number but there is a valid suffix, assume the
+ number is 1. The string is invalid otherwise. */
+ if (valid_suffixes && **p && strchr (valid_suffixes, **p))
+ tmp = 1;
+ else
+ return LONGINT_INVALID;
+ }
+ else if (errno != 0)
+ {
+ if (errno != ERANGE)
+ return LONGINT_INVALID;
+ err = LONGINT_OVERFLOW;
+ }
+
+ /* Let valid_suffixes == NULL mean `allow any suffix'. */
+ /* FIXME: update all callers except the ones that allow suffixes
+ after the number, changing last parameter NULL to `""'. */
+ if (!valid_suffixes)
+ {
+ *val = tmp;
+ return err;
+ }
+
+ if (**p != '\0')
+ {
+ int base = 1024;
+ int suffixes = 1;
+ strtol_error overflow;
+
+ if (!strchr (valid_suffixes, **p))
+ {
+ *val = tmp;
+ return err | LONGINT_INVALID_SUFFIX_CHAR;
+ }
+
+ if (strchr (valid_suffixes, '0'))
+ {
+ /* The ``valid suffix'' '0' is a special flag meaning that
+ an optional second suffix is allowed, which can change
+ the base. A suffix "B" (e.g. "100MB") stands for a power
+ of 1000, whereas a suffix "iB" (e.g. "100MiB") stands for
+ a power of 1024. If no suffix (e.g. "100M"), assume
+ power-of-1024. */
+
+ switch (p[0][1])
+ {
+ case 'i':
+ if (p[0][2] == 'B')
+ suffixes += 2;
+ break;
+
+ case 'B':
+ case 'D': /* 'D' is obsolescent */
+ base = 1000;
+ suffixes++;
+ break;
+ }
+ }
+
+ switch (**p)
+ {
+ case 'b':
+ overflow = bkm_scale (&tmp, 512);
+ break;
+
+ case 'B':
+ overflow = bkm_scale (&tmp, 1024);
+ break;
+
+ case 'c':
+ overflow = 0;
+ break;
+
+ case 'E': /* exa or exbi */
+ overflow = bkm_scale_by_power (&tmp, base, 6);
+ break;
+
+ case 'G': /* giga or gibi */
+ case 'g': /* 'g' is undocumented; for compatibility only */
+ overflow = bkm_scale_by_power (&tmp, base, 3);
+ break;
+
+ case 'k': /* kilo */
+ case 'K': /* kibi */
+ overflow = bkm_scale_by_power (&tmp, base, 1);
+ break;
+
+ case 'M': /* mega or mebi */
+ case 'm': /* 'm' is undocumented; for compatibility only */
+ overflow = bkm_scale_by_power (&tmp, base, 2);
+ break;
+
+ case 'P': /* peta or pebi */
+ overflow = bkm_scale_by_power (&tmp, base, 5);
+ break;
+
+ case 'T': /* tera or tebi */
+ case 't': /* 't' is undocumented; for compatibility only */
+ overflow = bkm_scale_by_power (&tmp, base, 4);
+ break;
+
+ case 'w':
+ overflow = bkm_scale (&tmp, 2);
+ break;
+
+ case 'Y': /* yotta or 2**80 */
+ overflow = bkm_scale_by_power (&tmp, base, 8);
+ break;
+
+ case 'Z': /* zetta or 2**70 */
+ overflow = bkm_scale_by_power (&tmp, base, 7);
+ break;
+
+ default:
+ *val = tmp;
+ return err | LONGINT_INVALID_SUFFIX_CHAR;
+ }
+
+ err |= overflow;
+ *p += suffixes;
+ if (**p)
+ err |= LONGINT_INVALID_SUFFIX_CHAR;
+ }
+
+ *val = tmp;
+ return err;
+}
+
+#ifdef TESTING_XSTRTO
+
+# include <stdio.h>
+# include "error.h"
+
+char *program_name;
+
+int
+main (int argc, char **argv)
+{
+ strtol_error s_err;
+ int i;
+
+ program_name = argv[0];
+ for (i=1; i<argc; i++)
+ {
+ char *p;
+ __strtol_t val;
+
+ s_err = __xstrtol (argv[i], &p, 0, &val, "bckmw");
+ if (s_err == LONGINT_OK)
+ {
+ printf ("%s->%lu (%s)\n", argv[i], val, p);
+ }
+ else
+ {
+ STRTOL_FATAL_ERROR (argv[i], "arg", s_err);
+ }
+ }
+ exit (0);
+}
+
+#endif /* TESTING_XSTRTO */
diff --git a/contrib/diff/lib/xstrtol.h b/contrib/diff/lib/xstrtol.h
new file mode 100644
index 0000000..5765c88
--- /dev/null
+++ b/contrib/diff/lib/xstrtol.h
@@ -0,0 +1,91 @@
+/* A more useful interface to strtol.
+
+ Copyright (C) 1995, 1996, 1998, 1999, 2001, 2002, 2003, 2004 Free
+ Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+#ifndef XSTRTOL_H_
+# define XSTRTOL_H_ 1
+
+# include "exitfail.h"
+
+/* Get uintmax_t. */
+# if HAVE_INTTYPES_H
+# include <inttypes.h>
+# else
+# if HAVE_STDINT_H
+# include <stdint.h>
+# endif
+# endif
+
+# ifndef _STRTOL_ERROR
+enum strtol_error
+ {
+ LONGINT_OK = 0,
+
+ /* These two values can be ORed together, to indicate that both
+ errors occurred. */
+ LONGINT_OVERFLOW = 1,
+ LONGINT_INVALID_SUFFIX_CHAR = 2,
+
+ LONGINT_INVALID_SUFFIX_CHAR_WITH_OVERFLOW = (LONGINT_INVALID_SUFFIX_CHAR
+ | LONGINT_OVERFLOW),
+ LONGINT_INVALID = 4
+ };
+typedef enum strtol_error strtol_error;
+# endif
+
+# define _DECLARE_XSTRTOL(name, type) \
+ strtol_error name (const char *, char **, int, type *, const char *);
+_DECLARE_XSTRTOL (xstrtol, long int)
+_DECLARE_XSTRTOL (xstrtoul, unsigned long int)
+_DECLARE_XSTRTOL (xstrtoimax, intmax_t)
+_DECLARE_XSTRTOL (xstrtoumax, uintmax_t)
+
+# define _STRTOL_ERROR(Exit_code, Str, Argument_type_string, Err) \
+ do \
+ { \
+ switch ((Err)) \
+ { \
+ default: \
+ abort (); \
+ \
+ case LONGINT_INVALID: \
+ error ((Exit_code), 0, "invalid %s `%s'", \
+ (Argument_type_string), (Str)); \
+ break; \
+ \
+ case LONGINT_INVALID_SUFFIX_CHAR: \
+ case LONGINT_INVALID_SUFFIX_CHAR | LONGINT_OVERFLOW: \
+ error ((Exit_code), 0, "invalid character following %s in `%s'", \
+ (Argument_type_string), (Str)); \
+ break; \
+ \
+ case LONGINT_OVERFLOW: \
+ error ((Exit_code), 0, "%s `%s' too large", \
+ (Argument_type_string), (Str)); \
+ break; \
+ } \
+ } \
+ while (0)
+
+# define STRTOL_FATAL_ERROR(Str, Argument_type_string, Err) \
+ _STRTOL_ERROR (exit_failure, Str, Argument_type_string, Err)
+
+# define STRTOL_FAIL_WARN(Str, Argument_type_string, Err) \
+ _STRTOL_ERROR (0, Str, Argument_type_string, Err)
+
+#endif /* not XSTRTOL_H_ */
diff --git a/contrib/diff/lib/xstrtoul.c b/contrib/diff/lib/xstrtoul.c
new file mode 100644
index 0000000..285f7b9
--- /dev/null
+++ b/contrib/diff/lib/xstrtoul.c
@@ -0,0 +1,6 @@
+#define __strtol strtoul
+#define __strtol_t unsigned long int
+#define __xstrtol xstrtoul
+#define STRTOL_T_MINIMUM 0
+#define STRTOL_T_MAXIMUM ULONG_MAX
+#include "xstrtol.c"
diff --git a/contrib/diff/lib/xstrtoumax.c b/contrib/diff/lib/xstrtoumax.c
new file mode 100644
index 0000000..a8858bf
--- /dev/null
+++ b/contrib/diff/lib/xstrtoumax.c
@@ -0,0 +1,37 @@
+/* xstrtoumax.c -- A more useful interface to strtoumax.
+ Copyright (C) 1999, 2003 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2, or (at your option)
+ any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software Foundation,
+ Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+
+/* Written by Paul Eggert. */
+
+#if HAVE_CONFIG_H
+# include <config.h>
+#endif
+
+#if HAVE_INTTYPES_H
+# include <inttypes.h>
+#elif HAVE_STDINT_H
+# include <stdint.h>
+#endif
+
+#define __strtol strtoumax
+#define __strtol_t uintmax_t
+#define __xstrtol xstrtoumax
+#ifdef UINTMAX_MAX
+# define STRTOL_T_MINIMUM 0
+# define STRTOL_T_MAXIMUM UINTMAX_MAX
+#endif
+#include "xstrtol.c"
OpenPOWER on IntegriCloud