lib/unilbrk/ulc-possible-linebreaks.c

   1 /* Line breaking of strings.
   2    Copyright (C) 2001-2003, 2006-2024 Free Software Foundation, Inc.
   3    Written by Bruno Haible <bruno@clisp.org>, 2001.
   4
   5    This file is free software.
   6    It is dual-licensed under "the GNU LGPLv3+ or the GNU GPLv2+".
   7    You can redistribute it and/or modify it under either
   8      - the terms of the GNU Lesser General Public License as published
   9        by the Free Software Foundation, either version 3, or (at your
  10        option) any later version, or
  11      - the terms of the GNU General Public License as published by the
  12        Free Software Foundation; either version 2, or (at your option)
  13        any later version, or
  14      - the same dual license "the GNU LGPLv3+ or the GNU GPLv2+".
  15
  16    This file is distributed in the hope that it will be useful,
  17    but WITHOUT ANY WARRANTY; without even the implied warranty of
  18    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19    Lesser General Public License and the GNU General Public License
  20    for more details.
  21
  22    You should have received a copy of the GNU Lesser General Public
  23    License and of the GNU General Public License along with this
  24    program.  If not, see <https://www.gnu.org/licenses/>.  */
  25
  26 #include <config.h>
  27
  28 /* Specification.  */
  29 #include "unilbrk.h"
  30
  31 #include <stdlib.h>
  32 #include <string.h>
  33
  34 #include "c-ctype.h"
  35 #include "uniconv.h"
  36 #include "unilbrk/internal.h"
  37 #include "unilbrk/lbrktables.h"
  38 #include "unilbrk/ulc-common.h"
  39
  40 /* Line breaking of a string in an arbitrary encoding.
  41
  42    We convert the input string to Unicode.
  43
  44    The standardized Unicode encodings are UTF-8, UCS-2, UCS-4, UTF-16,
  45    UTF-16BE, UTF-16LE, UTF-7.  UCS-2 supports only characters up to
  46    \U0000FFFF.  UTF-16 and variants support only characters up to
  47    \U0010FFFF.  UTF-7 is way too complex and not supported by glibc-2.1.
  48    UCS-4 specification leaves doubts about endianness and byte order mark.
  49    glibc currently interprets it as big endian without byte order mark,
  50    but this is not backed by an RFC.  So we use UTF-8. It supports
  51    characters up to \U7FFFFFFF and is unambiguously defined.  */
  52
  53 static void
  54 ulc_possible_linebreaks_internal (const char *s, size_t n, const char *encoding,
  55                                   int cr, char *p)
  56 {
  57   if (n > 0)
  58     {
  59       if (is_utf8_encoding (encoding))
  60         u8_possible_linebreaks_loop ((const uint8_t *) s, n, encoding, cr, p);
  61       else
  62         {
  63           /* Convert the string to UTF-8 and build a translation table
  64              from offsets into s to offsets into the translated string.  */
  65           size_t *offsets = (size_t *) malloc (n * sizeof (size_t));
  66
  67           if (offsets != NULL)
  68             {
  69               uint8_t *t;
  70               size_t m;
  71
  72               t = u8_conv_from_encoding (encoding, iconveh_question_mark,
  73                                          s, n, offsets, NULL, &m);
  74               if (t != NULL)
  75                 {
  76                   char *q = (char *) (m > 0 ? malloc (m) : NULL);
  77
  78                   if (m == 0 || q != NULL)
  79                     {
  80                       size_t i;
  81
  82                       /* Determine the possible line breaks of the UTF-8
  83                          string.  */
  84                       u8_possible_linebreaks_loop (t, m, encoding, cr, q);
  85
  86                       /* Translate the result back to the original string.  */
  87                       memset (p, UC_BREAK_PROHIBITED, n);
  88                       for (i = 0; i < n; i++)
  89                         if (offsets[i] != (size_t)(-1))
  90                           p[i] = q[offsets[i]];
  91
  92                       free (q);
  93                       free (t);
  94                       free (offsets);
  95                       return;
  96                     }
  97                   free (t);
  98                 }
  99               free (offsets);
 100             }
 101
 102           /* Impossible to convert.  */
 103 #if C_CTYPE_ASCII
 104           if (is_all_ascii (s, n))
 105             {
 106               /* ASCII is a subset of UTF-8.  */
 107               u8_possible_linebreaks_loop ((const uint8_t *) s, n, encoding, cr, p);
 108               return;
 109             }
 110 #endif
 111           /* We have a non-ASCII string and cannot convert it.
 112              Don't produce line breaks except those already present in the
 113              input string.  All we assume here is that the encoding is
 114              minimally ASCII compatible.  */
 115           {
 116             const char *s_end = s + n;
 117             while (s < s_end)
 118               {
 119                 *p = (*s == '\n'
 120                       ? UC_BREAK_MANDATORY
 121                       : ((cr >= 0
 122                           && *s == '\r'
 123                           && s + 1 < s_end
 124                           && *(s + 1) == '\n')
 125                          ? UC_BREAK_CR_BEFORE_LF
 126                          : UC_BREAK_PROHIBITED));
 127                 s++;
 128                 p++;
 129               }
 130           }
 131         }
 132     }
 133 }
 134
 135 #if defined IN_LIBUNISTRING
 136 /* For backward compatibility with older versions of libunistring.  */
 137
 138 # undef ulc_possible_linebreaks
 139
 140 void
 141 ulc_possible_linebreaks (const char *s, size_t n, const char *encoding,
 142                          char *p)
 143 {
 144   ulc_possible_linebreaks_internal (s, n, encoding, -1, p);
 145 }
 146
 147 #endif
 148
 149 void
 150 ulc_possible_linebreaks_v2 (const char *s, size_t n, const char *encoding,
 151                             char *p)
 152 {
 153   ulc_possible_linebreaks_internal (s, n, encoding, LBP_CR, p);
 154 }
 155
 156
 157 #ifdef TEST
 158
 159 #include <stdio.h>
 160 #include <locale.h>
 161 #include <string.h>
 162
 163 /* Read the contents of an input stream, and return it, terminated with a NUL
 164    byte. */
 165 char *
 166 read_file (FILE *stream)
 167 {
 168 #define BUFSIZE 4096
 169   char *buf = NULL;
 170   int alloc = 0;
 171   int size = 0;
 172   int count;
 173
 174   while (! feof (stream))
 175     {
 176       if (size + BUFSIZE > alloc)
 177         {
 178           alloc = alloc + alloc / 2;
 179           if (alloc < size + BUFSIZE)
 180             alloc = size + BUFSIZE;
 181           buf = realloc (buf, alloc);
 182           if (buf == NULL)
 183             {
 184               fprintf (stderr, "out of memory\n");
 185               exit (1);
 186             }
 187         }
 188       count = fread (buf + size, 1, BUFSIZE, stream);
 189       if (count == 0)
 190         {
 191           if (ferror (stream))
 192             {
 193               perror ("fread");
 194               exit (1);
 195             }
 196         }
 197       else
 198         size += count;
 199     }
 200   buf = realloc (buf, size + 1);
 201   if (buf == NULL)
 202     {
 203       fprintf (stderr, "out of memory\n");
 204       exit (1);
 205     }
 206   buf[size] = '\0';
 207   return buf;
 208 #undef BUFSIZE
 209 }
 210
 211 int
 212 main (int argc, char * argv[])
 213 {
 214   setlocale (LC_CTYPE, "");
 215   if (argc == 1)
 216     {
 217       /* Display all the break opportunities in the input string.  */
 218       char *input = read_file (stdin);
 219       int length = strlen (input);
 220       char *breaks = malloc (length);
 221       int i;
 222
 223       ulc_possible_linebreaks_v2 (input, length, locale_charset (), breaks);
 224
 225       for (i = 0; i < length; i++)
 226         {
 227           switch (breaks[i])
 228             {
 229             case UC_BREAK_POSSIBLE:
 230               putc ('|', stdout);
 231               break;
 232             case UC_BREAK_MANDATORY:
 233               break;
 234             case UC_BREAK_CR_BEFORE_LF:
 235               break;
 236             case UC_BREAK_PROHIBITED:
 237               break;
 238             default:
 239               abort ();
 240             }
 241           putc (input[i], stdout);
 242         }
 243
 244       free (breaks);
 245
 246       return 0;
 247     }
 248   else
 249     return 1;
 250 }
 251
 252 #endif /* TEST */