Prepare to handle state-dependent encodings. This mainly involves not

taking shortcuts when it comes to storing and passing around conversion
states.

12 files changed, 87 insertions, 71 deletions

diff --git a/lib/libc/locale/btowc.c b/lib/libc/locale/btowc.c
index b0b24f2..34439b8 100644
--- a/lib/libc/locale/btowc.c
+++ b/lib/libc/locale/btowc.c
@@ -28,28 +28,25 @@
 __FBSDID("$FreeBSD$");
 
 #include <stdio.h>
-#include <string.h>
 #include <wchar.h>
 
 wint_t
 btowc(int c)
 {
+	static const mbstate_t initial;
+	mbstate_t mbs = initial;
 	char cc;
 	wchar_t wc;
 
 	if (c == EOF)
 		return (WEOF);
-	cc = (char)c;
 	/*
 	 * We expect mbrtowc() to return 0 or 1, hence the check for n > 1
 	 * which detects error return values as well as "impossible" byte
 	 * counts.
-	 *
-	 * We pass NULL as the state pointer to mbrtowc() because we don't
-	 * support state-dependent encodings and don't want to waste time
-	 * creating a zeroed mbstate_t that will not be used.
 	 */
-	if (mbrtowc(&wc, &cc, 1, NULL) > 1)
+	cc = (char)c;
+	if (mbrtowc(&wc, &cc, 1, &mbs) > 1)
 		return (WEOF);
 	return (wc);
 }
diff --git a/lib/libc/locale/mblen.c b/lib/libc/locale/mblen.c
index 5025115..8db5335 100644
--- a/lib/libc/locale/mblen.c
+++ b/lib/libc/locale/mblen.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002, 2003 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -27,15 +27,29 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
+#include <errno.h>
+#include <limits.h>
 #include <stdlib.h>
+#include <wchar.h>
 
 int
 mblen(const char *s, size_t n)
 {
+	static const mbstate_t initial;
+	static mbstate_t mbs;
+	size_t rval;
 
-	/*
-	 * Calling mbtowc() is only legal because we don't support
-	 * state-dependent encodings.
-	 */
-	return (mbtowc(NULL, s, n));
+	if (s == NULL) {
+		/* No support for state dependent encodings. */
+		mbs = initial;
+		return (0);
+	}
+	rval = mbrtowc(NULL, s, n, &mbs);
+	if (rval == (size_t)-1 || rval == (size_t)-2)
+		return (-1);
+	if (rval > INT_MAX) {
+		errno = ERANGE;
+		return (-1);
+	}
+	return ((int)rval);
 }
diff --git a/lib/libc/locale/mbrlen.c b/lib/libc/locale/mbrlen.c
index a2ab65b..edad289 100644
--- a/lib/libc/locale/mbrlen.c
+++ b/lib/libc/locale/mbrlen.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -30,8 +30,11 @@ __FBSDID("$FreeBSD$");
 #include <wchar.h>
 
 size_t
-mbrlen(const char * __restrict s, size_t n, mbstate_t * __restrict ps __unused)
+mbrlen(const char * __restrict s, size_t n, mbstate_t * __restrict ps)
 {
+	static mbstate_t mbs;
 
-	return (mbrtowc(NULL, s, n, NULL));
+	if (ps == NULL)
+		ps = &mbs;
+	return (mbrtowc(NULL, s, n, ps));
 }
diff --git a/lib/libc/locale/mbrtowc.c b/lib/libc/locale/mbrtowc.c
index f3b12b4..22a9041 100644
--- a/lib/libc/locale/mbrtowc.c
+++ b/lib/libc/locale/mbrtowc.c
@@ -36,6 +36,9 @@ size_t
 mbrtowc(wchar_t * __restrict pwc, const char * __restrict s,
     size_t n, mbstate_t * __restrict ps)
 {
+	static mbstate_t mbs;
 
+	if (ps == NULL)
+		ps = &mbs;
 	return (__mbrtowc(pwc, s, n, ps));
 }
diff --git a/lib/libc/locale/mbsrtowcs.c b/lib/libc/locale/mbsrtowcs.c
index 3e7b3b8..8b3de2f 100644
--- a/lib/libc/locale/mbsrtowcs.c
+++ b/lib/libc/locale/mbsrtowcs.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -34,8 +34,9 @@ __FBSDID("$FreeBSD$");
 
 size_t
 mbsrtowcs(wchar_t * __restrict dst, const char ** __restrict src, size_t len,
-    mbstate_t * __restrict ps __unused)
+    mbstate_t * __restrict ps)
 {
+	static mbstate_t mbs;
 	const char *s;
 	size_t nchr;
 	wchar_t wc;
@@ -44,9 +45,11 @@ mbsrtowcs(wchar_t * __restrict dst, const char ** __restrict src, size_t len,
 	s = *src;
 	nchr = 0;
 
+	if (ps == NULL)
+		ps = &mbs;
 	if (dst == NULL) {
 		for (;;) {
-			if ((nb = (int)mbrtowc(&wc, s, MB_CUR_MAX, NULL)) < 0)
+			if ((nb = (int)mbrtowc(&wc, s, MB_CUR_MAX, ps)) < 0)
 				/* Invalid sequence - mbrtowc() sets errno. */
 				return ((size_t)-1);
 			else if (nb == 0)
@@ -58,7 +61,7 @@ mbsrtowcs(wchar_t * __restrict dst, const char ** __restrict src, size_t len,
 	}
 
 	while (len-- > 0) {
-		if ((nb = (int)mbrtowc(dst, s, MB_CUR_MAX, NULL)) < 0) {
+		if ((nb = (int)mbrtowc(dst, s, MB_CUR_MAX, ps)) < 0) {
 			*src = s;
 			return ((size_t)-1);
 		} else if (nb == 0) {
diff --git a/lib/libc/locale/mbstowcs.c b/lib/libc/locale/mbstowcs.c
index 419c1d3..b63f1f2 100644
--- a/lib/libc/locale/mbstowcs.c
+++ b/lib/libc/locale/mbstowcs.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002, 2003 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -28,17 +28,12 @@
 __FBSDID("$FreeBSD$");
 
 #include <stdlib.h>
-#include <string.h>
 #include <wchar.h>
 
 size_t
 mbstowcs(wchar_t * __restrict pwcs, const char * __restrict s, size_t n)
 {
+	static mbstate_t mbs;
 
-	/*
-	 * We pass NULL as the state pointer to mbsrtowcs() because we don't
-	 * support state-dependent encodings and don't want to waste time
-	 * creating a zeroed mbstate_t that will not be used.
-	 */
-	return (mbsrtowcs(pwcs, &s, n, NULL));
+	return (mbsrtowcs(pwcs, &s, n, &mbs));
 }
diff --git a/lib/libc/locale/mbtowc.c b/lib/libc/locale/mbtowc.c
index cc88566..465f45a 100644
--- a/lib/libc/locale/mbtowc.c
+++ b/lib/libc/locale/mbtowc.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002, 2003 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -36,17 +36,16 @@ __FBSDID("$FreeBSD$");
 int
 mbtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n)
 {
+	static const mbstate_t initial;
+	static mbstate_t mbs;
 	size_t rval;
 
-	if (s == NULL)
+	if (s == NULL) {
 		/* No support for state dependent encodings. */
+		mbs = initial;
 		return (0);
-	/*
-	 * We pass NULL as the state pointer to mbrtowc() because we don't
-	 * support state-dependent encodings and don't want to waste time
-	 * creating a zeroed mbstate_t that will not be used.
-	 */
-	rval = mbrtowc(pwc, s, n, NULL);
+	}
+	rval = mbrtowc(pwc, s, n, &mbs);
 	if (rval == (size_t)-1 || rval == (size_t)-2)
 		return (-1);
 	if (rval > INT_MAX) {
diff --git a/lib/libc/locale/wcrtomb.c b/lib/libc/locale/wcrtomb.c
index 722e5c1..5f79949 100644
--- a/lib/libc/locale/wcrtomb.c
+++ b/lib/libc/locale/wcrtomb.c
@@ -34,6 +34,9 @@ extern size_t (*__wcrtomb)(char * __restrict, wchar_t, mbstate_t * __restrict);
 size_t
 wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps)
 {
+	static mbstate_t mbs;
 
+	if (ps == NULL)
+		ps = &mbs;
 	return (__wcrtomb(s, wc, ps));
 }
diff --git a/lib/libc/locale/wcsrtombs.c b/lib/libc/locale/wcsrtombs.c
index ea3354b..4fa5ff8 100644
--- a/lib/libc/locale/wcsrtombs.c
+++ b/lib/libc/locale/wcsrtombs.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -27,7 +27,6 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include <errno.h>
 #include <limits.h>
 #include <stdlib.h>
 #include <string.h>
@@ -35,8 +34,10 @@ __FBSDID("$FreeBSD$");
 
 size_t
 wcsrtombs(char * __restrict dst, const wchar_t ** __restrict src, size_t len,
-    mbstate_t * __restrict ps __unused)
+    mbstate_t * __restrict ps)
 {
+	static mbstate_t mbs;
+	mbstate_t mbsbak;
 	char buf[MB_LEN_MAX];
 	const wchar_t *s;
 	size_t nbytes;
@@ -45,9 +46,11 @@ wcsrtombs(char * __restrict dst, const wchar_t ** __restrict src, size_t len,
 	s = *src;
 	nbytes = 0;
 
+	if (ps == NULL)
+		ps = &mbs;
 	if (dst == NULL) {
 		for (;;) {
-			if ((nb = (int)wcrtomb(buf, *s, NULL)) < 0)
+			if ((nb = (int)wcrtomb(buf, *s, ps)) < 0)
 				/* Invalid character - wcrtomb() sets errno. */
 				return ((size_t)-1);
 			else if (*s == L'\0')
@@ -61,19 +64,28 @@ wcsrtombs(char * __restrict dst, const wchar_t ** __restrict src, size_t len,
 	while (len > 0) {
 		if (len > (size_t)MB_CUR_MAX) {
 			/* Enough space to translate in-place. */
-			if ((nb = (int)wcrtomb(dst, *s, NULL)) < 0) {
+			if ((nb = (int)wcrtomb(dst, *s, ps)) < 0) {
 				*src = s;
 				return ((size_t)-1);
 			}
 		} else {
-			/* May not be enough space; use temp. buffer. */
-			if ((nb = (int)wcrtomb(buf, *s, NULL)) < 0) {
+			/*
+			 * May not be enough space; use temp. buffer.
+			 *
+			 * We need to save a copy of the conversion state
+			 * here so we can restore it if the multibyte
+			 * character is too long for the buffer.
+			 */
+			mbsbak = *ps;
+			if ((nb = (int)wcrtomb(buf, *s, ps)) < 0) {
 				*src = s;
 				return ((size_t)-1);
 			}
-			if (nb > (int)len)
+			if (nb > (int)len) {
 				/* MB sequence for character won't fit. */
+				*ps = mbsbak;
 				break;
+			}
 			memcpy(dst, buf, nb);
 		}
 		if (*s == L'\0') {
diff --git a/lib/libc/locale/wcstombs.c b/lib/libc/locale/wcstombs.c
index f415aed..cc5fab4 100644
--- a/lib/libc/locale/wcstombs.c
+++ b/lib/libc/locale/wcstombs.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002, 2003 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -28,17 +28,12 @@
 __FBSDID("$FreeBSD$");
 
 #include <stdlib.h>
-#include <string.h>
 #include <wchar.h>
 
 size_t
 wcstombs(char * __restrict s, const wchar_t * __restrict pwcs, size_t n)
 {
+	static mbstate_t mbs;
 
-	/*
-	 * We pass NULL as the state pointer to wcsrtombs() because we don't
-	 * support state-dependent encodings and don't want to waste time
-	 * creating a zeroed mbstate_t that will not be used.
-	 */
-	return (wcsrtombs(s, &pwcs, n, NULL));
+	return (wcsrtombs(s, &pwcs, n, &mbs));
 }
diff --git a/lib/libc/locale/wctob.c b/lib/libc/locale/wctob.c
index da7afcf..4f4d295 100644
--- a/lib/libc/locale/wctob.c
+++ b/lib/libc/locale/wctob.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002, 2003 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -29,20 +29,16 @@ __FBSDID("$FreeBSD$");
 
 #include <limits.h>
 #include <stdio.h>
-#include <string.h>
 #include <wchar.h>
 
 int
 wctob(wint_t c)
 {
+	static const mbstate_t initial;
+	mbstate_t mbs = initial;
 	char buf[MB_LEN_MAX];
 
-	/*
-	 * We pass NULL as the state pointer to wcrtomb() because we don't
-	 * support state-dependent encodings and don't want to waste time
-	 * creating a zeroed mbstate_t that will not be used.
-	 */
-	if (c == WEOF || wcrtomb(buf, c, NULL) != 1)
+	if (c == WEOF || wcrtomb(buf, c, &mbs) != 1)
 		return (EOF);
 	return ((unsigned char)*buf);
 }
diff --git a/lib/libc/locale/wctomb.c b/lib/libc/locale/wctomb.c
index e212b96..9e6f183 100644
--- a/lib/libc/locale/wctomb.c
+++ b/lib/libc/locale/wctomb.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2002, 2003 Tim J. Robbins.
+ * Copyright (c) 2002-2004 Tim J. Robbins.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -36,20 +36,16 @@ __FBSDID("$FreeBSD$");
 int
 wctomb(char *s, wchar_t wchar)
 {
-	char buf[MB_LEN_MAX];
+	static const mbstate_t initial;
+	static mbstate_t mbs;
 	size_t rval;
 
-	if (s == NULL)
+	if (s == NULL) {
 		/* No support for state dependent encodings. */
+		mbs = initial;
 		return (0);
-	if (s == NULL)
-		s = buf;
-	/*
-	 * We pass NULL as the state pointer to wcrtomb() because we don't
-	 * support state-dependent encodings and don't want to waste time
-	 * creating a zeroed mbstate_t that will not be used.
-	 */
-	if ((rval = wcrtomb(s, wchar, NULL)) == (size_t)-1)
+	}
+	if ((rval = wcrtomb(s, wchar, &mbs)) == (size_t)-1)
 		return (-1);
 	if (rval > INT_MAX) {
 		errno = ERANGE;