1 const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.25 2008/04/06 15:18:38 fabiankeil Exp $";
2 /*********************************************************************
4 * File : $Source: /cvsroot/ijbswa/current/urlmatch.c,v $
6 * Purpose : Declares functions to match URLs against URL
9 * Copyright : Written by and Copyright (C) 2001-2003, 2006-2008 the SourceForge
10 * Privoxy team. http://www.privoxy.org/
12 * Based on the Internet Junkbuster originally written
13 * by and Copyright (C) 1997 Anonymous Coders and
14 * Junkbusters Corporation. http://www.junkbusters.com
16 * This program is free software; you can redistribute it
17 * and/or modify it under the terms of the GNU General
18 * Public License as published by the Free Software
19 * Foundation; either version 2 of the License, or (at
20 * your option) any later version.
22 * This program is distributed in the hope that it will
23 * be useful, but WITHOUT ANY WARRANTY; without even the
24 * implied warranty of MERCHANTABILITY or FITNESS FOR A
25 * PARTICULAR PURPOSE. See the GNU General Public
26 * License for more details.
28 * The GNU General Public License should be included with
29 * this file. If not, you can view it at
30 * http://www.gnu.org/copyleft/gpl.html
31 * or write to the Free Software Foundation, Inc., 59
32 * Temple Place - Suite 330, Boston, MA 02111-1307, USA.
35 * $Log: urlmatch.c,v $
36 * Revision 1.25 2008/04/06 15:18:38 fabiankeil
37 * Oh well, rename the --enable-pcre-host-patterns option to
38 * --enable-extended-host-patterns as it's not really PCRE syntax.
40 * Revision 1.24 2008/04/06 14:54:26 fabiankeil
41 * Use PCRE syntax in host patterns when configured
42 * with --enable-pcre-host-patterns.
44 * Revision 1.23 2008/04/05 12:19:20 fabiankeil
45 * Factor compile_host_pattern() out of create_url_spec().
47 * Revision 1.22 2008/03/30 15:02:32 fabiankeil
48 * SZitify unknown_method().
50 * Revision 1.21 2007/12/24 16:34:23 fabiankeil
51 * Band-aid (and micro-optimization) that makes it less likely to run out of
52 * stack space with overly-complex path patterns. Probably masks the problem
53 * reported by Lee in #1856679. Hohoho.
55 * Revision 1.20 2007/09/02 15:31:20 fabiankeil
56 * Move match_portlist() from filter.c to urlmatch.c.
57 * It's used for url matching, not for filtering.
59 * Revision 1.19 2007/09/02 13:42:11 fabiankeil
60 * - Allow port lists in url patterns.
61 * - Ditch unused url_spec member pathlen.
63 * Revision 1.18 2007/07/30 16:42:21 fabiankeil
64 * Move the method check into unknown_method()
65 * and loop through the known methods instead
66 * of using a screen-long OR chain.
68 * Revision 1.17 2007/04/15 16:39:21 fabiankeil
69 * Introduce tags as alternative way to specify which
70 * actions apply to a request. At the moment tags can be
71 * created based on client and server headers.
73 * Revision 1.16 2007/02/13 13:59:24 fabiankeil
74 * Remove redundant log message.
76 * Revision 1.15 2007/01/28 16:11:23 fabiankeil
77 * Accept WebDAV methods for subversion
78 * in parse_http_request(). Closes FR 1581425.
80 * Revision 1.14 2007/01/06 14:23:56 fabiankeil
81 * Fix gcc43 warnings. Mark *csp as immutable
82 * for parse_http_url() and url_match().
83 * Replace a sprintf call with snprintf.
85 * Revision 1.13 2006/12/06 19:50:54 fabiankeil
86 * parse_http_url() now handles intercepted
87 * HTTP request lines as well. Moved parts
88 * of parse_http_url()'s code into
89 * init_domain_components() so that it can
90 * be reused in chat().
92 * Revision 1.12 2006/07/18 14:48:47 david__schmidt
93 * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
94 * with what was really the latest development (the v_3_0_branch branch)
96 * Revision 1.10.2.7 2003/05/17 15:57:24 oes
97 * - parse_http_url now checks memory allocation failure for
98 * duplication of "*" URL and rejects "*something" URLs
100 * - Added a comment to what might look like a bug in
101 * create_url_spec (see !bug #736931)
102 * - Comment cosmetics
104 * Revision 1.10.2.6 2003/05/07 12:39:48 oes
105 * Fix typo: Default port for https URLs is 443, not 143.
106 * Thanks to Scott Tregear for spotting this one.
108 * Revision 1.10.2.5 2003/02/28 13:09:29 oes
109 * Fixed a rare double free condition as per Bug #694713
111 * Revision 1.10.2.4 2003/02/28 12:57:44 oes
112 * Moved freeing of http request structure to its owner
113 * as per Dan Price's observations in Bug #694713
115 * Revision 1.10.2.3 2002/11/12 16:50:40 oes
116 * Fixed memory leak in parse_http_request() reported by Oliver Stoeneberg. Fixes bug #637073
118 * Revision 1.10.2.2 2002/09/25 14:53:15 oes
119 * Added basic support for OPTIONS and TRACE HTTP methods:
120 * parse_http_url now recognizes the "*" URI as well as
121 * the OPTIONS and TRACE method keywords.
123 * Revision 1.10.2.1 2002/06/06 19:06:44 jongfoster
124 * Adding support for proprietary Microsoft WebDAV extensions
126 * Revision 1.10 2002/05/12 21:40:37 jongfoster
127 * - Removing some unused code
129 * Revision 1.9 2002/04/04 00:36:36 gliptak
130 * always use pcre for matching
132 * Revision 1.8 2002/04/03 23:32:47 jongfoster
133 * Fixing memory leak on error
135 * Revision 1.7 2002/03/26 22:29:55 swa
136 * we have a new homepage!
138 * Revision 1.6 2002/03/24 13:25:43 swa
139 * name change related issues
141 * Revision 1.5 2002/03/13 00:27:05 jongfoster
144 * Revision 1.4 2002/03/07 03:46:17 oes
145 * Fixed compiler warnings
147 * Revision 1.3 2002/03/03 14:51:11 oes
148 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
150 * Revision 1.2 2002/01/21 00:14:09 jongfoster
151 * Correcting comment style
152 * Fixing an uninitialized memory bug in create_url_spec()
154 * Revision 1.1 2002/01/17 20:53:46 jongfoster
155 * Moving all our URL and URL pattern parsing code to the same file - it
156 * was scattered around in filters.c, loaders.c and parsers.c.
158 * Providing a single, simple url_match(pattern,url) function - rather than
159 * the 3-line match routine which was repeated all over the place.
161 * Renaming free_url to free_url_spec, since it frees a struct url_spec.
163 * Providing parse_http_url() so that URLs can be parsed without faking a
164 * HTTP request line for parse_http_request() or repeating the parsing
165 * code (both of which were techniques that were actually in use).
167 * Standardizing that struct http_request is used to represent a URL, and
168 * struct url_spec is used to represent a URL pattern. (Before, URLs were
169 * represented as seperate variables and a partially-filled-in url_spec).
172 *********************************************************************/
179 #include <sys/types.h>
187 #if !defined(_WIN32) && !defined(__OS2__)
192 #include "urlmatch.h"
194 #include "miscutil.h"
197 const char urlmatch_h_rcs[] = URLMATCH_H_VERSION;
200 /*********************************************************************
202 * Function : free_http_request
204 * Description : Freez a http_request structure
207 * 1 : http = points to a http_request structure to free
211 *********************************************************************/
212 void free_http_request(struct http_request *http)
221 freez(http->hostport);
224 freez(http->host_ip_addr_str);
225 freez(http->dbuffer);
231 /*********************************************************************
233 * Function : init_domain_components
235 * Description : Splits the domain name so we can compare it
236 * against wildcards. It used to be part of
237 * parse_http_url, but was separated because the
238 * same code is required in chat in case of
239 * intercepted requests.
242 * 1 : http = pointer to the http structure to hold elements.
244 * Returns : JB_ERR_OK on success
245 * JB_ERR_MEMORY on out of memory
246 * JB_ERR_PARSE on malformed command/URL
247 * or >100 domains deep.
249 *********************************************************************/
250 jb_err init_domain_components(struct http_request *http)
252 char *vec[BUFFER_SIZE];
256 http->dbuffer = strdup(http->host);
257 if (NULL == http->dbuffer)
259 return JB_ERR_MEMORY;
262 /* map to lower case */
263 for (p = http->dbuffer; *p ; p++)
265 *p = (char)tolower((int)(unsigned char)*p);
268 /* split the domain name into components */
269 http->dcount = ssplit(http->dbuffer, ".", vec, SZ(vec), 1, 1);
271 if (http->dcount <= 0)
274 * Error: More than SZ(vec) components in domain
275 * or: no components in domain
277 log_error(LOG_LEVEL_ERROR, "More than SZ(vec) components in domain or none at all.");
281 /* save a copy of the pointers in dvec */
282 size = (size_t)http->dcount * sizeof(*http->dvec);
284 http->dvec = (char **)malloc(size);
285 if (NULL == http->dvec)
287 return JB_ERR_MEMORY;
290 memcpy(http->dvec, vec, size);
296 /*********************************************************************
298 * Function : parse_http_url
300 * Description : Parse out the host and port from the URL. Find the
301 * hostname & path, port (if ':'), and/or password (if '@')
304 * 1 : url = URL (or is it URI?) to break down
305 * 2 : http = pointer to the http structure to hold elements.
306 * Will be zeroed before use. Note that this
307 * function sets the http->gpc and http->ver
309 * 3 : csp = Current client state (buffers, headers, etc...)
311 * Returns : JB_ERR_OK on success
312 * JB_ERR_MEMORY on out of memory
313 * JB_ERR_PARSE on malformed command/URL
314 * or >100 domains deep.
316 *********************************************************************/
317 jb_err parse_http_url(const char * url,
318 struct http_request *http,
319 const struct client_state *csp)
321 int host_available = 1; /* A proxy can dream. */
324 * Zero out the results structure
326 memset(http, '\0', sizeof(*http));
330 * Save our initial URL
332 http->url = strdup(url);
333 if (http->url == NULL)
335 return JB_ERR_MEMORY;
340 * Check for * URI. If found, we're done.
342 if (*http->url == '*')
344 if ( NULL == (http->path = strdup("*"))
345 || NULL == (http->hostport = strdup("")) )
347 return JB_ERR_MEMORY;
349 if (http->url[1] != '\0')
358 * Split URL into protocol,hostport,path.
368 return JB_ERR_MEMORY;
371 /* Find the start of the URL in our scratch space */
373 if (strncmpic(url_noproto, "http://", 7) == 0)
378 else if (strncmpic(url_noproto, "https://", 8) == 0)
383 else if (*url_noproto == '/')
386 * Short request line without protocol and host.
387 * Most likely because the client's request
388 * was intercepted and redirected into Privoxy.
399 url_path = strchr(url_noproto, '/');
400 if (url_path != NULL)
405 * NOTE: The following line ignores the path for HTTPS URLS.
406 * This means that you get consistent behaviour if you type a
407 * https URL in and it's parsed by the function. (When the
408 * URL is actually retrieved, SSL hides the path part).
410 http->path = strdup(http->ssl ? "/" : url_path);
412 http->hostport = strdup(url_noproto);
417 * Repair broken HTTP requests that don't contain a path,
418 * or CONNECT requests
420 http->path = strdup("/");
421 http->hostport = strdup(url_noproto);
426 if ( (http->path == NULL)
427 || (http->hostport == NULL))
429 return JB_ERR_MEMORY;
435 /* Without host, there is nothing left to do here */
440 * Split hostport into user/password (ignored), host, port.
447 buf = strdup(http->hostport);
450 return JB_ERR_MEMORY;
453 /* check if url contains username and/or password */
454 host = strchr(buf, '@');
457 /* Contains username/password, skip it and the @ sign. */
462 /* No username or password. */
466 /* check if url contains port */
467 port = strchr(host, ':');
471 /* Terminate hostname and point to start of port string */
473 http->port = atoi(port);
477 /* No port specified. */
478 http->port = (http->ssl ? 443 : 80);
481 http->host = strdup(host);
485 if (http->host == NULL)
487 return JB_ERR_MEMORY;
492 * Split domain name so we can compare it against wildcards
494 return init_domain_components(http);
499 /*********************************************************************
501 * Function : unknown_method
503 * Description : Checks whether a method is unknown.
506 * 1 : method = points to a http method
508 * Returns : TRUE if it's unknown, FALSE otherwise.
510 *********************************************************************/
511 static int unknown_method(const char *method)
513 static const char *known_http_methods[] = {
514 /* Basic HTTP request type */
515 "GET", "HEAD", "POST", "PUT", "DELETE", "OPTIONS", "TRACE", "CONNECT",
516 /* webDAV extensions (RFC2518) */
517 "PROPFIND", "PROPPATCH", "MOVE", "COPY", "MKCOL", "LOCK", "UNLOCK",
519 * Microsoft webDAV extension for Exchange 2000. See:
520 * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html
521 * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp
523 "BCOPY", "BMOVE", "BDELETE", "BPROPFIND", "BPROPPATCH",
525 * Another Microsoft webDAV extension for Exchange 2000. See:
526 * http://systems.cs.colorado.edu/grunwald/MobileComputing/Papers/draft-cohen-gena-p-base-00.txt
527 * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html
528 * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp
530 "SUBSCRIBE", "UNSUBSCRIBE", "NOTIFY", "POLL",
532 * Yet another WebDAV extension, this time for
533 * Web Distributed Authoring and Versioning (RFC3253)
535 "VERSION-CONTROL", "REPORT", "CHECKOUT", "CHECKIN", "UNCHECKOUT",
536 "MKWORKSPACE", "UPDATE", "LABEL", "MERGE", "BASELINE-CONTROL", "MKACTIVITY",
540 for (i = 0; i < SZ(known_http_methods); i++)
542 if (0 == strcmpic(method, known_http_methods[i]))
553 /*********************************************************************
555 * Function : parse_http_request
557 * Description : Parse out the host and port from the URL. Find the
558 * hostname & path, port (if ':'), and/or password (if '@')
561 * 1 : req = HTTP request line to break down
562 * 2 : http = pointer to the http structure to hold elements
563 * 3 : csp = Current client state (buffers, headers, etc...)
565 * Returns : JB_ERR_OK on success
566 * JB_ERR_MEMORY on out of memory
567 * JB_ERR_CGI_PARAMS on malformed command/URL
568 * or >100 domains deep.
570 *********************************************************************/
571 jb_err parse_http_request(const char *req,
572 struct http_request *http,
573 const struct client_state *csp)
576 char *v[10]; /* XXX: Why 10? We should only need three. */
581 memset(http, '\0', sizeof(*http));
586 return JB_ERR_MEMORY;
589 n = ssplit(buf, " \r\n", v, SZ(v), 1, 1);
597 * Fail in case of unknown methods
598 * which we might not handle correctly.
600 * XXX: There should be a config option
601 * to forward requests with unknown methods
602 * anyway. Most of them don't need special
605 if (unknown_method(v[0]))
607 log_error(LOG_LEVEL_ERROR, "Unknown HTTP method detected: %s", v[0]);
612 if (strcmpic(v[0], "CONNECT") == 0)
617 err = parse_http_url(v[1], http, csp);
625 * Copy the details into the structure
627 http->ssl = is_connect;
628 http->cmd = strdup(req);
629 http->gpc = strdup(v[0]);
630 http->ver = strdup(v[2]);
632 if ( (http->cmd == NULL)
633 || (http->gpc == NULL)
634 || (http->ver == NULL) )
637 return JB_ERR_MEMORY;
646 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
647 /*********************************************************************
649 * Function : compile_host_pattern
651 * Description : Parses and compiles a PCRE host pattern..
654 * 1 : url = Target url_spec to be filled in.
655 * 2 : host_pattern = Host pattern to compile.
657 * Returns : JB_ERR_OK - Success
658 * JB_ERR_MEMORY - Out of memory
659 * JB_ERR_PARSE - Cannot parse regex
661 *********************************************************************/
662 static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
665 char rebuf[BUFFER_SIZE];
667 assert(host_pattern);
668 assert(strlen(host_pattern) < sizeof(rebuf) - 2);
670 url->host_regex = zalloc(sizeof(*url->host_regex));
671 if (NULL == url->host_regex)
674 return JB_ERR_MEMORY;
677 snprintf(rebuf, sizeof(rebuf), "%s$", host_pattern);
679 errcode = regcomp(url->host_regex, rebuf,
680 (REG_EXTENDED|REG_NOSUB|REG_ICASE));
684 size_t errlen = regerror(errcode, url->host_regex, rebuf, sizeof(rebuf));
685 if (errlen > (sizeof(rebuf) - (size_t)1))
687 errlen = sizeof(rebuf) - (size_t)1;
689 rebuf[errlen] = '\0';
690 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s", url->spec, rebuf);
702 /*********************************************************************
704 * Function : compile_host_pattern
706 * Description : Parses and "compiles" an old-school host pattern.
709 * 1 : url = Target url_spec to be filled in.
710 * 2 : host_pattern = Host pattern to parse.
712 * Returns : JB_ERR_OK - Success
713 * JB_ERR_MEMORY - Out of memory
714 * JB_ERR_PARSE - Cannot parse regex
716 *********************************************************************/
717 static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
726 if (host_pattern[strlen(host_pattern) - 1] == '.')
728 url->unanchored |= ANCHOR_RIGHT;
730 if (host_pattern[0] == '.')
732 url->unanchored |= ANCHOR_LEFT;
736 * Split domain into components
738 url->dbuffer = strdup(host_pattern);
739 if (NULL == url->dbuffer)
742 return JB_ERR_MEMORY;
748 for (p = url->dbuffer; *p ; p++)
750 *p = (char)tolower((int)(unsigned char)*p);
754 * Split the domain name into components
756 url->dcount = ssplit(url->dbuffer, ".", v, SZ(v), 1, 1);
761 return JB_ERR_MEMORY;
763 else if (url->dcount != 0)
766 * Save a copy of the pointers in dvec
768 size = (size_t)url->dcount * sizeof(*url->dvec);
770 url->dvec = (char **)malloc(size);
771 if (NULL == url->dvec)
774 return JB_ERR_MEMORY;
777 memcpy(url->dvec, v, size);
780 * else dcount == 0 in which case we needn't do anything,
781 * since dvec will never be accessed and the pattern will
788 /*********************************************************************
790 * Function : simple_domaincmp
792 * Description : Domain-wise Compare fqdn's. The comparison is
793 * both left- and right-anchored. The individual
794 * domain names are compared with simplematch().
795 * This is only used by domain_match.
798 * 1 : pv = array of patterns to compare
799 * 2 : fv = array of domain components to compare
800 * 3 : len = length of the arrays (both arrays are the
801 * same length - if they weren't, it couldn't
802 * possibly be a match).
804 * Returns : 0 => domains are equivalent, else no match.
806 *********************************************************************/
807 static int simple_domaincmp(char **pv, char **fv, int len)
811 for (n = 0; n < len; n++)
813 if (simplematch(pv[n], fv[n]))
824 /*********************************************************************
826 * Function : domain_match
828 * Description : Domain-wise Compare fqdn's. Governed by the bimap in
829 * pattern->unachored, the comparison is un-, left-,
830 * right-anchored, or both.
831 * The individual domain names are compared with
835 * 1 : pattern = a domain that may contain a '*' as a wildcard.
836 * 2 : fqdn = domain name against which the patterns are compared.
838 * Returns : 0 => domains are equivalent, else no match.
840 *********************************************************************/
841 static int domain_match(const struct url_spec *pattern, const struct http_request *fqdn)
843 char **pv, **fv; /* vectors */
845 int unanchored = pattern->unanchored & (ANCHOR_RIGHT | ANCHOR_LEFT);
847 plen = pattern->dcount;
852 /* fqdn is too short to match this pattern */
859 if (unanchored == ANCHOR_LEFT)
864 * Convert this into a fully anchored pattern with
865 * the fqdn and pattern the same length
867 fv += (flen - plen); /* flen - plen >= 0 due to check above */
868 return simple_domaincmp(pv, fv, plen);
870 else if (unanchored == 0)
872 /* Fully anchored, check length */
877 return simple_domaincmp(pv, fv, plen);
879 else if (unanchored == ANCHOR_RIGHT)
881 /* Left anchored, ignore all extra in fqdn */
882 return simple_domaincmp(pv, fv, plen);
888 int maxn = flen - plen;
889 for (n = 0; n <= maxn; n++)
891 if (!simple_domaincmp(pv, fv, plen))
896 * Doesn't match from start of fqdn
897 * Try skipping first part of fqdn
905 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
908 /*********************************************************************
910 * Function : create_url_spec
912 * Description : Creates a "url_spec" structure from a string.
913 * When finished, free with free_url_spec().
916 * 1 : url = Target url_spec to be filled in. Will be
918 * 2 : buf = Source pattern, null terminated. NOTE: The
919 * contents of this buffer are destroyed by this
920 * function. If this function succeeds, the
921 * buffer is copied to url->spec. If this
922 * function fails, the contents of the buffer
925 * Returns : JB_ERR_OK - Success
926 * JB_ERR_MEMORY - Out of memory
927 * JB_ERR_PARSE - Cannot parse regex (Detailed message
928 * written to system log)
930 *********************************************************************/
931 jb_err create_url_spec(struct url_spec * url, const char * buf)
936 char rebuf[BUFFER_SIZE];
944 memset(url, '\0', sizeof(*url));
947 * Save a copy of the orignal specification
949 if ((url->spec = strdup(buf)) == NULL)
951 return JB_ERR_MEMORY;
954 /* Is it tag pattern? */
955 if (0 == strncmpic("TAG:", url->spec, 4))
957 if (NULL == (url->tag_regex = zalloc(sizeof(*url->tag_regex))))
960 return JB_ERR_MEMORY;
963 /* buf + 4 to skip "TAG:" */
964 errcode = regcomp(url->tag_regex, buf + 4, (REG_EXTENDED|REG_NOSUB|REG_ICASE));
967 errlen = regerror(errcode, url->preg, rebuf, sizeof(rebuf));
968 if (errlen > (sizeof(rebuf) - 1))
970 errlen = sizeof(rebuf) - 1;
972 rebuf[errlen] = '\0';
973 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s", url->spec, rebuf);
981 /* Only reached for URL patterns */
982 p = strchr(buf, '/');
985 url->path = strdup(p);
986 if (NULL == url->path)
989 return JB_ERR_MEMORY;
999 if (NULL == (url->preg = zalloc(sizeof(*url->preg))))
1002 return JB_ERR_MEMORY;
1005 snprintf(rebuf, sizeof(rebuf), "^(%s)", url->path);
1007 errcode = regcomp(url->preg, rebuf,
1008 (REG_EXTENDED|REG_NOSUB|REG_ICASE));
1011 errlen = regerror(errcode, url->preg, rebuf, sizeof(rebuf));
1013 if (errlen > (sizeof(rebuf) - (size_t)1))
1015 errlen = sizeof(rebuf) - (size_t)1;
1017 rebuf[errlen] = '\0';
1018 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s",
1022 return JB_ERR_PARSE;
1026 p = strchr(buf, ':');
1030 url->port_list = strdup(p);
1031 if (NULL == url->port_list)
1033 return JB_ERR_MEMORY;
1038 url->port_list = NULL;
1043 return compile_host_pattern(url, buf);
1051 /*********************************************************************
1053 * Function : free_url_spec
1055 * Description : Called from the "unloaders". Freez the url
1056 * structure elements.
1059 * 1 : url = pointer to a url_spec structure.
1063 *********************************************************************/
1064 void free_url_spec(struct url_spec *url)
1066 if (url == NULL) return;
1069 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1070 if (url->host_regex)
1072 regfree(url->host_regex);
1073 freez(url->host_regex);
1076 freez(url->dbuffer);
1079 #endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */
1081 freez(url->port_list);
1089 regfree(url->tag_regex);
1090 freez(url->tag_regex);
1095 /*********************************************************************
1097 * Function : url_match
1099 * Description : Compare a URL against a URL pattern.
1102 * 1 : pattern = a URL pattern
1103 * 2 : url = URL to match
1105 * Returns : Nonzero if the URL matches the pattern, else 0.
1107 *********************************************************************/
1108 int url_match(const struct url_spec *pattern,
1109 const struct http_request *url)
1111 /* XXX: these should probably be functions. */
1112 #define PORT_MATCHES ((NULL == pattern->port_list) || match_portlist(pattern->port_list, url->port))
1113 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1114 #define DOMAIN_MATCHES ((NULL == pattern->host_regex) || (0 == regexec(pattern->host_regex, url->host, 0, NULL, 0)))
1116 #define DOMAIN_MATCHES ((NULL == pattern->dbuffer) || (0 == domain_match(pattern, url)))
1118 #define PATH_MATCHES ((NULL == pattern->path) || (0 == regexec(pattern->preg, url->path, 0, NULL, 0)))
1120 if (pattern->tag_regex != NULL)
1122 /* It's a tag pattern and shouldn't be matched against URLs */
1126 return (PORT_MATCHES && DOMAIN_MATCHES && PATH_MATCHES);
1131 /*********************************************************************
1133 * Function : match_portlist
1135 * Description : Check if a given number is covered by a comma
1136 * separated list of numbers and ranges (a,b-c,d,..)
1139 * 1 : portlist = String with list
1140 * 2 : port = port to check
1142 * Returns : 0 => no match
1145 *********************************************************************/
1146 int match_portlist(const char *portlist, int port)
1148 char *min, *max, *next, *portlist_copy;
1150 min = next = portlist_copy = strdup(portlist);
1153 * Zero-terminate first item and remember offset for next
1155 if (NULL != (next = strchr(portlist_copy, (int) ',')))
1161 * Loop through all items, checking for match
1165 if (NULL == (max = strchr(min, (int) '-')))
1168 * No dash, check for equality
1170 if (port == atoi(min))
1172 free(portlist_copy);
1179 * This is a range, so check if between min and max,
1180 * or, if max was omitted, between min and 65K
1183 if(port >= atoi(min) && port <= (atoi(max) ? atoi(max) : 65535))
1185 free(portlist_copy);
1197 * Zero-terminate next item and remember offset for n+1
1199 if ((NULL != next) && (NULL != (next = strchr(next, (int) ','))))
1205 free(portlist_copy);