Detect input encodings from EXIF/IPTC data and convert to utf-8 as

appropriate.  This allows us to switch the exif value column back to
varchar and improves the way that we deal with non-utf8 data in our
embedded EXIF/IPTC data.

3 files changed, 16 insertions, 11 deletions

diff --git a/modules/exif/helpers/exif.php b/modules/exif/helpers/exif.php
index 9bdbc172..25d9eadd 100644
--- a/modules/exif/helpers/exif.php
+++ b/modules/exif/helpers/exif.php
@@ -37,10 +37,11 @@ class exif_Core {
       if (isset($exif_raw['ValidEXIFData'])) {
         foreach(self::_keys() as $field => $exifvar) {
           if (isset($exif_raw[$exifvar[0]][$exifvar[1]])) {
-            $data[] = sprintf(
-              "(%d, '%s', '%s')",
-              $item->id, $db->escape_str($field),
-              $db->escape_str($exif_raw[$exifvar[0]][$exifvar[1]]));
+            $value = $exif_raw[$exifvar[0]][$exifvar[1]];
+            if (mb_detect_encoding($value) != "UTF-8") {
+              $value = utf8_encode($value);
+            }
+            $data[] = sprintf("(%d, '%s', '%s')", $item->id, $field, $db->escape_str($value));
           }
         }
       }
@@ -49,10 +50,14 @@ class exif_Core {
         $iptc = iptcparse($info["APP13"]);
         foreach (array("Keywords" => "2#025", "Caption" => "2#120") as $keyword => $iptc_key) {
           if (!empty($iptc[$iptc_key])) {
+            $value = implode(" ", $iptc[$iptc_key]);
+            if (mb_detect_encoding($value) != "UTF-8") {
+              $value = utf8_encode($value);
+            }
             $data[] = sprintf(
               "(%d, '%s', '%s')",
-              $item->id, $db->escape_str($keyword),
-              $db->escape_str(implode(" ", $iptc[$iptc_key])));
+              $item->id, $keyword,
+              $db->escape_str($value));
           }
         }
       }
diff --git a/modules/exif/helpers/exif_installer.php b/modules/exif/helpers/exif_installer.php
index c5c0b027..4cd1a6b2 100644
--- a/modules/exif/helpers/exif_installer.php
+++ b/modules/exif/helpers/exif_installer.php
@@ -23,14 +23,11 @@ class exif_installer {
 
     if ($version == 0) {
       $db = Database::instance();
-      // we use blob as the data type for the value column because occasionally there's binary
-      // data embedded in the caption field for photos.  This preserves the binary data at the
-      // expense of natural language collation (which we don't use anyway).
       $db->query("CREATE TABLE IF NOT EXISTS {exif_keys} (
                    `id` int(9) NOT NULL auto_increment,
                    `item_id` int(9) NOT NULL,
                    `name` varchar(64) NOT NULL,
-                   `value` varbinary(1536) NOT NULL,
+                   `value` varchar(1536) NOT NULL,
                    PRIMARY KEY (`id`))
                  ENGINE=InnoDB DEFAULT CHARSET=utf8;");
       $db->query("CREATE TABLE IF NOT EXISTS {exif_records} (
diff --git a/modules/tag/helpers/tag_event.php b/modules/tag/helpers/tag_event.php
index a301c441..a4a6ec26 100644
--- a/modules/tag/helpers/tag_event.php
+++ b/modules/tag/helpers/tag_event.php
@@ -34,7 +34,10 @@ class tag_event_Core {
         if (!empty($iptc["2#025"])) {
           foreach($iptc["2#025"] as $tag) {
             $tag = str_replace("\0",  "", $tag);
-            $tags[$tag]= 1;
+            if (mb_detect_encoding($tag) != "UTF-8") {
+              $tag = utf8_encode($tag);
+            }
+            $tags[$tag] = 1;
           }
         }
       }