Add a UTF-8 encoding method, which will eventually replace the antique

"UTF2" method. Although UTF-8 and the old UTF2 encoding are compatible
for 16-bit characters, the new UTF-8 implementation is much more strict
about rejecting malformed input and also handles the full 31 bit range
of characters.

10 files changed, 349 insertions, 9 deletions

diff --git a/lib/libc/locale/Makefile.inc b/lib/libc/locale/Makefile.inc
index 66ea275..08f94ce 100644
--- a/lib/libc/locale/Makefile.inc
+++ b/lib/libc/locale/Makefile.inc
@@ -11,7 +11,8 @@ SRCS+=	big5.c btowc.c collate.c collcmp.c euc.c fix_grouping.c frune.c \
 	mbrtowc.c mbrune.c mbsinit.c mbsrtowcs.c mbtowc.c mbstowcs.c \
 	mskanji.c nl_langinfo.c nomacros.c none.c rune.c \
 	runetype.c setinvalidrune.c setlocale.c setrunelocale.c table.c \
-	tolower.c toupper.c utf2.c wcrtomb.c wcsrtombs.c wcsftime.c wcstod.c \
+	tolower.c toupper.c utf2.c utf8.c wcrtomb.c wcsrtombs.c wcsftime.c \
+	wcstod.c \
 	wcstoimax.c wcstol.c wcstoll.c \
 	wcstombs.c \
 	wcstoul.c wcstoull.c wcstoumax.c wctob.c wctomb.c wctrans.c wctype.c \
@@ -31,6 +32,7 @@ MAN+=	btowc.3 \
 	wcsrtombs.3 wcstod.3 wcstol.3 \
 	wctrans.3 wctype.3 wcwidth.3
 MAN+=	euc.4 utf2.4
+MAN+=	utf8.5
 
 MLINKS+=btowc.3 wctob.3
 MLINKS+=isdigit.3 isnumber.3
diff --git a/lib/libc/locale/mbrune.3 b/lib/libc/locale/mbrune.3
index 8e0aa08..f2598e7 100644
--- a/lib/libc/locale/mbrune.3
+++ b/lib/libc/locale/mbrune.3
@@ -161,7 +161,8 @@ does not appear in the string.
 .Xr rune 3 ,
 .Xr setlocale 3 ,
 .Xr euc 4 ,
-.Xr utf2 4
+.Xr utf2 4 ,
+.Xr utf8 5
 .Sh HISTORY
 The
 .Fn mbrune ,
diff --git a/lib/libc/locale/multibyte.3 b/lib/libc/locale/multibyte.3
index 4d4a7d7..8988eb6 100644
--- a/lib/libc/locale/multibyte.3
+++ b/lib/libc/locale/multibyte.3
@@ -232,7 +232,8 @@ both functions return \-1.
 .Xr wcrtomb 3 ,
 .Xr wcsrtombs 3 ,
 .Xr euc 4 ,
-.Xr utf2 4
+.Xr utf2 4 ,
+.Xr utf8 5
 .Sh STANDARDS
 The
 .Fn mblen ,
diff --git a/lib/libc/locale/rune.3 b/lib/libc/locale/rune.3
index c11dc89..755ac96 100644
--- a/lib/libc/locale/rune.3
+++ b/lib/libc/locale/rune.3
@@ -265,7 +265,8 @@ binary LC_CTYPE file for the locale
 .Xr mbrune 3 ,
 .Xr setlocale 3 ,
 .Xr euc 4 ,
-.Xr utf2 4
+.Xr utf2 4 ,
+.Xr utf8 5
 .Sh HISTORY
 These functions first appeared in
 .Bx 4.4 .
diff --git a/lib/libc/locale/setlocale.3 b/lib/libc/locale/setlocale.3
index 7580867..95996b7 100644
--- a/lib/libc/locale/setlocale.3
+++ b/lib/libc/locale/setlocale.3
@@ -334,7 +334,8 @@ and the category
 .Xr strcoll 3 ,
 .Xr strxfrm 3 ,
 .Xr euc 4 ,
-.Xr utf2 4
+.Xr utf2 4 ,
+.Xr utf8 5
 .Sh STANDARDS
 The
 .Fn setlocale
diff --git a/lib/libc/locale/setrunelocale.c b/lib/libc/locale/setrunelocale.c
index 042a06c..9cfe8d6 100644
--- a/lib/libc/locale/setrunelocale.c
+++ b/lib/libc/locale/setrunelocale.c
@@ -48,6 +48,7 @@ __FBSDID("$FreeBSD$");
 
 extern int		_none_init(_RuneLocale *);
 extern int		_UTF2_init(_RuneLocale *);
+extern int		_UTF8_init(_RuneLocale *);
 extern int		_EUC_init(_RuneLocale *);
 extern int		_BIG5_init(_RuneLocale *);
 extern int		_MSKanji_init(_RuneLocale *);
@@ -130,6 +131,8 @@ setrunelocale(char *encoding)
 		ret = _none_init(rl);
 	else if (strcmp(rl->encoding, "UTF2") == 0)
 		ret = _UTF2_init(rl);
+	else if (strcmp(rl->encoding, "UTF-8") == 0)
+		ret = _UTF8_init(rl);
 	else if (strcmp(rl->encoding, "EUC") == 0)
 		ret = _EUC_init(rl);
 	else if (strcmp(rl->encoding, "BIG5") == 0)
diff --git a/lib/libc/locale/utf2.4 b/lib/libc/locale/utf2.4
index 35eb3ad..f41c0f9 100644
--- a/lib/libc/locale/utf2.4
+++ b/lib/libc/locale/utf2.4
@@ -35,7 +35,7 @@
 .\"	@(#)utf2.4	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd June 4, 1993
+.Dd October 11, 2002
 .Dt UTF2 4
 .Os
 .Sh NAME
@@ -45,6 +45,11 @@
 .Nm ENCODING
 .Qq UTF2
 .Sh DESCRIPTION
+.Bf Em
+The UTF2 encoding has been deprecated in favour of UTF-8.
+.Ef
+New applications should not use UTF2.
+.Pp
 The
 .Nm UTF2
 encoding is based on a proposed X-Open multibyte
@@ -85,4 +90,5 @@ which provides for the entire proposed ISO-10646 31 bit standard are currently
 not implemented.
 .Sh "SEE ALSO"
 .Xr mklocale 1 ,
-.Xr setlocale 3
+.Xr setlocale 3 ,
+.Xr utf8 5
diff --git a/lib/libc/locale/utf2.5 b/lib/libc/locale/utf2.5
index 35eb3ad..f41c0f9 100644
--- a/lib/libc/locale/utf2.5
+++ b/lib/libc/locale/utf2.5
@@ -35,7 +35,7 @@
 .\"	@(#)utf2.4	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd June 4, 1993
+.Dd October 11, 2002
 .Dt UTF2 4
 .Os
 .Sh NAME
@@ -45,6 +45,11 @@
 .Nm ENCODING
 .Qq UTF2
 .Sh DESCRIPTION
+.Bf Em
+The UTF2 encoding has been deprecated in favour of UTF-8.
+.Ef
+New applications should not use UTF2.
+.Pp
 The
 .Nm UTF2
 encoding is based on a proposed X-Open multibyte
@@ -85,4 +90,5 @@ which provides for the entire proposed ISO-10646 31 bit standard are currently
 not implemented.
 .Sh "SEE ALSO"
 .Xr mklocale 1 ,
-.Xr setlocale 3
+.Xr setlocale 3 ,
+.Xr utf8 5
diff --git a/lib/libc/locale/utf8.5 b/lib/libc/locale/utf8.5
new file mode 100644
index 0000000..079e9ea
--- /dev/null
+++ b/lib/libc/locale/utf8.5
@@ -0,0 +1,115 @@
+.\" Copyright (c) 1993
+.\"	The Regents of the University of California.  All rights reserved.
+.\"
+.\" This code is derived from software contributed to Berkeley by
+.\" Paul Borman at Krystal Technologies.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"    notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"    notice, this list of conditions and the following disclaimer in the
+.\"    documentation and/or other materials provided with the distribution.
+.\" 3. All advertising materials mentioning features or use of this software
+.\"    must display the following acknowledgement:
+.\"	This product includes software developed by the University of
+.\"	California, Berkeley and its contributors.
+.\" 4. Neither the name of the University nor the names of its contributors
+.\"    may be used to endorse or promote products derived from this software
+.\"    without specific prior written permission.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\"	@(#)utf2.4	8.1 (Berkeley) 6/4/93
+.\" $FreeBSD$
+.\"
+.Dd October 10, 2002
+.Dt UTF8 5
+.Os
+.Sh NAME
+.Nm utf8
+.Nd "UTF-8, a transformation format of ISO 10646"
+.Sh SYNOPSIS
+.Nm ENCODING
+.Qq UTF-8
+.Sh DESCRIPTION
+The
+.Nm UTF-8
+encoding represents UCS-4 characters as a sequence of octets, using
+between 1 and 6 for each character.
+It is backwards compatible with
+.Tn ASCII ,
+so 0x00-0x7f refer to the
+.Tn ASCII
+character set.
+The multibyte encoding of non-
+.Tn ASCII
+characters
+consist entirely of bytes whose high order bit is set.
+The actual
+encoding is represented by the following table:
+.Bd -literal
+[0x00000000 - 0x0000007f] [00000000.0bbbbbbb] -> 0bbbbbbb
+[0x00000080 - 0x000007ff] [00000bbb.bbbbbbbb] -> 110bbbbb, 10bbbbbb
+[0x00000800 - 0x0000ffff] [bbbbbbbb.bbbbbbbb] ->
+	1110bbbb, 10bbbbbb, 10bbbbbb
+[0x00010000 - 0x001fffff] [00000000.000bbbbb.bbbbbbbb.bbbbbbbb] ->
+	11110bbb, 10bbbbbb, 10bbbbbb, 10bbbbbb
+[0x00200000 - 0x03ffffff] [000000bb.bbbbbbbb.bbbbbbbb.bbbbbbbb] ->
+	111110bb, 10bbbbbb, 10bbbbbb, 10bbbbbb, 10bbbbbb
+[0x04000000 - 0x7fffffff] [0bbbbbbb.bbbbbbbb.bbbbbbbb.bbbbbbbb] ->
+	1111110b, 10bbbbbb, 10bbbbbb, 10bbbbbb, 10bbbbbb, 10bbbbbb
+.Ed
+.Pp
+If more than a single representation of a value exists (for example,
+0x00; 0xC0 0x80; 0xE0 0x80 0x80) the shortest representation is always
+used.
+Longer ones are detected as an error as they pose a potential
+security risk, and destroy the 1:1 character:octet sequence mapping.
+.Sh COMPATIBILITY
+The
+.Nm
+encoding supersedes the
+.Xr utf2 4
+encoding.
+The only differences between the two are that
+.Nm
+handles the full 31-bit character set of
+.Tn ISO
+10646
+whereas
+.Xr utf2 4
+is limited to a 16-bit character set,
+and that
+.Xr utf2 4
+accepts redundant, non-"shortest form" representations of characters.
+.Sh SEE ALSO
+.Xr euc 4 ,
+.Xr utf2 4
+.Rs
+.%A "F. Yergeau"
+.%T "UTF-8, a transformation format of ISO 10646"
+.%O "RFC 2279"
+.%D "January 1998"
+.Re
+.Sh STANDARDS
+The
+.Nm
+encoding is compatible with RFC 2279.
+.Sh BUGS
+Byte order marker (BOM) characters are neither added nor removed
+from UTF-8-encoded wide character
+.Xr stdio 3
+streams.
diff --git a/lib/libc/locale/utf8.c b/lib/libc/locale/utf8.c
new file mode 100644
index 0000000..590bd43
--- /dev/null
+++ b/lib/libc/locale/utf8.c
@@ -0,0 +1,204 @@
+/*-
+ * Copyright (c) 2002 Tim J. Robbins
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <rune.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+rune_t	_UTF8_sgetrune(const char *, size_t, char const **);
+int	_UTF8_sputrune(rune_t, char *, size_t, char **);
+
+int
+_UTF8_init(_RuneLocale *rl)
+{
+
+	rl->sgetrune = _UTF8_sgetrune;
+	rl->sputrune = _UTF8_sputrune;
+	_CurrentRuneLocale = rl;
+	__mb_cur_max = 6;
+
+	return (0);
+}
+
+rune_t
+_UTF8_sgetrune(const char *string, size_t n, const char **result)
+{
+	int ch, len, mask, siglen;
+	rune_t lbound, wch;
+
+	if (n < 1) {
+		if (result != NULL)
+			*result = string;
+		return (_INVALID_RUNE);
+	}
+
+	/*
+	 * Determine the number of octets that make up this character from
+	 * the first octet, and a mask that extracts the interesting bits of
+	 * the first octet.
+	 *
+	 * We also specify a lower bound for the character code to detect
+	 * redundant, non-"shortest form" encodings. For example, the
+	 * sequence C0 80 is _not_ a legal representation of the null
+	 * character. This enforces a 1-to-1 mapping between character
+	 * codes and their multibyte representations.
+	 */
+	ch = (unsigned char)*string;
+	if ((ch & 0x80) == 0) {
+		mask = 0x7f;
+		len = 1;
+		lbound = 0;
+	} else if ((ch & 0xe0) == 0xc0) {
+		mask = 0x1f;
+		len = 2;
+		lbound = 0x80;
+	} else if ((ch & 0xf0) == 0xe0) {
+		mask = 0x0f;
+		len = 3;
+		lbound = 0x800;
+	} else if ((ch & 0xf8) == 0xf0) {
+		mask = 0x07;
+		len = 4;
+		lbound = 0x10000;
+	} else if ((ch & 0xfc) == 0xf8) {
+		mask = 0x03;
+		len = 5;
+		lbound = 0x200000;
+	} else if ((ch & 0xfc) == 0xfc) {
+		mask = 0x01;
+		len = 6;
+		lbound = 0x4000000;
+	} else {
+		/*
+		 * Malformed input; input is not UTF-8.
+		 */
+		if (result != NULL)
+			*result = string + 1;
+		return (_INVALID_RUNE);
+	}
+
+	if (n < len) {
+		/*
+		 * Truncated or partial input.
+		 */
+		if (result != NULL)
+			*result = string;
+		return (_INVALID_RUNE);
+	}
+
+	/*
+	 * Decode the octet sequence representing the character in chunks
+	 * of 6 bits, most significant first.
+	 */
+	wch = (unsigned char)*string++ & mask;
+	while (--len != 0) {
+		if ((*string & 0xc0) != 0x80) {
+			/*
+			 * Malformed input; bad characters in the middle
+			 * of a character.
+			 */
+			wch = _INVALID_RUNE;
+			if (result != NULL)
+				*result = string + 1;
+			return (_INVALID_RUNE);
+		}
+		wch <<= 6;
+		wch |= *string++ & 0x3f;
+	}
+	if (wch != _INVALID_RUNE && wch < lbound)
+		/*
+		 * Malformed input; redundant encoding.
+		 */
+		wch = _INVALID_RUNE;
+	if (result != NULL)
+		*result = string;
+	return (wch);
+}
+
+int
+_UTF8_sputrune(rune_t c, char *string, size_t n, char **result)
+{
+	unsigned char lead;
+	int i, len;
+
+	/*
+	 * Determine the number of octets needed to represent this character.
+	 * We always output the shortest sequence possible. Also specify the
+	 * first few bits of the first octet, which contains the information
+	 * about the sequence length.
+	 */
+	if ((c & ~0x7f) == 0) {
+		lead = 0;
+		len = 1;
+	} else if ((c & ~0x7ff) == 0) {
+		lead = 0xc0;
+		len = 2;
+	} else if ((c & ~0xffff) == 0) {
+		lead = 0xe0;
+		len = 3;
+	} else if ((c & ~0x1fffff) == 0) {
+		lead = 0xf0;
+		len = 4;
+	} else if ((c & ~0x3ffffff) == 0) {
+		lead = 0xf8;
+		len = 5;
+	} else if ((c & ~0x7fffffff) == 0) {
+		lead = 0xfc;
+		len = 6;
+	} else {
+		/*
+		 * Wide character code is out of range.
+		 */
+		if (result != NULL)
+			*result = NULL;
+		return (0);
+	}
+
+	if (n < len) {
+		if (result != NULL)
+			*result = NULL;
+	} else {
+		/*
+		 * Output the octets representing the character in chunks
+		 * of 6 bits, least significant last. The first octet is
+		 * a special case because it contains the sequence length
+		 * information.
+		 */
+		for (i = len - 1; i > 0; i--) {
+			string[i] = (c & 0x3f) | 0x80;
+			c >>= 6;
+		}
+		*string = (c & 0xff) | lead;
+		if (result != NULL)
+			*result = string + len;
+	}
+
+	return (len);
+}