Reimplement segment splitting in get_last_url() by using ssplit()

It's available on all platforms and doesn't require locking.

While at it, properly detect URLs that are concatenated together
without separator again. All the redirect regression tests should
work now.

diff --git a/configure.in b/configure.in
index 8845863..bec11de 100644 (file)
--- a/configure.in
+++ b/configure.in
@@ -1,6 +1,6 @@
 dnl Process this file with autoconf to produce a configure script.
 dnl
-dnl $Id: configure.in,v 1.160 2011/09/04 11:11:17 fabiankeil Exp $
+dnl $Id: configure.in,v 1.161 2011/10/30 16:20:12 fabiankeil Exp $
 dnl
 dnl Written by and Copyright (C) 2001-2010 the
 dnl Privoxy team. http://www.privoxy.org/
@@ -32,7 +32,7 @@ dnl =================================================================
 dnl AutoConf Initialization
 dnl =================================================================
 
-AC_REVISION($Revision: 1.160 $)
+AC_REVISION($Revision: 1.161 $)
 AC_INIT(jcc.c)
 
 if test ! -f config.h.in; then
@@ -704,7 +704,7 @@ dnl bcopy is for PCRE
 AC_CHECK_FUNCS([bcopy])
 AC_PROG_GCC_TRADITIONAL
 AC_TYPE_SIGNAL
-AC_CHECK_FUNCS([access atexit getcwd gethostbyaddr gethostbyaddr_r gethostbyname gethostbyname_r gettimeofday inet_ntoa localtime_r memchr memmove memset poll putenv random regcomp select setlocale snprintf socket strchr strdup strerror strtok strftime strlcat strlcpy strptime strstr strtoul timegm tzset])
+AC_CHECK_FUNCS([access atexit getcwd gethostbyaddr gethostbyaddr_r gethostbyname gethostbyname_r gettimeofday inet_ntoa localtime_r memchr memmove memset poll putenv random regcomp select setlocale snprintf socket strchr strdup strerror strftime strlcat strlcpy strptime strstr strtoul timegm tzset])
 
 dnl Checks for RFC 2553 resolver and socket functions
 AC_ARG_ENABLE(ipv6-support,

diff --git a/filters.c b/filters.c
index 3994638..4f4684e 100644 (file)
--- a/filters.c
+++ b/filters.c
@@ -1,4 +1,4 @@
-const char filters_rcs[] = "$Id: filters.c,v 1.155 2011/11/06 11:36:42 fabiankeil Exp $";
+const char filters_rcs[] = "$Id: filters.c,v 1.156 2011/11/06 11:45:28 fabiankeil Exp $";
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/filters.c,v $
@@ -78,11 +78,6 @@ const char filters_rcs[] = "$Id: filters.c,v 1.155 2011/11/06 11:36:42 fabiankei
 #include "urlmatch.h"
 #include "loaders.h"
 
-#ifdef HAVE_STRTOK
-/* Only used for locks */
-#include "jcc.h"
-#endif /* def HAVE_STRTOK */
-
 #ifdef _WIN32
 #include "win32.h"
 #endif
@@ -1115,62 +1110,66 @@ char *get_last_url(char *subject, const char *redirect_mode)
       log_error(LOG_LEVEL_REDIRECTS,
          "Checking \"%s\" for encoded redirects.", subject);
 
-#if defined(MUTEX_LOCKS_AVAILABLE) && defined(HAVE_STRTOK)
       /*
        * Check each parameter in the URL separately.
        * Sectionize the URL at "?" and "&",
-       * then URL-decode each component,
+       * go backwards through the segments, URL-decode them
        * and look for a URL in the decoded result.
-       * Keep the last one we spot.
+       * Stop the search after the first match.
+       */
+      char *url_segment = NULL;
+      /*
+       * XXX: This estimate is guaranteed to be high enough as we
+       *      let ssplit() ignore empty fields, but also a bit wasteful.
        */
-      char *found = NULL;
+      size_t max_segments = strlen(subject) / 2;
+      char **url_segments = malloc(max_segments * sizeof(char *));
+      int segments;
+
+      if (NULL == url_segments)
+      {
+         log_error(LOG_LEVEL_ERROR, "Out of memory while decoding URL: %s", new_url);
+         freez(subject);
+         return NULL;
+      }
+
+      segments = ssplit(subject, "?&", url_segments, max_segments, 1, 1);
 
-      privoxy_mutex_lock(&strtok_mutex);
-      char *token = strtok(subject, "?&");
-      while (token)
+      while (segments-- > 0)
       {
-         char *dtoken = url_decode(token);
+         char *dtoken = url_decode(url_segments[segments]);
          if (NULL == dtoken)
          {
-            log_error(LOG_LEVEL_ERROR, "Unable to decode \"%s\".", token);
+            log_error(LOG_LEVEL_ERROR, "Unable to decode \"%s\".", url_segments[segments]);
             continue;
          }
-         char *http_url = strstr(dtoken, "http://");
-         char *https_url = strstr(dtoken, "https://");
-         char *last_url = (http_url && https_url
-                          ? (http_url < https_url ? http_url : https_url)
-                          : (http_url ? http_url : https_url));
-         if (last_url)
+         url_segment = strstr(dtoken, "http://");
+         if (NULL == url_segment)
          {
-            freez(found);
-            found = strdup(last_url);
-            if (found == NULL)
+            url_segment = strstr(dtoken, "https://");
+         }
+         if (NULL != url_segment)
+         {
+            url_segment = strdup(url_segment);
+            freez(dtoken);
+            if (url_segment == NULL)
             {
                log_error(LOG_LEVEL_ERROR,
                   "Out of memory while searching for redirects.");
-               privoxy_mutex_unlock(&strtok_mutex);
                return NULL;
             }
+            break;
          }
          freez(dtoken);
-         token = strtok(NULL, "?&");
       }
-      privoxy_mutex_unlock(&strtok_mutex);
       freez(subject);
+      freez(url_segments);
 
-      return found;
-#else
-      new_url = url_decode(subject);
-      if (new_url != NULL)
-      {
-         freez(subject);
-         subject = new_url;
-      }
-      else
+      if (url_segment == NULL)
       {
-         log_error(LOG_LEVEL_ERROR, "Unable to decode \"%s\".", subject);
+         return NULL;
       }
-#endif /* defined(MUTEX_LOCKS_AVAILABLE) && defined(HAVE_STRTOK) */
+      subject = url_segment;
    }
 
    /* Else, just look for a URL inside this one, without decoding anything. */

diff --git a/jcc.c b/jcc.c
index 3efa390..946acb0 100644 (file)
--- a/jcc.c
+++ b/jcc.c
@@ -1,4 +1,4 @@
-const char jcc_rcs[] = "$Id: jcc.c,v 1.371 2011/10/23 11:24:33 fabiankeil Exp $";
+const char jcc_rcs[] = "$Id: jcc.c,v 1.372 2011/10/30 16:20:12 fabiankeil Exp $";
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/jcc.c,v $
@@ -198,10 +198,6 @@ privoxy_mutex_t localtime_mutex;
 privoxy_mutex_t rand_mutex;
 #endif /* ndef HAVE_RANDOM */
 
-#ifdef HAVE_STRTOK
-privoxy_mutex_t strtok_mutex;
-#endif /* def HAVE_STRTOK */
-
 #endif /* def MUTEX_LOCKS_AVAILABLE */
 
 #if defined(unix)
@@ -2852,9 +2848,6 @@ static void initialize_mutexes(void)
    privoxy_mutex_init(&rand_mutex);
 #endif /* ndef HAVE_RANDOM */
 
-#ifdef HAVE_STRTOK
-   privoxy_mutex_init(&strtok_mutex);
-#endif /* def HAVE_STRTOK */
 #endif /* def MUTEX_LOCKS_AVAILABLE */
 }
 

diff --git a/jcc.h b/jcc.h
index e871f5d..0035a3e 100644 (file)
--- a/jcc.h
+++ b/jcc.h
@@ -1,6 +1,6 @@
 #ifndef JCC_H_INCLUDED
 #define JCC_H_INCLUDED
-#define JCC_H_VERSION "$Id: jcc.h,v 1.30 2011/09/04 11:10:56 fabiankeil Exp $"
+#define JCC_H_VERSION "$Id: jcc.h,v 1.31 2011/10/30 16:20:12 fabiankeil Exp $"
 /*********************************************************************
  *
  * File        :  $Source: /cvsroot/ijbswa/current/jcc.h,v $
@@ -99,10 +99,6 @@ extern privoxy_mutex_t resolver_mutex;
 extern privoxy_mutex_t rand_mutex;
 #endif /* ndef HAVE_RANDOM */
 
-#ifdef HAVE_STRTOK
-extern privoxy_mutex_t strtok_mutex;
-#endif /* ndef HAVE_STRTOK */
-
 #endif /* FEATURE_PTHREAD */
 
 /* Functions */