libguile/regex-posix.c

   1 /*      Copyright (C) 1997, 1998, 1999, 2000, 2001, 2004, 2006, 2007, 2010 Free Software Foundation, Inc.
   2  *
   3  * This library is free software; you can redistribute it and/or
   4  * modify it under the terms of the GNU Lesser General Public License
   5  * as published by the Free Software Foundation; either version 3 of
   6  * the License, or (at your option) any later version.
   7  *
   8  * This library is distributed in the hope that it will be useful, but
   9  * WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  11  * Lesser General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU Lesser General Public
  14  * License along with this library; if not, write to the Free Software
  15  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
  16  * 02110-1301 USA
  17  */
  18
  19
  20 \f
  21
  22 /* regex-posix.c -- POSIX regular expression support.
  23
  24    This code was written against Henry Spencer's famous regex package.
  25    The principal reference for POSIX behavior was the man page for this
  26    library, not the 1003.2 document itself.  Ergo, other `POSIX'
  27    libraries which do not agree with the Spencer implementation may
  28    produce varying behavior.  Sigh. */
  29
  30 #ifdef HAVE_CONFIG_H
  31 #  include <config.h>
  32 #endif
  33
  34 #include <sys/types.h>
  35
  36 #include "libguile/_scm.h"
  37
  38 /* Supposedly, this file is never compiled unless we know we have
  39    POSIX regular expressions.  But we still put this in an #ifdef so
  40    the file is CPP'able (for dependency scanning) even on systems that
  41    don't have a <regex.h> header.  */
  42 #ifdef HAVE_REGCOMP
  43 #ifdef HAVE_REGEX_H
  44 #include <regex.h>
  45 #else
  46 #ifdef HAVE_RXPOSIX_H
  47 #include <rxposix.h>            /* GNU Rx library */
  48 #else
  49 #ifdef HAVE_RX_RXPOSIX_H
  50 #include <rx/rxposix.h>         /* GNU Rx library on Linux */
  51 #endif
  52 #endif
  53 #endif
  54 #endif
  55
  56 #ifdef HAVE_WCHAR_H
  57 #include <wchar.h>
  58 #endif
  59
  60 #include "libguile/async.h"
  61 #include "libguile/smob.h"
  62 #include "libguile/symbols.h"
  63 #include "libguile/vectors.h"
  64 #include "libguile/strports.h"
  65 #include "libguile/ports.h"
  66 #include "libguile/feature.h"
  67 #include "libguile/strings.h"
  68
  69 #include "libguile/validate.h"
  70 #include "libguile/regex-posix.h"
  71
  72 /* This is defined by some regex libraries and omitted by others. */
  73 #ifndef REG_BASIC
  74 #define REG_BASIC 0
  75 #endif
  76
  77 scm_t_bits scm_tc16_regex;
  78
  79 static size_t
  80 regex_free (SCM obj)
  81 {
  82   regfree (SCM_RGX (obj));
  83   scm_gc_free (SCM_RGX (obj), sizeof(regex_t), "regex");
  84   return 0;
  85 }
  86
  87 \f
  88
  89 SCM_SYMBOL (scm_regexp_error_key, "regular-expression-syntax");
  90
  91 static SCM
  92 scm_regexp_error_msg (int regerrno, regex_t *rx)
  93 {
  94   char *errmsg;
  95   int l;
  96
  97   errmsg = scm_malloc (80);
  98   l = regerror (regerrno, rx, errmsg, 80);
  99   if (l > 80)
 100     {
 101       free (errmsg);
 102       errmsg = scm_malloc (l);
 103       regerror (regerrno, rx, errmsg, l);
 104     }
 105   return scm_take_locale_string (errmsg);
 106 }
 107
 108 SCM_DEFINE (scm_regexp_p, "regexp?", 1, 0, 0,
 109             (SCM obj),
 110             "Return @code{#t} if @var{obj} is a compiled regular expression,\n"
 111             "or @code{#f} otherwise.")
 112 #define FUNC_NAME s_scm_regexp_p
 113 {
 114   return scm_from_bool(SCM_RGXP (obj));
 115 }
 116 #undef FUNC_NAME
 117
 118 SCM_DEFINE (scm_make_regexp, "make-regexp", 1, 0, 1,
 119             (SCM pat, SCM flags),
 120             "Compile the regular expression described by @var{pat}, and\n"
 121             "return the compiled regexp structure.  If @var{pat} does not\n"
 122             "describe a legal regular expression, @code{make-regexp} throws\n"
 123             "a @code{regular-expression-syntax} error.\n"
 124             "\n"
 125             "The @var{flags} arguments change the behavior of the compiled\n"
 126             "regular expression.  The following flags may be supplied:\n"
 127             "\n"
 128             "@table @code\n"
 129             "@item regexp/icase\n"
 130             "Consider uppercase and lowercase letters to be the same when\n"
 131             "matching.\n"
 132             "@item regexp/newline\n"
 133             "If a newline appears in the target string, then permit the\n"
 134             "@samp{^} and @samp{$} operators to match immediately after or\n"
 135             "immediately before the newline, respectively.  Also, the\n"
 136             "@samp{.} and @samp{[^...]} operators will never match a newline\n"
 137             "character.  The intent of this flag is to treat the target\n"
 138             "string as a buffer containing many lines of text, and the\n"
 139             "regular expression as a pattern that may match a single one of\n"
 140             "those lines.\n"
 141             "@item regexp/basic\n"
 142             "Compile a basic (``obsolete'') regexp instead of the extended\n"
 143             "(``modern'') regexps that are the default.  Basic regexps do\n"
 144             "not consider @samp{|}, @samp{+} or @samp{?} to be special\n"
 145             "characters, and require the @samp{@{...@}} and @samp{(...)}\n"
 146             "metacharacters to be backslash-escaped (@pxref{Backslash\n"
 147             "Escapes}).  There are several other differences between basic\n"
 148             "and extended regular expressions, but these are the most\n"
 149             "significant.\n"
 150             "@item regexp/extended\n"
 151             "Compile an extended regular expression rather than a basic\n"
 152             "regexp.  This is the default behavior; this flag will not\n"
 153             "usually be needed.  If a call to @code{make-regexp} includes\n"
 154             "both @code{regexp/basic} and @code{regexp/extended} flags, the\n"
 155             "one which comes last will override the earlier one.\n"
 156             "@end table")
 157 #define FUNC_NAME s_scm_make_regexp
 158 {
 159   SCM flag;
 160   regex_t *rx;
 161   int status, cflags;
 162   char *c_pat;
 163
 164   SCM_VALIDATE_STRING (1, pat);
 165   SCM_VALIDATE_REST_ARGUMENT (flags);
 166
 167   /* Examine list of regexp flags.  If REG_BASIC is supplied, then
 168      turn off REG_EXTENDED flag (on by default). */
 169   cflags = REG_EXTENDED;
 170   flag = flags;
 171   while (!scm_is_null (flag))
 172     {
 173       if (scm_to_int (SCM_CAR (flag)) == REG_BASIC)
 174         cflags &= ~REG_EXTENDED;
 175       else
 176         cflags |= scm_to_int (SCM_CAR (flag));
 177       flag = SCM_CDR (flag);
 178     }
 179
 180   rx = scm_gc_malloc_pointerless (sizeof (regex_t), "regex");
 181   c_pat = scm_to_locale_string (pat);
 182   status = regcomp (rx, c_pat,
 183                     /* Make sure they're not passing REG_NOSUB;
 184                        regexp-exec assumes we're getting match data.  */
 185                     cflags & ~REG_NOSUB);
 186   free (c_pat);
 187   if (status != 0)
 188     {
 189       SCM errmsg = scm_regexp_error_msg (status, rx);
 190       scm_gc_free (rx, sizeof(regex_t), "regex");
 191       scm_error_scm (scm_regexp_error_key,
 192                      scm_from_locale_string (FUNC_NAME),
 193                      errmsg,
 194                      SCM_BOOL_F,
 195                      scm_list_1 (pat));
 196
 197       /* never returns */
 198     }
 199   SCM_RETURN_NEWSMOB (scm_tc16_regex, rx);
 200 }
 201 #undef FUNC_NAME
 202
 203 #ifdef HAVE_WCHAR_H
 204 /*
 205  * While regexec does respect the current locale, it returns byte
 206  * offsets instead of character offsets. This routine fixes up the
 207  * regmatch_t structures to refer to characters instead. See "Converting
 208  * a Character" in the libc manual, for more details.
 209  */
 210 static void
 211 fixup_multibyte_match (regmatch_t *matches, int nmatches, char *str)
 212 {
 213   mbstate_t state;
 214   int i;
 215   size_t char_idx, byte_idx;
 216   size_t nbytes = 1; /* just to kick off the for loop */
 217
 218   memset (&state, '\0', sizeof (state));
 219
 220   for (char_idx = byte_idx = 0; nbytes > 0; char_idx++, byte_idx += nbytes)
 221     {
 222       for (i = 0; i < nmatches; ++i)
 223         {
 224           if (matches[i].rm_so == byte_idx)
 225             matches[i].rm_so = char_idx;
 226           if (matches[i].rm_eo == byte_idx)
 227             matches[i].rm_eo = char_idx;
 228         }
 229
 230       nbytes = mbrlen (str + byte_idx, MB_LEN_MAX, &state);
 231     }
 232
 233   if (nbytes >= (size_t) -2)
 234     /* Something is wrong. Shouldn't be possible, as the regex match
 235        succeeded.  */
 236     abort ();
 237 }
 238 #endif
 239
 240 SCM_DEFINE (scm_regexp_exec, "regexp-exec", 2, 2, 0,
 241             (SCM rx, SCM str, SCM start, SCM flags),
 242             "Match the compiled regular expression @var{rx} against\n"
 243             "@code{str}.  If the optional integer @var{start} argument is\n"
 244             "provided, begin matching from that position in the string.\n"
 245             "Return a match structure describing the results of the match,\n"
 246             "or @code{#f} if no match could be found.\n"
 247             "\n"
 248             "The @var{flags} arguments change the matching behavior.\n"
 249             "The following flags may be supplied:\n"
 250             "\n"
 251             "@table @code\n"
 252             "@item regexp/notbol\n"
 253             "Operator @samp{^} always fails (unless @code{regexp/newline}\n"
 254             "is used).  Use this when the beginning of the string should\n"
 255             "not be considered the beginning of a line.\n"
 256             "@item regexp/noteol\n"
 257             "Operator @samp{$} always fails (unless @code{regexp/newline}\n"
 258             "is used).  Use this when the end of the string should not be\n"
 259             "considered the end of a line.\n"
 260             "@end table")
 261 #define FUNC_NAME s_scm_regexp_exec
 262 {
 263   /* We used to have an SCM_DEFER_INTS, and then later an
 264      SCM_CRITICAL_SECTION_START, around the regexec() call.  Can't quite
 265      remember what defer ints was for, but a critical section would only be
 266      wanted now if we think regexec() is not thread-safe.  The posix spec
 267
 268      http://www.opengroup.org/onlinepubs/009695399/functions/regcomp.html
 269
 270      reads like regexec is meant to be both thread safe and reentrant
 271      (mentioning simultaneous use in threads, and in signal handlers).  So
 272      for now believe no protection needed.  */
 273
 274   int status, nmatches, offset;
 275   regmatch_t *matches;
 276   char *c_str;
 277   SCM mvec = SCM_BOOL_F;
 278   SCM substr;
 279
 280   SCM_VALIDATE_RGXP (1, rx);
 281   SCM_VALIDATE_STRING (2, str);
 282
 283   if (SCM_UNBNDP (start))
 284     {
 285       substr = str;
 286       offset = 0;
 287     }
 288   else
 289     {
 290       substr = scm_substring (str, start, SCM_UNDEFINED);
 291       offset = scm_to_int (start);
 292     }
 293
 294   if (SCM_UNBNDP (flags))
 295     flags = SCM_INUM0;
 296
 297   /* re_nsub doesn't account for the `subexpression' representing the
 298      whole regexp, so add 1 to nmatches. */
 299
 300   c_str = scm_to_locale_string (substr);
 301
 302   nmatches = SCM_RGX(rx)->re_nsub + 1;
 303   matches = scm_malloc (sizeof (regmatch_t) * nmatches);
 304   status = regexec (SCM_RGX (rx), c_str, nmatches, matches,
 305                     scm_to_int (flags));
 306
 307 #ifdef HAVE_WCHAR_H
 308   if (!status)
 309     fixup_multibyte_match (matches, nmatches, c_str);
 310 #endif
 311
 312   free (c_str);
 313
 314   if (!status)
 315     {
 316       int i;
 317       /* The match vector must include a cell for the string that was matched,
 318          so add 1. */
 319       mvec = scm_c_make_vector (nmatches + 1, SCM_UNSPECIFIED);
 320       SCM_SIMPLE_VECTOR_SET(mvec,0, str);
 321       for (i = 0; i < nmatches; ++i)
 322         if (matches[i].rm_so == -1)
 323           SCM_SIMPLE_VECTOR_SET(mvec, i+1,
 324                          scm_cons (scm_from_int (-1), scm_from_int (-1)));
 325         else
 326           SCM_SIMPLE_VECTOR_SET(mvec, i+1,
 327                          scm_cons (scm_from_long (matches[i].rm_so + offset),
 328                                    scm_from_long (matches[i].rm_eo + offset)));
 329     }
 330   free (matches);
 331
 332   if (status != 0 && status != REG_NOMATCH)
 333     scm_error_scm (scm_regexp_error_key,
 334                    scm_from_locale_string (FUNC_NAME),
 335                    scm_regexp_error_msg (status, SCM_RGX (rx)),
 336                    SCM_BOOL_F, SCM_BOOL_F);
 337   return mvec;
 338 }
 339 #undef FUNC_NAME
 340
 341 void
 342 scm_init_regex_posix ()
 343 {
 344   scm_tc16_regex = scm_make_smob_type ("regexp", sizeof (regex_t));
 345   scm_set_smob_free (scm_tc16_regex, regex_free);
 346
 347   /* Compilation flags.  */
 348   scm_c_define ("regexp/basic",    scm_from_int (REG_BASIC));
 349   scm_c_define ("regexp/extended", scm_from_int (REG_EXTENDED));
 350   scm_c_define ("regexp/icase",    scm_from_int (REG_ICASE));
 351   scm_c_define ("regexp/newline",  scm_from_int (REG_NEWLINE));
 352
 353   /* Execution flags.  */
 354   scm_c_define ("regexp/notbol", scm_from_int (REG_NOTBOL));
 355   scm_c_define ("regexp/noteol", scm_from_int (REG_NOTEOL));
 356
 357 #include "libguile/regex-posix.x"
 358
 359   scm_add_feature ("regex");
 360 }
 361
 362 /*
 363   Local Variables:
 364   c-file-style: "gnu"
 365   End:
 366 */