1 const char urlmatch_rcs[] = "$Id: urlmatch.c,v 1.26 2008/04/07 16:57:18 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 1.26 2008/04/07 16:57:18 fabiankeil
37 * - Use free_url_spec() more consistently.
38 * - Let it reset url->dcount just in case.
40 * Revision 1.25 2008/04/06 15:18:38 fabiankeil
41 * Oh well, rename the --enable-pcre-host-patterns option to
42 * --enable-extended-host-patterns as it's not really PCRE syntax.
44 * Revision 1.24 2008/04/06 14:54:26 fabiankeil
45 * Use PCRE syntax in host patterns when configured
46 * with --enable-pcre-host-patterns.
48 * Revision 1.23 2008/04/05 12:19:20 fabiankeil
49 * Factor compile_host_pattern() out of create_url_spec().
51 * Revision 1.22 2008/03/30 15:02:32 fabiankeil
52 * SZitify unknown_method().
54 * Revision 1.21 2007/12/24 16:34:23 fabiankeil
55 * Band-aid (and micro-optimization) that makes it less likely to run out of
56 * stack space with overly-complex path patterns. Probably masks the problem
57 * reported by Lee in #1856679. Hohoho.
59 * Revision 1.20 2007/09/02 15:31:20 fabiankeil
60 * Move match_portlist() from filter.c to urlmatch.c.
61 * It's used for url matching, not for filtering.
63 * Revision 1.19 2007/09/02 13:42:11 fabiankeil
64 * - Allow port lists in url patterns.
65 * - Ditch unused url_spec member pathlen.
67 * Revision 1.18 2007/07/30 16:42:21 fabiankeil
68 * Move the method check into unknown_method()
69 * and loop through the known methods instead
70 * of using a screen-long OR chain.
72 * Revision 1.17 2007/04/15 16:39:21 fabiankeil
73 * Introduce tags as alternative way to specify which
74 * actions apply to a request. At the moment tags can be
75 * created based on client and server headers.
77 * Revision 1.16 2007/02/13 13:59:24 fabiankeil
78 * Remove redundant log message.
80 * Revision 1.15 2007/01/28 16:11:23 fabiankeil
81 * Accept WebDAV methods for subversion
82 * in parse_http_request(). Closes FR 1581425.
84 * Revision 1.14 2007/01/06 14:23:56 fabiankeil
85 * Fix gcc43 warnings. Mark *csp as immutable
86 * for parse_http_url() and url_match().
87 * Replace a sprintf call with snprintf.
89 * Revision 1.13 2006/12/06 19:50:54 fabiankeil
90 * parse_http_url() now handles intercepted
91 * HTTP request lines as well. Moved parts
92 * of parse_http_url()'s code into
93 * init_domain_components() so that it can
94 * be reused in chat().
96 * Revision 1.12 2006/07/18 14:48:47 david__schmidt
97 * Reorganizing the repository: swapping out what was HEAD (the old 3.1 branch)
98 * with what was really the latest development (the v_3_0_branch branch)
100 * Revision 1.10.2.7 2003/05/17 15:57:24 oes
101 * - parse_http_url now checks memory allocation failure for
102 * duplication of "*" URL and rejects "*something" URLs
104 * - Added a comment to what might look like a bug in
105 * create_url_spec (see !bug #736931)
106 * - Comment cosmetics
108 * Revision 1.10.2.6 2003/05/07 12:39:48 oes
109 * Fix typo: Default port for https URLs is 443, not 143.
110 * Thanks to Scott Tregear for spotting this one.
112 * Revision 1.10.2.5 2003/02/28 13:09:29 oes
113 * Fixed a rare double free condition as per Bug #694713
115 * Revision 1.10.2.4 2003/02/28 12:57:44 oes
116 * Moved freeing of http request structure to its owner
117 * as per Dan Price's observations in Bug #694713
119 * Revision 1.10.2.3 2002/11/12 16:50:40 oes
120 * Fixed memory leak in parse_http_request() reported by Oliver Stoeneberg. Fixes bug #637073
122 * Revision 1.10.2.2 2002/09/25 14:53:15 oes
123 * Added basic support for OPTIONS and TRACE HTTP methods:
124 * parse_http_url now recognizes the "*" URI as well as
125 * the OPTIONS and TRACE method keywords.
127 * Revision 1.10.2.1 2002/06/06 19:06:44 jongfoster
128 * Adding support for proprietary Microsoft WebDAV extensions
130 * Revision 1.10 2002/05/12 21:40:37 jongfoster
131 * - Removing some unused code
133 * Revision 1.9 2002/04/04 00:36:36 gliptak
134 * always use pcre for matching
136 * Revision 1.8 2002/04/03 23:32:47 jongfoster
137 * Fixing memory leak on error
139 * Revision 1.7 2002/03/26 22:29:55 swa
140 * we have a new homepage!
142 * Revision 1.6 2002/03/24 13:25:43 swa
143 * name change related issues
145 * Revision 1.5 2002/03/13 00:27:05 jongfoster
148 * Revision 1.4 2002/03/07 03:46:17 oes
149 * Fixed compiler warnings
151 * Revision 1.3 2002/03/03 14:51:11 oes
152 * Fixed CLF logging: Added ocmd member for client's request to struct http_request
154 * Revision 1.2 2002/01/21 00:14:09 jongfoster
155 * Correcting comment style
156 * Fixing an uninitialized memory bug in create_url_spec()
158 * Revision 1.1 2002/01/17 20:53:46 jongfoster
159 * Moving all our URL and URL pattern parsing code to the same file - it
160 * was scattered around in filters.c, loaders.c and parsers.c.
162 * Providing a single, simple url_match(pattern,url) function - rather than
163 * the 3-line match routine which was repeated all over the place.
165 * Renaming free_url to free_url_spec, since it frees a struct url_spec.
167 * Providing parse_http_url() so that URLs can be parsed without faking a
168 * HTTP request line for parse_http_request() or repeating the parsing
169 * code (both of which were techniques that were actually in use).
171 * Standardizing that struct http_request is used to represent a URL, and
172 * struct url_spec is used to represent a URL pattern. (Before, URLs were
173 * represented as seperate variables and a partially-filled-in url_spec).
176 *********************************************************************/
183 #include <sys/types.h>
191 #if !defined(_WIN32) && !defined(__OS2__)
196 #include "urlmatch.h"
198 #include "miscutil.h"
201 const char urlmatch_h_rcs[] = URLMATCH_H_VERSION;
204 /*********************************************************************
206 * Function : free_http_request
208 * Description : Freez a http_request structure
211 * 1 : http = points to a http_request structure to free
215 *********************************************************************/
216 void free_http_request(struct http_request *http)
225 freez(http->hostport);
228 freez(http->host_ip_addr_str);
229 freez(http->dbuffer);
235 /*********************************************************************
237 * Function : init_domain_components
239 * Description : Splits the domain name so we can compare it
240 * against wildcards. It used to be part of
241 * parse_http_url, but was separated because the
242 * same code is required in chat in case of
243 * intercepted requests.
246 * 1 : http = pointer to the http structure to hold elements.
248 * Returns : JB_ERR_OK on success
249 * JB_ERR_MEMORY on out of memory
250 * JB_ERR_PARSE on malformed command/URL
251 * or >100 domains deep.
253 *********************************************************************/
254 jb_err init_domain_components(struct http_request *http)
256 char *vec[BUFFER_SIZE];
260 http->dbuffer = strdup(http->host);
261 if (NULL == http->dbuffer)
263 return JB_ERR_MEMORY;
266 /* map to lower case */
267 for (p = http->dbuffer; *p ; p++)
269 *p = (char)tolower((int)(unsigned char)*p);
272 /* split the domain name into components */
273 http->dcount = ssplit(http->dbuffer, ".", vec, SZ(vec), 1, 1);
275 if (http->dcount <= 0)
278 * Error: More than SZ(vec) components in domain
279 * or: no components in domain
281 log_error(LOG_LEVEL_ERROR, "More than SZ(vec) components in domain or none at all.");
285 /* save a copy of the pointers in dvec */
286 size = (size_t)http->dcount * sizeof(*http->dvec);
288 http->dvec = (char **)malloc(size);
289 if (NULL == http->dvec)
291 return JB_ERR_MEMORY;
294 memcpy(http->dvec, vec, size);
300 /*********************************************************************
302 * Function : parse_http_url
304 * Description : Parse out the host and port from the URL. Find the
305 * hostname & path, port (if ':'), and/or password (if '@')
308 * 1 : url = URL (or is it URI?) to break down
309 * 2 : http = pointer to the http structure to hold elements.
310 * Will be zeroed before use. Note that this
311 * function sets the http->gpc and http->ver
313 * 3 : csp = Current client state (buffers, headers, etc...)
315 * Returns : JB_ERR_OK on success
316 * JB_ERR_MEMORY on out of memory
317 * JB_ERR_PARSE on malformed command/URL
318 * or >100 domains deep.
320 *********************************************************************/
321 jb_err parse_http_url(const char * url,
322 struct http_request *http,
323 const struct client_state *csp)
325 int host_available = 1; /* A proxy can dream. */
328 * Zero out the results structure
330 memset(http, '\0', sizeof(*http));
334 * Save our initial URL
336 http->url = strdup(url);
337 if (http->url == NULL)
339 return JB_ERR_MEMORY;
344 * Check for * URI. If found, we're done.
346 if (*http->url == '*')
348 if ( NULL == (http->path = strdup("*"))
349 || NULL == (http->hostport = strdup("")) )
351 return JB_ERR_MEMORY;
353 if (http->url[1] != '\0')
362 * Split URL into protocol,hostport,path.
372 return JB_ERR_MEMORY;
375 /* Find the start of the URL in our scratch space */
377 if (strncmpic(url_noproto, "http://", 7) == 0)
382 else if (strncmpic(url_noproto, "https://", 8) == 0)
387 else if (*url_noproto == '/')
390 * Short request line without protocol and host.
391 * Most likely because the client's request
392 * was intercepted and redirected into Privoxy.
403 url_path = strchr(url_noproto, '/');
404 if (url_path != NULL)
409 * NOTE: The following line ignores the path for HTTPS URLS.
410 * This means that you get consistent behaviour if you type a
411 * https URL in and it's parsed by the function. (When the
412 * URL is actually retrieved, SSL hides the path part).
414 http->path = strdup(http->ssl ? "/" : url_path);
416 http->hostport = strdup(url_noproto);
421 * Repair broken HTTP requests that don't contain a path,
422 * or CONNECT requests
424 http->path = strdup("/");
425 http->hostport = strdup(url_noproto);
430 if ( (http->path == NULL)
431 || (http->hostport == NULL))
433 return JB_ERR_MEMORY;
439 /* Without host, there is nothing left to do here */
444 * Split hostport into user/password (ignored), host, port.
451 buf = strdup(http->hostport);
454 return JB_ERR_MEMORY;
457 /* check if url contains username and/or password */
458 host = strchr(buf, '@');
461 /* Contains username/password, skip it and the @ sign. */
466 /* No username or password. */
470 /* check if url contains port */
471 port = strchr(host, ':');
475 /* Terminate hostname and point to start of port string */
477 http->port = atoi(port);
481 /* No port specified. */
482 http->port = (http->ssl ? 443 : 80);
485 http->host = strdup(host);
489 if (http->host == NULL)
491 return JB_ERR_MEMORY;
496 * Split domain name so we can compare it against wildcards
498 return init_domain_components(http);
503 /*********************************************************************
505 * Function : unknown_method
507 * Description : Checks whether a method is unknown.
510 * 1 : method = points to a http method
512 * Returns : TRUE if it's unknown, FALSE otherwise.
514 *********************************************************************/
515 static int unknown_method(const char *method)
517 static const char *known_http_methods[] = {
518 /* Basic HTTP request type */
519 "GET", "HEAD", "POST", "PUT", "DELETE", "OPTIONS", "TRACE", "CONNECT",
520 /* webDAV extensions (RFC2518) */
521 "PROPFIND", "PROPPATCH", "MOVE", "COPY", "MKCOL", "LOCK", "UNLOCK",
523 * Microsoft webDAV extension for Exchange 2000. See:
524 * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html
525 * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp
527 "BCOPY", "BMOVE", "BDELETE", "BPROPFIND", "BPROPPATCH",
529 * Another Microsoft webDAV extension for Exchange 2000. See:
530 * http://systems.cs.colorado.edu/grunwald/MobileComputing/Papers/draft-cohen-gena-p-base-00.txt
531 * http://lists.w3.org/Archives/Public/w3c-dist-auth/2002JanMar/0001.html
532 * http://msdn.microsoft.com/library/en-us/wss/wss/_webdav_methods.asp
534 "SUBSCRIBE", "UNSUBSCRIBE", "NOTIFY", "POLL",
536 * Yet another WebDAV extension, this time for
537 * Web Distributed Authoring and Versioning (RFC3253)
539 "VERSION-CONTROL", "REPORT", "CHECKOUT", "CHECKIN", "UNCHECKOUT",
540 "MKWORKSPACE", "UPDATE", "LABEL", "MERGE", "BASELINE-CONTROL", "MKACTIVITY",
544 for (i = 0; i < SZ(known_http_methods); i++)
546 if (0 == strcmpic(method, known_http_methods[i]))
557 /*********************************************************************
559 * Function : parse_http_request
561 * Description : Parse out the host and port from the URL. Find the
562 * hostname & path, port (if ':'), and/or password (if '@')
565 * 1 : req = HTTP request line to break down
566 * 2 : http = pointer to the http structure to hold elements
567 * 3 : csp = Current client state (buffers, headers, etc...)
569 * Returns : JB_ERR_OK on success
570 * JB_ERR_MEMORY on out of memory
571 * JB_ERR_CGI_PARAMS on malformed command/URL
572 * or >100 domains deep.
574 *********************************************************************/
575 jb_err parse_http_request(const char *req,
576 struct http_request *http,
577 const struct client_state *csp)
580 char *v[10]; /* XXX: Why 10? We should only need three. */
585 memset(http, '\0', sizeof(*http));
590 return JB_ERR_MEMORY;
593 n = ssplit(buf, " \r\n", v, SZ(v), 1, 1);
601 * Fail in case of unknown methods
602 * which we might not handle correctly.
604 * XXX: There should be a config option
605 * to forward requests with unknown methods
606 * anyway. Most of them don't need special
609 if (unknown_method(v[0]))
611 log_error(LOG_LEVEL_ERROR, "Unknown HTTP method detected: %s", v[0]);
616 if (strcmpic(v[0], "CONNECT") == 0)
621 err = parse_http_url(v[1], http, csp);
629 * Copy the details into the structure
631 http->ssl = is_connect;
632 http->cmd = strdup(req);
633 http->gpc = strdup(v[0]);
634 http->ver = strdup(v[2]);
636 if ( (http->cmd == NULL)
637 || (http->gpc == NULL)
638 || (http->ver == NULL) )
641 return JB_ERR_MEMORY;
650 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
651 /*********************************************************************
653 * Function : compile_host_pattern
655 * Description : Parses and compiles a PCRE host pattern..
658 * 1 : url = Target url_spec to be filled in.
659 * 2 : host_pattern = Host pattern to compile.
661 * Returns : JB_ERR_OK - Success
662 * JB_ERR_MEMORY - Out of memory
663 * JB_ERR_PARSE - Cannot parse regex
665 *********************************************************************/
666 static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
669 char rebuf[BUFFER_SIZE];
671 assert(host_pattern);
672 assert(strlen(host_pattern) < sizeof(rebuf) - 2);
674 url->host_regex = zalloc(sizeof(*url->host_regex));
675 if (NULL == url->host_regex)
678 return JB_ERR_MEMORY;
681 snprintf(rebuf, sizeof(rebuf), "%s$", host_pattern);
683 errcode = regcomp(url->host_regex, rebuf,
684 (REG_EXTENDED|REG_NOSUB|REG_ICASE));
688 size_t errlen = regerror(errcode, url->host_regex, rebuf, sizeof(rebuf));
689 if (errlen > (sizeof(rebuf) - (size_t)1))
691 errlen = sizeof(rebuf) - (size_t)1;
693 rebuf[errlen] = '\0';
694 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s", url->spec, rebuf);
706 /*********************************************************************
708 * Function : compile_host_pattern
710 * Description : Parses and "compiles" an old-school host pattern.
713 * 1 : url = Target url_spec to be filled in.
714 * 2 : host_pattern = Host pattern to parse.
716 * Returns : JB_ERR_OK - Success
717 * JB_ERR_MEMORY - Out of memory
718 * JB_ERR_PARSE - Cannot parse regex
720 *********************************************************************/
721 static jb_err compile_host_pattern(struct url_spec *url, const char *host_pattern)
730 if (host_pattern[strlen(host_pattern) - 1] == '.')
732 url->unanchored |= ANCHOR_RIGHT;
734 if (host_pattern[0] == '.')
736 url->unanchored |= ANCHOR_LEFT;
740 * Split domain into components
742 url->dbuffer = strdup(host_pattern);
743 if (NULL == url->dbuffer)
746 return JB_ERR_MEMORY;
752 for (p = url->dbuffer; *p ; p++)
754 *p = (char)tolower((int)(unsigned char)*p);
758 * Split the domain name into components
760 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)
778 return JB_ERR_MEMORY;
781 memcpy(url->dvec, v, size);
784 * else dcount == 0 in which case we needn't do anything,
785 * since dvec will never be accessed and the pattern will
792 /*********************************************************************
794 * Function : simple_domaincmp
796 * Description : Domain-wise Compare fqdn's. The comparison is
797 * both left- and right-anchored. The individual
798 * domain names are compared with simplematch().
799 * This is only used by domain_match.
802 * 1 : pv = array of patterns to compare
803 * 2 : fv = array of domain components to compare
804 * 3 : len = length of the arrays (both arrays are the
805 * same length - if they weren't, it couldn't
806 * possibly be a match).
808 * Returns : 0 => domains are equivalent, else no match.
810 *********************************************************************/
811 static int simple_domaincmp(char **pv, char **fv, int len)
815 for (n = 0; n < len; n++)
817 if (simplematch(pv[n], fv[n]))
828 /*********************************************************************
830 * Function : domain_match
832 * Description : Domain-wise Compare fqdn's. Governed by the bimap in
833 * pattern->unachored, the comparison is un-, left-,
834 * right-anchored, or both.
835 * The individual domain names are compared with
839 * 1 : pattern = a domain that may contain a '*' as a wildcard.
840 * 2 : fqdn = domain name against which the patterns are compared.
842 * Returns : 0 => domains are equivalent, else no match.
844 *********************************************************************/
845 static int domain_match(const struct url_spec *pattern, const struct http_request *fqdn)
847 char **pv, **fv; /* vectors */
849 int unanchored = pattern->unanchored & (ANCHOR_RIGHT | ANCHOR_LEFT);
851 plen = pattern->dcount;
856 /* fqdn is too short to match this pattern */
863 if (unanchored == ANCHOR_LEFT)
868 * Convert this into a fully anchored pattern with
869 * the fqdn and pattern the same length
871 fv += (flen - plen); /* flen - plen >= 0 due to check above */
872 return simple_domaincmp(pv, fv, plen);
874 else if (unanchored == 0)
876 /* Fully anchored, check length */
881 return simple_domaincmp(pv, fv, plen);
883 else if (unanchored == ANCHOR_RIGHT)
885 /* Left anchored, ignore all extra in fqdn */
886 return simple_domaincmp(pv, fv, plen);
892 int maxn = flen - plen;
893 for (n = 0; n <= maxn; n++)
895 if (!simple_domaincmp(pv, fv, plen))
900 * Doesn't match from start of fqdn
901 * Try skipping first part of fqdn
909 #endif /* def FEATURE_EXTENDED_HOST_PATTERNS */
912 /*********************************************************************
914 * Function : create_url_spec
916 * Description : Creates a "url_spec" structure from a string.
917 * When finished, free with free_url_spec().
920 * 1 : url = Target url_spec to be filled in. Will be
922 * 2 : buf = Source pattern, null terminated. NOTE: The
923 * contents of this buffer are destroyed by this
924 * function. If this function succeeds, the
925 * buffer is copied to url->spec. If this
926 * function fails, the contents of the buffer
929 * Returns : JB_ERR_OK - Success
930 * JB_ERR_MEMORY - Out of memory
931 * JB_ERR_PARSE - Cannot parse regex (Detailed message
932 * written to system log)
934 *********************************************************************/
935 jb_err create_url_spec(struct url_spec * url, const char * buf)
940 char rebuf[BUFFER_SIZE];
948 memset(url, '\0', sizeof(*url));
951 * Save a copy of the orignal specification
953 if ((url->spec = strdup(buf)) == NULL)
955 return JB_ERR_MEMORY;
958 /* Is it tag pattern? */
959 if (0 == strncmpic("TAG:", url->spec, 4))
961 if (NULL == (url->tag_regex = zalloc(sizeof(*url->tag_regex))))
964 return JB_ERR_MEMORY;
967 /* buf + 4 to skip "TAG:" */
968 errcode = regcomp(url->tag_regex, buf + 4, (REG_EXTENDED|REG_NOSUB|REG_ICASE));
971 errlen = regerror(errcode, url->preg, rebuf, sizeof(rebuf));
972 if (errlen > (sizeof(rebuf) - 1))
974 errlen = sizeof(rebuf) - 1;
976 rebuf[errlen] = '\0';
977 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s", url->spec, rebuf);
985 /* Only reached for URL patterns */
986 p = strchr(buf, '/');
991 url->path = strdup(p);
992 if (NULL == url->path)
995 return JB_ERR_MEMORY;
1001 * The path pattern is a single slash and can
1002 * be ignored as it won't affect the result.
1004 assert(NULL == url->path);
1015 if (NULL == (url->preg = zalloc(sizeof(*url->preg))))
1018 return JB_ERR_MEMORY;
1021 snprintf(rebuf, sizeof(rebuf), "^(%s)", url->path);
1023 errcode = regcomp(url->preg, rebuf,
1024 (REG_EXTENDED|REG_NOSUB|REG_ICASE));
1027 errlen = regerror(errcode, url->preg, rebuf, sizeof(rebuf));
1029 if (errlen > (sizeof(rebuf) - (size_t)1))
1031 errlen = sizeof(rebuf) - (size_t)1;
1033 rebuf[errlen] = '\0';
1034 log_error(LOG_LEVEL_ERROR, "error compiling %s: %s",
1038 return JB_ERR_PARSE;
1042 p = strchr(buf, ':');
1046 url->port_list = strdup(p);
1047 if (NULL == url->port_list)
1049 return JB_ERR_MEMORY;
1054 url->port_list = NULL;
1059 return compile_host_pattern(url, buf);
1067 /*********************************************************************
1069 * Function : free_url_spec
1071 * Description : Called from the "unloaders". Freez the url
1072 * structure elements.
1075 * 1 : url = pointer to a url_spec structure.
1079 *********************************************************************/
1080 void free_url_spec(struct url_spec *url)
1082 if (url == NULL) return;
1085 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1086 if (url->host_regex)
1088 regfree(url->host_regex);
1089 freez(url->host_regex);
1092 freez(url->dbuffer);
1095 #endif /* ndef FEATURE_EXTENDED_HOST_PATTERNS */
1097 freez(url->port_list);
1105 regfree(url->tag_regex);
1106 freez(url->tag_regex);
1111 /*********************************************************************
1113 * Function : url_match
1115 * Description : Compare a URL against a URL pattern.
1118 * 1 : pattern = a URL pattern
1119 * 2 : url = URL to match
1121 * Returns : Nonzero if the URL matches the pattern, else 0.
1123 *********************************************************************/
1124 int url_match(const struct url_spec *pattern,
1125 const struct http_request *url)
1127 /* XXX: these should probably be functions. */
1128 #define PORT_MATCHES ((NULL == pattern->port_list) || match_portlist(pattern->port_list, url->port))
1129 #ifdef FEATURE_EXTENDED_HOST_PATTERNS
1130 #define DOMAIN_MATCHES ((NULL == pattern->host_regex) || (0 == regexec(pattern->host_regex, url->host, 0, NULL, 0)))
1132 #define DOMAIN_MATCHES ((NULL == pattern->dbuffer) || (0 == domain_match(pattern, url)))
1134 #define PATH_MATCHES ((NULL == pattern->path) || (0 == regexec(pattern->preg, url->path, 0, NULL, 0)))
1136 if (pattern->tag_regex != NULL)
1138 /* It's a tag pattern and shouldn't be matched against URLs */
1142 return (PORT_MATCHES && DOMAIN_MATCHES && PATH_MATCHES);
1147 /*********************************************************************
1149 * Function : match_portlist
1151 * Description : Check if a given number is covered by a comma
1152 * separated list of numbers and ranges (a,b-c,d,..)
1155 * 1 : portlist = String with list
1156 * 2 : port = port to check
1158 * Returns : 0 => no match
1161 *********************************************************************/
1162 int match_portlist(const char *portlist, int port)
1164 char *min, *max, *next, *portlist_copy;
1166 min = next = portlist_copy = strdup(portlist);
1169 * Zero-terminate first item and remember offset for next
1171 if (NULL != (next = strchr(portlist_copy, (int) ',')))
1177 * Loop through all items, checking for match
1181 if (NULL == (max = strchr(min, (int) '-')))
1184 * No dash, check for equality
1186 if (port == atoi(min))
1188 free(portlist_copy);
1195 * This is a range, so check if between min and max,
1196 * or, if max was omitted, between min and 65K
1199 if(port >= atoi(min) && port <= (atoi(max) ? atoi(max) : 65535))
1201 free(portlist_copy);
1213 * Zero-terminate next item and remember offset for n+1
1215 if ((NULL != next) && (NULL != (next = strchr(next, (int) ','))))
1221 free(portlist_copy);