locale/FreeBSD/utf8.c.patch

   1 --- utf8.c.bsdnew       2009-11-09 15:05:25.000000000 -0800
   2 +++ utf8.c      2009-11-09 17:35:23.000000000 -0800
   3 @@ -27,6 +27,8 @@
   4  #include <sys/param.h>
   5  __FBSDID("$FreeBSD: src/lib/libc/locale/utf8.c,v 1.16 2007/10/15 09:51:30 ache Exp $");
   6
   7 +#include "xlocale_private.h"
   8 +
   9  #include <errno.h>
  10  #include <limits.h>
  11  #include <runetype.h>
  12 @@ -35,18 +37,18 @@ __FBSDID("$FreeBSD: src/lib/libc/locale/
  13  #include <wchar.h>
  14  #include "mblocal.h"
  15
  16 -extern int __mb_sb_limit;
  17 +#define UTF8_MB_CUR_MAX                6
  18
  19  static size_t  _UTF8_mbrtowc(wchar_t * __restrict, const char * __restrict,
  20 -                   size_t, mbstate_t * __restrict);
  21 -static int     _UTF8_mbsinit(const mbstate_t *);
  22 +                   size_t, mbstate_t * __restrict, locale_t);
  23 +static int     _UTF8_mbsinit(const mbstate_t *, locale_t);
  24  static size_t  _UTF8_mbsnrtowcs(wchar_t * __restrict,
  25                     const char ** __restrict, size_t, size_t,
  26 -                   mbstate_t * __restrict);
  27 +                   mbstate_t * __restrict, locale_t);
  28  static size_t  _UTF8_wcrtomb(char * __restrict, wchar_t,
  29 -                   mbstate_t * __restrict);
  30 +                   mbstate_t * __restrict, locale_t);
  31  static size_t  _UTF8_wcsnrtombs(char * __restrict, const wchar_t ** __restrict,
  32 -                   size_t, size_t, mbstate_t * __restrict);
  33 +                   size_t, size_t, mbstate_t * __restrict, locale_t);
  34
  35  typedef struct {
  36         wchar_t ch;
  37 @@ -54,29 +56,28 @@ typedef struct {
  38         wchar_t lbound;
  39  } _UTF8State;
  40
  41 -int
  42 -_UTF8_init(_RuneLocale *rl)
  43 +__private_extern__ int
  44 +_UTF8_init(struct __xlocale_st_runelocale *xrl)
  45  {
  46
  47 -       __mbrtowc = _UTF8_mbrtowc;
  48 -       __wcrtomb = _UTF8_wcrtomb;
  49 -       __mbsinit = _UTF8_mbsinit;
  50 -       __mbsnrtowcs = _UTF8_mbsnrtowcs;
  51 -       __wcsnrtombs = _UTF8_wcsnrtombs;
  52 -       _CurrentRuneLocale = rl;
  53 -       __mb_cur_max = 6;
  54 +       xrl->__mbrtowc = _UTF8_mbrtowc;
  55 +       xrl->__wcrtomb = _UTF8_wcrtomb;
  56 +       xrl->__mbsinit = _UTF8_mbsinit;
  57 +       xrl->__mbsnrtowcs = _UTF8_mbsnrtowcs;
  58 +       xrl->__wcsnrtombs = _UTF8_wcsnrtombs;
  59 +       xrl->__mb_cur_max = UTF8_MB_CUR_MAX;
  60         /*
  61          * UCS-4 encoding used as the internal representation, so
  62          * slots 0x0080-0x00FF are occuped and must be excluded
  63          * from the single byte ctype by setting the limit.
  64          */
  65 -       __mb_sb_limit = 128;
  66 +       xrl->__mb_sb_limit = 128;
  67
  68         return (0);
  69  }
  70
  71  static int
  72 -_UTF8_mbsinit(const mbstate_t *ps)
  73 +_UTF8_mbsinit(const mbstate_t *ps, locale_t loc)
  74  {
  75
  76         return (ps == NULL || ((const _UTF8State *)ps)->want == 0);
  77 @@ -84,7 +85,7 @@ _UTF8_mbsinit(const mbstate_t *ps)
  78
  79  static size_t
  80  _UTF8_mbrtowc(wchar_t * __restrict pwc, const char * __restrict s, size_t n,
  81 -    mbstate_t * __restrict ps)
  82 +    mbstate_t * __restrict ps, locale_t loc)
  83  {
  84         _UTF8State *us;
  85         int ch, i, mask, want;
  86 @@ -206,7 +207,7 @@ _UTF8_mbrtowc(wchar_t * __restrict pwc,
  87
  88  static size_t
  89  _UTF8_mbsnrtowcs(wchar_t * __restrict dst, const char ** __restrict src,
  90 -    size_t nms, size_t len, mbstate_t * __restrict ps)
  91 +    size_t nms, size_t len, mbstate_t * __restrict ps, locale_t loc)
  92  {
  93         _UTF8State *us;
  94         const char *s;
  95 @@ -236,7 +237,7 @@ _UTF8_mbsnrtowcs(wchar_t * __restrict ds
  96                                  * excluding NUL.
  97                                  */
  98                                 nb = 1;
  99 -                       else if ((nb = _UTF8_mbrtowc(&wc, s, nms, ps)) ==
 100 +                       else if ((nb = _UTF8_mbrtowc(&wc, s, nms, ps, loc)) ==
 101                             (size_t)-1)
 102                                 /* Invalid sequence - mbrtowc() sets errno. */
 103                                 return ((size_t)-1);
 104 @@ -266,7 +267,7 @@ _UTF8_mbsnrtowcs(wchar_t * __restrict ds
 105                          */
 106                         *dst = (wchar_t)*s;
 107                         nb = 1;
 108 -               } else if ((nb = _UTF8_mbrtowc(dst, s, nms, ps)) ==
 109 +               } else if ((nb = _UTF8_mbrtowc(dst, s, nms, ps, loc)) ==
 110                     (size_t)-1) {
 111                         *src = s;
 112                         return ((size_t)-1);
 113 @@ -287,7 +288,7 @@ _UTF8_mbsnrtowcs(wchar_t * __restrict ds
 114  }
 115
 116  static size_t
 117 -_UTF8_wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps)
 118 +_UTF8_wcrtomb(char * __restrict s, wchar_t wc, mbstate_t * __restrict ps, locale_t loc)
 119  {
 120         _UTF8State *us;
 121         unsigned char lead;
 122 @@ -356,7 +357,7 @@ _UTF8_wcrtomb(char * __restrict s, wchar
 123
 124  static size_t
 125  _UTF8_wcsnrtombs(char * __restrict dst, const wchar_t ** __restrict src,
 126 -    size_t nwc, size_t len, mbstate_t * __restrict ps)
 127 +    size_t nwc, size_t len, mbstate_t * __restrict ps, locale_t loc)
 128  {
 129         _UTF8State *us;
 130         char buf[MB_LEN_MAX];
 131 @@ -379,7 +380,7 @@ _UTF8_wcsnrtombs(char * __restrict dst,
 132                         if (0 <= *s && *s < 0x80)
 133                                 /* Fast path for plain ASCII characters. */
 134                                 nb = 1;
 135 -                       else if ((nb = _UTF8_wcrtomb(buf, *s, ps)) ==
 136 +                       else if ((nb = _UTF8_wcrtomb(buf, *s, ps, loc)) ==
 137                             (size_t)-1)
 138                                 /* Invalid character - wcrtomb() sets errno. */
 139                                 return ((size_t)-1);
 140 @@ -396,9 +397,9 @@ _UTF8_wcsnrtombs(char * __restrict dst,
 141                         /* Fast path for plain ASCII characters. */
 142                         nb = 1;
 143                         *dst = *s;
 144 -               } else if (len > (size_t)MB_CUR_MAX) {
 145 +               } else if (len > (size_t)UTF8_MB_CUR_MAX) {
 146                         /* Enough space to translate in-place. */
 147 -                       if ((nb = _UTF8_wcrtomb(dst, *s, ps)) == (size_t)-1) {
 148 +                       if ((nb = _UTF8_wcrtomb(dst, *s, ps, loc)) == (size_t)-1) {
 149                                 *src = s;
 150                                 return ((size_t)-1);
 151                         }
 152 @@ -406,7 +407,7 @@ _UTF8_wcsnrtombs(char * __restrict dst,
 153                         /*
 154                          * May not be enough space; use temp. buffer.
 155                          */
 156 -                       if ((nb = _UTF8_wcrtomb(buf, *s, ps)) == (size_t)-1) {
 157 +                       if ((nb = _UTF8_wcrtomb(buf, *s, ps, loc)) == (size_t)-1) {
 158                                 *src = s;
 159                                 return ((size_t)-1);
 160                         }