Various thumb.php fixes.

* Made wfThumbHandle404() handle full REQUEST_URI urls.
* Made wfExtractThumbParams() handle protocal relative zone urls.
* Added FileRepo::getZoneHandlerUrl() and site zone configuration to simplify wmf rewrite rules.
* Renamed some variables to be less misleading.

Change-Id: Ic3e23c2a623c1241ee22a9811aee073fb07aa68c

diff --git a/includes/DefaultSettings.php b/includes/DefaultSettings.php
index d7feab8..abf25dc 100644 (file)
--- a/includes/DefaultSettings.php
+++ b/includes/DefaultSettings.php
@@ -358,9 +358,11 @@ $wgImgAuthPublicTest = true;
  *
  * For most core repos:
  *   - zones            Associative array of zone names that each map to an array with:
- *                          container : backend container name the zone is in
- *                          directory : root path within container for the zone
- *                          url       : base URL to the root of the zone
+ *                          container  : backend container name the zone is in
+ *                          directory  : root path within container for the zone
+ *                          url        : base URL to the root of the zone
+ *                          handlerUrl : base script handled URL to the root of the zone
+ *                                       (see FileRepo::getZoneHandlerUrl() function)
  *                      Zones default to using "<repo name>-<zone name>" as the container name
  *                      and default to using the container root as the zone's root directory.
  *                      Nesting of zone locations within other zones should be avoided.

diff --git a/includes/filerepo/FileRepo.php b/includes/filerepo/FileRepo.php
index 5a55096..32ef5a9 100644 (file)
--- a/includes/filerepo/FileRepo.php
+++ b/includes/filerepo/FileRepo.php
@@ -216,6 +216,28 @@ class FileRepo {
                }
        }
 
+       /**
+        * Get the thumb zone URL configured to be handled by scripts like thumb_handler.php.
+        * This is probably only useful for internal requests, such as from a fast frontend server
+        * to a slower backend server.
+        *
+        * Large sites may use a different host name for uploads than for wikis. In any case, the
+        * wiki configuration is needed in order to use thumb.php. To avoid extracting the wiki ID
+        * from the URL path, one can configure thumb_handler.php to recognize a special path on the
+        * same host name as the wiki that is used for viewing thumbnails.
+        *
+        * @param $zone String: one of: public, deleted, temp, thumb
+        * @return String or false
+        */
+       public function getZoneHandlerUrl( $zone ) {
+               if ( isset( $this->zones[$zone]['handlerUrl'] )
+                       && in_array( $zone, array( 'public', 'temp', 'thumb' ) ) )
+               {
+                       return $this->zones[$zone]['handlerUrl'];
+               }
+               return false;
+       }
+
        /**
         * Get the backend storage path corresponding to a virtual URL.
         * Use this function wisely.

diff --git a/thumb.php b/thumb.php
index 50d3754..9cfdae4 100644 (file)
--- a/thumb.php
+++ b/thumb.php
@@ -65,19 +65,22 @@ function wfThumbHandle404() {
        # that to the 404 handler, and puts the original request in REDIRECT_URL.
        if ( isset( $_SERVER['REDIRECT_URL'] ) ) {
                # The URL is un-encoded, so put it back how it was
-               $uri = str_replace( "%2F", "/", urlencode( $_SERVER['REDIRECT_URL'] ) );
-               # Just get the URI path (REDIRECT_URL is either a full URL or a path)
-               if ( $uri[0] !== '/' ) {
-                       $bits = wfParseUrl( $uri );
-                       if ( $bits && isset( $bits['path'] ) ) {
-                               $uri = $bits['path'];
-                       }
-               }
+               $uriPath = str_replace( "%2F", "/", urlencode( $_SERVER['REDIRECT_URL'] ) );
        } else {
-               $uri = $_SERVER['REQUEST_URI'];
+               $uriPath = $_SERVER['REQUEST_URI'];
+       }
+       # Just get the URI path (REDIRECT_URL/REQUEST_URI is either a full URL or a path)
+       if ( substr( $uriPath, 0, 1 ) !== '/' ) {
+               $bits = wfParseUrl( $uriPath );
+               if ( $bits && isset( $bits['path'] ) ) {
+                       $uriPath = $bits['path'];
+               } else {
+                       wfThumbError( 404, 'The source file for the specified thumbnail does not exist.' );
+                       return;
+               }
        }
 
-       $params = wfExtractThumbParams( $uri ); // basic wiki URL param extracting
+       $params = wfExtractThumbParams( $uriPath ); // basic wiki URL param extracting
        if ( $params == null ) {
                wfThumbError( 404, 'The source file for the specified thumbnail does not exist.' );
                return;
@@ -265,22 +268,22 @@ function wfStreamThumb( array $params ) {
  * Extract the required params for thumb.php from the thumbnail request URI.
  * At least 'width' and 'f' should be set if the result is an array.
  *
- * @param $uri String Thumbnail request URI path
+ * @param $uriPath String Thumbnail request URI path
  * @return Array|null associative params array or null
  */
-function wfExtractThumbParams( $uri ) {
+function wfExtractThumbParams( $uriPath ) {
        $repo = RepoGroup::singleton()->getLocalRepo();
 
-       $zoneURI = $repo->getZoneUrl( 'thumb' );
-       if ( substr( $zoneURI, 0, 1 ) !== '/' ) {
-               $bits = wfParseUrl( $zoneURI );
-               if ( $bits && isset( $bits['path'] ) ) {
-                       $zoneURI = $bits['path'];
-               } else {
-                       return null;
-               }
+       $zoneUriPath = $repo->getZoneHandlerUrl( 'thumb' )
+               ? $repo->getZoneHandlerUrl( 'thumb' ) // custom URL
+               : $repo->getZoneUrl( 'thumb' ); // default to main URL
+       // URL might be relative ("/images") or protocol-relative ("//lang.site/image")
+       $bits = wfParseUrl( wfExpandUrl( $zoneUriPath, PROTO_INTERNAL ) );
+       if ( $bits && isset( $bits['path'] ) ) {
+               $zoneUriPath = $bits['path'];
+       } else {
+               return null;
        }
-       $zoneUrlRegex = preg_quote( $zoneURI );
 
        $hashDirRegex = $subdirRegex = '';
        for ( $i = 0; $i < $repo->getHashLevels(); $i++ ) {
@@ -288,10 +291,11 @@ function wfExtractThumbParams( $uri ) {
                $hashDirRegex .= "$subdirRegex/";
        }
 
-       $thumbUrlRegex = "!^$zoneUrlRegex/((archive/|temp/)?$hashDirRegex([^/]*)/([^/]*))$!";
+       $thumbPathRegex = "!^" . preg_quote( $zoneUriPath ) .
+               "/((archive/|temp/)?$hashDirRegex([^/]*)/([^/]*))$!";
 
        // Check if this is a valid looking thumbnail request...
-       if ( preg_match( $thumbUrlRegex, $uri, $matches ) ) {
+       if ( preg_match( $thumbPathRegex, $uriPath, $matches ) ) {
                list( /* all */, $rel, $archOrTemp, $filename, $thumbname ) = $matches;
                $filename = urldecode( $filename );
                $thumbname = urldecode( $thumbname );