lib/iconv_open1.h

   1 /*
   2  * Copyright (C) 1999-2008, 2011, 2018, 2020, 2023 Free Software Foundation, Inc.
   3  * This file is part of the GNU LIBICONV Library.
   4  *
   5  * The GNU LIBICONV Library is free software; you can redistribute it
   6  * and/or modify it under the terms of the GNU Lesser General Public
   7  * License as published by the Free Software Foundation; either version 2.1
   8  * of the License, or (at your option) any later version.
   9  *
  10  * The GNU LIBICONV Library is distributed in the hope that it will be
  11  * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Lesser General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Lesser General Public
  16  * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
  17  * If not, see <https://www.gnu.org/licenses/>.
  18  */
  19
  20 /* Part 1 of iconv_open.
  21    Input: const char* tocode, const char* fromcode.
  22    Output:
  23      unsigned int from_index;
  24      int from_wchar;
  25      unsigned int from_surface;
  26      unsigned int to_index;
  27      int to_wchar;
  28      unsigned int to_surface;
  29      int transliterate;
  30      int discard_ilseq;
  31    Jumps to 'invalid' in case of errror.
  32  */
  33 {
  34   char buf[MAX_WORD_LENGTH+9+9+1];
  35   const char* cp;
  36   char* bp;
  37   const struct alias * ap;
  38   unsigned int count;
  39
  40   from_surface = ICONV_SURFACE_NONE;
  41   to_surface = ICONV_SURFACE_NONE;
  42   transliterate = 0;
  43   discard_ilseq = 0;
  44
  45   /* Before calling aliases_lookup, convert the input string to upper case,
  46    * and check whether it's entirely ASCII (we call gperf with option "-7"
  47    * to achieve a smaller table) and non-empty. If it's not entirely ASCII,
  48    * or if it's too long, it is not a valid encoding name.
  49    */
  50   for (to_wchar = 0;;) {
  51     /* Search tocode in the table. */
  52     for (cp = tocode, bp = buf, count = MAX_WORD_LENGTH+9+9+1; ; cp++, bp++) {
  53       unsigned char c = (unsigned char) *cp;
  54       if (c >= 0x80)
  55         goto invalid;
  56       if (c >= 'a' && c <= 'z')
  57         c -= 'a'-'A';
  58       *bp = c;
  59       if (c == '\0')
  60         break;
  61       if (--count == 0)
  62         goto invalid;
  63     }
  64     for (;;) {
  65       char *sp = bp;
  66       int parsed_translit = 0;
  67       int parsed_ignore = 0;
  68       if (sp-buf > 9 && memcmp(sp-9,"/TRANSLIT",9)==0) {
  69         sp = sp - 9;
  70         parsed_translit = 1;
  71       } else if (sp-buf > 7 && memcmp(sp-7,"/IGNORE",7)==0) {
  72         sp = sp - 7;
  73         parsed_ignore = 1;
  74       }
  75       if (sp > buf && memcmp(sp-1,"/",1) == 0) {
  76         bp = sp - 1;
  77       } else if (sp-buf >= 9 && memcmp(sp-9,"/ZOS_UNIX",9)==0) {
  78         bp = sp - 9;
  79         to_surface = ICONV_SURFACE_EBCDIC_ZOS_UNIX;
  80       } else
  81         break;
  82       *bp = '\0';
  83       if (parsed_translit)
  84         transliterate = 1;
  85       if (parsed_ignore)
  86         discard_ilseq = 1;
  87       break;
  88     }
  89     if (buf[0] == '\0') {
  90       tocode = locale_charset();
  91       /* Avoid an endless loop that could occur when using an older version
  92          of localcharset.c. */
  93       if (tocode[0] == '\0')
  94         goto invalid;
  95       continue;
  96     }
  97     ap = aliases_lookup(buf,bp-buf);
  98     if (ap == NULL) {
  99       ap = aliases2_lookup(buf);
 100       if (ap == NULL)
 101         goto invalid;
 102     }
 103     if (ap->encoding_index == ei_local_char) {
 104       tocode = locale_charset();
 105       /* Avoid an endless loop that could occur when using an older version
 106          of localcharset.c. */
 107       if (tocode[0] == '\0')
 108         goto invalid;
 109       continue;
 110     }
 111     if (ap->encoding_index == ei_local_wchar_t) {
 112       /* On systems which define __STDC_ISO_10646__, wchar_t is Unicode.
 113          This is also the case on native Woe32 systems and Cygwin >= 1.7, where
 114          we know that it is UTF-16.  */
 115 #if (defined _WIN32 && !defined __CYGWIN__) || (defined __CYGWIN__ && CYGWIN_VERSION_DLL_MAJOR >= 1007)
 116       if (sizeof(wchar_t) == 4) {
 117         to_index = ei_ucs4internal;
 118         break;
 119       }
 120       if (sizeof(wchar_t) == 2) {
 121 # if WORDS_LITTLEENDIAN
 122         to_index = ei_utf16le;
 123 # else
 124         to_index = ei_utf16be;
 125 # endif
 126         break;
 127       }
 128 #elif __STDC_ISO_10646__
 129       if (sizeof(wchar_t) == 4) {
 130         to_index = ei_ucs4internal;
 131         break;
 132       }
 133       if (sizeof(wchar_t) == 2) {
 134         to_index = ei_ucs2internal;
 135         break;
 136       }
 137       if (sizeof(wchar_t) == 1) {
 138         to_index = ei_iso8859_1;
 139         break;
 140       }
 141 #endif
 142 #if HAVE_MBRTOWC
 143       to_wchar = 1;
 144       tocode = locale_charset();
 145       continue;
 146 #endif
 147       goto invalid;
 148     }
 149     to_index = ap->encoding_index;
 150     break;
 151   }
 152   for (from_wchar = 0;;) {
 153     /* Search fromcode in the table. */
 154     for (cp = fromcode, bp = buf, count = MAX_WORD_LENGTH+9+9+1; ; cp++, bp++) {
 155       unsigned char c = (unsigned char) *cp;
 156       if (c >= 0x80)
 157         goto invalid;
 158       if (c >= 'a' && c <= 'z')
 159         c -= 'a'-'A';
 160       *bp = c;
 161       if (c == '\0')
 162         break;
 163       if (--count == 0)
 164         goto invalid;
 165     }
 166     for (;;) {
 167       char *sp = bp;
 168       int parsed_translit = 0;
 169       int parsed_ignore = 0;
 170       if (sp-buf > 9 && memcmp(sp-9,"/TRANSLIT",9)==0) {
 171         sp = sp - 9;
 172         parsed_translit = 1;
 173       } else if (sp-buf > 7 && memcmp(sp-7,"/IGNORE",7)==0) {
 174         sp = sp - 7;
 175         parsed_ignore = 1;
 176       }
 177       if (sp > buf && memcmp(sp-1,"/",1) == 0) {
 178         bp = sp - 1;
 179       } else if (sp-buf >= 9 && memcmp(sp-9,"/ZOS_UNIX",9)==0) {
 180         bp = sp - 9;
 181         from_surface = ICONV_SURFACE_EBCDIC_ZOS_UNIX;
 182       } else
 183         break;
 184       *bp = '\0';
 185       if (parsed_translit)
 186         transliterate = 1;
 187       if (parsed_ignore)
 188         discard_ilseq = 1;
 189       break;
 190     }
 191     if (buf[0] == '\0') {
 192       fromcode = locale_charset();
 193       /* Avoid an endless loop that could occur when using an older version
 194          of localcharset.c. */
 195       if (fromcode[0] == '\0')
 196         goto invalid;
 197       continue;
 198     }
 199     ap = aliases_lookup(buf,bp-buf);
 200     if (ap == NULL) {
 201       ap = aliases2_lookup(buf);
 202       if (ap == NULL)
 203         goto invalid;
 204     }
 205     if (ap->encoding_index == ei_local_char) {
 206       fromcode = locale_charset();
 207       /* Avoid an endless loop that could occur when using an older version
 208          of localcharset.c. */
 209       if (fromcode[0] == '\0')
 210         goto invalid;
 211       continue;
 212     }
 213     if (ap->encoding_index == ei_local_wchar_t) {
 214       /* On systems which define __STDC_ISO_10646__, wchar_t is Unicode.
 215          This is also the case on native Woe32 systems and Cygwin >= 1.7, where
 216          we know that it is UTF-16.  */
 217 #if (defined _WIN32 && !defined __CYGWIN__) || (defined __CYGWIN__ && CYGWIN_VERSION_DLL_MAJOR >= 1007)
 218       if (sizeof(wchar_t) == 4) {
 219         from_index = ei_ucs4internal;
 220         break;
 221       }
 222       if (sizeof(wchar_t) == 2) {
 223 # if WORDS_LITTLEENDIAN
 224         from_index = ei_utf16le;
 225 # else
 226         from_index = ei_utf16be;
 227 # endif
 228         break;
 229       }
 230 #elif __STDC_ISO_10646__
 231       if (sizeof(wchar_t) == 4) {
 232         from_index = ei_ucs4internal;
 233         break;
 234       }
 235       if (sizeof(wchar_t) == 2) {
 236         from_index = ei_ucs2internal;
 237         break;
 238       }
 239       if (sizeof(wchar_t) == 1) {
 240         from_index = ei_iso8859_1;
 241         break;
 242       }
 243 #endif
 244 #if HAVE_WCRTOMB
 245       from_wchar = 1;
 246       fromcode = locale_charset();
 247       continue;
 248 #endif
 249       goto invalid;
 250     }
 251     from_index = ap->encoding_index;
 252     break;
 253   }
 254 }