sl@0: /*
sl@0: * Copyright (c) 1997-2009 Nokia Corporation and/or its subsidiary(-ies).
sl@0: * All rights reserved.
sl@0: * This component and the accompanying materials are made available
sl@0: * under the terms of "Eclipse Public License v1.0"
sl@0: * which accompanies this distribution, and is available
sl@0: * at the URL "http://www.eclipse.org/legal/epl-v10.html".
sl@0: *
sl@0: * Initial Contributors:
sl@0: * Nokia Corporation - initial contribution.
sl@0: *
sl@0: * Contributors:
sl@0: *
sl@0: * Description: 
sl@0: *
sl@0: */
sl@0: 
sl@0: 
sl@0: #include <e32std.h>
sl@0: #include <convdata.h>
sl@0: 
sl@0: #define ARRAY_LENGTH(aArray) (sizeof(aArray)/sizeof((aArray)[0]))
sl@0: 
sl@0: LOCAL_D const TUint16 keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_1[]=
sl@0: 	{
sl@0: 	0x201a,
sl@0: 	0x0192,
sl@0: 	0x201e,
sl@0: 	0x2026,
sl@0: 	0x2020,
sl@0: 	0x2021,
sl@0: 	0x02c6,
sl@0: 	0x2030,
sl@0: 	0x0160,
sl@0: 	0x2039,
sl@0: 	0x0152
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const TUint16 keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_2[]=
sl@0: 	{
sl@0: 	0x2018,
sl@0: 	0x2019,
sl@0: 	0x201c,
sl@0: 	0x201d,
sl@0: 	0x2022,
sl@0: 	0x2013,
sl@0: 	0x2014,
sl@0: 	0x02dc,
sl@0: 	0x2122,
sl@0: 	0x0161,
sl@0: 	0x203a,
sl@0: 	0x0153
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const TUint16 keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_3[]=
sl@0: 	{
sl@0: 	0x017e,
sl@0: 	0x0178
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const SCnvConversionData::SOneDirectionData::SRange::UData::SKeyedTable16OfIndexedTables16::SKeyedEntry keyedTables16OfIndexedTables16_keyedEntries_codePage1252ToUnicode_1[]=
sl@0: 	{
sl@0: 		{
sl@0: 		0x82,
sl@0: 		0x8c,
sl@0: 		keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_1
sl@0: 		},
sl@0: 		{
sl@0: 		0x91,
sl@0: 		0x9c,
sl@0: 		keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_2
sl@0: 		},
sl@0: 		{
sl@0: 		0x9e,
sl@0: 		0x9f,
sl@0: 		keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_3
sl@0: 		}
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const SCnvConversionData::SOneDirectionData::SRange::UData::SKeyedTable1616::SEntry keyedTable1616_unicodeToCodePage1252_1[]=
sl@0: 	{
sl@0: 		{
sl@0: 		0x0152,
sl@0: 		0x8c
sl@0: 		},
sl@0: 		{
sl@0: 		0x0153,
sl@0: 		0x9c
sl@0: 		},
sl@0: 		{
sl@0: 		0x0160,
sl@0: 		0x8a
sl@0: 		},
sl@0: 		{
sl@0: 		0x0161,
sl@0: 		0x9a
sl@0: 		},
sl@0: 		{
sl@0: 		0x0178,
sl@0: 		0x9f
sl@0: 		},
sl@0: 		{
sl@0: 		0x017d,
sl@0: 		0x8e
sl@0: 		},
sl@0: 		{
sl@0: 		0x017e,
sl@0: 		0x9e
sl@0: 		},
sl@0: 		{
sl@0: 		0x0192,
sl@0: 		0x83
sl@0: 		},
sl@0: 		{
sl@0: 		0x02c6,
sl@0: 		0x88
sl@0: 		},
sl@0: 		{
sl@0: 		0x02dc,
sl@0: 		0x98
sl@0: 		},
sl@0: 		{
sl@0: 		0x2013,
sl@0: 		0x96
sl@0: 		},
sl@0: 		{
sl@0: 		0x2014,
sl@0: 		0x97
sl@0: 		},
sl@0: 		{
sl@0: 		0x2018,
sl@0: 		0x91
sl@0: 		},
sl@0: 		{
sl@0: 		0x2019,
sl@0: 		0x92
sl@0: 		},
sl@0: 		{
sl@0: 		0x201a,
sl@0: 		0x82
sl@0: 		},
sl@0: 		{
sl@0: 		0x201c,
sl@0: 		0x93
sl@0: 		},
sl@0: 		{
sl@0: 		0x201d,
sl@0: 		0x94
sl@0: 		},
sl@0: 		{
sl@0: 		0x201e,
sl@0: 		0x84
sl@0: 		},
sl@0: 		{
sl@0: 		0x2020,
sl@0: 		0x86
sl@0: 		},
sl@0: 		{
sl@0: 		0x2021,
sl@0: 		0x87
sl@0: 		},
sl@0: 		{
sl@0: 		0x2022,
sl@0: 		0x95
sl@0: 		},
sl@0: 		{
sl@0: 		0x2026,
sl@0: 		0x85
sl@0: 		},
sl@0: 		{
sl@0: 		0x2030,
sl@0: 		0x89
sl@0: 		},
sl@0: 		{
sl@0: 		0x2039,
sl@0: 		0x8b
sl@0: 		},
sl@0: 		{
sl@0: 		0x203a,
sl@0: 		0x9b
sl@0: 		},
sl@0: 		{
sl@0: 		0x20ac,
sl@0: 		0x80
sl@0: 		},
sl@0: 		{
sl@0: 		0x2122,
sl@0: 		0x99
sl@0: 		}
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const SCnvConversionData::SVariableByteData::SRange codePage1252VariableByteDataRanges[]=
sl@0: 	{
sl@0: 		{
sl@0: 		0x00,
sl@0: 		0xff,
sl@0: 		0,
sl@0: 		0
sl@0: 		}
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const SCnvConversionData::SOneDirectionData::SRange codePage1252ToUnicodeDataRanges[]=
sl@0: 	{
sl@0: 		{
sl@0: 		0x00,
sl@0: 		0x7f,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EDirect,
sl@0: 		0,
sl@0: 		0,
sl@0: 			{
sl@0: 			0,
sl@0: 			0
sl@0: 			}
sl@0: 		},
sl@0: 		{
sl@0: 		0xa0,
sl@0: 		0xff,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EDirect,
sl@0: 		0,
sl@0: 		0,
sl@0: 			{
sl@0: 			0,
sl@0: 			0
sl@0: 			}
sl@0: 		},
sl@0: 		{
sl@0: 		0x80,
sl@0: 		0x80,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EOffset,
sl@0: 		0,
sl@0: 		0,
sl@0: 			{
sl@0: 			STATIC_CAST(TUint, 8236),
sl@0: 			0
sl@0: 			}
sl@0: 		},
sl@0: 		{
sl@0: 		0x8e,
sl@0: 		0x8e,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EOffset,
sl@0: 		0,
sl@0: 		0,
sl@0: 			{
sl@0: 			STATIC_CAST(TUint, 239),
sl@0: 			0
sl@0: 			}
sl@0: 		},
sl@0: 		{
sl@0: 		0x82,
sl@0: 		0x9f,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EKeyedTable16OfIndexedTables16,
sl@0: 		0,
sl@0: 		0,
sl@0: 			{
sl@0: 			UData_SKeyedTable16OfIndexedTables16(keyedTables16OfIndexedTables16_keyedEntries_codePage1252ToUnicode_1)
sl@0: 			}
sl@0: 		}
sl@0: 	};
sl@0: 
sl@0: LOCAL_D const SCnvConversionData::SOneDirectionData::SRange unicodeToCodePage1252DataRanges[]=
sl@0: 	{
sl@0: 		{
sl@0: 		0x0000,
sl@0: 		0x007f,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EDirect,
sl@0: 		1,
sl@0: 		0,
sl@0: 			{
sl@0: 			0,
sl@0: 			0
sl@0: 			}
sl@0: 		},
sl@0: 		{
sl@0: 		0x00a0,
sl@0: 		0x00ff,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EDirect,
sl@0: 		1,
sl@0: 		0,
sl@0: 			{
sl@0: 			0,
sl@0: 			0
sl@0: 			}
sl@0: 		},
sl@0: 		{
sl@0: 		0x0152,
sl@0: 		0x2122,
sl@0: 		SCnvConversionData::SOneDirectionData::SRange::EKeyedTable1616,
sl@0: 		1,
sl@0: 		0,
sl@0: 			{
sl@0: 			UData_SKeyedTable1616(keyedTable1616_unicodeToCodePage1252_1)
sl@0: 			}
sl@0: 		}
sl@0: 	};
sl@0: 
sl@0: GLREF_D const SCnvConversionData codePage1252ConversionData=
sl@0: 	{
sl@0: 	SCnvConversionData::EUnspecified,
sl@0: 		{
sl@0: 		ARRAY_LENGTH(codePage1252VariableByteDataRanges),
sl@0: 		codePage1252VariableByteDataRanges
sl@0: 		},
sl@0: 		{
sl@0: 		ARRAY_LENGTH(codePage1252ToUnicodeDataRanges),
sl@0: 		codePage1252ToUnicodeDataRanges
sl@0: 		},
sl@0: 		{
sl@0: 		ARRAY_LENGTH(unicodeToCodePage1252DataRanges),
sl@0: 		unicodeToCodePage1252DataRanges
sl@0: 		},
sl@0: 	NULL,
sl@0: 	NULL
sl@0: 	};
sl@0: 
sl@0: GLREF_C void IsCharacterSetCP1252(TInt& aConfidenceLevel, const TDesC8& aSample)
sl@0: 	{
sl@0: 	aConfidenceLevel = 60;
sl@0: 	TInt sampleLength = aSample.Length();
sl@0: 
sl@0: 	for (TInt i=0; i<sampleLength; ++i)
sl@0: 		{
sl@0: 		// CP1252 includes ASCII as well
sl@0: 		// first check if the char is in the range 0x80 - 0x9f (controls codes in ISO88591)
sl@0: 		// If it is in that range then the likelihood that it's CP1252 is a bit higher
sl@0: 		if ((aSample[i] >= 0x80) && (aSample[i] <= 0x9f))
sl@0: 			{
sl@0: 			if((aSample[i]==0x81)||(aSample[i]==0x8D)||(aSample[i]==0x8f)||
sl@0: 				(aSample[i]==0x90)||(aSample[i]==0x9d))
sl@0: 				{
sl@0: 				// These code values are not supported by the Codepage CP1252
sl@0: 				aConfidenceLevel = 0;
sl@0: 				break;
sl@0: 				}
sl@0: 			else
sl@0: 				{
sl@0: 				// problem: UTF8 uses the values 0x80-0x9f in more than 50% of it's multibyte representation
sl@0: 				// so if the text was UTF8 .... the confidence here would hit the roof. Could check to make 
sl@0: 				// sure that this is not UTF8
sl@0: 				aConfidenceLevel+=1;
sl@0: 				}
sl@0: 			}
sl@0: 		TInt increment1 = i+1;
sl@0: 		TInt decrement1 = i-1;
sl@0: 		// 0xf7 is the division symbol in CP1252.
sl@0: 		// 0xd7 is the division symbol in CP1252.If char on either side of the division
sl@0: 		// symbol is a number then the confidence that it's ISO88591 increases
sl@0: 		if( decrement1>= 0 && ((aSample[i]==0xf7) || (aSample[i]==0xd7)) && increment1<sampleLength)
sl@0: 			{
sl@0: 			
sl@0: 			if (increment1 >= sampleLength)
sl@0: 				break;
sl@0: 			if ( (aSample[decrement1] >= 0x30) && (aSample[decrement1] <= 0x39) &&  // char before is a number
sl@0: 				 (aSample[increment1] >= 0x30) && (aSample[increment1] <= 0x39) )   // char after is a number
sl@0: 				{
sl@0: 				aConfidenceLevel+=5;
sl@0: 				}
sl@0: 			}
sl@0: 		// Can also use the currency symbol to increase confidence if the char after a 
sl@0: 		// currency symbol is numeric
sl@0: 		if((aSample[i]>=0xa2) && (aSample[i] <= 0xa5) && increment1<sampleLength)
sl@0: 			{
sl@0: 			if ((aSample[increment1] >= 0x30) && (aSample[increment1] <= 0x39))
sl@0: 				{
sl@0: 				aConfidenceLevel+=5; 
sl@0: 				}
sl@0: 			}
sl@0: 		} // for loop
sl@0: 	aConfidenceLevel =(aConfidenceLevel >0)? ((aConfidenceLevel > 100)? 100: aConfidenceLevel): 0;
sl@0: 	}