mmlibs/mmfw/Recogniser/src/matparser.cpp
changeset 0 b8ed18f6c07b
equal deleted inserted replaced
-1:000000000000 0:b8ed18f6c07b
       
     1 // Copyright (c) 2006-2009 Nokia Corporation and/or its subsidiary(-ies).
       
     2 // All rights reserved.
       
     3 // This component and the accompanying materials are made available
       
     4 // under the terms of "Eclipse Public License v1.0"
       
     5 // which accompanies this distribution, and is available
       
     6 // at the URL "http://www.eclipse.org/legal/epl-v10.html".
       
     7 //
       
     8 // Initial Contributors:
       
     9 // Nokia Corporation - initial contribution.
       
    10 //
       
    11 // Contributors:
       
    12 //
       
    13 // Description:
       
    14 //
       
    15 
       
    16 #include "constants.h"
       
    17 #include "parsers.h"
       
    18 
       
    19 //
       
    20 // Matroska 'XML' tags.
       
    21 //
       
    22 static const TUint32 KEBML = 0x1A45DFA3;
       
    23 static const TUint32 KSegment = 0x18538067;
       
    24 static const TUint32 KTracks = 0x1654AE6B;
       
    25 static const TUint32 KTrackEntry = 0xAE;
       
    26 static const TUint32 KTrackType = 0x83;
       
    27 static const TUint32 KVoid = 0xEC;
       
    28 static const TUint32 KSeekHead = 0x114D9B74;
       
    29 static const TUint32 KSegmentInfo = 0x1549A966;
       
    30 static const TUint32 KCluster = 0x1F43B675;
       
    31 static const TUint32 KCues = 0x1C53BB6B;
       
    32 static const TUint32 KAttachments = 0x1941A469;
       
    33 static const TUint32 KChapters = 0x1043A770;
       
    34 static const TUint32 KTags = 0x1254C367;
       
    35 
       
    36 //
       
    37 // Matroska Track Types.
       
    38 //
       
    39 static const TUint8 KTrackTypeVideo = 0x01;
       
    40 
       
    41 #define KMatroskaConfidenceMask	0x07	// 00000111
       
    42 #define KMatroskaEBMLBit		KBit1
       
    43 #define KMatroskaSegmentBit		KBit2
       
    44 #define KMatroskaVideoBit		KBit3
       
    45 
       
    46 
       
    47 //
       
    48 // This truth table maps the following flags to confidence levels.
       
    49 //
       
    50 // A: Extension match.
       
    51 // B: EBML tag found.
       
    52 // C: Segment tag found.
       
    53 //
       
    54 // C B A -> Confidence
       
    55 // --------------------
       
    56 // 0 0 0 -> NotRecognised
       
    57 // 0 0 1 -> EPossible
       
    58 // 0 1 0 -> EPossible
       
    59 // 0 1 1 -> EProbable
       
    60 // 1 0 0 -> ENotRecognised (EBML should be present)
       
    61 // 1 0 1 -> ENotRecognised (EBML should be present)
       
    62 // 1 1 0 -> EProbable
       
    63 // 1 1 1 -> ECertain
       
    64 //
       
    65 static const TInt KMatroskaFlagsToConfidence[8] =
       
    66 	{
       
    67 	KConfNotRecognised,
       
    68 	KConfPossible,
       
    69 	KConfPossible,
       
    70 	KConfProbable,
       
    71 	KConfNotRecognised,
       
    72 	KConfNotRecognised,
       
    73 	KConfProbable,
       
    74 	KConfCertain
       
    75 	};
       
    76 	
       
    77 static const TInt KMatExtensionOnlyIndex = 1; // See truth table.
       
    78 
       
    79 
       
    80 typedef struct
       
    81 	{
       
    82 	const TText* iExt;
       
    83 	const TText8* iMime;
       
    84 	}
       
    85 TMatroskaExt;
       
    86 
       
    87 //
       
    88 // Known Matroska extensions and their corresponding MIME-types.
       
    89 //
       
    90 static const TMatroskaExt KMatroskaExt[] =
       
    91 	{
       
    92 		{KExtMAT_A, KMimeMAT_A },
       
    93 		{KExtMAT_V, KMimeMAT_V }
       
    94 	};
       
    95 
       
    96 static const TInt KMatroskaExtCount = sizeof(KMatroskaExt) / sizeof(TMatroskaExt);
       
    97 
       
    98 
       
    99 //
       
   100 //
       
   101 //
       
   102 TMatroskaParser::TMatroskaParser(CReader& aReader, TFlags& aFlags)
       
   103  :	iReader(aReader),
       
   104  	iFlags(aFlags)
       
   105 	{
       
   106 	}
       
   107 
       
   108 
       
   109 //
       
   110 //
       
   111 //
       
   112 const TText8* TMatroskaParser::MatchExtension(const TDesC& aExt)
       
   113 	{
       
   114 	for (TInt i = 0; i < KMatroskaExtCount; i++)
       
   115 		{
       
   116 		if (aExt.MatchF(TPtrC(KMatroskaExt[i].iExt)) != KErrNotFound)
       
   117 			{
       
   118 			return KMatroskaExt[i].iMime;
       
   119 			}
       
   120 		}
       
   121 	
       
   122 	return NULL;
       
   123 	}
       
   124 
       
   125 
       
   126 //
       
   127 // This function calls the parser and turns the results
       
   128 // into a MIME-type.
       
   129 //
       
   130 void TMatroskaParser::DoRecognise(const TDesC& aExt, CReader& aReader, TMatch& aMatch)
       
   131 	{
       
   132 	TFlags flags;
       
   133 	TMatroskaParser parser(aReader, flags);
       
   134 	
       
   135 	// Try to match the extension.
       
   136 	const TText8* extMime = parser.MatchExtension(aExt);
       
   137 	if (extMime != NULL)
       
   138 		{
       
   139 		flags.SetExtensionFlag();
       
   140 		}
       
   141 
       
   142 	// Try to parse the content.
       
   143 	TRAP_IGNORE(parser.ParseL());
       
   144 	TInt confIndex = flags.GetBitField(KMatroskaConfidenceMask);
       
   145 	aMatch.iConfidence = KMatroskaFlagsToConfidence[confIndex];
       
   146 	if (aMatch.iConfidence != KConfNotRecognised)
       
   147 		{
       
   148 		// If any header data has been recognised trust that,
       
   149 		// otherwise go with the extension.
       
   150 		if (confIndex == KMatExtensionOnlyIndex)
       
   151 			{
       
   152 			aMatch.iMime = extMime;
       
   153 			}
       
   154 		else
       
   155 			{
       
   156 			aMatch.iMime = (flags.GetBitField(KMatroskaVideoBit) ? KMimeMAT_V : KMimeMAT_A);
       
   157 			}
       
   158 		}
       
   159 	}
       
   160 
       
   161 
       
   162 //
       
   163 // This function does the parsing.
       
   164 //
       
   165 void TMatroskaParser::ParseL()
       
   166 	{
       
   167 	TUint64 id;
       
   168 	TInt64 size;
       
   169 	TBool haveElement = EFalse;
       
   170 	
       
   171 	// Assume there's video content if we only have buffer data.
       
   172 	if (iReader.Type() == CReader::EBuffer)
       
   173 		{
       
   174 		iFlags.SetBit(KMatroskaVideoBit);
       
   175 		}
       
   176 		
       
   177 	FOREVER
       
   178 		{
       
   179 		if (!haveElement)
       
   180 			{
       
   181 			ReadElementL(id, size);
       
   182 			}
       
   183 		
       
   184 		switch (id)
       
   185 			{
       
   186 			case KEBML:
       
   187 				iFlags.SetBit(KMatroskaEBMLBit);
       
   188 				break;
       
   189 				
       
   190 			case KSegment:
       
   191 				haveElement = ReadSegmentL(id, size);
       
   192 				break;
       
   193 				
       
   194 			default:
       
   195 				// Skip it.
       
   196 				iReader.SeekL(size);
       
   197 			}
       
   198 		}
       
   199 	}
       
   200 
       
   201 
       
   202 //
       
   203 // This function returns ETrue if the aNextID and aNextSize parameters
       
   204 // contain valid values, EFalse otherwise.
       
   205 //
       
   206 TBool TMatroskaParser::ReadSegmentL(TUint64& aNextID, TInt64& aNextSize)
       
   207 	{
       
   208 	TUint64 id;
       
   209 	TInt64 size;
       
   210 	TBool videoContent = EFalse;
       
   211 	
       
   212 	aNextID = 0;
       
   213 	aNextSize = 0;
       
   214 	
       
   215 	iFlags.SetBit(KMatroskaSegmentBit);
       
   216 	
       
   217 	while (!videoContent)
       
   218 		{
       
   219 		ReadElementL(id, size);
       
   220 		
       
   221 		switch (id)
       
   222 			{
       
   223 			case KTracks:
       
   224 				videoContent = ReadTrackL(size);
       
   225 				break;
       
   226 			
       
   227 			case KSeekHead:
       
   228 			case KSegmentInfo:
       
   229 			case KCluster:
       
   230 			case KCues:
       
   231 			case KAttachments: 
       
   232 			case KChapters:
       
   233 			case KTags:
       
   234 				iReader.SeekL(size);
       
   235 				break;
       
   236 				
       
   237 			default:
       
   238 				// Unrecognised element id. Pass it back to the caller
       
   239 				aNextID = id;
       
   240 				aNextSize = size;
       
   241 				return ETrue;
       
   242 			}
       
   243 		}
       
   244 	
       
   245 	// Tell the caller they must read the next element themselves.
       
   246 	return EFalse;
       
   247 	}
       
   248 
       
   249 
       
   250 //
       
   251 // The Track. This lets us know if there's video content present.
       
   252 //
       
   253 TBool TMatroskaParser::ReadTrackL(TInt64 aTrackSize)
       
   254 	{
       
   255 	TUint64 id;
       
   256 	TInt64 size;
       
   257 	TInt startPos = iReader.Position();
       
   258 	
       
   259 	while (iReader.Position() - startPos < aTrackSize)
       
   260 		{
       
   261 		ReadElementL(id, size);
       
   262 
       
   263 		switch (id)
       
   264 			{
       
   265 			case KTrackEntry:
       
   266 				break;
       
   267 				
       
   268 			case KTrackType:
       
   269 				TUint8 trackType;
       
   270 				iReader.ReadByteL(trackType);
       
   271 				if (trackType == KTrackTypeVideo)
       
   272 					{
       
   273 					// We found video content so we can stop parsing.
       
   274 					iFlags.SetBit(KMatroskaVideoBit);
       
   275 					return ETrue;
       
   276 					}
       
   277 				break;
       
   278 				
       
   279 			default:
       
   280 				iReader.SeekL(size);
       
   281 			}
       
   282 		}
       
   283 		
       
   284 	return EFalse;
       
   285 	}
       
   286 
       
   287 
       
   288 //
       
   289 //
       
   290 //
       
   291 void TMatroskaParser::ReadElementL(TUint64& aElementID, TInt64& aSize)
       
   292 	{
       
   293 	do
       
   294 		{
       
   295 		aElementID = ReadDataL();
       
   296 		aSize = ReadDataL(ETrue);
       
   297 
       
   298 		// Void elements are used for padding and
       
   299 		// can be ignored.
       
   300 		if (aElementID == KVoid)
       
   301 			{
       
   302 			iReader.SeekL(aSize);
       
   303 			}
       
   304 		}
       
   305 	while (aElementID == KVoid);
       
   306 	}
       
   307 
       
   308 
       
   309 //
       
   310 //
       
   311 //
       
   312 TUint64 TMatroskaParser::ReadDataL(TBool aTurnOffHighestSetBit)
       
   313 	{
       
   314 	TUint64 retval;
       
   315 	TUint8 byte;
       
   316 	TUint8 mask = 0x80; // [1000 0000]. It will be shifted right 1 in each 'i' iteration.
       
   317 	TUint8 size = 1;	// It will be incremented in each 'i' iteration.
       
   318 	
       
   319 	iReader.ReadByteL(byte);
       
   320 	
       
   321 	for (TInt i = 0; i < sizeof(TUint64); i++)
       
   322 		{
       
   323 		if (byte & mask)
       
   324 			{
       
   325 			retval = byte;
       
   326 			if (aTurnOffHighestSetBit)
       
   327 				{
       
   328 				retval &= ~mask; // Turn off the highest set bit.
       
   329 				}
       
   330 			
       
   331 			// Now read the real data.
       
   332 			// Start from 1 because we've already read a byte.
       
   333 			for (TInt j = 1; j < size; j++)
       
   334 				{
       
   335 				iReader.ReadByteL(byte);
       
   336 				retval <<= 8;
       
   337 				retval |= byte;
       
   338 				}
       
   339 				
       
   340 			return retval;
       
   341 			}
       
   342 		else
       
   343 			{
       
   344 			mask >>= 1;
       
   345 			size++;
       
   346 			}
       
   347 		}
       
   348 
       
   349 	User::Leave(KErrCorrupt);
       
   350 	return 0; // Keep the compiler happy.
       
   351 	}
       
   352