[Digikam-devel] [Bug 114211] image comments encoding unreadable after moving an album

[Marcel Wiesweg]

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
         
http://bugs.kde.org/show_bug.cgi?id=114211         




------- Additional Comments From marcel.wiesweg gmx de  2006-05-21 17:32 -------
SVN commit 543272 by mwiesweg:

Add some autodetection magic for charset support

- DMetadata::detectEncodingAndDecode will check if a given string
  is in UTF8. If not, it will leave it to QTextCodec to decide
  if the local charset or latin1 will be used
- use detectEncodingAndDecode when reading the JFIF comment
  and for Exif comments with undefined encoding
- When writing the Exif comment, use UCS-2 only when
  necessary. Check with QTextCodec::canEncode if plain
  latin1 is enough.

I have tested this successfully with some Arabian and cyrillic characters.
But please test this with some more pictures. UTF-8 should be no problem,
but the local8Bit vs. latin1 decision may be.

CCBUGS: 120241, 114211



 M  +75 -15    dmetadata.cpp  
 M  +3 -0      dmetadata.h  


--- trunk/extragear/graphics/digikam/libs/dmetadata/dmetadata.cpp #543271:543272
 @ -33,7 +33,9  @
 
 // KDE includes.
 
+#include <kapplication.h>
 #include <kdebug.h>
+#include <kstringhandler.h>
 #include <ktempfile.h>
 
 // Exiv2 includes.
 @ -714,7 +716,7  @
 
         // In first we trying to get image comments, outside of Exif and IPTC.
 
-        QString comments = QString::fromUtf8(d->imageComments.c_str());
+        QString comments = detectEncodingAndDecode(d->imageComments);
 
         if (!comments.isEmpty())
             return comments;
 @ -780,18 +782,32  @
 
         // In Second we write comments into Exif.
 
-        // Be aware that we are dealing with a UCS-2 string.
-        // Null termination means \0\0, strlen does not work,
-        // do not use any const-char*-only methods,
-        // pass a std::string and not a const char * to ExifDatum::operator=().
-        const unsigned short *ucs2 = comment.ucs2();
-        std::string exifComment("charset=\"Unicode\" ");
-        exifComment.append((const char*)ucs2, sizeof(unsigned short) * comment.length());
-        d->exifMetadata["Exif.Photo.UserComment"] = exifComment;
-        //d->exifMetadata["Exif.Photo.UserComment"] = comment.latin1();
+        // Write as Unicode only when necessary.
+        QTextCodec *latin1Codec = QTextCodec::codecForName("iso8859-1");
+        if (latin1Codec->canEncode(comment))
+        {
+            // write as ASCII
+            std::string exifComment("charset=\"Ascii\" ");
+            exifComment += comment.latin1();
+            d->exifMetadata["Exif.Photo.UserComment"] = exifComment;
+        }
+        else
+        {
+            // write as Unicode (UCS-2)
 
-        // In Third we write comments into Iptc. Note that Caption IPTC tag is limited to 2000 char.
+            // Be aware that we are dealing with a UCS-2 string.
+            // Null termination means \0\0, strlen does not work,
+            // do not use any const-char*-only methods,
+            // pass a std::string and not a const char * to ExifDatum::operator=().
+            const unsigned short *ucs2 = comment.ucs2();
+            std::string exifComment("charset=\"Unicode\" ");
+            exifComment.append((const char*)ucs2, sizeof(unsigned short) * comment.length());
+            d->exifMetadata["Exif.Photo.UserComment"] = exifComment;
+        }
 
+        // In Third we write comments into Iptc.
+        // Note that Caption IPTC tag is limited to 2000 char and ASCII charset.
+
         QString commentIptc = comment;
         commentIptc.truncate(2000);
         d->iptcMetadata["Iptc.Application2.Caption"] = commentIptc.latin1();
 @ -815,7 +831,7  @
     {
         std::string comment = exifDatum.toString();
         std::string charset;
-    
+
         // libexiv2 will prepend "charset=\"SomeCharset\" " if charset is specified
         // Before conversion to QString, we must know the charset, so we stay with std::string for a while
         if (comment.length() > 8 && comment.substr(0, 8) == "charset=")
 @ -830,7 +846,7  @
                 comment = comment.substr(pos+1);
             }
         }
-    
+
         if (charset == "\"Unicode\"")
         {
             // QString expects a null-terminated UCS-2 string.
 @ -849,8 +865,7  @
         }
         else
         {
-            // or from local8bit ??
-            return QString::fromLatin1(comment.c_str());
+            return detectEncodingAndDecode(comment);
         }
     }
     catch( Exiv2::Error &e )
 @ -863,6 +878,51  @
     return QString();
 }
 
+QString DMetadata::detectEncodingAndDecode(const std::string &value)
+{
+    // For charset autodetection, we could use sophisticated code
+    // (Mozilla chardet, KHTML's autodetection, QTextCodec::codecForContent),
+    // but that is probably too much.
+    // We check for UTF8, Local encoding and ASCII.
+
+    if (value.empty())
+        return QString();
+
+#if KDE_IS_VERSION(3,2,0)
+    if (KStringHandler::isUtf8(value.c_str()))
+    {
+        return QString::fromUtf8(value.c_str());
+    }
+#else
+    // anyone who is still running KDE 3.0 or 3.1 is missing so many features
+    // that he will have to accept this missing feature.
+    return QString::fromUtf8(value.c_str());
+#endif
+
+    // Utf8 has a pretty unique byte pattern.
+    // Thats not true for ASCII, it is not possible
+    // to reliably autodetect different ISO-8859 charsets.
+    // We try if QTextCodec can decide here, otherwise we use Latin1.
+    // Or use local8Bit as default?
+
+    // load QTextCodecs
+    QTextCodec *latin1Codec = QTextCodec::codecForName("iso8859-1");
+    //QTextCodec *utf8Codec   = QTextCodec::codecForName("utf8");
+    QTextCodec *localCodec  = QTextCodec::codecForLocale();
+
+    // make heuristic match
+    int latin1Score = latin1Codec->heuristicContentMatch(value.c_str(), value.length());
+    int localScore  = localCodec->heuristicContentMatch(value.c_str(), value.length());
+
+    // convert string:
+    // Use whatever has the larger score, local or ASCII
+    if (localScore >= 0 && localScore >= latin1Score)
+        return localCodec->toUnicode(value.c_str(), value.length());
+    else
+        return QString::fromLatin1(value.c_str());
+}
+
+
 /*
 Iptc.Application2.Urgency <==> digiKam Rating links:
 
--- trunk/extragear/graphics/digikam/libs/dmetadata/dmetadata.h #543271:543272
 @ -21,6 +21,8  @
 #ifndef DMETADATA_H
 #define DMETADATA_H
 
+#include <string>
+
 // QT includes.
 
 #include <qcstring.h>
 @ -108,6 +110,7  @
     PhotoInfoContainer getPhotographInformations() const;
 
     static QString convertCommentValue(const Exiv2::Exifdatum &comment);
+    static QString detectEncodingAndDecode(const std::string &value);
 
 private:
_______________________________________________
Digikam-devel mailing list
[hidden email]
https://mail.kde.org/mailman/listinfo/digikam-devel