Changeset 17536 in webkit

Timestamp:

Nov 2, 2006 8:13:35 AM (17 years ago)

Author:

spadma

Message:

2006-11-01 w3liu <​wei.liu@nokia.com>

Reviewed by <​yongjun.zhang@nokia.com>
DESC: Thai text rendering fails when using HTTPS.
​http://bugs.webkit.org/show_bug.cgi?id=11486
WARNING: NO TEST CASES ADDED OR CHANGED

• ChangeLog:
• kwq/KWQTextCodec.cpp: (KWQTextDecoder::KWQTextDecoder): (KWQTextDecoder::~KWQTextDecoder): (QTextCodec::~QTextCodec): (QTextCodec::makeDecoder): (QTextCodec::toUnicode): (KWQTextDecoder::toUnicode): (QTextCodec::PrepareForExit):
• kwq/KWQTextCodec.h:

Merge from r17535 from main branch

Location:

S60/branches/3.1m/WebCore

Files:

• ChangeLog (modified) (1 diff)
• kwq/KWQTextCodec.cpp (modified) (14 diffs)
• kwq/KWQTextCodec.h (modified) (3 diffs)

S60/branches/3.1m/WebCore/ChangeLog

@@ +1 @@
+2006-11-01  w3liu  <wei.liu@nokia.com>
+
+        Reviewed by <yongjun.zhang@nokia.com>
+        DESC: Thai text rendering fails when using HTTPS.
+        http://bugs.webkit.org/show_bug.cgi?id=11486
+        WARNING: NO TEST CASES ADDED OR CHANGED
+
+        * ChangeLog:
+        * kwq/KWQTextCodec.cpp:
+        (KWQTextDecoder::KWQTextDecoder):
+        (KWQTextDecoder::~KWQTextDecoder):
+        (QTextCodec::~QTextCodec):
+        (QTextCodec::makeDecoder):
+        (QTextCodec::toUnicode):
+        (KWQTextDecoder::toUnicode):
+        (QTextCodec::PrepareForExit):
+        * kwq/KWQTextCodec.h:
+        
+        Merge from r17535 from main branch
+
+2006-11-01  w3liu  <wei.liu@nokia.com>
+
+        Reviewed by <yongjun.zhang@nokia.com>
+        DESC: 
+        http://bugs.webkit.org/show_bug.cgi?id=11486
+        Thai text rendering fails when using HTTPS 
+
+        WARNING: NO TEST CASES ADDED OR CHANGED
+
+        * kwq/KWQTextCodec.cpp:
+        (KWQTextDecoder::KWQTextDecoder):
+        (QTextCodec::~QTextCodec):
+        (QTextCodec::makeDecoder):
+        (QTextCodec::toUnicode):
+        (KWQTextDecoder::toUnicode):
+        * kwq/KWQTextCodec.h:
+
 2006-10-25  raalexan  <rasvan.alexandrescu@nokia.com>
 

S60/branches/3.1m/WebCore/kwq/KWQTextCodec.cpp

@@ -33 +33 @@
 #include "KWQKDebug.h"
 
+#define _DEBUG1 0
+#ifdef _DEBUG1
+#include <flogger.h>
+#endif
 #define BrDefaultEnc KCharacterSetIdentifierIso88591;
 #define BrDefaultEncStr "iso-8859-1";
@@ -46 +50 @@
 friend class QTextCodec;
 public:
-    KWQTextDecoder(TUint aEnc) : iEncoding(aEnc)  {}
-    ~KWQTextDecoder()                 {}
+    KWQTextDecoder(TUint aEnc) : iEncoding(aEnc), iRemainderBuf(NULL)   {}
+    ~KWQTextDecoder()                 { delete iRemainderBuf; }
 
     QString toUnicode(const char *chs, int len, bool flush);
     QCString fromUnicode(const QString &);
     bool IsConversionSupportedL();
-  QString windows125xToUnicode( const char *chs, int len );
+    QString windows125xToUnicode( const char *chs, int len );
 
 private:
@@ -60 +64 @@
     static CArrayFix<CCnvCharacterSetConverter::SCharacterSet>* iArrayOfCharacterSetsAvailable;
     TUint iEncoding;
+    HBufC8 * iRemainderBuf;
 };
 
@@ -88 +93 @@
     }
 }
+
+QTextCodec::~QTextCodec()
+{ delete iQTextDecoder; }
 
 bool validateUtf8(const char* chs, int len, int& validMultiByteChars)
@@ -221 +229 @@
 QTextDecoder *QTextCodec::makeDecoder() const
 {
-    return new KWQTextDecoder( iEncoding );
+    return iQTextDecoder? iQTextDecoder : new KWQTextDecoder( iEncoding );
 }
 
@@ -231 +239 @@
 QString QTextCodec::toUnicode(const char *chs, int len) const
 {
-    return KWQTextDecoder( iEncoding ).toUnicode(chs, len, true);
+    return makeDecoder()->toUnicode(chs, len, true);
 }
 
@@ -308 +316 @@
     }
 
-    /**
+ /**
  * Converts the text buffer from one encoding to Unicode
  * If the indicated charset is not supported, or its converter is not available,
@@ -315 +323 @@
 QString KWQTextDecoder::toUnicode(const char *chs, int len, bool flush)
     {
-    // make sure the converter is available
-
-  int bomLength = 0;
-  TUint encoding = iEncoding;
-  CCnvCharacterSetConverter::TEndianness endianness = CCnvCharacterSetConverter::ELittleEndian;
-
-  if (len > 2 && chs[0] == FF && chs[1] == FE) {
-        encoding = KCharacterSetIdentifierUnicodeLittle;
+
+    bool lastChunk = (len == 0)? true : false;
+    int bomLength = 0;
+    HBufC8 * inHBuf = NULL;     // buffer to hold combination of remainder and next chunk
+    CCnvCharacterSetConverter::TEndianness endianness = CCnvCharacterSetConverter::ELittleEndian;
+
+    if (len > 2 && chs[0] == FF && chs[1] == FE)
+        {
+        iEncoding = KCharacterSetIdentifierUnicodeLittle;
         bomLength = 2;
-    } else if (len > 2 && chs[0] == FE && chs[1] == FF) {
-        encoding = KCharacterSetIdentifierUnicodeBig;
-    endianness = CCnvCharacterSetConverter::EBigEndian;
+        }
+    else if (len > 2 && chs[0] == FE && chs[1] == FF)
+        {
+        iEncoding = KCharacterSetIdentifierUnicodeBig;
+        endianness = CCnvCharacterSetConverter::EBigEndian;
         bomLength = 2;
-    } else if (len > 3 && chs[0] == EF && chs[1] == BB && chs[2] == BF) {
-        encoding = KCharacterSetIdentifierUtf8;
+        }
+    else if (len > 3 && chs[0] == EF && chs[1] == BB && chs[2] == BF)
+        {
+        iEncoding = KCharacterSetIdentifierUtf8;
         bomLength = 3;
-    }
+        }
 
     if (!IsConversionSupportedL())
-    {
-    iEncoding = KCharacterSetIdentifierIso88591;
-    }
+        {
+        iEncoding = KCharacterSetIdentifierIso88591;
+        }
 
     // check if it really looks like valid utf-8
-    if (iEncoding == KCharacterSetIdentifierUtf8 && bomLength==0) {
+    if (iEncoding == KCharacterSetIdentifierUtf8 && bomLength==0)
+        {
         int validMultibyteChars;
         bool validUtf8 = validateUtf8(chs, Min(len,32*1024),validMultibyteChars);
@@ -345 +359 @@
         if (!validUtf8 && validMultibyteChars<3)
             iEncoding = KCharacterSetIdentifierIso88591;
-    }
+        }
 
     if (iEncoding == KCharacterSetIdentifierIso88591)
+        {
         return convertLatin1(reinterpret_cast<const unsigned char *>(chs), len);
-  else if(iEncoding == KCharacterSetIdentifierWindow1250 ||
-      iEncoding == KCharacterSetIdentifierWindow1251 ||
-      iEncoding == KCharacterSetIdentifierWindow1253 ||
-      iEncoding == KCharacterSetIdentifierWindow1256 ||
-      iEncoding == KCharacterSetIdentifierWindow1257
-      ) // work around for Symbian converter hanging when converting windows-125x charset
-    {
-    return windows125xToUnicode( chs, len );
-    }
+        }
+    else if(iEncoding == KCharacterSetIdentifierWindow1250 ||
+        iEncoding == KCharacterSetIdentifierWindow1251 ||
+        iEncoding == KCharacterSetIdentifierWindow1253 ||
+        iEncoding == KCharacterSetIdentifierWindow1256 ||
+        iEncoding == KCharacterSetIdentifierWindow1257
+        ) // work around for Symbian converter hanging when converting windows-125x charset
+        {
+        return windows125xToUnicode( chs, len );
+        }
+
+#if _DEBUG1
+    // print out the incoming buffer to a file for debugging purpose
+    if (len)
+        RFileLogger::HexDump( _L("Browser"), _L("text_chunks.txt"), EFileLoggingModeAppend,
+        NULL, NULL, (const TUint8*)chs, len );
+#endif // _DEBUG1
 
     CCnvCharacterSetConverter* charConv = KWQGlobalServices::InstanceL()->CharConv();
@@ -363 +386 @@
 
     // check if the target charset is supported
-    charConv->PrepareToConvertToOrFromL( encoding, *iArrayOfCharacterSetsAvailable, fileSession );
+    charConv->PrepareToConvertToOrFromL( iEncoding, *iArrayOfCharacterSetsAvailable, fileSession );
     charConv->SetDefaultEndiannessOfForeignCharacters( endianness );
     TBuf16<32> tmpBuf;
     TBuf16<32> outBuf;
 
-  len = len - bomLength ;
+    len = len - bomLength ;
     TPtrC8  inBuf( (const TUint8*)chs + bomLength, len );
-    TInt state=CCnvCharacterSetConverter::KStateDefault;
+    TInt state = CCnvCharacterSetConverter::KStateDefault;
 
     // guess the output size
@@ -381 +404 @@
     // reserve function reserves 3*len of bytes in momery, to achieve 2*len, times factor of 2/3
     result.reserve( len*2/3 + 1 );
+
+    if (iRemainderBuf && iRemainderBuf->Length())
+        {
+        inHBuf = HBufC8::NewL(iRemainderBuf->Length() + len);
+        inHBuf->Des().Append(iRemainderBuf->Des());
+        inHBuf->Des().Append((const unsigned char *)chs, len);
+        delete iRemainderBuf;
+        iRemainderBuf = NULL;
+        inBuf.Set(inHBuf->Des());
+        }
+
     for(;;)
         {
@@ -387 +421 @@
         TInt ret = charConv->ConvertToUnicode( tmpBuf, inBuf, state, rep, rIndx );
         if (ret==CCnvCharacterSetConverter::EErrorIllFormedInput )
+            {
+            // in case the error is caused by incomplete character at end of chunk
+            if (inBuf.Length() < 3 && !lastChunk)
+                {
+                iRemainderBuf = HBufC8::NewL(inBuf.Length());
+                iRemainderBuf->Des().Append(inBuf);
+                }
             break;
+            }
 
         // copy result to out buffer
-
-    if(ret == -1 )
-      {
-      // if there are no sifficient character to be converted to unicode,
-      // just append the rest to the characters to the result string.
-      result.append( (char *)inBuf.Ptr(), inBuf.Length() );
-      }
-    else
-      {
-      // ### FIXME NOKIA: is it correct to remove the replacement 0xFFFD chars here?
-      // Maybe we need to replace it with space character
-
-      TPtr outPtr( 0, 0 );
-
-      // there is some illegal/unconvertible chars
-      // try to create a buffer with all 0xFFFD chars trimmed
-      if (rep != 0 && rIndx < tmpBuf.Length() )
-        {
-        TInt i=rIndx;
-        // first copy the part of buffer before the first 0xFFFD char
-        outBuf.Append(tmpBuf.Left(i++));
-        rep--;
-        // search for more occurance of 0xFFFD chars
-        // and append valid chars before the last 0xFFFD char
-        // if there is only one 0xFFFD char, wouldn't enter the loop
-        for( ; rep && i<tmpBuf.Length(); i++ )
-          {
-          if( tmpBuf[i] != 0xFFFD )
-            outBuf.Append( tmpBuf[i] );
-          else
-            rep--;
-          }
-        // append the chars after last 0xFFFD char
-        outBuf.Append(tmpBuf.Right(tmpBuf.Length() - i));
-        outPtr.Set( (TUint16*)outBuf.Ptr(), outBuf.Length(), outBuf.Length() );
-        outBuf.Zero();
-
-        }
-      else outPtr.Set( (TUint16*)tmpBuf.Ptr(), tmpBuf.Length(), tmpBuf.Length()  );
+        TPtr outPtr( 0, 0 );
+
+        // there is some illegal/unconvertible chars
+        // try to create a buffer with all 0xFFFD chars trimmed
+        if (rep != 0)
+            {
+            if (inBuf.Length() <= 64 && !lastChunk)
+                {
+                iRemainderBuf = HBufC8::NewL(inBuf.Length());
+                iRemainderBuf->Des().Append(inBuf);
+                break;
+                }
+
+            TInt i= (rIndx -1 )/2;
+            // first copy the part of buffer before the first 0xFFFD char
+            outBuf.Append(tmpBuf.Left(i));
+            // search for more occurance of 0xFFFD chars
+            // and append valid chars before the last 0xFFFD char
+            // if there is only one 0xFFFD char, wouldn't enter the loop
+            for( ; rep && i<tmpBuf.Length(); i++ )
+                {
+                if( tmpBuf[i] != 0xFFFD )
+                    outBuf.Append( tmpBuf[i] );
+                else
+                    rep--;
+                }
+            // append the chars after last 0xFFFD char
+            outBuf.Append(tmpBuf.Right(tmpBuf.Length() - i));
+            outPtr.Set( (TUint16*)outBuf.Ptr(), outBuf.Length(), outBuf.Length() );
+            outBuf.Zero();
+            }
+        else outPtr.Set( (TUint16*)tmpBuf.Ptr(), tmpBuf.Length(), tmpBuf.Length() );
 
         // append the buffers to final result, will reincrease the size in needed
         result.append( (const QChar*)(outPtr.Ptr()), outPtr.Length() );
-      }
 
         if( ret == 0  || ret == -1 ) break;
@@ -440 +474 @@
 
     OOM_POST_CHECK_FAILED( return result; )
+
+    delete inHBuf;
     return result;
     }
@@ -530 +566 @@
 void QTextCodec::PrepareForExit()
 {
-    CArrayFix<CCnvCharacterSetConverter::SCharacterSet>* arr 
+    CArrayFix<CCnvCharacterSetConverter::SCharacterSet>* arr
                 = KWQTextDecoder::iArrayOfCharacterSetsAvailable;
 

S60/branches/3.1m/WebCore/kwq/KWQTextCodec.h

@@ -43 +43 @@
 
     QTextCodec( TUint aEnc, KWQEncodingFlags f = NoEncodingFlags ) : iEncoding( aEnc ), _flags(f)          {}
+    ~QTextCodec();
     const char *name() const;
     bool usesVisualOrdering() const { return _flags & VisualOrdering; }
@@ -62 +63 @@
 
 private:
-        TUint           iEncoding;
+    TUint       iEncoding;
     KWQEncodingFlags _flags;
+    QTextDecoder * iQTextDecoder;
 };
 
@@ -72 +74 @@
 {
 public:
-    virtual ~QTextDecoder()                                                                                     {}
+    virtual ~QTextDecoder()                                         {}
     virtual QString toUnicode(const char *, int, bool flush = false) = 0;
 };