-const char filters_rcs[] = "$Id: filters.c,v 1.38 2001/10/23 21:32:33 jongfoster Exp $";
+const char filters_rcs[] = "$Id: filters.c,v 1.42 2002/01/17 21:00:32 jongfoster Exp $";
/*********************************************************************
*
* File : $Source: /cvsroot/ijbswa/current/filters.c,v $
* Purpose : Declares functions to parse/crunch headers and pages.
* Functions declared include:
* `acl_addr', `add_stats', `block_acl', `block_imageurl',
- * `block_url', `url_actions', `domaincmp', `dsplit',
+ * `block_url', `url_actions', `domain_split',
* `filter_popups', `forward_url', 'redirect_url',
* `ij_untrusted_url', `intercept_url', `pcrs_filter_respose',
- * `show_proxy_args', 'ijb_send_banner', and `trust_url'
+ * 'ijb_send_banner', and `trust_url'
*
* Copyright : Written by and Copyright (C) 2001 the SourceForge
* IJBSWA team. http://ijbswa.sourceforge.net
*
* Revisions :
* $Log: filters.c,v $
+ * Revision 1.42 2002/01/17 21:00:32 jongfoster
+ * Moving all our URL and URL pattern parsing code to urlmatch.c.
+ *
+ * Using a single, simple url_match(pattern,url) function - rather than
+ * the 3-line match routine which was repeated all over the place.
+ *
+ * Renaming free_url to free_url_spec, since it frees a struct url_spec.
+ *
+ * Using parse_http_url() to parse URLs without faking a HTTP
+ * request line for parse_http_request().
+ *
+ * Revision 1.41 2001/11/13 00:14:07 jongfoster
+ * Fixing stupid bug now I've figured out what || means.
+ * (It always returns 0 or 1, not one of it's paramaters.)
+ *
+ * Revision 1.40 2001/10/26 17:37:55 oes
+ * - Re-enabled Netscape 200/404 bug workaround in block_url():
+ * - Removed OS/2 special case
+ * - Made block_url() independant from sed() having been run
+ * - Made trust_url independant from sed() having been run
+ * - Made is_imageurl independant from sed() having been run.
+ * It now checks User-Agent: and Accept: by itself.
+ *
+ *
+ * Revision 1.39 2001/10/25 03:40:48 david__schmidt
+ * Change in porting tactics: OS/2's EMX porting layer doesn't allow multiple
+ * threads to call select() simultaneously. So, it's time to do a real, live,
+ * native OS/2 port. See defines for __EMX__ (the porting layer) vs. __OS2__
+ * (native). Both versions will work, but using __OS2__ offers multi-threading.
+ *
* Revision 1.38 2001/10/23 21:32:33 jongfoster
* Adding error-checking to selected functions
*
#include "cgi.h"
#include "list.h"
#include "deanimate.h"
+#include "urlmatch.h"
#ifdef _WIN32
#include "win32.h"
jb_err err;
struct map * exports;
- /* FIXME */
-#ifdef __EMX__
- /*
- * The entire OS/2 community will hit the stupid Netscape bug
- * (all three of us! :-) so we'll just keep ourselves out
- * of this contentious debate and special-case ourselves.
- * The problem is... a this point in parsing, we don't know
- * what the csp->http->user_agent is (yet). So we can't use
- * it to decide if we should work around the NS bug or not.
- */
- rsp->status = strdup("200 Request for blocked URL");
-#else
/*
* Workaround for stupid Netscape bug which prevents
* pages from being displayed if loading a referenced
* JavaScript or style sheet fails. So make it appear
* as if it succeeded.
*/
- if (csp->http->user_agent
- && !strncmpic(csp->http->user_agent, "mozilla", 7)
- && !strstr(csp->http->user_agent, "compatible")
- && !strstr(csp->http->user_agent, "Opera"))
+ if ( NULL != (p = get_header_value(csp->headers, "User-Agent:"))
+ && !strncmpic(p, "mozilla", 7) /* Catch Netscape but */
+ && !strstr(p, "Gecko") /* save Mozilla, */
+ && !strstr(p, "compatible") /* MSIE */
+ && !strstr(p, "Opera")) /* and Opera. */
{
rsp->status = strdup("200 Request for blocked URL");
}
{
rsp->status = strdup("404 Request for blocked URL");
}
-#endif /* __EMX__ */
+
if (rsp->status == NULL)
{
free_http_response(rsp);
err = map_block_killer(exports, "force-support");
#endif /* ndef FEATURE_FORCE_LOAD */
- err = err || map(exports, "hostport", 1, csp->http->hostport, 1);
- err = err || map(exports, "hostport-html", 1, html_encode(csp->http->hostport), 0);
- err = err || map(exports, "path", 1, csp->http->path, 1);
- err = err || map(exports, "path-html", 1, html_encode(csp->http->path), 0);
+ if (!err) err = map(exports, "hostport", 1, html_encode(csp->http->hostport), 0);
+ if (!err) err = map(exports, "path", 1, html_encode(csp->http->path), 0);
if (err)
{
/*
* Export the host, port, and referrer information
*/
- err = map(exports, "hostport", 1, csp->http->hostport, 1)
- || map(exports, "path", 1, csp->http->path, 1)
- || map(exports, "hostport-html", 1, html_encode(csp->http->hostport), 0)
- || map(exports, "path-html", 1, html_encode(csp->http->path), 0);
+ err = map(exports, "hostport", 1, csp->http->hostport, 1);
+ if (!err) err = map(exports, "path", 1, csp->http->path, 1);
- if (csp->referrer && strlen(csp->referrer) > 9)
+ if (NULL != (p = get_header_value(csp->headers, "Referer:")))
{
- err = err || map(exports, "referrer", 1, csp->referrer + 9, 1);
- err = err || map(exports, "referrer-html", 1, html_encode(csp->referrer + 9), 0);
+ if (!err) err = map(exports, "referrer", 1, html_encode(p), 0);
}
else
{
- err = err || map(exports, "referrer", 1, "unknown", 1);
- err = err || map(exports, "referrer-html", 1, "unknown", 1);
+ if (!err) err = map(exports, "referrer", 1, "unknown", 1);
}
if (err)
int is_imageurl(struct client_state *csp)
{
#ifdef FEATURE_IMAGE_DETECT_MSIE
- if ((csp->accept_types
- & (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_IMAGE|ACCEPT_TYPE_MSIE_HTML))
- == (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_IMAGE))
- {
- return 1;
- }
- else if ((csp->accept_types
- & (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_IMAGE|ACCEPT_TYPE_MSIE_HTML))
- == (ACCEPT_TYPE_IS_MSIE|ACCEPT_TYPE_MSIE_HTML))
+ char *tmp;
+
+ tmp = get_header_value(csp->headers, "User-Agent:");
+ if (tmp && strstr(tmp, "MSIE"))
{
- return 0;
+ tmp = get_header_value(csp->headers, "Accept:");
+ if (tmp && strstr(tmp, "image/gif"))
+ {
+ /* Client will accept HTML. If this seems counterintuitive,
+ * blame Microsoft.
+ */
+ return(0);
+ }
+ else
+ {
+ return(1);
+ }
}
#endif /* def FEATURE_IMAGE_DETECT_MSIE */
{
struct file_list *fl;
struct block_spec *b;
- struct url_spec url[1], **tl, *t;
+ struct url_spec **trusted_url;
struct http_request rhttp[1];
- char *p, *h;
+ const char * referer;
+ jb_err err;
/*
* If we don't have a trustlist, we trust everybody
*/
if (((fl = csp->tlist) == NULL) || ((b = fl->f) == NULL))
{
- return(0);
+ return 0;
}
+ memset(rhttp, '\0', sizeof(*rhttp));
/*
* Do we trust the request URL itself?
*/
- *url = dsplit(csp->http->host);
-
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL) return(0);
-
- memset(rhttp, '\0', sizeof(*rhttp));
-
for (b = b->next; b ; b = b->next)
{
- if ((b->url->port == 0) || (b->url->port == csp->http->port))
+ if (url_match(b->url, csp->http))
{
- if ((b->url->domain[0] == '\0') || (domaincmp(b->url, url) == 0))
- {
- if ((b->url->path == NULL) ||
-#ifdef REGEX
- (regexec(b->url->preg, csp->http->path, 0, NULL, 0) == 0)
-#else
- (strncmp(b->url->path, csp->http->path, b->url->pathlen) == 0)
-#endif
- )
- {
- freez(url->dbuf);
- freez(url->dvec);
-
- if (b->reject == 0) return(0);
-
- return(1);
- }
- }
+ return b->reject;
}
}
- freez(url->dbuf);
- freez(url->dvec);
-
- if ((csp->referrer == NULL)|| (strlen(csp->referrer) <= 9))
+ if (NULL == (referer = get_header_value(csp->headers, "Referer:")))
{
/* no referrer was supplied */
- return(1);
+ return 1;
}
- /* forge a URL from the referrer so we can use
- * convert_url() to parse it into its components.
+ /*
+ * If not, do we maybe trust its referrer?
*/
- p = NULL;
- p = strsav(p, "GET ");
- p = strsav(p, csp->referrer + 9); /* skip over "Referer: " */
- p = strsav(p, " HTTP/1.0");
-
- parse_http_request(p, rhttp, csp);
- freez(p);
-
- if (rhttp->cmd == NULL)
- {
- return(1);
- }
-
/*
- * If not, do we maybe trust its referrer?
+ * Parse the URL from the referrer
*/
- *url = dsplit(rhttp->host);
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL) return(1);
+ err = parse_http_url(referer, rhttp, csp);
+ if (err)
+ {
+ return 1;
+ }
- for (tl = csp->config->trust_list; (t = *tl) ; tl++)
+ for (trusted_url = csp->config->trust_list; *trusted_url != NULL; trusted_url++)
{
- if ((t->port == 0) || (t->port == rhttp->port))
+ if (url_match(*trusted_url, rhttp))
{
- if ((t->domain[0] == '\0') || domaincmp(t, url) == 0)
- {
- if ((t->path == NULL) ||
-#ifdef REGEX
- (regexec(t->preg, rhttp->path, 0, NULL, 0) == 0)
-#else
- (strncmp(t->path, rhttp->path, t->pathlen) == 0)
-#endif
- )
- {
- /* if the URL's referrer is from a trusted referrer, then
- * add the target spec to the trustfile as an unblocked
- * domain and return NULL (which means it's OK).
- */
+ /* if the URL's referrer is from a trusted referrer, then
+ * add the target spec to the trustfile as an unblocked
+ * domain and return NULL (which means it's OK).
+ */
- FILE *fp;
+ FILE *fp;
- freez(url->dbuf);
- freez(url->dvec);
+ if ((fp = fopen(csp->config->trustfile, "a")))
+ {
+ char * path;
+ char * path_end;
+ char * new_entry = strdup("~");
- if ((fp = fopen(csp->config->trustfile, "a")))
+ string_append(&new_entry, csp->http->hostport);
+
+ path = csp->http->path;
+ if ( (path[0] == '/')
+ && (path[1] == '~')
+ && ((path_end = strchr(path + 2, '/')) != NULL))
+ {
+ /* since this path points into a user's home space
+ * be sure to include this spec in the trustfile.
+ */
+ int path_len = path_end - path; /* save offset */
+ path = strdup(path); /* Copy string */
+ if (path != NULL)
{
- h = NULL;
-
- h = strsav(h, "~");
- h = strsav(h, csp->http->hostport);
-
- p = csp->http->path;
- if ((*p++ == '/')
- && (*p++ == '~'))
- {
- /* since this path points into a user's home space
- * be sure to include this spec in the trustfile.
- */
- if ((p = strchr(p, '/')))
- {
- *p = '\0';
- h = strsav(h, csp->http->path); /* FIXME: p?! */
- h = strsav(h, "/");
- }
- }
-
- fprintf(fp, "%s\n", h);
- freez(h);
- fclose(fp);
+ path_end = path + path_len; /* regenerate ptr to new buffer */
+ *(path_end + 1) = '\0'; /* Truncate path after '/' */
}
- return(0);
+ string_join(&new_entry, path);
+ }
+
+ if (new_entry != NULL)
+ {
+ fprintf(fp, "%s\n", new_entry);
+ free(new_entry);
+ }
+ else
+ {
+ /* FIXME: No way to handle out-of memory, so mostly ignoring it */
+ log_error(LOG_LEVEL_ERROR, "Out of memory adding pattern to trust file");
}
+
+ fclose(fp);
}
+ return 0;
}
}
- return(1);
+ return 1;
}
#endif /* def FEATURE_COOKIE_JAR */
struct http_request *http,
struct url_actions *b)
{
- struct url_spec url[1];
-
if (b == NULL)
{
/* Should never happen */
return;
}
- *url = dsplit(http->host);
-
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL)
- {
- return;
- }
-
for (b = b->next; NULL != b; b = b->next)
{
- if ((b->url->port == 0) || (b->url->port == http->port))
+ if (url_match(b->url, http))
{
- if ((b->url->domain[0] == '\0') || (domaincmp(b->url, url) == 0))
- {
- if ((b->url->path == NULL) ||
-#ifdef REGEX
- (regexec(b->url->preg, http->path, 0, NULL, 0) == 0)
-#else
- (strncmp(b->url->path, http->path, b->url->pathlen) == 0)
-#endif
- )
- {
- merge_current_action(action, b->action);
- }
- }
+ merge_current_action(action, b->action);
}
}
-
- freez(url->dbuf);
- freez(url->dvec);
}
{
static const struct forward_spec fwd_default[1] = { FORWARD_SPEC_INITIALIZER };
struct forward_spec *fwd = csp->config->forward;
- struct url_spec url[1];
if (fwd == NULL)
{
- return(fwd_default);
- }
-
- *url = dsplit(http->host);
-
- /* if splitting the domain fails, punt */
- if (url->dbuf == NULL)
- {
- return(fwd_default);
+ return fwd_default;
}
while (fwd != NULL)
{
- if ((fwd->url->port == 0) || (fwd->url->port == http->port))
+ if (url_match(fwd->url, http))
{
- if ((fwd->url->domain[0] == '\0') || (domaincmp(fwd->url, url) == 0))
- {
- if ((fwd->url->path == NULL) ||
-#ifdef REGEX
- (regexec(fwd->url->preg, http->path, 0, NULL, 0) == 0)
-#else
- (strncmp(fwd->url->path, http->path, fwd->url->pathlen) == 0)
-#endif
- )
- {
- freez(url->dbuf);
- freez(url->dvec);
- return(fwd);
- }
- }
+ return fwd;
}
fwd = fwd->next;
}
- freez(url->dbuf);
- freez(url->dvec);
- return(fwd_default);
-
-}
-
-
-/*********************************************************************
- *
- * Function : dsplit
- *
- * Description : Takes a domain and returns a pointer to a url_spec
- * structure populated with dbuf, dcnt and dvec. The
- * other fields in the structure that is returned are zero.
- *
- * Parameters :
- * 1 : domain = a URL address
- *
- * Returns : url_spec structure populated with dbuf, dcnt and dvec.
- * On error, the dbuf field will be set to NULL. (As
- * will all the others, but you don't need to check
- * them).
- *
- * FIXME: Returning a structure is horribly inefficient, please can
- * this structure take a (struct url_spec * dest)
- * pointer instead?
- *
- *********************************************************************/
-struct url_spec dsplit(char *domain)
-{
- struct url_spec ret[1];
- char *v[BUFFER_SIZE];
- int size;
- char *p;
-
- memset(ret, '\0', sizeof(*ret));
-
- if (domain[strlen(domain) - 1] == '.')
- {
- ret->unanchored |= ANCHOR_RIGHT;
- }
-
- if (domain[0] == '.')
- {
- ret->unanchored |= ANCHOR_LEFT;
- }
-
- ret->dbuf = strdup(domain);
- if (NULL == ret->dbuf)
- {
- return *ret;
- }
-
- /* map to lower case */
- for (p = ret->dbuf; *p ; p++)
- {
- *p = tolower((int)(unsigned char)*p);
- }
-
- /* split the domain name into components */
- ret->dcnt = ssplit(ret->dbuf, ".", v, SZ(v), 1, 1);
-
- if (ret->dcnt < 0)
- {
- free(ret->dbuf);
- memset(ret, '\0', sizeof(ret));
- return *ret;
- }
- else if (ret->dcnt == 0)
- {
- return *ret;
- }
-
- /* save a copy of the pointers in dvec */
- size = ret->dcnt * sizeof(*ret->dvec);
-
- ret->dvec = (char **)malloc(size);
- if (NULL == ret->dvec)
- {
- free(ret->dbuf);
- memset(ret, '\0', sizeof(ret));
- return *ret;
- }
-
- memcpy(ret->dvec, v, size);
-
- return *ret;
-
-}
-
-
-/*********************************************************************
- *
- * Function : simple_domaincmp
- *
- * Description : Domain-wise Compare fqdn's. The comparison is
- * both left- and right-anchored. The individual
- * domain names are compared with simplematch().
- * This is only used by domaincmp.
- *
- * Parameters :
- * 1 : pv = array of patterns to compare
- * 2 : fv = array of domain components to compare
- * 3 : len = length of the arrays (both arrays are the
- * same length - if they weren't, it couldn't
- * possibly be a match).
- *
- * Returns : 0 => domains are equivalent, else no match.
- *
- *********************************************************************/
-static int simple_domaincmp(char **pv, char **fv, int len)
-{
- int n;
-
- for (n = 0; n < len; n++)
- {
- if (simplematch(pv[n], fv[n]))
- {
- return 1;
- }
- }
-
- return 0;
-
-}
-
-
-/*********************************************************************
- *
- * Function : domaincmp
- *
- * Description : Domain-wise Compare fqdn's. Governed by the bimap in
- * pattern->unachored, the comparison is un-, left-,
- * right-anchored, or both.
- * The individual domain names are compared with
- * simplematch().
- *
- * Parameters :
- * 1 : pattern = a domain that may contain a '*' as a wildcard.
- * 2 : fqdn = domain name against which the patterns are compared.
- *
- * Returns : 0 => domains are equivalent, else no match.
- *
- *********************************************************************/
-int domaincmp(struct url_spec *pattern, struct url_spec *fqdn)
-{
- char **pv, **fv; /* vectors */
- int plen, flen;
- int unanchored = pattern->unanchored & (ANCHOR_RIGHT | ANCHOR_LEFT);
-
- plen = pattern->dcnt;
- flen = fqdn->dcnt;
-
- if (flen < plen)
- {
- /* fqdn is too short to match this pattern */
- return 1;
- }
-
- pv = pattern->dvec;
- fv = fqdn->dvec;
-
- if (unanchored == ANCHOR_LEFT)
- {
- /*
- * Right anchored.
- *
- * Convert this into a fully anchored pattern with
- * the fqdn and pattern the same length
- */
- fv += (flen - plen); /* flen - plen >= 0 due to check above */
- return simple_domaincmp(pv, fv, plen);
- }
- else if (unanchored == 0)
- {
- /* Fully anchored, check length */
- if (flen != plen)
- {
- return 1;
- }
- return simple_domaincmp(pv, fv, plen);
- }
- else if (unanchored == ANCHOR_RIGHT)
- {
- /* Left anchored, ignore all extra in fqdn */
- return simple_domaincmp(pv, fv, plen);
- }
- else
- {
- /* Unanchored */
- int n;
- int maxn = flen - plen;
- for (n = 0; n <= maxn; n++)
- {
- if (!simple_domaincmp(pv, fv, plen))
- {
- return 0;
- }
- /*
- * Doesn't match from start of fqdn
- * Try skipping first part of fqdn
- */
- fv++;
- }
- return 1;
- }
-
+ return fwd_default;
}