lib/regex-quote.c

   1 /* Construct a regular expression from a literal string.
   2    Copyright (C) 1995, 2010-2024 Free Software Foundation, Inc.
   3    Written by Bruno Haible <haible@clisp.cons.org>, 2010.
   4
   5    This program is free software: you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published by
   7    the Free Software Foundation, either version 3 of the License, or
   8    (at your option) any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  17
  18 #include <config.h>
  19
  20 /* Specification.  */
  21 #include "regex-quote.h"
  22
  23 #include <string.h>
  24
  25 #if GNULIB_MCEL_PREFER
  26 # include "mcel.h"
  27 #else
  28 # include "mbuiter.h"
  29 #endif
  30
  31 #include "xalloc.h"
  32
  33 /* Characters that are special in a BRE.  */
  34 static const char bre_special[] = "$^.*[\\";
  35
  36 /* Characters that are special in an ERE.  */
  37 static const char ere_special[] = "$^.*[\\+?{()|";
  38
  39 struct regex_quote_spec
  40 regex_quote_spec_posix (int cflags, bool anchored)
  41 {
  42   struct regex_quote_spec result;
  43
  44   strcpy (result.special, cflags != 0 ? ere_special : bre_special);
  45   result.multibyte = true;
  46   result.anchored = anchored;
  47
  48   return result;
  49 }
  50
  51 /* Syntax bit values, defined in GNU <regex.h>.  We don't include it here,
  52    otherwise this module would need to depend on gnulib module 'regex'.  */
  53 #define RE_BK_PLUS_QM    0x00000002
  54 #define RE_INTERVALS     0x00000200
  55 #define RE_LIMITED_OPS   0x00000400
  56 #define RE_NEWLINE_ALT   0x00000800
  57 #define RE_NO_BK_BRACES  0x00001000
  58 #define RE_NO_BK_PARENS  0x00002000
  59 #define RE_NO_BK_VBAR    0x00008000
  60
  61 struct regex_quote_spec
  62 regex_quote_spec_gnu (unsigned long /*reg_syntax_t*/ syntax, bool anchored)
  63 {
  64   struct regex_quote_spec result;
  65   char *p;
  66
  67   p = result.special;
  68   memcpy (p, bre_special, sizeof (bre_special) - 1);
  69   p += sizeof (bre_special) - 1;
  70   if ((syntax & RE_LIMITED_OPS) == 0 && (syntax & RE_BK_PLUS_QM) == 0)
  71     {
  72       *p++ = '+';
  73       *p++ = '?';
  74     }
  75   if ((syntax & RE_INTERVALS) != 0 && (syntax & RE_NO_BK_BRACES) != 0)
  76     {
  77       *p++ = '{';
  78       *p++ = '}';
  79     }
  80   if ((syntax & RE_NO_BK_PARENS) != 0)
  81     {
  82       *p++ = '(';
  83       *p++ = ')';
  84     }
  85   if ((syntax & RE_LIMITED_OPS) == 0 && (syntax & RE_NO_BK_VBAR) != 0)
  86     *p++ = '|';
  87   if ((syntax & RE_NEWLINE_ALT) != 0)
  88     *p++ = '\n';
  89   *p = '\0';
  90
  91   result.multibyte = true;
  92   result.anchored = anchored;
  93
  94   return result;
  95 }
  96
  97 /* Characters that are special in a PCRE.  */
  98 static const char pcre_special[] = "$^.*[]\\+?{}()|";
  99
 100 /* Options bit values, defined in <pcre.h>.  We don't include it here, because
 101    it is not a standard header.  */
 102 #define PCRE_ANCHORED 0x00000010
 103 #define PCRE_EXTENDED 0x00000008
 104
 105 struct regex_quote_spec
 106 regex_quote_spec_pcre (int options, bool anchored)
 107 {
 108   struct regex_quote_spec result;
 109   char *p;
 110
 111   p = result.special;
 112   memcpy (p, pcre_special, sizeof (pcre_special) - 1);
 113   p += sizeof (pcre_special) - 1;
 114   if (options & PCRE_EXTENDED)
 115     {
 116       *p++ = ' ';
 117       *p++ = '\t';
 118       *p++ = '\n';
 119       *p++ = '\v';
 120       *p++ = '\f';
 121       *p++ = '\r';
 122       *p++ = '#';
 123     }
 124   *p = '\0';
 125
 126   /* PCRE regular expressions consist of UTF-8 characters of options contains
 127      PCRE_UTF8 and of single bytes otherwise.  */
 128   result.multibyte = false;
 129   /* If options contains PCRE_ANCHORED, the anchoring is implicit.  */
 130   result.anchored = (options & PCRE_ANCHORED ? 0 : anchored);
 131
 132   return result;
 133 }
 134
 135 size_t
 136 regex_quote_length (const char *string, const struct regex_quote_spec *spec)
 137 {
 138   const char *special = spec->special;
 139   size_t length;
 140
 141   length = 0;
 142   if (spec->anchored)
 143     length += 2; /* for '^' at the beginning and '$' at the end */
 144   if (spec->multibyte)
 145     {
 146 #if GNULIB_MCEL_PREFER
 147       char const *iter = string;
 148       for (mcel_t g; *iter; iter += g.len)
 149         {
 150           g = mcel_scanz (iter);
 151           /* We know that special contains only ASCII characters.  */
 152           length += g.len == 1 && strchr (special, *iter);
 153         }
 154       length += iter - string;
 155 #else
 156       mbui_iterator_t iter;
 157
 158       for (mbui_init (iter, string); mbui_avail (iter); mbui_advance (iter))
 159         {
 160           /* We know that special contains only ASCII characters.  */
 161           if (mb_len (mbui_cur (iter)) == 1
 162               && strchr (special, * mbui_cur_ptr (iter)))
 163             length += 1;
 164           length += mb_len (mbui_cur (iter));
 165         }
 166 #endif
 167     }
 168   else
 169     {
 170       const char *iter;
 171
 172       for (iter = string; *iter != '\0'; iter++)
 173         {
 174           if (strchr (special, *iter))
 175             length += 1;
 176           length += 1;
 177         }
 178     }
 179
 180   return length;
 181 }
 182
 183 char *
 184 regex_quote_copy (char *p, const char *string, const struct regex_quote_spec *spec)
 185 {
 186   const char *special = spec->special;
 187
 188   if (spec->anchored)
 189     *p++ = '^';
 190   if (spec->multibyte)
 191     {
 192 #if GNULIB_MCEL_PREFER
 193       for (char const *iter = string; *iter; )
 194         {
 195           mcel_t g = mcel_scanz (iter);
 196           *p = '\\';
 197           /* We know that special contains only ASCII characters.  */
 198           p += g.len == 1 && strchr (special, *iter);
 199           p = mempcpy (p, iter, g.len);
 200           iter += g.len;
 201         }
 202 #else
 203       mbui_iterator_t iter;
 204
 205       for (mbui_init (iter, string); mbui_avail (iter); mbui_advance (iter))
 206         {
 207           /* We know that special contains only ASCII characters.  */
 208           if (mb_len (mbui_cur (iter)) == 1
 209               && strchr (special, * mbui_cur_ptr (iter)))
 210             *p++ = '\\';
 211           memcpy (p, mbui_cur_ptr (iter), mb_len (mbui_cur (iter)));
 212           p += mb_len (mbui_cur (iter));
 213         }
 214 #endif
 215     }
 216   else
 217     {
 218       const char *iter;
 219
 220       for (iter = string; *iter != '\0'; iter++)
 221         {
 222           if (strchr (special, *iter))
 223             *p++ = '\\';
 224           *p++ = *iter++;
 225         }
 226     }
 227   if (spec->anchored)
 228     *p++ = '$';
 229
 230   return p;
 231 }
 232
 233 char *
 234 regex_quote (const char *string, const struct regex_quote_spec *spec)
 235 {
 236   size_t length = regex_quote_length (string, spec);
 237   char *result = XNMALLOC (length + 1, char);
 238   char *p;
 239
 240   p = result;
 241   p = regex_quote_copy (p, string, spec);
 242   *p = '\0';
 243   return result;
 244 }