lib/mbrtoc16.c

   1 /* Convert multibyte character and return next 16-bit wide character.
   2    Copyright (C) 2020-2024 Free Software Foundation, Inc.
   3
   4    This file is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU Lesser General Public License as
   6    published by the Free Software Foundation; either version 2.1 of the
   7    License, or (at your option) any later version.
   8
   9    This file is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser General Public License
  15    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  16
  17 /* Written by Bruno Haible <bruno@clisp.org>, 2023.  */
  18
  19 #include <config.h>
  20
  21 /* Specification.  */
  22 #include <uchar.h>
  23
  24 #include <stdlib.h>
  25 #include <wchar.h>
  26
  27 /* We must find room for a two-bytes char16_t in an mbstate_t, without
  28    interfering with the existing use of the mbstate_t in mbrtoc32.  */
  29 static_assert (sizeof (mbstate_t) >= 4);
  30
  31 #if GNULIB_defined_mbstate_t                              /* AIX, IRIX */
  32 /* mbstate_t has at least 4 bytes.  They are used as coded in
  33    gnulib/lib/mbrtowc.c.  */
  34 # define SET_EXTRA_STATE(ps, c16) \
  35     (((char *)(ps))[0] = 8, \
  36      ((char *)(ps))[1] = (unsigned char) ((c16) >> 8), \
  37      ((char *)(ps))[2] = (unsigned char) ((c16) & 0xff))
  38 # define GET_EXTRA_STATE(ps) \
  39     (((char *)(ps))[0] == 8 \
  40      ? ((unsigned char) ((char *)(ps))[1] << 8) | (unsigned char) ((char *)(ps))[2] \
  41      : 0)
  42 # define RESET_EXTRA_STATE(ps) \
  43     (((char *)(ps))[0] = 0)
  44 #elif __GLIBC__ >= 2
  45 /* mbstate_t is defined in <bits/types/__mbstate_t.h>.
  46    For more details, see glibc/iconv/skeleton.c.  */
  47 # define SET_EXTRA_STATE(ps, c16) \
  48     ((ps)->__count |= (c16 << 16))
  49 # define GET_EXTRA_STATE(ps) \
  50     (((unsigned int) (ps)->__count) >> 16)
  51 # define RESET_EXTRA_STATE(ps) \
  52     ((ps)->__count &= 0xffff)
  53 #elif (defined __APPLE__ && defined __MACH__) || defined __FreeBSD__ || defined __NetBSD__ || defined __OpenBSD__ || defined __minix
  54 /* macOS, FreeBSD, NetBSD, OpenBSD, Minix */
  55 /* On macOS, mbstate_t is defined in <machine/_types.h>.
  56    It is an opaque aligned 128-byte struct, of which at most the first
  57    12 bytes are used.
  58    For more details, see the __mbsinit implementations in
  59    Libc-<version>/locale/FreeBSD/
  60    {ascii,none,euc,mskanji,big5,gb2312,gbk,gb18030,utf8,utf2}.c.  */
  61 /* On FreeBSD, mbstate_t is defined in src/sys/sys/_types.h.
  62    It is an opaque aligned 128-byte struct, of which at most the first
  63    12 bytes are used.
  64    For more details, see the __mbsinit implementations in
  65    src/lib/libc/locale/
  66    {ascii,none,euc,mskanji,big5,gb2312,gbk,gb18030,utf8}.c.  */
  67 /* On NetBSD, mbstate_t is defined in src/sys/sys/ansi.h.
  68    It is an opaque aligned 128-byte struct, of which at most the first
  69    28 bytes are used.
  70    For more details, see the *State types in
  71    src/lib/libc/citrus/modules/citrus_*.c
  72    (ignoring citrus_{hz,iso2022,utf7,viqr,zw}.c, since these implement
  73    stateful encodings, not usable as locale encodings).  */
  74 /* On OpenBSD, mbstate_t is defined in src/sys/sys/_types.h.
  75    It is an opaque aligned 128-byte struct, of which at most the first
  76    12 bytes are used.
  77    For more details, see src/lib/libc/citrus/citrus_*.c.  */
  78 /* Minix has borrowed its mbstate_t type and mbrtowc implementation from the
  79    BSDs.  */
  80 # define SET_EXTRA_STATE(ps, c16) \
  81     (((unsigned short *)(ps))[16] = (c16))
  82 # define GET_EXTRA_STATE(ps) \
  83     (((unsigned short *)(ps))[16])
  84 # define RESET_EXTRA_STATE(ps) \
  85     (((unsigned short *)(ps))[16] = 0)
  86 #elif defined __sun                                       /* Solaris */
  87 /* On Solaris, mbstate_t is defined in <wchar_impl.h>.
  88    It is an opaque aligned 24-byte or 32-byte struct, of which at most the first
  89    20 or 28 bytes are used.
  90    For more details on OpenSolaris derivatives, see the *State types in
  91    illumos-gate/usr/src/lib/libc/port/locale/
  92    {none,euc,mskanji,big5,gb2312,gbk,gb18030,utf8}.c.  */
  93 # define SET_EXTRA_STATE(ps, c16) \
  94     (((unsigned short *)(ps))[10] = (c16))
  95 # define GET_EXTRA_STATE(ps) \
  96     (((unsigned short *)(ps))[10])
  97 # define RESET_EXTRA_STATE(ps) \
  98     (((unsigned short *)(ps))[10] = 0)
  99 #elif defined __CYGWIN__
 100 /* On Cygwin, mbstate_t is defined in <sys/_types.h>.
 101    For more details, see newlib/libc/stdlib/mbtowc_r.c and
 102    winsup/cygwin/strfuncs.cc.  */
 103 # define SET_EXTRA_STATE(ps, c16) \
 104     ((ps)->__count = 8, \
 105      (ps)->__value.__wch = (c16))
 106 # define GET_EXTRA_STATE(ps) \
 107     ((ps)->__count == 8 ? (ps)->__value.__wch : 0)
 108 # define RESET_EXTRA_STATE(ps) \
 109     ((ps)->__count = 0)
 110 #elif defined _WIN32 && !defined __CYGWIN__               /* Native Windows.  */
 111 /* MSVC defines 'mbstate_t' as an aligned 8-byte struct.
 112    On mingw, 'mbstate_t' is sometimes defined as 'int', sometimes defined
 113    as an aligned 8-byte struct, of which the first 4 bytes matter.  */
 114 # define SET_EXTRA_STATE(ps, c16) \
 115     (((char *)(ps))[3] = 4, \
 116      ((unsigned short *)(ps))[0] = (c16))
 117 # define GET_EXTRA_STATE(ps) \
 118     (((char *)(ps))[3] == 4 \
 119      ? ((unsigned short *)(ps))[0] \
 120      : 0)
 121 # define RESET_EXTRA_STATE(ps) \
 122     (((char *)(ps))[3] = 0, \
 123      ((unsigned short *)(ps))[0] = 0)
 124 #elif defined __ANDROID__                                 /* Android */
 125 /* Android defines 'mbstate_t' in <bits/mbstate_t.h>.
 126    It is an opaque 4-byte or 8-byte struct.
 127    For more details, see
 128    bionic/libc/private/bionic_mbstate.h
 129    bionic/libc/bionic/mbrtoc32.cpp
 130    bionic/libc/bionic/mbrtoc16.cpp
 131  */
 132 # define SET_EXTRA_STATE(ps, c16) \
 133     (((char *)(ps))[3] = 4, \
 134      ((char *)(ps))[0] = (unsigned char) ((c16) & 0xff), \
 135      ((char *)(ps))[1] = (unsigned char) ((c16) >> 8))
 136 # define GET_EXTRA_STATE(ps) \
 137     (((char *)(ps))[3] == 4 \
 138      ? ((unsigned char) ((char *)(ps))[1] << 8) | (unsigned char) ((char *)(ps))[0] \
 139      : 0)
 140 # define RESET_EXTRA_STATE(ps) \
 141     (((char *)(ps))[0] = ((char *)(ps))[1] = ((char *)(ps))[2] = ((char *)(ps))[3] = 0)
 142 #else
 143 /* This is just a wild guess, for other platforms.  It likely causes unit test
 144    failures.  */
 145 # define SET_EXTRA_STATE(ps, c16) \
 146     (((char *)(ps))[1] = (unsigned char) ((c16) >> 8), \
 147      ((char *)(ps))[2] = (unsigned char) ((c16) & 0xff))
 148 # define GET_EXTRA_STATE(ps) \
 149     (((unsigned char) ((char *)(ps))[1] << 8) | (unsigned char) ((char *)(ps))[2])
 150 # define RESET_EXTRA_STATE(ps) \
 151     (((char *)(ps))[1] = ((char *)(ps))[2] = 0)
 152 #endif
 153
 154 static mbstate_t internal_state;
 155
 156 size_t
 157 mbrtoc16 (char16_t *pwc, const char *s, size_t n, mbstate_t *ps)
 158 #undef mbrtoc16
 159 {
 160   /* It's simpler to handle the case s == NULL upfront, than to worry about
 161      this case later, before every test of pwc and n.  */
 162   if (s == NULL)
 163     {
 164       pwc = NULL;
 165       s = "";
 166       n = 1;
 167     }
 168
 169   if (ps == NULL)
 170     ps = &internal_state;
 171
 172   if (GET_EXTRA_STATE (ps) == 0)
 173     {
 174       if (n == 0)
 175         return (size_t) -2;
 176
 177       char32_t c32;
 178       size_t ret = mbrtoc32 (&c32, s, n, ps);
 179       if (ret == (size_t)(-1) || ret == (size_t)(-2))
 180         ;
 181       else if (ret == (size_t)(-3))
 182         {
 183           /* When mbrtoc32 returns several char32_t values for a single
 184              multibyte character, they are all in the Unicode BMP range.  */
 185           if (c32 >= 0x10000)
 186             abort ();
 187           if (pwc != NULL)
 188             *pwc = c32;
 189         }
 190       else if (c32 < 0x10000)
 191         {
 192           if (pwc != NULL)
 193             *pwc = c32;
 194         }
 195       else
 196         {
 197           if (c32 >= 0x110000)
 198             abort ();
 199           /* Decompose a Unicode character into a high surrogate and a low
 200              surrogate.  */
 201           char16_t surr1 = 0xd800 + ((c32 - 0x10000) >> 10);
 202           char16_t surr2 = 0xdc00 + ((c32 - 0x10000) & 0x3ff);
 203           if (pwc != NULL)
 204             *pwc = surr1;
 205           SET_EXTRA_STATE (ps, surr2);
 206         }
 207       return ret;
 208     }
 209   else
 210     {
 211       if (pwc != NULL)
 212         *pwc = GET_EXTRA_STATE (ps);
 213       RESET_EXTRA_STATE (ps);
 214       return (size_t)(-3);
 215     }
 216 }