1 const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.24 2008/04/06 14:54:26 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 1.24 2008/04/06 14:54:26 fabiankeil
37 * Use PCRE syntax in host patterns when configured
38 * with --enable-pcre-host-patterns.
40 * Revision 1.23 2008/04/05 12:19:20 fabiankeil
41 * Factor compile_host_pattern() out of create_url_spec().
43 * Revision 1.22 2008/03/30 15:02:32 fabiankeil
44 * SZitify unknown_method().
46 * Revision 1.21 2007/12/24 16:34:23 fabiankeil
47 * Band-aid (and micro-optimization) that makes it less likely to run out of
48 * stack space with overly-complex path patterns. Probably masks the problem
49 * reported by Lee in #1856679. Hohoho.
51 * Revision 1.20 2007/09/02 15:31:20 fabiankeil
52 * Move match_portlist() from filter.c to urlmatch.c.
53 * It's used for url matching, not for filtering.
55 * Revision 1.19 2007/09/02 13:42:11 fabiankeil
56 * - Allow port lists in url patterns.
57 * - Ditch unused url_spec member pathlen.
59 * Revision 1.18 2007/07/30 16:42:21 fabiankeil
60 * Move the method check into unknown_method()
61 * and loop through the known methods instead
62 * of using a screen-long OR chain.
64 * Revision 1.17 2007/04/15 16:39:21 fabiankeil
65 * Introduce tags as alternative way to specify which
66 * actions apply to a request. At the moment tags can be
67 * created based on client and server headers.
69 * Revision 1.16 2007/02/13 13:59:24 fabiankeil
70 * Remove redundant log message.
72 * Revision 1.15 2007/01/28 16:11:23 fabiankeil
73 * Accept WebDAV methods for subversion
74 * in parse_http_request(). Closes FR 1581425.
76 * Revision 1.14 2007/01/06 14:23:56 fabiankeil
77 * Fix gcc43 warnings. Mark *csp as immutable
78 * for parse_http_url() and url_match().
79 * Replace a sprintf call with snprintf.
81 * Revision 1.13 2006/12/06 19:50:54 fabiankeil
82 * parse_http_url() now handles intercepted
83 * HTTP request lines as well. Moved parts
84 * of parse_http_url()'s code into
85 * init_domain_components() so that it can
86 * be reused in chat().
88 * Revision 1.12 2006/07/18 14:48:47 david__schmidt
89 * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
90 * with what was really the latest development (the v_3_0_branch branch)
92 * Revision 1.10.2.7 2003/05/17 15:57:24 oes
93 * - parse_http_url now checks memory allocation failure for
94 * duplication of "*" URL and rejects "*something" URLs
96 * - Added a comment to what might look like a bug in
97 * create_url_spec (see !bug #736931)
100 * Revision 1.10.2.6 2003/05/07 12:39:48 oes
101 * Fix typo: Default port for https URLs is 443, not 143.
102 * Thanks to Scott Tregear for spotting this one.
104 * Revision 1.10.2.5 2003/02/28 13:09:29 oes
105 * Fixed a rare double free condition as per Bug #694713
107 * Revision 1.10.2.4 2003/02/28 12:57:44 oes
108 * Moved freeing of http request structure to its owner
109 * as per Dan Price's observations in Bug #694713
111 * Revision 1.10.2.3 2002/11/12 16:50:40 oes
112 * Fixed memory leak in parse_http_request() reported by Oliver Stoeneberg. Fixes bug #637073
114 * Revision 1.10.2.2 2002/09/25 14:53:15 oes
115 * Added basic support for OPTIONS and TRACE HTTP methods:
116 * parse_http_url now recognizes the "*" URI as well as
117 * the OPTIONS and TRACE method keywords.
119 * Revision 1.10.2.1 2002/06/06 19:06:44 jongfoster
120 * Adding support for proprietary Microsoft WebDAV extensions
122 * Revision 1.10 2002/05/12 21:40:37 jongfoster
123 * - Removing some unused code
125 * Revision 1.9 2002/04/04 00:36:36 gliptak
126 * always use pcre for matching
128 * Revision 1.8 2002/04/03 23:32:47 jongfoster
129 * Fixing memory leak on error
131 * Revision 1.7 2002/03/26 22:29:55 swa
132 * we have a new homepage!
134 * Revision 1.6 2002/03/24 13:25:43 swa
135 * name change related issues
137 * Revision 1.5 2002/03/13 00:27:05 jongfoster
140 * Revision 1.4 2002/03/07 03:46:17 oes
141 * Fixed compiler warnings
143 * Revision 1.3 2002/03/03 14:51:11 oes
144 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
146 * Revision 1.2 2002/01/21 00:14:09 jongfoster
147 * Correcting comment style
148 * Fixing an uninitialized memory bug in create_url_spec()
150 * Revision 1.1 2002/01/17 20:53:46 jongfoster
151 * Moving all our URL and URL pattern parsing code to the same file - it
152 * was scattered around in filters.c, loaders.c and parsers.c.
154 * Providing a single, simple url_match(pattern,url) function - rather than
155 * the 3-line match routine which was repeated all over the place.
157 * Renaming free_url to free_url_spec, since it frees a struct url_spec.
159 * Providing parse_http_url() so that URLs can be parsed without faking a
160 * HTTP request line for parse_http_request() or repeating the parsing
161 * code (both of which were techniques that were actually in use).
163 * Standardizing that struct http_request is used to represent a URL, and
164 * struct url_spec is used to represent a URL pattern. (Before, URLs were
165 * represented as seperate variables and a partially-filled-in url_spec).
168 *********************************************************************/
175 #include <sys/types.h>
183 #if !defined(_WIN32) && !defined(__OS2__)
188 #include "urlmatch.h"
190 #include "miscutil.h"
193 const char urlmatch_h_rcs[] = URLMATCH_H_VERSION;
196 /*********************************************************************
198 * Function : free_http_request
200 * Description : Freez a http_request structure
203 * 1 : http = points to a http_request structure to free
207 *********************************************************************/
208 void free_http_request(struct http_request *http)
217 freez(http->hostport);
220 freez(http->host_ip_addr_str);
221 freez(http->dbuffer);
227 /*********************************************************************
229 * Function : init_domain_components
231 * Description : Splits the domain name so we can compare it
232 * against wildcards. It used to be part of
233 * parse_http_url, but was separated because the
234 * same code is required in chat in case of
235 * intercepted requests.
238 * 1 : http = pointer to the http structure to hold elements.
240 * Returns : JB_ERR_OK on success
241 * JB_ERR_MEMORY on out of memory
242 * JB_ERR_PARSE on malformed command/URL
243 * or >100 domains deep.
245 *********************************************************************/
246 jb_err init_domain_components(struct http_request *http)
248 char *vec[BUFFER_SIZE];
252 http->dbuffer = strdup(http->host);
253 if (NULL == http->dbuffer)
255 return JB_ERR_MEMORY;
258 /* map to lower case */
259 for (p = http->dbuffer; *p ; p++)
261 *p = (char)tolower((int)(unsigned char)*p);
264 /* split the domain name into components */
265 http->dcount = ssplit(http->dbuffer, ".", vec, SZ(vec), 1, 1);
267 if (http->dcount <= 0)
270 * Error: More than SZ(vec) components in domain
271 * or: no components in domain
273 log_error(LOG_LEVEL_ERROR, "More than SZ(vec) components in domain or none at all.");
277 /* save a copy of the pointers in dvec */
278 size = (size_t)http->dcount * sizeof(*http->dvec);
280 http->dvec = (char **)malloc(size);
281 if (NULL == http->dvec)
283 return JB_ERR_MEMORY;
286 memcpy(http->dvec, vec, size);
292 /*********************************************************************
294 * Function : parse_http_url
296 * Description : Parse out the host and port from the URL. Find the
297 * hostname & path, port (if ':'), and/or password (if '@')
300 * 1 : url = URL (or is it URI?) to break down
301 * 2 : http = pointer to the http structure to hold elements.
302 * Will be zeroed before use. Note that this
303 * function sets the http->gpc and http->ver
305 * 3 : csp = Current client state (buffers, headers, etc...)
307 * Returns : JB_ERR_OK on success
308 * JB_ERR_MEMORY on out of memory
309 * JB_ERR_PARSE on malformed command/URL
310 * or >100 domains deep.
312 *********************************************************************/
313 jb_err parse_http_url(const char * url,
314 struct http_request *http,
315 const struct client_state *csp)
317 int host_available = 1; /* A proxy can dream. */
320 * Zero out the results structure
322 memset(http, '\0', sizeof(*http));
326 * Save our initial URL
328 http->url = strdup(url);
329 if (http->url == NULL)
331 return JB_ERR_MEMORY;
336 * Check for * URI. If found, we're done.
338 if (*http->url == '*')
340 if ( NULL == (http->path = strdup("*"))
341 || NULL == (http->hostport = strdup("")) )
343 return JB_ERR_MEMORY;
345 if (http->url[1] != '\0')
354 * Split URL into protocol,hostport,path.
364 return JB_ERR_MEMORY;
367 /* Find the start of the URL in our scratch space */
369 if (strncmpic(url_noproto, "http://", 7) == 0)
374 else if (strncmpic(url_noproto, "https://", 8) == 0)
379 else if (*url_noproto == '/')
382 * Short request line without protocol and host.
383 * Most likely because the client's request
384 * was intercepted and redirected into Privoxy.
395 url_path = strchr(url_noproto, '/');
396 if (url_path != NULL)
401 * NOTE: The following line ignores the path for HTTPS URLS.
402 * This means that you get consistent behaviour if you type a
403 * https URL in and it's parsed by the function. (When the
404 * URL is actually retrieved, SSL hides the path part).
406 http->path = strdup(http->ssl ? "/" : url_path);
408 http->hostport = strdup(url_noproto);
413 * Repair broken HTTP requests that don't contain a path,
414 * or CONNECT requests
416 http->path = strdup("/");
417 http->hostport = strdup(url_noproto);
422 if ( (http->path == NULL)
423 || (http->hostport == NULL))
425 return JB_ERR_MEMORY;
431 /* Without host, there is nothing left to do here */
436 * Split hostport into user/password (ignored), host, port.
443 buf = strdup(http->hostport);
446 return JB_ERR_MEMORY;
449 /* check if url contains username and/or password */
450 host = strchr(buf, '@');
453 /* Contains username/password, skip it and the @ sign. */
458 /* No username or password. */
462 /* check if url contains port */
463 port = strchr(host, ':');
467 /* Terminate hostname and point to start of port string */
469 http->port = atoi(port);
473 /* No port specified. */
474 http->port = (http->ssl ? 443 : 80);
477 http->host = strdup(host);
481 if (http->host == NULL)
483 return JB_ERR_MEMORY;
488 * Split domain name so we can compare it against wildcards
490 return init_domain_components(http);
495 /*********************************************************************
497 * Function : unknown_method
499 * Description : Checks whether a method is unknown.
502 * 1 : method = points to a http method
504 * Returns : TRUE if it's unknown, FALSE otherwise.
506 *********************************************************************/
507 static int unknown_method(const char *method)
509 static const char *known_http_methods[] = {
510 /* Basic HTTP request type */
511 "GET", "HEAD", "POST", "PUT", "DELETE", "OPTIONS", "TRACE", "CONNECT",
512 /* webDAV extensions (RFC2518) */
513 "PROPFIND", "PROPPATCH", "MOVE", "COPY", "MKCOL", "LOCK", "UNLOCK",
515 * Microsoft webDAV extension for Exchange 2000. See:
516 * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html
517 * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp
519 "BCOPY", "BMOVE", "BDELETE", "BPROPFIND", "BPROPPATCH",
521 * Another Microsoft webDAV extension for Exchange 2000. See:
522 * http://systems.cs.colorado.edu/grunwald/MobileComputing/Papers/draft-cohen-gena-p-base-00.txt
523 * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html
524 * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp
526 "SUBSCRIBE", "UNSUBSCRIBE", "NOTIFY", "POLL",
528 * Yet another WebDAV extension, this time for
529 * Web Distributed Authoring and Versioning (RFC3253)
531 "VERSION-CONTROL", "REPORT", "CHECKOUT", "CHECKIN", "UNCHECKOUT",
532 "MKWORKSPACE", "UPDATE", "LABEL", "MERGE", "BASELINE-CONTROL", "MKACTIVITY",
536 for (i = 0; i < SZ(known_http_methods); i++)
538 if (0 == strcmpic(method, known_http_methods[i]))
549 /*********************************************************************
551 * Function : parse_http_request
553 * Description : Parse out the host and port from the URL. Find the
554 * hostname & path, port (if ':'), and/or password (if '@')
557 * 1 : req = HTTP request line to break down
558 * 2 : http = pointer to the http structure to hold elements
559 * 3 : csp = Current client state (buffers, headers, etc...)
561 * Returns : JB_ERR_OK on success
562 * JB_ERR_MEMORY on out of memory
563 * JB_ERR_CGI_PARAMS on malformed command/URL
564 * or >100 domains deep.
566 *********************************************************************/
567 jb_err parse_http_request(const char *req,
568 struct http_request *http,
569 const struct client_state *csp)
572 char *v[10]; /* XXX: Why 10? We should only need three. */
577 memset(http, '\0', sizeof(*http));
582 return JB_ERR_MEMORY;
585 n = ssplit(buf, " \r\n", v, SZ(v), 1, 1);
593 * Fail in case of unknown methods
594 * which we might not handle correctly.
596 * XXX: There should be a config option
597 * to forward requests with unknown methods
598 * anyway. Most of them don't need special
601 if (unknown_method(v[0]))
603 log_error(LOG_LEVEL_ERROR, "Unknown HTTP method detected: %s", v[0]);
608 if (strcmpic(v[0], "CONNECT") == 0)
613 err = parse_http_url(v[1], http, csp);
621 * Copy the details into the structure
623 http->ssl = is_connect;
624 http->cmd = strdup(req);
625 http->gpc = strdup(v[0]);
626 http->ver = strdup(v[2]);
628 if ( (http->cmd == NULL)
629 || (http->gpc == NULL)
630 || (http->ver == NULL) )
633 return JB_ERR_MEMORY;
642 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
643 /*********************************************************************
645 * Function : compile_host_pattern
647 * Description : Parses and compiles a PCRE host pattern..
650 * 1 : url = Target url_spec to be filled in.
651 * 2 : host_pattern = Host pattern to compile.
653 * Returns : JB_ERR_OK - Success
654 * JB_ERR_MEMORY - Out of memory
655 * JB_ERR_PARSE - Cannot parse regex
657 *********************************************************************/
658 static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
661 char rebuf[BUFFER_SIZE];
663 assert(host_pattern);
664 assert(strlen(host_pattern) < sizeof(rebuf) - 2);
666 url->host_regex = zalloc(sizeof(*url->host_regex));
667 if (NULL == url->host_regex)
670 return JB_ERR_MEMORY;
673 snprintf(rebuf, sizeof(rebuf), "%s$", host_pattern);
675 errcode = regcomp(url->host_regex, rebuf,
676 (REG_EXTENDED|REG_NOSUB|REG_ICASE));
680 size_t errlen = regerror(errcode, url->host_regex, rebuf, sizeof(rebuf));
681 if (errlen > (sizeof(rebuf) - (size_t)1))
683 errlen = sizeof(rebuf) - (size_t)1;
685 rebuf[errlen] = '\0';
686 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s", url->spec, rebuf);
698 /*********************************************************************
700 * Function : compile_host_pattern
702 * Description : Parses and "compiles" an old-school host pattern.
705 * 1 : url = Target url_spec to be filled in.
706 * 2 : host_pattern = Host pattern to parse.
708 * Returns : JB_ERR_OK - Success
709 * JB_ERR_MEMORY - Out of memory
710 * JB_ERR_PARSE - Cannot parse regex
712 *********************************************************************/
713 static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
722 if (host_pattern[strlen(host_pattern) - 1] == '.')
724 url->unanchored |= ANCHOR_RIGHT;
726 if (host_pattern[0] == '.')
728 url->unanchored |= ANCHOR_LEFT;
732 * Split domain into components
734 url->dbuffer = strdup(host_pattern);
735 if (NULL == url->dbuffer)
741 return JB_ERR_MEMORY;
747 for (p = url->dbuffer; *p ; p++)
749 *p = (char)tolower((int)(unsigned char)*p);
753 * Split the domain name into components
755 url->dcount = ssplit(url->dbuffer, ".", v, SZ(v), 1, 1);
765 return JB_ERR_MEMORY;
767 else if (url->dcount != 0)
770 * Save a copy of the pointers in dvec
772 size = (size_t)url->dcount * sizeof(*url->dvec);
774 url->dvec = (char **)malloc(size);
775 if (NULL == url->dvec)
783 return JB_ERR_MEMORY;
786 memcpy(url->dvec, v, size);
789 * else dcount == 0 in which case we needn't do anything,
790 * since dvec will never be accessed and the pattern will
797 /*********************************************************************
799 * Function : simple_domaincmp
801 * Description : Domain-wise Compare fqdn's. The comparison is
802 * both left- and right-anchored. The individual
803 * domain names are compared with simplematch().
804 * This is only used by domain_match.
807 * 1 : pv = array of patterns to compare
808 * 2 : fv = array of domain components to compare
809 * 3 : len = length of the arrays (both arrays are the
810 * same length - if they weren't, it couldn't
811 * possibly be a match).
813 * Returns : 0 => domains are equivalent, else no match.
815 *********************************************************************/
816 static int simple_domaincmp(char **pv, char **fv, int len)
820 for (n = 0; n < len; n++)
822 if (simplematch(pv[n], fv[n]))
833 /*********************************************************************
835 * Function : domain_match
837 * Description : Domain-wise Compare fqdn's. Governed by the bimap in
838 * pattern->unachored, the comparison is un-, left-,
839 * right-anchored, or both.
840 * The individual domain names are compared with
844 * 1 : pattern = a domain that may contain a '*' as a wildcard.
845 * 2 : fqdn = domain name against which the patterns are compared.
847 * Returns : 0 => domains are equivalent, else no match.
849 *********************************************************************/
850 static int domain_match(const struct url_spec *pattern, const struct http_request *fqdn)
852 char **pv, **fv; /* vectors */
854 int unanchored = pattern->unanchored & (ANCHOR_RIGHT | ANCHOR_LEFT);
856 plen = pattern->dcount;
861 /* fqdn is too short to match this pattern */
868 if (unanchored == ANCHOR_LEFT)
873 * Convert this into a fully anchored pattern with
874 * the fqdn and pattern the same length
876 fv += (flen - plen); /* flen - plen >= 0 due to check above */
877 return simple_domaincmp(pv, fv, plen);
879 else if (unanchored == 0)
881 /* Fully anchored, check length */
886 return simple_domaincmp(pv, fv, plen);
888 else if (unanchored == ANCHOR_RIGHT)
890 /* Left anchored, ignore all extra in fqdn */
891 return simple_domaincmp(pv, fv, plen);
897 int maxn = flen - plen;
898 for (n = 0; n <= maxn; n++)
900 if (!simple_domaincmp(pv, fv, plen))
905 * Doesn't match from start of fqdn
906 * Try skipping first part of fqdn
914 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
917 /*********************************************************************
919 * Function : create_url_spec
921 * Description : Creates a "url_spec" structure from a string.
922 * When finished, free with free_url_spec().
925 * 1 : url = Target url_spec to be filled in. Will be
927 * 2 : buf = Source pattern, null terminated. NOTE: The
928 * contents of this buffer are destroyed by this
929 * function. If this function succeeds, the
930 * buffer is copied to url->spec. If this
931 * function fails, the contents of the buffer
934 * Returns : JB_ERR_OK - Success
935 * JB_ERR_MEMORY - Out of memory
936 * JB_ERR_PARSE - Cannot parse regex (Detailed message
937 * written to system log)
939 *********************************************************************/
940 jb_err create_url_spec(struct url_spec * url, const char * buf)
945 char rebuf[BUFFER_SIZE];
953 memset(url, '\0', sizeof(*url));
956 * Save a copy of the orignal specification
958 if ((url->spec = strdup(buf)) == NULL)
960 return JB_ERR_MEMORY;
963 /* Is it tag pattern? */
964 if (0 == strncmpic("TAG:", url->spec, 4))
966 if (NULL == (url->tag_regex = zalloc(sizeof(*url->tag_regex))))
969 return JB_ERR_MEMORY;
972 /* buf + 4 to skip "TAG:" */
973 errcode = regcomp(url->tag_regex, buf + 4, (REG_EXTENDED|REG_NOSUB|REG_ICASE));
976 errlen = regerror(errcode, url->preg, rebuf, sizeof(rebuf));
977 if (errlen > (sizeof(rebuf) - 1))
979 errlen = sizeof(rebuf) - 1;
981 rebuf[errlen] = '\0';
983 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s", url->spec, rebuf);
986 regfree(url->tag_regex);
987 freez(url->tag_regex);
994 /* Only reached for URL patterns */
995 p = strchr(buf, '/');
998 url->path = strdup(p);
999 if (NULL == url->path)
1002 return JB_ERR_MEMORY;
1012 if (NULL == (url->preg = zalloc(sizeof(*url->preg))))
1016 return JB_ERR_MEMORY;
1019 snprintf(rebuf, sizeof(rebuf), "^(%s)", url->path);
1021 errcode = regcomp(url->preg, rebuf,
1022 (REG_EXTENDED|REG_NOSUB|REG_ICASE));
1025 errlen = regerror(errcode, url->preg, rebuf, sizeof(rebuf));
1027 if (errlen > (sizeof(rebuf) - (size_t)1))
1029 errlen = sizeof(rebuf) - (size_t)1;
1031 rebuf[errlen] = '\0';
1033 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s",
1041 return JB_ERR_PARSE;
1045 p = strchr(buf, ':');
1049 url->port_list = strdup(p);
1050 if (NULL == url->port_list)
1052 return JB_ERR_MEMORY;
1057 url->port_list = NULL;
1062 return compile_host_pattern(url, buf);
1070 /*********************************************************************
1072 * Function : free_url_spec
1074 * Description : Called from the "unloaders". Freez the url
1075 * structure elements.
1078 * 1 : url = pointer to a url_spec structure.
1082 *********************************************************************/
1083 void free_url_spec(struct url_spec *url)
1085 if (url == NULL) return;
1088 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1089 if (url->host_regex)
1091 regfree(url->host_regex);
1092 freez(url->host_regex);
1095 freez(url->dbuffer);
1097 #endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */
1099 freez(url->port_list);
1107 regfree(url->tag_regex);
1108 freez(url->tag_regex);
1113 /*********************************************************************
1115 * Function : url_match
1117 * Description : Compare a URL against a URL pattern.
1120 * 1 : pattern = a URL pattern
1121 * 2 : url = URL to match
1123 * Returns : Nonzero if the URL matches the pattern, else 0.
1125 *********************************************************************/
1126 int url_match(const struct url_spec *pattern,
1127 const struct http_request *url)
1129 /* XXX: these should probably be functions. */
1130 #define PORT_MATCHES ((NULL == pattern->port_list) || match_portlist(pattern->port_list, url->port))
1131 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1132 #define DOMAIN_MATCHES ((NULL == pattern->host_regex) || (0 == regexec(pattern->host_regex, url->host, 0, NULL, 0)))
1134 #define DOMAIN_MATCHES ((NULL == pattern->dbuffer) || (0 == domain_match(pattern, url)))
1136 #define PATH_MATCHES ((NULL == pattern->path) || (0 == regexec(pattern->preg, url->path, 0, NULL, 0)))
1138 if (pattern->tag_regex != NULL)
1140 /* It's a tag pattern and shouldn't be matched against URLs */
1144 return (PORT_MATCHES && DOMAIN_MATCHES && PATH_MATCHES);
1149 /*********************************************************************
1151 * Function : match_portlist
1153 * Description : Check if a given number is covered by a comma
1154 * separated list of numbers and ranges (a,b-c,d,..)
1157 * 1 : portlist = String with list
1158 * 2 : port = port to check
1160 * Returns : 0 => no match
1163 *********************************************************************/
1164 int match_portlist(const char *portlist, int port)
1166 char *min, *max, *next, *portlist_copy;
1168 min = next = portlist_copy = strdup(portlist);
1171 * Zero-terminate first item and remember offset for next
1173 if (NULL != (next = strchr(portlist_copy, (int) ',')))
1179 * Loop through all items, checking for match
1183 if (NULL == (max = strchr(min, (int) '-')))
1186 * No dash, check for equality
1188 if (port == atoi(min))
1190 free(portlist_copy);
1197 * This is a range, so check if between min and max,
1198 * or, if max was omitted, between min and 65K
1201 if(port >= atoi(min) && port <= (atoi(max) ? atoi(max) : 65535))
1203 free(portlist_copy);
1215 * Zero-terminate next item and remember offset for n+1
1217 if ((NULL != next) && (NULL != (next = strchr(next, (int) ','))))
1223 free(portlist_copy);