Symaptic: os/textandloc/charconvfw/charconv_fw/src/charconv/iso88591.cpp@260cb5ec6c19

     1 /*

     2 * Copyright (c) 1997-2009 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 "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 */

    19 #include <e32std.h>

    20 #include <convdata.h>

    22 #define ARRAY_LENGTH(aArray) (sizeof(aArray)/sizeof((aArray)[0]))

    24 LOCAL_D const SCnvConversionData::SVariableByteData::SRange iso88591VariableByteDataRanges[]=

    25 	{

    26 		{

    27 		0x00,

    28 		0xff,

    29 		0,

    30 		0

    31 		}

    32 	};

    34 LOCAL_D const SCnvConversionData::SOneDirectionData::SRange iso88591ToUnicodeDataRanges[]=

    35 	{

    36 		{

    37 		0x00,

    38 		0x7f,

    39 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

    40 		0,

    41 		0,

    42 			{

    43 			0,

    44 			0

    45 			}

    46 		},

    47 		{

    48 		0x81,

    49 		0xff,

    50 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

    51 		0,

    52 		0,

    53 			{

    54 			0,

    55 			0

    56 			}

    57 		},

    58 		{

    59 		0x80,

    60 		0x80,

    61 		SCnvConversionData::SOneDirectionData::SRange::EOffset,

    62 		0,

    63 		0,

    64 			{

    65 			STATIC_CAST(TUint, 8236)

    66 			}

    67 		}

    68 	};

    70 LOCAL_D const SCnvConversionData::SOneDirectionData::SRange unicodeToIso88591DataRanges[]=

    71 	{

    72 		{

    73 		0x0000,

    74 		0x007f,

    75 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

    76 		1,

    77 		0,

    78 			{

    79 			0,

    80 			0

    81 			}

    82 		},

    83 		{

    84 		0x0081,

    85 		0x00ff,

    86 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

    87 		1,

    88 		0,

    89 			{

    90 			0,

    91 			0

    92 			}

    93 		},

    94 		{

    95 		0x20ac,

    96 		0x20ac,

    97 		SCnvConversionData::SOneDirectionData::SRange::EOffset,

    98 		1,

    99 		0,

   100 			{

   101 			STATIC_CAST(TUint, -8236)

   102 			}

   103 		}

   104 	};

   106 GLREF_D const SCnvConversionData iso88591ConversionData=

   107 	{

   108 	SCnvConversionData::EUnspecified,

   109 		{

   110 		ARRAY_LENGTH(iso88591VariableByteDataRanges),

   111 		iso88591VariableByteDataRanges

   112 		},

   113 		{

   114 		ARRAY_LENGTH(iso88591ToUnicodeDataRanges),

   115 		iso88591ToUnicodeDataRanges

   116 		},

   117 		{

   118 		ARRAY_LENGTH(unicodeToIso88591DataRanges),

   119 		unicodeToIso88591DataRanges

   120 		},

   121 	NULL,

   122 	NULL

   123 	};

   125 GLREF_C void IsCharacterSetISO88591(TInt& aConfidenceLevel, const TDesC8& aSample)

   126 	{

   127 	TInt sampleLength = aSample.Length();

   128 	aConfidenceLevel = 75;

   130 	for (TInt i=0; i<sampleLength; ++i)

   131 		{

   132 		// ISO88591 includes ASCII as well

   133 		// first check if the char is in the range 0x80 - 0x9f (controls codes)

   134 		// If it is in that range then it's not ISO88591

   135 		if ((aSample[i] >= 0x80) && (aSample[i] <= 0x9f))

   136 			{

   137 			aConfidenceLevel=0;

   138 			break;

   139 			}

   140 		// 0xf7 is the division symbol in ISO88591.

   141 		// 0xd7 is the division symbol in ISO88591.If char on either side of the division

   142 		// symbol is a number then the confidence that it's ISO88591 increases

   143 		if( i>0 && ((aSample[i]==0xf7) || (aSample[i]==0xd7)) && ((i+1)<sampleLength) )

   144 			{

   145 			if ( (aSample[i-1] >= 0x30) && (aSample[i-1] <= 0x39) &&  // char before is a number

   146 				 (aSample[i+1] >= 0x30) && (aSample[i+1] <= 0x39) )   // char after is a number

   147 				{

   148 				aConfidenceLevel+=5;

   149 				}

   150 			}

   151 		// Can also use the currency symbol to increase confidence if the char after a

   152 		// currency symbol is numeric

   153 		if((aSample[i]>=0xa2) && (aSample[i] <= 0xa5) && ((i+1)<sampleLength))

   154 			{

   155 			if ((aSample[i+1] >= 0x30) && (aSample[i+1] <= 0x39))

   156 				{

   157 				aConfidenceLevel+=5;

   158 				}

   159 			}

   160 		} // for loop

   161 	aConfidenceLevel =(aConfidenceLevel >0)? ((aConfidenceLevel > 100)? 100: aConfidenceLevel): 0;

   162 	}

author	sl
	Tue, 10 Jun 2014 14:32:02 +0200
changeset 1	260cb5ec6c19
permissions	-rw-r--r--