Fix XMP parser errors due to trailing nullchar

JPEG files can have trailing \0 chars at the end of the XMP value. Use
trim() to remove these from the string value.

Bug: T118799
Change-Id: Id4ab223ef432e5d2c0dd3b4e332320db02422700

diff --git a/includes/media/JpegMetadataExtractor.php b/includes/media/JpegMetadataExtractor.php
index 8a26f60..dffb1f9 100644 (file)
--- a/includes/media/JpegMetadataExtractor.php
+++ b/includes/media/JpegMetadataExtractor.php
@@ -122,14 +122,17 @@ class JpegMetadataExtractor {
                                $temp = self::jpegExtractMarker( $fh );
                                // check what type of app segment this is.
                                if ( substr( $temp, 0, 29 ) === "http://ns.adobe.com/xap/1.0/\x00" && $showXMP ) {
-                                       $segments["XMP"] = substr( $temp, 29 );
+                                       // use trim to remove trailing \0 chars
+                                       $segments["XMP"] = trim( substr( $temp, 29 ) );
                                } elseif ( substr( $temp, 0, 35 ) === "http://ns.adobe.com/xmp/extension/\x00" && $showXMP ) {
-                                       $segments["XMP_ext"][] = substr( $temp, 35 );
+                                       // use trim to remove trailing \0 chars
+                                       $segments["XMP_ext"][] = trim( substr( $temp, 35 ) );
                                } elseif ( substr( $temp, 0, 29 ) === "XMP\x00://ns.adobe.com/xap/1.0/\x00" && $showXMP ) {
                                        // Some images (especially flickr images) seem to have this.
                                        // I really have no idea what the deal is with them, but
                                        // whatever...
-                                       $segments["XMP"] = substr( $temp, 29 );
+                                       // use trim to remove trailing \0 chars
+                                       $segments["XMP"] = trim( substr( $temp, 29 ) );
                                        wfDebug( __METHOD__ . ' Found XMP section with wrong app identifier '
                                                . "Using anyways.\n" );
                                } elseif ( substr( $temp, 0, 6 ) === "Exif\0\0" ) {

diff --git a/tests/phpunit/data/media/jpeg-xmp-nullchar.jpg b/tests/phpunit/data/media/jpeg-xmp-nullchar.jpg
new file mode 100644 (file)
index 0000000..76ae2b4
Binary files /dev/null and b/tests/phpunit/data/media/jpeg-xmp-nullchar.jpg differ

diff --git a/tests/phpunit/includes/media/JpegMetadataExtractorTest.php b/tests/phpunit/includes/media/JpegMetadataExtractorTest.php
index 6c56510..4a0a6d0 100644 (file)
--- a/tests/phpunit/includes/media/JpegMetadataExtractorTest.php
+++ b/tests/phpunit/includes/media/JpegMetadataExtractorTest.php
@@ -1,7 +1,7 @@
 <?php
 /**
  * @todo Could use a test of extended XMP segments. Hard to find programs that
- * create example files, and creating my own in vim propbably wouldn't
+ * create example files, and creating my own in vim probably wouldn't
  * serve as a very good "test". (Adobe photoshop probably creates such files
  * but it costs money). The implementation of it currently in MediaWiki is based
  * solely on reading the standard, without any real world test files.
@@ -76,6 +76,12 @@ class JpegMetadataExtractorTest extends MediaWikiIntegrationTestCase {
                $this->assertEquals( $expected, bin2hex( $res['PSIR'][0] ) );
        }
 
+       public function testXMPExtractionNullChar() {
+               $res = JpegMetadataExtractor::segmentSplitter( $this->filePath . 'jpeg-xmp-nullchar.jpg' );
+               $expected = file_get_contents( $this->filePath . 'jpeg-xmp-psir.xmp' );
+               $this->assertEquals( $expected, $res['XMP'] );
+       }
+
        public function testXMPExtractionAltAppId() {
                $res = JpegMetadataExtractor::segmentSplitter( $this->filePath . 'jpeg-xmp-alt.jpg' );
                $expected = file_get_contents( $this->filePath . 'jpeg-xmp-psir.xmp' );