Remove some of the weirder methods of EncodingDetector and make the one method public that we actually want: analyze(). analyze() also does not alter its input...

svn path=/branches/kdepim/enterprise/kdepim/; revision=783567

encodingdetector.cpp

@@ -809,34 +809,6 @@ static int findXMLEncoding(const QCString &str, int &encodingLength)
     return pos;
 }
 
-bool EncodingDetector::processNull(char *data, int len)
-{
-    bool bin=false;
-    if(is16Bit(d->m_codec))
-    {
-        for (int i=1; i < len; i+=2)
-        {
-            if ((data[i]=='\0') && (data[i-1]=='\0'))
-            {
-                bin=true;
-                data[i]=' ';
-            }
-        }
-        return bin;
-    }
-    // replace '\0' by spaces, for buggy pages
-    int i = len-1;
-    while(--i>=0)
-    {
-        if(data[i]==0)
-        {
-            bin=true;
-            data[i]=' ';
-        }
-    }
-    return bin;
-}
-
 
 bool EncodingDetector::errorsIfUtf8 (const char* data, int length)
 {
@@ -902,7 +874,6 @@ static const unsigned char highest5Bits = 0xF8;
     return false;
 }
 
-
 EncodingDetector::EncodingDetector() : d(new EncodingDetectorPrivate)
 {
 }
@@ -1007,111 +978,9 @@ bool EncodingDetector::setEncoding(const char *_encoding, EncodingChoiceSource t
     return true;
 }
 
-QString EncodingDetector::decode(const char *data, int len)
-{
-    processNull(const_cast<char *>(data),len);
-    if (!d->m_analyzeCalled)
-    {
-        analyze(data,len);
-        d->m_analyzeCalled=true;
-    }
-
-    return d->m_decoder->toUnicode(data,len);
-}
-
-QString EncodingDetector::decode(const QByteArray &data)
-{
-    processNull(const_cast<char *>(data.data()),data.size());
-    if (!d->m_analyzeCalled)
-    {
-        analyze(data.data(),data.size());
-        d->m_analyzeCalled=true;
-    }
-
-    return d->m_decoder->toUnicode(data, data.size());
-}
-
-QString EncodingDetector::decodeWithBuffering(const char *data, int len)
-{
-#ifdef DECODE_DEBUG
-        kWarning() << "EncodingDetector: decoding "<<len<<" bytes";
-#endif
-    if (d->m_writtingHappened)
-    {
-#ifdef DECODE_DEBUG
-        kWarning() << "EncodingDetector: d->m_writtingHappened "<< d->m_codec->name();
-#endif
-        processNull(const_cast<char *>(data),len);
-        return d->m_decoder->toUnicode(data, len);
-    }
-    else
-    {
-        if (d->m_bufferForDefferedEncDetection.isEmpty())
-        {
-            if (analyze(data,len)
-                && (d->m_seenBody
-                    || !(d->m_source==AutoDetectedEncoding
-                        ||d->m_source==DefaultEncoding
-                        )
-                   )
-               )//dontWannaSeeHead()
-            {
-#ifdef DECODE_DEBUG
-                kWarning() << "EncodingDetector: m_writtingHappened first time "<< d->m_codec->name();
-#endif
-                processNull(const_cast<char *>(data),len);
-                d->m_writtingHappened=true;
-                return d->m_decoder->toUnicode(data, len);
-            }
-            else
-            {
-#ifdef DECODE_DEBUG
-                kWarning() << "EncodingDetector: begin deffer";
-#endif
-                d->m_bufferForDefferedEncDetection=data;
-            }
-        }
-        else
-        {
-            d->m_bufferForDefferedEncDetection+=data;
-            if ( (analyze(data,len)
-                  && (d->m_seenBody
-                     || !(d->m_source==AutoDetectedEncoding
-                        ||d->m_source==DefaultEncoding
-                         )
-                     )
-                 ) || d->m_bufferForDefferedEncDetection.length()>MAX_BUFFER
-               )//dontWannaSeeHead()
-            {
-                d->m_writtingHappened=true;
-                d->m_bufferForDefferedEncDetection.replace('\0',' ');
-                QString result(d->m_decoder->toUnicode(d->m_bufferForDefferedEncDetection,
-                                                       d->m_bufferForDefferedEncDetection.size()));
-                d->m_bufferForDefferedEncDetection.resize(0);
-#ifdef DECODE_DEBUG
-                kWarning() << "EncodingDetector: m_writtingHappened in the middle " << d->m_codec->name();
-#endif
-                return result;
-            }
-        }
-    }
-
-    return QString();
-}
-
-QString EncodingDetector::flush()
+bool EncodingDetector::analyze(const QByteArray &data)
 {
-    if (d->m_bufferForDefferedEncDetection.isEmpty())
-        return QString();
-
-    d->m_bufferForDefferedEncDetection.replace('\0',' ');
-    QString result(d->m_decoder->toUnicode(d->m_bufferForDefferedEncDetection,
-                                           d->m_bufferForDefferedEncDetection.size()));
-    d->m_bufferForDefferedEncDetection.resize(0);
-#ifdef DECODE_DEBUG
-    kWarning() << "EncodingDetector:flush() "<< d->m_bufferForDefferedEncDetection.length()<<" bytes "<< d->m_codec->name();
-#endif
-    return result;
+    return analyze( data.data(), data.size() );
 }
 
 bool EncodingDetector::analyze(const char *data, int len)

encodingdetector.h

@@ -125,32 +125,16 @@ public:
     EncodingChoiceSource encodingChoiceSource() const;
 
     /**
-    * The main class method
-    *
-    * Calls protected analyze() only the first time of the whole object life
-    *
-    * Replaces all null chars with spaces.
-    */
-    QString decode(const char *data, int len);
-    QString decode(const QByteArray &data);
-
-    //* You don't need to call analyze() if you use this method.
-    /**
-    * Convenience method that uses buffering. It waits for full html head to be buffered
-    * (i.e. calls analyze every time until it returns true).
-    *
-    * Replaces all null chars with spaces.
-    *
-    * @returns Decoded data, or empty string, if there was not enough data for accurate detection
-    * @see flush()
+    * Analyze text data.
+    * @returns true if there was enough data for accurate detection
     */
-    QString decodeWithBuffering(const char *data, int len);
+    bool analyze( const char *data, int len );
 
     /**
-    * Convenience method to be used with decodeForHtml. Flushes buffer.
-    * @see decodeForHtml()
+    * Analyze text data.
+    * @returns true if there was enough data for accurate detection
     */
-    QString flush();
+    bool analyze( const QByteArray &data );
 
     /**
      * Takes lang name _after_ it were i18n()'ed
@@ -161,12 +145,6 @@ public:
     static bool hasAutoDetectionForScript(AutoDetectScript);
 
 protected:
-    /**
-     * This nice method will kill all 0 bytes (or double bytes)
-     * and remember if this was a binary or not ;)
-     */
-    bool processNull(char* data,int length);
-
     /**
      * Check if we are really utf8. Taken from kate
      *
@@ -176,12 +154,6 @@ protected:
      */
     bool errorsIfUtf8 (const char* data, int length);
 
-    /**
-    * Analyze text data.
-    * @returns true if there was enough data for accurate detection
-    */
-    bool analyze (const char *data, int len);
-
     /**
     * @returns QTextDecoder for detected encoding
     */

kmcomposewin.cpp

@@ -2758,12 +2758,9 @@ void KMComposeWin::slotAttachFileResult(KIO::Job *job)
     EncodingDetector ed;
     KLocale *loc = KGlobal::locale();
     ed.setAutoDetectLanguage( EncodingDetector::scriptForLanguageCode ( loc->language() ) );
-    // ### FIXME decode(_const_ QByteArray &) replaces 0x00 bytes by 0x20 in the parameter
-    // despite being const, which obviously breaks binary attachments.
-    QByteArray copy = (*it).data.copy();
-    ed.decode( copy );
+    ed.analyze( (*it).data );
     partCharset = ed.encoding();
-    if (partCharset.isEmpty()) //shouldn't happen
+    if ( partCharset.isEmpty() ) //shouldn't happen
       partCharset = mCharset;
   }