From 2d1eae820f0915a561494c1a12fd2dec1759e6fd Mon Sep 17 00:00:00 2001
From: Andrew Nacin <wp@andrewnacin.com>
Date: Tue, 18 Sep 2012 23:38:25 +0000
Subject: [PATCH] Avoid mangling UTF-8 strings that may be present in image
 metadata. props SergeyBiryukov for the unit tests [UT665]. fixes #9417.

git-svn-id: http://core.svn.wordpress.org/trunk@21905 1a063a9b-81f0-0310-95a4-ce76da25c4cd
---
 wp-admin/includes/image.php | 37 +++++++++++++++++++++----------------
 1 file changed, 21 insertions(+), 16 deletions(-)

diff --git a/wp-admin/includes/image.php b/wp-admin/includes/image.php
index 2de1a7ee3b..5af65465d9 100644
--- a/wp-admin/includes/image.php
+++ b/wp-admin/includes/image.php
@@ -237,13 +237,13 @@ function wp_read_image_metadata( $file ) {
 
 			// headline, "A brief synopsis of the caption."
 			if ( ! empty( $iptc['2#105'][0] ) )
-				$meta['title'] = utf8_encode( trim( $iptc['2#105'][0] ) );
+				$meta['title'] = trim( $iptc['2#105'][0] );
 			// title, "Many use the Title field to store the filename of the image, though the field may be used in many ways."
 			elseif ( ! empty( $iptc['2#005'][0] ) )
-				$meta['title'] = utf8_encode( trim( $iptc['2#005'][0] ) );
+				$meta['title'] = trim( $iptc['2#005'][0] );
 
 			if ( ! empty( $iptc['2#120'][0] ) ) { // description / legacy caption
-				$caption = utf8_encode( trim( $iptc['2#120'][0] ) );
+				$caption = trim( $iptc['2#120'][0] );
 				if ( empty( $meta['title'] ) ) {
 					// Assume the title is stored in 2:120 if it's short.
 					if ( strlen( $caption ) < 80 )
@@ -256,15 +256,15 @@ function wp_read_image_metadata( $file ) {
 			}
 
 			if ( ! empty( $iptc['2#110'][0] ) ) // credit
-				$meta['credit'] = utf8_encode(trim($iptc['2#110'][0]));
+				$meta['credit'] = trim( $iptc['2#110'][0] );
 			elseif ( ! empty( $iptc['2#080'][0] ) ) // creator / legacy byline
-				$meta['credit'] = utf8_encode(trim($iptc['2#080'][0]));
+				$meta['credit'] = trim( $iptc['2#080'][0] );
 
 			if ( ! empty( $iptc['2#055'][0] ) and ! empty( $iptc['2#060'][0] ) ) // created date and time
 				$meta['created_timestamp'] = strtotime( $iptc['2#055'][0] . ' ' . $iptc['2#060'][0] );
 
 			if ( ! empty( $iptc['2#116'][0] ) ) // copyright
-				$meta['copyright'] = utf8_encode( trim( $iptc['2#116'][0] ) );
+				$meta['copyright'] = trim( $iptc['2#116'][0] );
 		 }
 	}
 
@@ -273,44 +273,49 @@ function wp_read_image_metadata( $file ) {
 		$exif = @exif_read_data( $file );
 
 		if ( !empty( $exif['Title'] ) )
-			$meta['title'] = utf8_encode( trim( $exif['Title'] ) );
+			$meta['title'] = trim( $exif['Title'] );
 
 		if ( ! empty( $exif['ImageDescription'] ) ) {
 			if ( empty( $meta['title'] ) && strlen( $exif['ImageDescription'] ) < 80 ) {
 				// Assume the title is stored in ImageDescription
-				$meta['title'] = utf8_encode( trim( $exif['ImageDescription'] ) );
+				$meta['title'] = trim( $exif['ImageDescription'] );
 				if ( ! empty( $exif['COMPUTED']['UserComment'] ) && trim( $exif['COMPUTED']['UserComment'] ) != $meta['title'] )
-					$meta['caption'] = utf8_encode( trim( $exif['COMPUTED']['UserComment'] ) );
+					$meta['caption'] = trim( $exif['COMPUTED']['UserComment'] );
 			} elseif ( trim( $exif['ImageDescription'] ) != $meta['title'] ) {
-				$meta['caption'] = utf8_encode( trim( $exif['ImageDescription'] ) );
+				$meta['caption'] = trim( $exif['ImageDescription'] );
 			}
 		} elseif ( ! empty( $exif['Comments'] ) && trim( $exif['Comments'] ) != $meta['title'] ) {
-			$meta['caption'] = utf8_encode( trim( $exif['Comments'] ) );
+			$meta['caption'] = trim( $exif['Comments'] );
 		}
 
 		if ( ! empty( $exif['Artist'] ) )
-			$meta['credit'] = utf8_encode( trim( $exif['Artist'] ) );
+			$meta['credit'] = trim( $exif['Artist'] );
 		elseif ( ! empty($exif['Author'] ) )
-			$meta['credit'] = utf8_encode( trim( $exif['Author'] ) );
+			$meta['credit'] = trim( $exif['Author'] );
 
 		if ( ! empty( $exif['Copyright'] ) )
-			$meta['copyright'] = utf8_encode( trim( $exif['Copyright'] ) );
+			$meta['copyright'] = trim( $exif['Copyright'] );
 		if ( ! empty($exif['FNumber'] ) )
 			$meta['aperture'] = round( wp_exif_frac2dec( $exif['FNumber'] ), 2 );
 		if ( ! empty($exif['Model'] ) )
-			$meta['camera'] = utf8_encode( trim( $exif['Model'] ) );
+			$meta['camera'] = trim( $exif['Model'] );
 		if ( ! empty($exif['DateTimeDigitized'] ) )
 			$meta['created_timestamp'] = wp_exif_date2ts($exif['DateTimeDigitized'] );
 		if ( ! empty($exif['FocalLength'] ) )
 			$meta['focal_length'] = wp_exif_frac2dec( $exif['FocalLength'] );
 		if ( ! empty($exif['ISOSpeedRatings'] ) ) {
 			$meta['iso'] = is_array( $exif['ISOSpeedRatings'] ) ? reset( $exif['ISOSpeedRatings'] ) : $exif['ISOSpeedRatings'];
-			$meta['iso'] = utf8_encode( trim( $meta['iso'] ) );
+			$meta['iso'] = trim( $meta['iso'] );
 		}
 		if ( ! empty($exif['ExposureTime'] ) )
 			$meta['shutter_speed'] = wp_exif_frac2dec( $exif['ExposureTime'] );
 	}
 
+	foreach ( array( 'title', 'caption', 'credit', 'copyright', 'camera', 'iso' ) as $key ) {
+		if ( $meta[ $key ] && ! seems_utf8( $meta[ $key ] ) )
+			$meta[ $key ] = utf8_encode( $meta[ $key ] );
+	}
+
 	return apply_filters( 'wp_read_image_metadata', $meta, $file, $sourceImageType );
 
 }