|
1 /* |
|
2 * Copyright (c) 2006 Nokia Corporation and/or its subsidiary(-ies). |
|
3 * All rights reserved. |
|
4 * This component and the accompanying materials are made available |
|
5 * under the terms of the License "Eclipse Public License v1.0" |
|
6 * which accompanies this distribution, and is available |
|
7 * at the URL "http://www.eclipse.org/legal/epl-v10.html". |
|
8 * |
|
9 * Initial Contributors: |
|
10 * Nokia Corporation - initial contribution. |
|
11 * |
|
12 * Contributors: |
|
13 * |
|
14 * Description: |
|
15 * |
|
16 */ |
|
17 |
|
18 |
|
19 #include "config.h" |
|
20 #include "StreamingTextCodecSymbian.h" |
|
21 #include "CharsetData.h" |
|
22 #include "CString.h" |
|
23 #include "PlatformString.h" |
|
24 #include "TextEncoding.h" |
|
25 #include "TextEncodingRegistry.h" |
|
26 #include <EIKENV.H> |
|
27 #include <wtf/Assertions.h> |
|
28 |
|
29 using std::auto_ptr; |
|
30 using std::min; |
|
31 |
|
32 static const size_t ConversionBufferSize = 16384; |
|
33 |
|
34 namespace WebCore { |
|
35 |
|
36 // static data |
|
37 CArrayFix<CCnvCharacterSetConverter::SCharacterSet>* TextCodecSymbian::s_availableCharsets = 0; |
|
38 |
|
39 static auto_ptr<TextCodec> newTextCodecSymbian(const TextEncoding&, const void* additionalData) |
|
40 { |
|
41 return auto_ptr<TextCodec>(new TextCodecSymbian(*static_cast<const TUint*>(additionalData))); |
|
42 } |
|
43 |
|
44 void TextCodecSymbian::registerEncodingNames(EncodingNameRegistrar registrar) |
|
45 { |
|
46 TInt lastEncoding = -1; |
|
47 const char* lastName = 0; |
|
48 |
|
49 for (size_t i = 0; CharsetTable[i].name; ++i) { |
|
50 if (CharsetTable[i].encoding != lastEncoding) { |
|
51 lastEncoding = CharsetTable[i].encoding; |
|
52 lastName = CharsetTable[i].name; |
|
53 } |
|
54 registrar(CharsetTable[i].name, lastName); |
|
55 } |
|
56 } |
|
57 |
|
58 void TextCodecSymbian::registerCodecs(TextCodecRegistrar registrar) |
|
59 { |
|
60 TInt lastEncoding = -1; |
|
61 |
|
62 for (size_t i = 0; CharsetTable[i].name; ++i) |
|
63 if (CharsetTable[i].encoding != lastEncoding) { |
|
64 registrar(CharsetTable[i].name, newTextCodecSymbian, &CharsetTable[i].encoding); |
|
65 lastEncoding = CharsetTable[i].encoding; |
|
66 } |
|
67 } |
|
68 |
|
69 TextCodecSymbian::TextCodecSymbian(TUint encoding) |
|
70 : m_converter(0) |
|
71 , m_state(CCnvCharacterSetConverter::KStateDefault) |
|
72 , m_encoding(encoding) |
|
73 , m_firstChunkDecoded(EFalse) |
|
74 , m_numBufferedBytes(0) |
|
75 { |
|
76 } |
|
77 |
|
78 TextCodecSymbian::~TextCodecSymbian() |
|
79 { |
|
80 delete m_converter; |
|
81 m_fallbackCodec.clear(); |
|
82 } |
|
83 |
|
84 bool TextCodecSymbian::initializeConverter() |
|
85 { |
|
86 if( m_firstChunkDecoded ) return true; |
|
87 |
|
88 // Removed the call to CEikonEnv::Static()->FsSession() which should not be |
|
89 // called in a thread where UiKon static env does not exist. This gets called |
|
90 // by the icon database thread which does not have it. Also, the file |
|
91 // server session is not needed anymore. It exists only for backward compatibility |
|
92 RFs fsSession; |
|
93 |
|
94 // FIXME: find a way to free the global availableCharsets array |
|
95 if( !m_converter ) { |
|
96 // first time call to charconv |
|
97 if( ( m_converter = |
|
98 CCnvCharacterSetConverter::NewL() ) == NULL ) |
|
99 return false; |
|
100 |
|
101 // caching the available charsets table |
|
102 if( !s_availableCharsets ) |
|
103 s_availableCharsets = CCnvCharacterSetConverter::CreateArrayOfCharacterSetsAvailableL( fsSession ); |
|
104 } |
|
105 |
|
106 // check decoder avialability |
|
107 checkSupportedEncoding(); |
|
108 |
|
109 // prepare character conversion |
|
110 m_converter->PrepareToConvertToOrFromL( m_encoding, *s_availableCharsets, fsSession ); |
|
111 |
|
112 return true; |
|
113 } |
|
114 |
|
115 String TextCodecSymbian::decode(const char* chs, size_t length, bool flush) |
|
116 { |
|
117 if( !initializeConverter() ) |
|
118 return String(); |
|
119 |
|
120 // FIXME: how to handle flush flag in Symbian converter? |
|
121 if( m_encoding == KCharacterSetIdentifierIso88591 ) { |
|
122 if( !m_fallbackCodec ) { |
|
123 m_fallbackCodec.set(newTextCodec(Latin1Encoding()).release()); |
|
124 } |
|
125 |
|
126 return m_fallbackCodec->decode( chs, length, flush ); |
|
127 } |
|
128 |
|
129 Vector<UChar> result; |
|
130 |
|
131 // prepare the input buffer for character converter, |
|
132 // the leftover bytes need to be decoded first. |
|
133 TPtrC8 inBuf(0,0); |
|
134 char* inputString = 0; |
|
135 if( m_numBufferedBytes != 0 ) { |
|
136 const int fullLength = m_numBufferedBytes + length; |
|
137 inputString = (char*)malloc( fullLength ); |
|
138 if( !inputString ) return String(); |
|
139 memcpy( inputString, m_bufferedBytes, m_numBufferedBytes ); |
|
140 memcpy( inputString + m_numBufferedBytes, chs, length ); |
|
141 inBuf.Set( (const TUint8*)inputString, fullLength ); |
|
142 m_numBufferedBytes = 0; |
|
143 } |
|
144 else { |
|
145 inBuf.Set( (const TUint8*)chs, length ); |
|
146 } |
|
147 |
|
148 // output buffer |
|
149 UChar* buffer = (UChar*)malloc( ConversionBufferSize*sizeof(UChar) ); |
|
150 if( !buffer ) return String(); |
|
151 TPtr16 outBuf( (TUint16*)buffer, 0, ConversionBufferSize ); |
|
152 |
|
153 for(;;) { |
|
154 int rep = 0; |
|
155 int rIndx = 0; |
|
156 |
|
157 int retValue = m_converter->ConvertToUnicode( outBuf, inBuf, m_state, rep, rIndx ); |
|
158 |
|
159 if( retValue == CCnvCharacterSetConverter::EErrorIllFormedInput ) { |
|
160 if (inBuf.Length() < strlen((char*) m_bufferedBytes)) |
|
161 { |
|
162 memcpy( m_bufferedBytes, inBuf.Ptr(), inBuf.Length() ); |
|
163 m_numBufferedBytes = inBuf.Length(); |
|
164 } |
|
165 break; |
|
166 } |
|
167 else if( retValue == 0 ) { |
|
168 // decoded fully, exit |
|
169 appendOmittingBOM( result, buffer, outBuf.Length() ); |
|
170 break; |
|
171 } |
|
172 else { |
|
173 // prepare for next round |
|
174 appendOmittingBOM( result, buffer, outBuf.Length() ); |
|
175 outBuf.SetLength(0); |
|
176 inBuf.Set( inBuf.Right(retValue) ); |
|
177 } |
|
178 } |
|
179 |
|
180 free( buffer ); |
|
181 free( inputString ); |
|
182 |
|
183 if(!m_firstChunkDecoded ) |
|
184 m_firstChunkDecoded = true; |
|
185 |
|
186 String resultString = String::adopt(result); |
|
187 return resultString; |
|
188 } |
|
189 |
|
190 CString TextCodecSymbian::encode(const UChar* data, size_t length, bool allowEntities) |
|
191 { |
|
192 if( !initializeConverter() ) |
|
193 return CString(); |
|
194 |
|
195 // FIXME: how to handle allowEntities flag in Symbian converter? |
|
196 // We should figure out a good way to use SetReplacementForUnconvertibleUnicodeCharactersL() |
|
197 if( m_encoding == KCharacterSetIdentifierIso88591 ) { |
|
198 if( !m_fallbackCodec ) { |
|
199 m_fallbackCodec.set(newTextCodec(Latin1Encoding()).release()); |
|
200 } |
|
201 |
|
202 return m_fallbackCodec->encode( data, length, allowEntities ); |
|
203 } |
|
204 |
|
205 Vector<char> result; |
|
206 |
|
207 // input buffer |
|
208 TPtrC inBuf( (const TUint16*)data, length ); |
|
209 |
|
210 // output buffer |
|
211 char* buffer = (char*)malloc( ConversionBufferSize*sizeof(char) ); |
|
212 if( !buffer ) return CString(); |
|
213 size_t size = 0; |
|
214 TPtr8 outBuf( (TUint8*)buffer, 0, ConversionBufferSize ); |
|
215 |
|
216 for(;;) { |
|
217 TInt retValue = m_converter->ConvertFromUnicode( outBuf, inBuf ); |
|
218 const int olen = outBuf.Length(); |
|
219 |
|
220 if( retValue == CCnvCharacterSetConverter::EErrorIllFormedInput ) |
|
221 break; |
|
222 |
|
223 result.resize( size + olen ); |
|
224 memcpy( result.data() + size, buffer, olen ); |
|
225 size += olen; |
|
226 |
|
227 if( retValue == 0 ) |
|
228 break; |
|
229 else { |
|
230 outBuf.SetLength(0); |
|
231 inBuf.Set( inBuf.Right(retValue) ); |
|
232 } |
|
233 } |
|
234 |
|
235 free( buffer ); |
|
236 |
|
237 if( !m_firstChunkDecoded ) |
|
238 m_firstChunkDecoded = true; |
|
239 |
|
240 return CString( result.data(), size ); |
|
241 } |
|
242 |
|
243 void TextCodecSymbian::checkSupportedEncoding() |
|
244 { |
|
245 int count = s_availableCharsets->Count(); |
|
246 |
|
247 for (int index = 0; index < count; index++) { |
|
248 if (m_encoding == ((CCnvCharacterSetConverter::SCharacterSet)(s_availableCharsets->At(index))).Identifier()) |
|
249 return; |
|
250 } |
|
251 |
|
252 // otherwise fall back to iso88591 |
|
253 m_encoding = KCharacterSetIdentifierIso88591; |
|
254 } |
|
255 |
|
256 |
|
257 } |
|
258 |
|
259 // END OF FILE |