1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/os/persistentdata/persistentstorage/store/pcstore/src/unicodecompression.h Fri Jun 15 03:10:57 2012 +0200
1.3 @@ -0,0 +1,334 @@
1.4 +// Copyright (c) 2006-2009 Nokia Corporation and/or its subsidiary(-ies).
1.5 +// All rights reserved.
1.6 +// This component and the accompanying materials are made available
1.7 +// under the terms of "Eclipse Public License v1.0"
1.8 +// which accompanies this distribution, and is available
1.9 +// at the URL "http://www.eclipse.org/legal/epl-v10.html".
1.10 +//
1.11 +// Initial Contributors:
1.12 +// Nokia Corporation - initial contribution.
1.13 +//
1.14 +// Contributors:
1.15 +//
1.16 +// Description:
1.17 +// Classes defined in this file are used for Unicode compression and decompression.
1.18 +// Their code is borrowed from Symbian, only with some changes such as the "Panic" function
1.19 +// is changed to exit the program. The Symbian coding standard will be kept in the code.
1.20 +//
1.21 +//
1.22 +
1.23 +
1.24 +#if !defined(__UNICODECOMPRESSION_H__)
1.25 +#define __UNICODECOMPRESSION_H__
1.26 +
1.27 +#include <stdio.h>
1.28 +#include <pcstore/pcstoredef.h>
1.29 +#include <pcstore/storereadstream.h>
1.30 +#include <pcstore/storewritestream.h>
1.31 +#include "pcstoreconst.h"
1.32 +
1.33 +namespace PCStore
1.34 +{
1.35 +/**
1.36 +@internalComponent
1.37 +*/
1.38 +class TUnicodeCompressionState
1.39 + {
1.40 + public:
1.41 + TUnicodeCompressionState();
1.42 + void Reset();
1.43 + static TInt StaticWindowIndex(TUint16 aCode);
1.44 + static TInt DynamicWindowOffsetIndex(TUint16 aCode);
1.45 + static TUint32 DynamicWindowBase(TInt aOffsetIndex);
1.46 + static TBool EncodeAsIs(TUint16 aCode);
1.47 +
1.48 + enum TPanic
1.49 + {
1.50 + EUnhandledByte, // expander code fails to handle all possible byte codes
1.51 + ENotUnicode, // expander can't handle Unicode values outside range 0x0..0x10FFFF;
1.52 + // that is, 16-bit codes plus 32-bit codes that can be expressed using
1.53 + // 16-bit surrogates
1.54 + EOutputBufferOverflow // output buffer is not big enough
1.55 + };
1.56 +
1.57 + static void Panic(TPanic aPanic);
1.58 +
1.59 + protected:
1.60 +
1.61 + enum
1.62 + {
1.63 + EStaticWindows = 8,
1.64 + EDynamicWindows = 8,
1.65 + ESpecialBases = 7
1.66 + };
1.67 +
1.68 + TBool iUnicodeMode; // TRUE if in Unicode mode as opposed to single-byte mode
1.69 + TUint32 iActiveWindowBase; // base of the active window - bases are 32-bit because they
1.70 + // can be set to the surrogate area, which represents codes
1.71 + // from 0x00010000 to 0x0010FFFF - planes 1-16 of ISO-10646.
1.72 + static const TUint32 iStaticWindow[EStaticWindows]; // bases of the static windows
1.73 + static const TUint32 iDynamicWindowDefault[EDynamicWindows]; // default bases of the dynamic windows
1.74 + static const TUint16 iSpecialBase[ESpecialBases]; // bases for window offsets F9..FF
1.75 +
1.76 + TUint32 iDynamicWindow[EDynamicWindows]; // bases of the dynamic windows
1.77 + TInt iUnicodeWords; // Unicode words processed; read by compressor, written by expander
1.78 + TInt iMaxUnicodeWords; // maximum number of Unicode words to read or write
1.79 + TInt iCompressedBytes; // compressed bytes processed: read by expander, written by compressor
1.80 + TInt iMaxCompressedBytes; // maximum number of compressed bytes to read or write
1.81 + };
1.82 +
1.83 +/**
1.84 +@internalComponent
1.85 +*/
1.86 +class MUnicodeSource
1.87 + {
1.88 + public:
1.89 + virtual TUint16 ReadUnicodeValueL() = 0;
1.90 + };
1.91 +
1.92 +/**
1.93 +@internalComponent
1.94 +
1.95 +A class to read Unicode values directly from memory.
1.96 +*/
1.97 +class TMemoryUnicodeSource: public MUnicodeSource
1.98 + {
1.99 + public:
1.100 + inline TMemoryUnicodeSource(const TUint16* aPtr);
1.101 + inline TUint16 ReadUnicodeValueL();
1.102 +
1.103 + private:
1.104 + const TUint16* iPtr;
1.105 + };
1.106 +
1.107 +/**
1.108 +@internalComponent
1.109 +
1.110 +A class to read Unicode values from a stream built on a memory object.
1.111 +*/
1.112 +class TMemoryStreamUnicodeSource: public MUnicodeSource
1.113 + {
1.114 + public:
1.115 + inline TMemoryStreamUnicodeSource(CStoreReadStream& aStream);
1.116 + inline TUint16 ReadUnicodeValueL();
1.117 +
1.118 + private:
1.119 + CStoreReadStream& iStream;
1.120 + };
1.121 +
1.122 +/**
1.123 +@internalComponent
1.124 +
1.125 +*/
1.126 +class MUnicodeSink
1.127 + {
1.128 + public:
1.129 + virtual void WriteUnicodeValueL(TUint16 aValue) = 0;
1.130 + };
1.131 +
1.132 +/**
1.133 +@internalComponent
1.134 +
1.135 +A class to write Unicode values directly to memory.
1.136 +*/
1.137 +class TMemoryUnicodeSink: public MUnicodeSink
1.138 + {
1.139 + public:
1.140 + inline TMemoryUnicodeSink(TUint16* aPtr);
1.141 + inline void WriteUnicodeValueL(TUint16 aValue);
1.142 +
1.143 + private:
1.144 + TUint16* iPtr;
1.145 + };
1.146 +
1.147 +/**
1.148 +@internalComponent
1.149 +
1.150 +A class to write Unicode values to a stream built on a memory object.
1.151 +*/
1.152 +class TMemoryStreamUnicodeSink: public MUnicodeSink
1.153 + {
1.154 + public:
1.155 + inline TMemoryStreamUnicodeSink(CStoreWriteStream& aStream);
1.156 + inline void WriteUnicodeValueL(TUint16 aValue);
1.157 +
1.158 + private:
1.159 + CStoreWriteStream& iStream;
1.160 + };
1.161 +
1.162 +/**
1.163 +@internalComponent
1.164 +
1.165 +A class to hold functions to compress text using the Standard Compression Scheme for Unicode.
1.166 +*/
1.167 +class TUnicodeCompressor: public TUnicodeCompressionState
1.168 + {
1.169 + public:
1.170 + TUnicodeCompressor();
1.171 + void CompressL(CStoreWriteStream& aOutput,MUnicodeSource& aInput,
1.172 + TInt aMaxOutputBytes = KMaxTInt,TInt aMaxInputWords = KMaxTInt,
1.173 + TInt* aOutputBytes = static_cast<TInt*>(NULL),TInt* aInputWords = static_cast<TInt*>(NULL));
1.174 + void CompressL(TUint8* aOutput,MUnicodeSource& aInput,
1.175 + TInt aMaxOutputBytes = KMaxTInt,TInt aMaxInputWords = KMaxTInt,
1.176 + TInt* aOutputBytes = static_cast<TInt*>(NULL),TInt* aInputWords = static_cast<TInt*>(NULL));
1.177 + TInt FlushL(CStoreWriteStream& aOutput,TInt aMaxOutputBytes,TInt& aOutputBytes);
1.178 + TInt FlushL(TUint8* aOutput,TInt aMaxOutputBytes,TInt& aOutputBytes);
1.179 + static TInt CompressedSizeL(MUnicodeSource& aInput,TInt aInputWords);
1.180 +
1.181 + private:
1.182 +
1.183 + // A structure to store a character and its treatment code
1.184 + struct TAction
1.185 + {
1.186 + // Treatment codes: static and dynamic window numbers, plain ASCII or plain Unicode
1.187 + enum
1.188 + {
1.189 + EPlainUnicode = -2, // character cannot be expressed as ASCII or using static or dynamic windows
1.190 + EPlainASCII = -1, // character can be emitted as an ASCII code
1.191 + EFirstDynamic = 0, // values 0..255 are for dynamic windows with offsets at these places in the offset table
1.192 + ELastDynamic = 255,
1.193 + EFirstStatic = 256, // values 256..263 are for static windows 0..7
1.194 + ELastStatic = 263
1.195 + };
1.196 +
1.197 + inline TAction();
1.198 + TAction(TUint16 aCode);
1.199 +
1.200 + TUint16 iCode; // Unicode value of the character
1.201 + TInt iTreatment; // treatment code: see above
1.202 + };
1.203 +
1.204 + void DoCompressL(CStoreWriteStream* aOutputStream,TUint8* aOutputPointer,MUnicodeSource* aInput,
1.205 + TInt aMaxCompressedBytes,TInt aMaxUnicodeWords,
1.206 + TInt* aCompressedBytes,TInt* aUnicodeWords);
1.207 + void FlushInputBufferL();
1.208 + void FlushOutputBufferL();
1.209 + void WriteRunL();
1.210 + void WriteCharacter(const TAction& aAction);
1.211 + void WriteSCharacter(const TAction& aAction);
1.212 + void WriteUCharacter(TUint16 aCode);
1.213 + void WriteByte(TUint aByte);
1.214 + void WriteCharacterFromBuffer();
1.215 + void SelectTreatment(TInt aTreatment);
1.216 +
1.217 + enum
1.218 + {
1.219 + EMaxInputBufferSize = 4,
1.220 + EMaxOutputBufferSize = EMaxInputBufferSize * 3 // no Unicode character can be encoded as more than three bytes
1.221 + };
1.222 + TAction iInputBuffer[EMaxInputBufferSize]; // circular buffer; queue of Unicode characters to be processed
1.223 + TInt iInputBufferStart; // position of first Unicode character to be processed
1.224 + TInt iInputBufferSize; // characters in the input buffer
1.225 + TUint8 iOutputBuffer[EMaxOutputBufferSize]; // circular buffer; queue of compressed bytes to be output
1.226 + TInt iOutputBufferStart; // position of first compressed byte to be output
1.227 + TInt iOutputBufferSize; // characters in the output buffer
1.228 + TInt iDynamicWindowIndex; // index of the current dynamic window
1.229 + CStoreWriteStream* iOutputStream; // if non-null, output is to this stream
1.230 + TUint8* iOutputPointer; // if non-null, output is to memory
1.231 + MUnicodeSource* iInput; // input object
1.232 + };
1.233 +
1.234 +/**
1.235 +@internalComponent
1.236 +
1.237 +A class to hold functions to expand text using the Standard Compression Scheme for Unicode.
1.238 +*/
1.239 +class TUnicodeExpander: public TUnicodeCompressionState
1.240 + {
1.241 + public:
1.242 + TUnicodeExpander();
1.243 + void ExpandL(MUnicodeSink& aOutput,CStoreReadStream& aInput,
1.244 + TInt aMaxOutputWords = KMaxTInt,TInt aMaxInputBytes = KMaxTInt,
1.245 + TInt* aOutputWords = static_cast<TInt*>(NULL),TInt* aInputBytes = static_cast<TInt*>(NULL));
1.246 + void ExpandL(MUnicodeSink& aOutput,const TUint8* aInput,
1.247 + TInt aMaxOutputWords = KMaxTInt,TInt aMaxInputBytes = KMaxTInt,
1.248 + TInt* aOutputWords = static_cast<TInt*>(NULL),TInt* aInputBytes = static_cast<TInt*>(NULL));
1.249 + TInt FlushL(MUnicodeSink& aOutput,TInt aMaxOutputWords,TInt& aOutputWords);
1.250 + static TInt ExpandedSizeL(CStoreReadStream& aInput,TInt aInputBytes);
1.251 + static TInt ExpandedSizeL(const TUint8* aInput,TInt aInputBytes);
1.252 +
1.253 + private:
1.254 + void DoExpandL(MUnicodeSink* aOutput,CStoreReadStream* aInputStream,const TUint8* aInputPointer,
1.255 + TInt aMaxOutputWords,TInt aMaxInputBytes,
1.256 + TInt* aOutputWords,TInt* aInputBytes);
1.257 + void HandleByteL();
1.258 + void FlushOutputBufferL();
1.259 + TBool HandleSByteL(TUint8 aByte);
1.260 + TBool HandleUByteL(TUint8 aByte);
1.261 + TBool ReadByteL(TUint8& aByte);
1.262 + TBool QuoteUnicodeL();
1.263 + TBool DefineWindowL(TInt aIndex);
1.264 + TBool DefineExpansionWindowL();
1.265 + void WriteChar(TText aChar);
1.266 + void WriteChar32(TUint aChar);
1.267 +
1.268 + enum
1.269 + {
1.270 + EMaxInputBufferSize = 3, // no Unicode character can be encoded as more than 3 bytes
1.271 + EMaxOutputBufferSize = 2 // no byte can be expanded into more than 2 Unicode characters
1.272 + };
1.273 + TUint8 iInputBuffer[EMaxInputBufferSize]; // buffer containing a group of compressed bytes representing
1.274 + // a single operation; when an input source ends in the
1.275 + // middle of an operation, this buffer enables the next
1.276 + // expansion to start in the correct state
1.277 + TInt iInputBufferStart; // next read position in the input buffer
1.278 + TInt iInputBufferSize; // bytes in the input buffer
1.279 + TUint16 iOutputBuffer[EMaxOutputBufferSize]; // circular buffer; queue of Unicode characters to be output
1.280 + TInt iOutputBufferStart; // position of first Unicode character to be output
1.281 + TInt iOutputBufferSize; // characters in the output buffer
1.282 + MUnicodeSink* iOutput; // output object
1.283 + CStoreReadStream* iInputStream; // if non-null, input is from this stream
1.284 + const TUint8* iInputPointer; // if non-null, input is from memory
1.285 + };
1.286 +
1.287 +// inline functions start here
1.288 +
1.289 +inline TMemoryUnicodeSource::TMemoryUnicodeSource(const TUint16* aPtr):
1.290 + iPtr(aPtr)
1.291 + {
1.292 + }
1.293 +
1.294 +inline TUint16 TMemoryUnicodeSource::ReadUnicodeValueL()
1.295 + {
1.296 + return *iPtr++;
1.297 + }
1.298 +
1.299 +inline TMemoryStreamUnicodeSource::TMemoryStreamUnicodeSource(CStoreReadStream& aStream):
1.300 + iStream(aStream)
1.301 + {
1.302 + }
1.303 +
1.304 +inline TUint16 TMemoryStreamUnicodeSource::ReadUnicodeValueL()
1.305 + {
1.306 + TUint16 x;
1.307 + iStream.Read(reinterpret_cast<TUint8*>(&x),sizeof(TUint16));
1.308 + return x;
1.309 + }
1.310 +
1.311 +inline TMemoryUnicodeSink::TMemoryUnicodeSink(TUint16* aPtr):
1.312 + iPtr(aPtr)
1.313 + {
1.314 + }
1.315 +
1.316 +inline void TMemoryUnicodeSink::WriteUnicodeValueL(TUint16 aValue)
1.317 + {
1.318 + *iPtr++ = aValue;
1.319 + }
1.320 +
1.321 +inline TMemoryStreamUnicodeSink::TMemoryStreamUnicodeSink(CStoreWriteStream& aStream):
1.322 + iStream(aStream)
1.323 + {
1.324 + }
1.325 +
1.326 +inline void TMemoryStreamUnicodeSink::WriteUnicodeValueL(TUint16 aValue)
1.327 + {
1.328 + iStream.Write(reinterpret_cast<TUint8*>(&aValue),sizeof(TUint16));
1.329 + }
1.330 +
1.331 +inline TUnicodeCompressor::TAction::TAction():
1.332 + iCode(0),
1.333 + iTreatment(EPlainUnicode)
1.334 + {
1.335 + }
1.336 +}
1.337 +#endif // !defined(__UNICODECOMPRESSION_H__)