libguile/regex-posix.c

   1 /*      Copyright (C) 1997, 1998, 1999, 2000, 2001, 2004, 2006, 2007 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 #include "libguile/async.h"
  57 #include "libguile/smob.h"
  58 #include "libguile/symbols.h"
  59 #include "libguile/vectors.h"
  60 #include "libguile/strports.h"
  61 #include "libguile/ports.h"
  62 #include "libguile/feature.h"
  63 #include "libguile/strings.h"
  64
  65 #include "libguile/validate.h"
  66 #include "libguile/regex-posix.h"
  67
  68 /* This is defined by some regex libraries and omitted by others. */
  69 #ifndef REG_BASIC
  70 #define REG_BASIC 0
  71 #endif
  72
  73 scm_t_bits scm_tc16_regex;
  74
  75 static size_t
  76 regex_free (SCM obj)
  77 {
  78   regfree (SCM_RGX (obj));
  79   scm_gc_free (SCM_RGX (obj), sizeof(regex_t), "regex");
  80   return 0;
  81 }
  82
  83 \f
  84
  85 SCM_SYMBOL (scm_regexp_error_key, "regular-expression-syntax");
  86
  87 static SCM
  88 scm_regexp_error_msg (int regerrno, regex_t *rx)
  89 {
  90   char *errmsg;
  91   int l;
  92
  93   errmsg = scm_malloc (80);
  94   l = regerror (regerrno, rx, errmsg, 80);
  95   if (l > 80)
  96     {
  97       free (errmsg);
  98       errmsg = scm_malloc (l);
  99       regerror (regerrno, rx, errmsg, l);
 100     }
 101   return scm_take_locale_string (errmsg);
 102 }
 103
 104 SCM_DEFINE (scm_regexp_p, "regexp?", 1, 0, 0,
 105             (SCM obj),
 106             "Return @code{#t} if @var{obj} is a compiled regular expression,\n"
 107             "or @code{#f} otherwise.")
 108 #define FUNC_NAME s_scm_regexp_p
 109 {
 110   return scm_from_bool(SCM_RGXP (obj));
 111 }
 112 #undef FUNC_NAME
 113
 114 SCM_DEFINE (scm_make_regexp, "make-regexp", 1, 0, 1,
 115             (SCM pat, SCM flags),
 116             "Compile the regular expression described by @var{pat}, and\n"
 117             "return the compiled regexp structure.  If @var{pat} does not\n"
 118             "describe a legal regular expression, @code{make-regexp} throws\n"
 119             "a @code{regular-expression-syntax} error.\n"
 120             "\n"
 121             "The @var{flags} arguments change the behavior of the compiled\n"
 122             "regular expression.  The following flags may be supplied:\n"
 123             "\n"
 124             "@table @code\n"
 125             "@item regexp/icase\n"
 126             "Consider uppercase and lowercase letters to be the same when\n"
 127             "matching.\n"
 128             "@item regexp/newline\n"
 129             "If a newline appears in the target string, then permit the\n"
 130             "@samp{^} and @samp{$} operators to match immediately after or\n"
 131             "immediately before the newline, respectively.  Also, the\n"
 132             "@samp{.} and @samp{[^...]} operators will never match a newline\n"
 133             "character.  The intent of this flag is to treat the target\n"
 134             "string as a buffer containing many lines of text, and the\n"
 135             "regular expression as a pattern that may match a single one of\n"
 136             "those lines.\n"
 137             "@item regexp/basic\n"
 138             "Compile a basic (``obsolete'') regexp instead of the extended\n"
 139             "(``modern'') regexps that are the default.  Basic regexps do\n"
 140             "not consider @samp{|}, @samp{+} or @samp{?} to be special\n"
 141             "characters, and require the @samp{@{...@}} and @samp{(...)}\n"
 142             "metacharacters to be backslash-escaped (@pxref{Backslash\n"
 143             "Escapes}).  There are several other differences between basic\n"
 144             "and extended regular expressions, but these are the most\n"
 145             "significant.\n"
 146             "@item regexp/extended\n"
 147             "Compile an extended regular expression rather than a basic\n"
 148             "regexp.  This is the default behavior; this flag will not\n"
 149             "usually be needed.  If a call to @code{make-regexp} includes\n"
 150             "both @code{regexp/basic} and @code{regexp/extended} flags, the\n"
 151             "one which comes last will override the earlier one.\n"
 152             "@end table")
 153 #define FUNC_NAME s_scm_make_regexp
 154 {
 155   SCM flag;
 156   regex_t *rx;
 157   int status, cflags;
 158   char *c_pat;
 159
 160   SCM_VALIDATE_STRING (1, pat);
 161   SCM_VALIDATE_REST_ARGUMENT (flags);
 162
 163   /* Examine list of regexp flags.  If REG_BASIC is supplied, then
 164      turn off REG_EXTENDED flag (on by default). */
 165   cflags = REG_EXTENDED;
 166   flag = flags;
 167   while (!scm_is_null (flag))
 168     {
 169       if (scm_to_int (SCM_CAR (flag)) == REG_BASIC)
 170         cflags &= ~REG_EXTENDED;
 171       else
 172         cflags |= scm_to_int (SCM_CAR (flag));
 173       flag = SCM_CDR (flag);
 174     }
 175
 176   rx = scm_gc_malloc (sizeof(regex_t), "regex");
 177   c_pat = scm_to_locale_string (pat);
 178   status = regcomp (rx, c_pat,
 179                     /* Make sure they're not passing REG_NOSUB;
 180                        regexp-exec assumes we're getting match data.  */
 181                     cflags & ~REG_NOSUB);
 182   free (c_pat);
 183   if (status != 0)
 184     {
 185       SCM errmsg = scm_regexp_error_msg (status, rx);
 186       scm_gc_free (rx, sizeof(regex_t), "regex");
 187       scm_error_scm (scm_regexp_error_key,
 188                      scm_from_locale_string (FUNC_NAME),
 189                      errmsg,
 190                      SCM_BOOL_F,
 191                      scm_list_1 (pat));
 192
 193       /* never returns */
 194     }
 195   SCM_RETURN_NEWSMOB (scm_tc16_regex, rx);
 196 }
 197 #undef FUNC_NAME
 198
 199 SCM_DEFINE (scm_regexp_exec, "regexp-exec", 2, 2, 0,
 200             (SCM rx, SCM str, SCM start, SCM flags),
 201             "Match the compiled regular expression @var{rx} against\n"
 202             "@code{str}.  If the optional integer @var{start} argument is\n"
 203             "provided, begin matching from that position in the string.\n"
 204             "Return a match structure describing the results of the match,\n"
 205             "or @code{#f} if no match could be found.\n"
 206             "\n"
 207             "The @var{flags} arguments change the matching behavior.\n"
 208             "The following flags may be supplied:\n"
 209             "\n"
 210             "@table @code\n"
 211             "@item regexp/notbol\n"
 212             "Operator @samp{^} always fails (unless @code{regexp/newline}\n"
 213             "is used).  Use this when the beginning of the string should\n"
 214             "not be considered the beginning of a line.\n"
 215             "@item regexp/noteol\n"
 216             "Operator @samp{$} always fails (unless @code{regexp/newline}\n"
 217             "is used).  Use this when the end of the string should not be\n"
 218             "considered the end of a line.\n"
 219             "@end table")
 220 #define FUNC_NAME s_scm_regexp_exec
 221 {
 222   /* We used to have an SCM_DEFER_INTS, and then later an
 223      SCM_CRITICAL_SECTION_START, around the regexec() call.  Can't quite
 224      remember what defer ints was for, but a critical section would only be
 225      wanted now if we think regexec() is not thread-safe.  The posix spec
 226
 227      http://www.opengroup.org/onlinepubs/009695399/functions/regcomp.html
 228
 229      reads like regexec is meant to be both thread safe and reentrant
 230      (mentioning simultaneous use in threads, and in signal handlers).  So
 231      for now believe no protection needed.  */
 232
 233   int status, nmatches, offset;
 234   regmatch_t *matches;
 235   char *c_str;
 236   SCM mvec = SCM_BOOL_F;
 237   SCM substr;
 238
 239   SCM_VALIDATE_RGXP (1, rx);
 240   SCM_VALIDATE_STRING (2, str);
 241
 242   if (SCM_UNBNDP (start))
 243     {
 244       substr = str;
 245       offset = 0;
 246     }
 247   else
 248     {
 249       substr = scm_substring (str, start, SCM_UNDEFINED);
 250       offset = scm_to_int (start);
 251     }
 252
 253   if (SCM_UNBNDP (flags))
 254     flags = SCM_INUM0;
 255
 256   /* re_nsub doesn't account for the `subexpression' representing the
 257      whole regexp, so add 1 to nmatches. */
 258
 259   nmatches = SCM_RGX(rx)->re_nsub + 1;
 260   matches = scm_malloc (sizeof (regmatch_t) * nmatches);
 261   c_str = scm_to_locale_string (substr);
 262   status = regexec (SCM_RGX (rx), c_str, nmatches, matches,
 263                     scm_to_int (flags));
 264   free (c_str);
 265
 266   if (!status)
 267     {
 268       int i;
 269       /* The match vector must include a cell for the string that was matched,
 270          so add 1. */
 271       mvec = scm_c_make_vector (nmatches + 1, SCM_UNSPECIFIED);
 272       SCM_SIMPLE_VECTOR_SET(mvec,0, str);
 273       for (i = 0; i < nmatches; ++i)
 274         if (matches[i].rm_so == -1)
 275           SCM_SIMPLE_VECTOR_SET(mvec, i+1,
 276                          scm_cons (scm_from_int (-1), scm_from_int (-1)));
 277         else
 278           SCM_SIMPLE_VECTOR_SET(mvec, i+1,
 279                          scm_cons (scm_from_long (matches[i].rm_so + offset),
 280                                    scm_from_long (matches[i].rm_eo + offset)));
 281     }
 282   free (matches);
 283
 284   if (status != 0 && status != REG_NOMATCH)
 285     scm_error_scm (scm_regexp_error_key,
 286                    scm_from_locale_string (FUNC_NAME),
 287                    scm_regexp_error_msg (status, SCM_RGX (rx)),
 288                    SCM_BOOL_F, SCM_BOOL_F);
 289   return mvec;
 290 }
 291 #undef FUNC_NAME
 292
 293 void
 294 scm_init_regex_posix ()
 295 {
 296   scm_tc16_regex = scm_make_smob_type ("regexp", sizeof (regex_t));
 297   scm_set_smob_free (scm_tc16_regex, regex_free);
 298
 299   /* Compilation flags.  */
 300   scm_c_define ("regexp/basic",    scm_from_int (REG_BASIC));
 301   scm_c_define ("regexp/extended", scm_from_int (REG_EXTENDED));
 302   scm_c_define ("regexp/icase",    scm_from_int (REG_ICASE));
 303   scm_c_define ("regexp/newline",  scm_from_int (REG_NEWLINE));
 304
 305   /* Execution flags.  */
 306   scm_c_define ("regexp/notbol", scm_from_int (REG_NOTBOL));
 307   scm_c_define ("regexp/noteol", scm_from_int (REG_NOTEOL));
 308
 309 #include "libguile/regex-posix.x"
 310
 311   scm_add_feature ("regex");
 312 }
 313
 314 /*
 315   Local Variables:
 316   c-file-style: "gnu"
 317   End:
 318 */