gzipped-xml-types-1.21.diff

commit 39bae5c32f39dfdf06f6b8494df2bc40459f3848
Author: Vitaliy Filippov <vitalif@yourcmc.ru>
Date:   Thu Jan 16 20:07:21 2014 +0400

    Patch: gzipped-xml-types
    Type: feature
    
    SVGZ and generic XML-GZIP mime-types support. CustIS Bug 63356

diff --git a/includes/DefaultSettings.php b/includes/DefaultSettings.php
index ee77e25..2494fa2 100644
--- a/includes/DefaultSettings.php
+++ b/includes/DefaultSettings.php
@@ -777,6 +777,7 @@ $wgTrustedMediaFormats = array(
 	MEDIATYPE_AUDIO, //all audio formats
 	MEDIATYPE_VIDEO, //all plain video formats
 	"image/svg+xml", //svg (only needed if inline rendering of svg is not supported)
+	"image/svg+xml+gzip", //gzipped svg
 	"application/pdf", //PDF files
 	#"application/x-shockwave-flash", //flash/shockwave movie
 );
@@ -795,6 +796,7 @@ $wgMediaHandlers = array(
 	'image/x-xcf' => 'XCFHandler',
 	'image/svg+xml' => 'SvgHandler', // official
 	'image/svg' => 'SvgHandler', // compat
+	'image/svg+xml+gzip' => 'SvgHandler', // gzipped
 	'image/vnd.djvu' => 'DjVuHandler', // official
 	'image/x.djvu' => 'DjVuHandler', // compat
 	'image/x-djvu' => 'DjVuHandler', // compat
@@ -1106,6 +1108,11 @@ $wgXMLMimeTypes = array(
 	'html' => 'text/html', // application/xhtml+xml?
 );
 
+$wgCompressedXMLTypes = array(
+	'image/svg+xml' => 'image/svg+xml+gzip',
+	'application/x-dia-diagram' => 'application/x-dia-diagram',
+);
+
 /**
  * Limit images on image description pages to a user-selectable limit. In order
  * to reduce disk usage, limits can only be selected from a list.
diff --git a/includes/MimeMagic.php b/includes/MimeMagic.php
index edabd54..616d1e1 100644
--- a/includes/MimeMagic.php
+++ b/includes/MimeMagic.php
@@ -483,7 +483,7 @@ class MimeMagic {
 			'webp',
 
 			// XML formats we sure hope we recognize reliably
-			'svg',
+			'svg', 'svgz',
 		);
 		return in_array( strtolower( $extension ), $types );
 	}
@@ -675,11 +675,16 @@ class MimeMagic {
 		 */
 		$xml = new XmlTypeCheck( $file );
 		if ( $xml->wellFormed ) {
-			global $wgXMLMimeTypes;
-			if ( isset( $wgXMLMimeTypes[$xml->getRootElement()] ) ) {
-				return $wgXMLMimeTypes[$xml->getRootElement()];
-			} else {
-				return 'application/xml';
+			global $wgXMLMimeTypes, $wgCompressedXMLTypes;
+			$t = $wgXMLMimeTypes[$xml->getRootElement()];
+			if ( !$xml->compressed ) {
+				if ( $t ) {
+					return $t;
+				} else {
+					return 'application/xml';
+				}
+			} elseif ( isset( $wgCompressedXMLTypes[$t] ) ) {
+				return $wgCompressedXMLTypes[$t];
 			}
 		}
 
diff --git a/includes/XmlTypeCheck.php b/includes/XmlTypeCheck.php
index 2e18460..e6a8a7a 100644
--- a/includes/XmlTypeCheck.php
+++ b/includes/XmlTypeCheck.php
@@ -40,7 +40,13 @@ class XmlTypeCheck {
 	public $rootElement = '';
 
 	/**
-	 * @param string $file filename
+	 * Name of file compression type (can be only 'gzip' by now),
+	 * or FALSE if the file is uncompressed.
+	 */
+	public $compressed = false;
+
+	/**
+	 * @param $file string filename
 	 * @param $filterCallback callable (optional)
 	 *        Function to call to do additional custom validity checks from the
 	 *        SAX element handler event. This gives you access to the element
@@ -75,6 +81,19 @@ class XmlTypeCheck {
 		if ( file_exists( $fname ) ) {
 			$file = fopen( $fname, "rb" );
 			if ( $file ) {
+				$gz = fread( $file, 2 );
+				if ( $gz == "\x1F\x8B" ) {
+					if ( function_exists( 'gzopen' ) ) {
+						fclose( $file );
+						$this->compressed = 'gzip';
+						$file = gzopen( $fname, "rb" );
+					} else {
+						return;
+					}
+				} else {
+					fseek( $file, 0, SEEK_SET );
+				}
+
 				do {
 					$chunk = fread( $file, 32768 );
 					$ret = xml_parse( $parser, $chunk, feof( $file ) );
diff --git a/includes/media/SVGMetadataExtractor.php b/includes/media/SVGMetadataExtractor.php
index 0de212b..409697a 100644
--- a/includes/media/SVGMetadataExtractor.php
+++ b/includes/media/SVGMetadataExtractor.php
@@ -64,16 +64,24 @@ class SVGReader {
 			throw new MWException( "Error getting filesize of SVG." );
 		}
 
-		if ( $size > $wgSVGMetadataCutoff ) {
-			$this->debug( "SVG is $size bytes, which is bigger than $wgSVGMetadataCutoff. Truncating." );
-			$contents = file_get_contents( $source, false, null, -1, $wgSVGMetadataCutoff );
-			if ( $contents === false ) {
-				throw new MWException( 'Error reading SVG file.' );
+		$fp = fopen( $source, 'rb' );
+		if ( !$fp ) {
+			throw new MWException( 'Error reading SVG file.' );
+		}
+		if ( function_exists( 'gzopen' ) && fread( $fp, 3 ) == "\x1f\x8b\x08" ) {
+			fclose( $fp );
+			$fp = gzopen( $source, 'rb' );
+			if ( !$fp ) {
+				throw new MWException( 'Error reading gzip-compressed SVG file.' );
 			}
-			$this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
+			$this->metadata['compressed'] = true;
 		} else {
-			$this->reader->open( $source, null, LIBXML_NOERROR | LIBXML_NOWARNING );
+			fseek( $fp, 0, 0 );
+			$this->metadata['compressed'] = false;
 		}
+		$contents = stream_get_contents( $fp, $wgSVGMetadataCutoff );
+		fclose( $fp );
+		$this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
 
 		// Expand entities, since Adobe Illustrator uses them for xmlns
 		// attributes (bug 31719). Note that libxml2 has some protection
@@ -88,6 +96,7 @@ class SVGReader {
 
 		$this->metadata['width'] = self::DEFAULT_WIDTH;
 		$this->metadata['height'] = self::DEFAULT_HEIGHT;
+		$this->metadata['viewBox'] = '';
 
 		// The size in the units specified by the SVG file
 		// (for the metadata box)
@@ -294,7 +303,8 @@ class SVGReader {
 
 		if( $this->reader->getAttribute( 'viewBox' ) ) {
 			// min-x min-y width height
-			$viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute( 'viewBox' ) ) );
+			$this->metadata['viewBox'] = trim( $this->reader->getAttribute('viewBox') );
+			$viewBox = preg_split( '/\s+/', $this->metadata['viewBox'] );
 			if( count( $viewBox ) == 4 ) {
 				$viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 				$viewHeight = $this->scaleSVGUnit( $viewBox[3] );
diff --git a/includes/mime.types b/includes/mime.types
index a89d229..e9856f5 100644
--- a/includes/mime.types
+++ b/includes/mime.types
@@ -85,6 +85,7 @@ image/jp2 j2k jp2 jpg2
 image/jpeg jpeg jpg jpe
 image/png png apng
 image/svg+xml svg
+image/svg+xml+gzip svgz
 image/tiff tiff tif
 image/vnd.djvu djvu djv
 image/vnd.wap.wbmp wbmp