Untested image snapshot feature, fixed complete breakage of default skin by Rob

diff --git a/maintenance/dumpHTML.inc b/maintenance/dumpHTML.inc
index 6adcb8d..35a2ce7 100644 (file)
--- a/maintenance/dumpHTML.inc
+++ b/maintenance/dumpHTML.inc
@@ -21,7 +21,10 @@ class DumpHTML {
        var $depth = 3;
 
        # Directory that commons images are copied into
-       var $sharedStaticPath;
+       var $sharedStaticDirectory;
+
+       # Directory that the images are in, after copying
+       var $destUploadDirectory;
 
        # Relative path to image directory
        var $imageRel = 'upload';
@@ -29,6 +32,9 @@ class DumpHTML {
        # Copy commons images instead of symlinking
        var $forceCopy = false;
 
+       # Make a copy of all images encountered
+       var $makeSnapshot = false;
+
        # Make links assuming the script path is in the same directory as
        # the destination
        var $alternateScriptPath = false;
@@ -43,7 +49,7 @@ class DumpHTML {
        var $rawPages;
 
        # Skin to use
-       var $skin = 'dumphtml';
+       var $skin = 'htmldump';
 
        function DumpHTML( $settings ) {
                foreach ( $settings as $var => $value ) {
@@ -156,8 +162,8 @@ class DumpHTML {
                $i = 0;
                for ( $hash = 0; $hash < 256; $hash++ ) {
                        $dir = sprintf( "%01x/%02x", intval( $hash / 16 ), $hash );
-                       $paths = array_merge( glob( "{$this->sharedStaticPath}/$dir/*" ),
-                               glob( "{$this->sharedStaticPath}/thumb/$dir/*" ) );
+                       $paths = array_merge( glob( "{$this->sharedStaticDirectory}/$dir/*" ),
+                               glob( "{$this->sharedStaticDirectory}/thumb/$dir/*" ) );
 
                        foreach ( $paths as $path ) {
                                $file = basename( $path );
@@ -334,8 +340,6 @@ class DumpHTML {
                $wgCopyrightIcon = str_replace( 'src="/images',
                        'src="' . htmlspecialchars( $wgScriptPath ) . '/images', $this->oldCopyrightIcon );
 
-
-
                $wgStylePath = "$wgScriptPath/skins";
                $wgUploadPath = "$wgScriptPath/{$this->imageRel}";
                $wgSharedUploadPath = "$wgUploadPath/shared";
@@ -353,7 +357,14 @@ class DumpHTML {
                $wgUser->setOption( 'skin', $this->skin );
                $wgUser->setOption( 'editsection', 0 );
 
-               $this->sharedStaticPath = "$wgUploadDirectory/shared";
+               if ( $this->makeSnapshot ) {
+                       $this->destUploadDirectory = "{$this->dest}/{$this->imageRel}";
+                       if ( realpath( $this->destUploadDirectory == $wgUploadDirectory ) ) {
+                               $this->makeSnapshot = false;
+                       }
+               }
+
+               $this->sharedStaticDirectory = "{$this->destUploadDirectory}/shared";
 
                $this->setupDone = true;
        }
@@ -433,67 +444,72 @@ ENDTEXT;
                return $wgDumpImages;
        }
 
+       /**
+        * Copy a file specified by a URL to a given directory
+        * 
+        * @param string $srcPath The source URL
+        * @param string $srcPathBase The base directory of the source URL
+        * @param string $srcDirBase The base filesystem directory of the source URL
+        * @param string $destDirBase The base filesystem directory of the destination URL
+        */
+       function relativeCopy( $srcPath, $srcPathBase, $srcDirBase, $destDirBase ) {
+               $rel = substr( $srcPath, strlen( $srcPathBase ) + 1 ); // +1 for slash
+               $sourceLoc = "$srcDirBase/$rel";
+               $destLoc = "$destDirBase/$rel";
+               #print "Copying $sourceLoc to $staticLoc\n";
+               if ( !file_exists( $destLoc ) ) {
+                       wfMkdirParents( dirname( $destLoc ), 0755 );
+                       if ( function_exists( 'symlink' ) && !$this->forceCopy ) {
+                               symlink( $sourceLoc, $destLoc );
+                       } else {
+                               copy( $sourceLoc, $destLoc );
+                       }
+               }
+       }
+
+       /**
+        * Copy an image, and if it is a thumbnail, copy its parent image too
+        */
+       function copyImage( $srcPath, $srcPathBase, $srcDirBase, $destDirBase ) {
+               global $wgUploadPath, $wgUploadDirectory, $wgSharedUploadPath;
+               $this->relativeCopy( $srcPath, $srcPathBase, $srcDirBase, $destDirBase );
+               if ( substr( $srcPath, strlen( $srcPathBase ) + 1, 6 ) == 'thumb/' ) {
+                       # The image was a thumbnail
+                       # Copy the source image as well
+                       $rel = substr( $srcPath, strlen( $srcPathBase ) + 1 );
+                       $parts = explode( '/', $rel );
+                       $rel = "{$parts[1]}/{$parts[2]}/{$parts[3]}";
+                       $newSrc = "$srcPathBase/$rel";
+                       $this->relativeCopy( $newSrc, $srcPathBase, $srcDirBase, $destDirBase );
+               }
+       }
+       
        /**
         * Copy images (or create symlinks) from commons to a static directory.
         * This is necessary even if you intend to distribute all of commons, because
         * the directory contents is used to work out which image description pages
         * are needed.
         *
-        * Also copies math images
+        * Also copies math images, and full-sized images if the makeSnapshot option 
+        * is specified.
         *
         */
        function copyImages( $images ) {
-               global $wgSharedUploadPath, $wgSharedUploadDirectory, $wgMathPath, $wgMathDirectory;
+               global $wgUploadPath, $wgUploadDirectory, $wgSharedUploadPath, $wgSharedUploadDirectory, 
+                       $wgMathPath, $wgMathDirectory;
                # Find shared uploads and copy them into the static directory
                $sharedPathLength = strlen( $wgSharedUploadPath );
                $mathPathLength = strlen( $wgMathPath );
+               $uploadPathLength = strlen( $wgUploadPath );
                foreach ( $images as $escapedImage => $dummy ) {
                        $image = urldecode( $escapedImage );
 
-                       # Is it shared?
-                       if ( substr( $image, 0, $sharedPathLength ) == $wgSharedUploadPath ) {
-                               # Reconstruct full filename
-                               $rel = substr( $image, $sharedPathLength + 1 ); // +1 for slash
-                               $sourceLoc = "$wgSharedUploadDirectory/$rel";
-                               $staticLoc = "{$this->sharedStaticPath}/$rel";
-                               #print "Copying $sourceLoc to $staticLoc\n";
-                               # Copy to static directory
-                               if ( !file_exists( $staticLoc ) ) {
-                                       wfMkdirParents( dirname( $staticLoc ), 0755 );
-                                       if ( function_exists( 'symlink' ) && !$this->forceCopy ) {
-                                               symlink( $sourceLoc, $staticLoc );
-                                       } else {
-                                               copy( $sourceLoc, $staticLoc );
-                                       }
-                               }
-
-                               if ( substr( $rel, 0, 6 ) == 'thumb/' ) {
-                                       # That was a thumbnail
-                                       # We will also copy the real image
-                                       $parts = explode( '/', $rel );
-                                       $rel = "{$parts[1]}/{$parts[2]}/{$parts[3]}";
-                                       $sourceLoc = "$wgSharedUploadDirectory/$rel";
-                                       $staticLoc = "{$this->sharedStaticPath}/$rel";
-                                       #print "Copying $sourceLoc to $staticLoc\n";
-                                       if ( !file_exists( $staticLoc ) ) {
-                                               wfMkdirParents( dirname( $staticLoc ), 0755 );
-                                               if ( function_exists( 'symlink' ) && !$this->forceCopy ) {
-                                                       symlink( $sourceLoc, $staticLoc );
-                                               } else {
-                                                       copy( $sourceLoc, $staticLoc );
-                                               }
-                                       }
-                               }
-                       } else
-                       # Is it math?
-                       if ( substr( $image, 0, $mathPathLength ) == $wgMathPath ) {
-                               $rel = substr( $image, $mathPathLength + 1 ); // +1 for slash
-                               $source = "$wgMathDirectory/$rel";
-                               $dest = "{$this->dest}/math/$rel";
-                               @mkdir( "{$this->dest}/math", 0755 );
-                               if ( !file_exists( $dest ) ) {
-                                       copy( $source, $dest );
-                               }
+                       if ( $this->makeSnapshot && substr( $image, 0, $uploadPathLength ) == $wgUploadPath ) {
+                               $this->copyImage( $image, $wgUploadPath, $wgUploadDirectory, $this->destUploadDirectory );
+                       } elseif ( substr( $image, 0, $sharedPathLength ) == $wgSharedUploadPath ) {
+                               $this->copyImage( $image, $wgSharedUploadPath, $wgSharedUploadDirectory, $this->sharedStaticDirectory );
+                       } elseif ( substr( $image, 0, $mathPathLength ) == $wgMathPath ) {
+                               $this->relativeCopy( $image, $wgMathPath, $wgMathDirectory, "{$this->dest}/math" );
                        }
                }
        }

diff --git a/maintenance/dumpHTML.php b/maintenance/dumpHTML.php
index 37a4646..cd70140 100644 (file)
--- a/maintenance/dumpHTML.php
+++ b/maintenance/dumpHTML.php
@@ -9,16 +9,17 @@
  * Usage:
  * php dumpHTML.php [options...]
  *
- * -d <dest>      destination directory
- * -s <start>     start ID
- * -e <end>       end ID
- * -k <skin>     skin to use (defaults to dumphtml)
- * --images       only do image description pages
- * --categories   only do category pages
- * --redirects    only do redirects
- * --special      only do miscellaneous stuff
- * --force-copy   copy commons instead of symlink, needed for Wikimedia
- * --interlang    allow interlanguage links
+ * -d <dest>          destination directory
+ * -s <start>         start ID
+ * -e <end>           end ID
+ * -k <skin>          skin to use (defaults to htmldump)
+ * --images           only do image description pages
+ * --categories       only do category pages
+ * --redirects        only do redirects
+ * --special          only do miscellaneous stuff
+ * --force-copy       copy commons instead of symlink, needed for Wikimedia
+ * --interlang        allow interlanguage links
+ * --image-snapshot   copy all images used to the destination directory
  */
 
 
@@ -61,7 +62,7 @@ if ( !empty( $options['d'] ) ) {
        $dest = 'static';
 }
 
-$skin = isset( $options['k'] ) ? $options['k'] : 'dumphtml';
+$skin = isset( $options['k'] ) ? $options['k'] : 'htmldump';
 
 $wgHTMLDump = new DumpHTML( array(
        'dest' => $dest,
@@ -69,6 +70,7 @@ $wgHTMLDump = new DumpHTML( array(
        'alternateScriptPath' => $options['interlang'],
        'interwiki' => $options['interlang'],
        'skin' => $skin,
+       'makeSnapshot' => $options['snapshot'],
 ));