--- utf2.c.orig Fri Feb 18 15:49:55 2005
+++ utf2.c Fri Feb 18 15:52:07 2005
@@ -25,8 +25,11 @@
*/
#include <sys/param.h>
+/* dumb down UTF-8 to do UTF2 */
__FBSDID("$FreeBSD: src/lib/libc/locale/utf8.c,v 1.11 2004/07/27 06:29:48 tjr Exp $");
+#include "xlocale_private.h"
+
#include <errno.h>
#include <limits.h>
#include <runetype.h>
@@ -35,54 +38,55 @@
#include <wchar.h>
#include "mblocal.h"
-size_t _UTF8_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
- mbstate_t * __restrict);
-int _UTF8_mbsinit(const mbstate_t *);
-size_t _UTF8_mbsnrtowcs(wchar_t * __restrict, const char ** __restrict,
- size_t, size_t, mbstate_t * __restrict);
-size_t _UTF8_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict);
-size_t _UTF8_wcsnrtombs(char * __restrict, const wchar_t ** __restrict,
- size_t, size_t, mbstate_t * __restrict);
+#define UTF2_MB_CUR_MAX 3
+
+static size_t _UTF2_mbrtowc(wchar_t * __restrict, const char * __restrict, size_t,
+ mbstate_t * __restrict, locale_t);
+static int _UTF2_mbsinit(const mbstate_t *, locale_t);
+static size_t _UTF2_mbsnrtowcs(wchar_t * __restrict, const char ** __restrict,
+ size_t, size_t, mbstate_t * __restrict, locale_t);
+static size_t _UTF2_wcrtomb(char * __restrict, wchar_t, mbstate_t * __restrict, locale_t);
+static size_t _UTF2_wcsnrtombs(char * __restrict, const wchar_t ** __restrict,
+ size_t, size_t, mbstate_t * __restrict, locale_t);
typedef struct {
wchar_t ch;
int want;
wchar_t lbound;
-} _UTF8State;
+} _UTF2State;
-int
-_UTF8_init(_RuneLocale *rl)
+__private_extern__ int
+_UTF2_init(struct __xlocale_st_runelocale *xrl)
{
- __mbrtowc = _UTF8_mbrtowc;
- __wcrtomb = _UTF8_wcrtomb;
- __mbsinit = _UTF8_mbsinit;
- __mbsnrtowcs = _UTF8_mbsnrtowcs;
- __wcsnrtombs = _UTF8_wcsnrtombs;
- _CurrentRuneLocale = rl;
- __mb_cur_max = 6;
+ xrl->__mbrtowc = _UTF2_mbrtowc;
+ xrl->__wcrtomb = _UTF2_wcrtomb;
+ xrl->__mbsinit = _UTF2_mbsinit;
+ xrl->__mbsnrtowcs = _UTF2_mbsnrtowcs;
+ xrl->__wcsnrtombs = _UTF2_wcsnrtombs;
+ xrl->__mb_cur_max = UTF2_MB_CUR_MAX;
return (0);
}
-int
-_UTF8_mbsinit(const mbstate_t *ps)
+static int
+_UTF2_mbsinit(const mbstate_t *ps, locale_t loc)
{
- return (ps == NULL || ((const _UTF8State *)ps)->want == 0);
+ return (ps == NULL || ((const _UTF2State *)ps)->want == 0);
}
-size_t
-_UTF8_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
- mbstate_t * __restrict ps)
+static size_t
+_UTF2_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
+ mbstate_t * __restrict ps, locale_t loc)
{
- _UTF8State *us;
+ _UTF2State *us;
int ch, i, mask, want;
wchar_t lbound, wch;
- us = (_UTF8State *)ps;
+ us = (_UTF2State *)ps;
- if (us->want < 0 || us->want > 6) {
+ if (us->want < 0 || us->want > 3) {
errno = EINVAL;
return ((size_t)-1);
}
@@ -130,21 +134,9 @@
mask = 0x0f;
want = 3;
lbound = 0x800;
- } else if ((ch & 0xf8) == 0xf0) {
- mask = 0x07;
- want = 4;
- lbound = 0x10000;
- } else if ((ch & 0xfc) == 0xf8) {
- mask = 0x03;
- want = 5;
- lbound = 0x200000;
- } else if ((ch & 0xfc) == 0xfc) {
- mask = 0x01;
- want = 6;
- lbound = 0x4000000;
} else {
/*
- * Malformed input; input is not UTF-8.
+ * Malformed input; input is not UTF2.
*/
errno = EILSEQ;
return ((size_t)-1);
@@ -194,17 +186,17 @@
return (wch == L'\0' ? 0 : want);
}
-size_t
-_UTF8_mbsnrtowcs(wchar_t * __restrict dst, const char ** __restrict src,
- size_t nms, size_t len, mbstate_t * __restrict ps)
+static size_t
+_UTF2_mbsnrtowcs(wchar_t * __restrict dst, const char ** __restrict src,
+ size_t nms, size_t len, mbstate_t * __restrict ps, locale_t loc)
{
- _UTF8State *us;
+ _UTF2State *us;
const char *s;
size_t nchr;
wchar_t wc;
size_t nb;
- us = (_UTF8State *)ps;
+ us = (_UTF2State *)ps;
s = *src;
nchr = 0;
@@ -226,7 +218,7 @@
* excluding NUL.
*/
nb = 1;
- else if ((nb = _UTF8_mbrtowc(&wc, s, nms, ps)) ==
+ else if ((nb = _UTF2_mbrtowc(&wc, s, nms, ps, loc)) ==
(size_t)-1)
/* Invalid sequence - mbrtowc() sets errno. */
return ((size_t)-1);
@@ -256,7 +248,7 @@
*/
*dst = (wchar_t)*s;
nb = 1;
- } else if ((nb = _UTF8_mbrtowc(dst, s, nms, ps)) ==
+ } else if ((nb = _UTF2_mbrtowc(dst, s, nms, ps, loc)) ==
(size_t)-1) {
*src = s;
return ((size_t)-1);
@@ -276,14 +268,14 @@
return (nchr);
}
-size_t
-_UTF8_wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps)
+static size_t
+_UTF2_wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps, locale_t loc)
{
- _UTF8State *us;
+ _UTF2State *us;
unsigned char lead;
int i, len;
- us = (_UTF8State *)ps;
+ us = (_UTF2State *)ps;
if (us->want != 0) {
errno = EINVAL;
@@ -315,15 +307,6 @@
} else if ((wc & ~0xffff) == 0) {
lead = 0xe0;
len = 3;
- } else if ((wc & ~0x1fffff) == 0) {
- lead = 0xf0;
- len = 4;
- } else if ((wc & ~0x3ffffff) == 0) {
- lead = 0xf8;
- len = 5;
- } else if ((wc & ~0x7fffffff) == 0) {
- lead = 0xfc;
- len = 6;
} else {
errno = EILSEQ;
return ((size_t)-1);
@@ -344,17 +327,17 @@
return (len);
}
-size_t
-_UTF8_wcsnrtombs(char * __restrict dst, const wchar_t ** __restrict src,
- size_t nwc, size_t len, mbstate_t * __restrict ps)
+static size_t
+_UTF2_wcsnrtombs(char * __restrict dst, const wchar_t ** __restrict src,
+ size_t nwc, size_t len, mbstate_t * __restrict ps, locale_t loc)
{
- _UTF8State *us;
+ _UTF2State *us;
char buf[MB_LEN_MAX];
const wchar_t *s;
size_t nbytes;
size_t nb;
- us = (_UTF8State *)ps;
+ us = (_UTF2State *)ps;
if (us->want != 0) {
errno = EINVAL;
@@ -369,7 +352,7 @@
if (0 <= *s && *s < 0x80)
/* Fast path for plain ASCII characters. */
nb = 1;
- else if ((nb = _UTF8_wcrtomb(buf, *s, ps)) ==
+ else if ((nb = _UTF2_wcrtomb(buf, *s, ps, loc)) ==
(size_t)-1)
/* Invalid character - wcrtomb() sets errno. */
return ((size_t)-1);
@@ -386,9 +369,9 @@
/* Fast path for plain ASCII characters. */
nb = 1;
*dst = *s;
- } else if (len > (size_t)MB_CUR_MAX) {
+ } else if (len > (size_t)UTF2_MB_CUR_MAX) {
/* Enough space to translate in-place. */
- if ((nb = (int)_UTF8_wcrtomb(dst, *s, ps)) < 0) {
+ if ((nb = (int)_UTF2_wcrtomb(dst, *s, ps, loc)) < 0) {
*src = s;
return ((size_t)-1);
}
@@ -396,7 +379,7 @@
/*
* May not be enough space; use temp. buffer.
*/
- if ((nb = (int)_UTF8_wcrtomb(buf, *s, ps)) < 0) {
+ if ((nb = (int)_UTF2_wcrtomb(buf, *s, ps, loc)) < 0) {
*src = s;
return ((size_t)-1);
}