Symaptic: os/textandloc/charconvfw/charconv_fw/src/charconv/cp1252.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 TUint16 keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_1[]=

    25 	{

    26 	0x201a,

    27 	0x0192,

    28 	0x201e,

    29 	0x2026,

    30 	0x2020,

    31 	0x2021,

    32 	0x02c6,

    33 	0x2030,

    34 	0x0160,

    35 	0x2039,

    36 	0x0152

    37 	};

    39 LOCAL_D const TUint16 keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_2[]=

    40 	{

    41 	0x2018,

    42 	0x2019,

    43 	0x201c,

    44 	0x201d,

    45 	0x2022,

    46 	0x2013,

    47 	0x2014,

    48 	0x02dc,

    49 	0x2122,

    50 	0x0161,

    51 	0x203a,

    52 	0x0153

    53 	};

    55 LOCAL_D const TUint16 keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_3[]=

    56 	{

    57 	0x017e,

    58 	0x0178

    59 	};

    61 LOCAL_D const SCnvConversionData::SOneDirectionData::SRange::UData::SKeyedTable16OfIndexedTables16::SKeyedEntry keyedTables16OfIndexedTables16_keyedEntries_codePage1252ToUnicode_1[]=

    62 	{

    63 		{

    64 		0x82,

    65 		0x8c,

    66 		keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_1

    67 		},

    68 		{

    69 		0x91,

    70 		0x9c,

    71 		keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_2

    72 		},

    73 		{

    74 		0x9e,

    75 		0x9f,

    76 		keyedTables16OfIndexedTables16_indexedEntries_codePage1252ToUnicode_3

    77 		}

    78 	};

    80 LOCAL_D const SCnvConversionData::SOneDirectionData::SRange::UData::SKeyedTable1616::SEntry keyedTable1616_unicodeToCodePage1252_1[]=

    81 	{

    82 		{

    83 		0x0152,

    84 		0x8c

    85 		},

    86 		{

    87 		0x0153,

    88 		0x9c

    89 		},

    90 		{

    91 		0x0160,

    92 		0x8a

    93 		},

    94 		{

    95 		0x0161,

    96 		0x9a

    97 		},

    98 		{

    99 		0x0178,

   100 		0x9f

   101 		},

   102 		{

   103 		0x017d,

   104 		0x8e

   105 		},

   106 		{

   107 		0x017e,

   108 		0x9e

   109 		},

   110 		{

   111 		0x0192,

   112 		0x83

   113 		},

   114 		{

   115 		0x02c6,

   116 		0x88

   117 		},

   118 		{

   119 		0x02dc,

   120 		0x98

   121 		},

   122 		{

   123 		0x2013,

   124 		0x96

   125 		},

   126 		{

   127 		0x2014,

   128 		0x97

   129 		},

   130 		{

   131 		0x2018,

   132 		0x91

   133 		},

   134 		{

   135 		0x2019,

   136 		0x92

   137 		},

   138 		{

   139 		0x201a,

   140 		0x82

   141 		},

   142 		{

   143 		0x201c,

   144 		0x93

   145 		},

   146 		{

   147 		0x201d,

   148 		0x94

   149 		},

   150 		{

   151 		0x201e,

   152 		0x84

   153 		},

   154 		{

   155 		0x2020,

   156 		0x86

   157 		},

   158 		{

   159 		0x2021,

   160 		0x87

   161 		},

   162 		{

   163 		0x2022,

   164 		0x95

   165 		},

   166 		{

   167 		0x2026,

   168 		0x85

   169 		},

   170 		{

   171 		0x2030,

   172 		0x89

   173 		},

   174 		{

   175 		0x2039,

   176 		0x8b

   177 		},

   178 		{

   179 		0x203a,

   180 		0x9b

   181 		},

   182 		{

   183 		0x20ac,

   184 		0x80

   185 		},

   186 		{

   187 		0x2122,

   188 		0x99

   189 		}

   190 	};

   192 LOCAL_D const SCnvConversionData::SVariableByteData::SRange codePage1252VariableByteDataRanges[]=

   193 	{

   194 		{

   195 		0x00,

   196 		0xff,

   197 		0,

   198 		0

   199 		}

   200 	};

   202 LOCAL_D const SCnvConversionData::SOneDirectionData::SRange codePage1252ToUnicodeDataRanges[]=

   203 	{

   204 		{

   205 		0x00,

   206 		0x7f,

   207 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

   208 		0,

   209 		0,

   210 			{

   211 			0,

   212 			0

   213 			}

   214 		},

   215 		{

   216 		0xa0,

   217 		0xff,

   218 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

   219 		0,

   220 		0,

   221 			{

   222 			0,

   223 			0

   224 			}

   225 		},

   226 		{

   227 		0x80,

   228 		0x80,

   229 		SCnvConversionData::SOneDirectionData::SRange::EOffset,

   230 		0,

   231 		0,

   232 			{

   233 			STATIC_CAST(TUint, 8236),

   234 			0

   235 			}

   236 		},

   237 		{

   238 		0x8e,

   239 		0x8e,

   240 		SCnvConversionData::SOneDirectionData::SRange::EOffset,

   241 		0,

   242 		0,

   243 			{

   244 			STATIC_CAST(TUint, 239),

   245 			0

   246 			}

   247 		},

   248 		{

   249 		0x82,

   250 		0x9f,

   251 		SCnvConversionData::SOneDirectionData::SRange::EKeyedTable16OfIndexedTables16,

   252 		0,

   253 		0,

   254 			{

   255 			UData_SKeyedTable16OfIndexedTables16(keyedTables16OfIndexedTables16_keyedEntries_codePage1252ToUnicode_1)

   256 			}

   257 		}

   258 	};

   260 LOCAL_D const SCnvConversionData::SOneDirectionData::SRange unicodeToCodePage1252DataRanges[]=

   261 	{

   262 		{

   263 		0x0000,

   264 		0x007f,

   265 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

   266 		1,

   267 		0,

   268 			{

   269 			0,

   270 			0

   271 			}

   272 		},

   273 		{

   274 		0x00a0,

   275 		0x00ff,

   276 		SCnvConversionData::SOneDirectionData::SRange::EDirect,

   277 		1,

   278 		0,

   279 			{

   280 			0,

   281 			0

   282 			}

   283 		},

   284 		{

   285 		0x0152,

   286 		0x2122,

   287 		SCnvConversionData::SOneDirectionData::SRange::EKeyedTable1616,

   288 		1,

   289 		0,

   290 			{

   291 			UData_SKeyedTable1616(keyedTable1616_unicodeToCodePage1252_1)

   292 			}

   293 		}

   294 	};

   296 GLREF_D const SCnvConversionData codePage1252ConversionData=

   297 	{

   298 	SCnvConversionData::EUnspecified,

   299 		{

   300 		ARRAY_LENGTH(codePage1252VariableByteDataRanges),

   301 		codePage1252VariableByteDataRanges

   302 		},

   303 		{

   304 		ARRAY_LENGTH(codePage1252ToUnicodeDataRanges),

   305 		codePage1252ToUnicodeDataRanges

   306 		},

   307 		{

   308 		ARRAY_LENGTH(unicodeToCodePage1252DataRanges),

   309 		unicodeToCodePage1252DataRanges

   310 		},

   311 	NULL,

   312 	NULL

   313 	};

   315 GLREF_C void IsCharacterSetCP1252(TInt& aConfidenceLevel, const TDesC8& aSample)

   316 	{

   317 	aConfidenceLevel = 60;

   318 	TInt sampleLength = aSample.Length();

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

   321 		{

   322 		// CP1252 includes ASCII as well

   323 		// first check if the char is in the range 0x80 - 0x9f (controls codes in ISO88591)

   324 		// If it is in that range then the likelihood that it's CP1252 is a bit higher

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

   326 			{

   327 			if((aSample[i]==0x81)||(aSample[i]==0x8D)||(aSample[i]==0x8f)||

   328 				(aSample[i]==0x90)||(aSample[i]==0x9d))

   329 				{

   330 				// These code values are not supported by the Codepage CP1252

   331 				aConfidenceLevel = 0;

   332 				break;

   333 				}

   334 			else

   335 				{

   336 				// problem: UTF8 uses the values 0x80-0x9f in more than 50% of it's multibyte representation

   337 				// so if the text was UTF8 .... the confidence here would hit the roof. Could check to make

   338 				// sure that this is not UTF8

   339 				aConfidenceLevel+=1;

   340 				}

   341 			}

   342 		TInt increment1 = i+1;

   343 		TInt decrement1 = i-1;

   344 		// 0xf7 is the division symbol in CP1252.

   345 		// 0xd7 is the division symbol in CP1252.If char on either side of the division

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

   347 		if( decrement1>= 0 && ((aSample[i]==0xf7) || (aSample[i]==0xd7)) && increment1<sampleLength)

   348 			{

   350 			if (increment1 >= sampleLength)

   351 				break;

   352 			if ( (aSample[decrement1] >= 0x30) && (aSample[decrement1] <= 0x39) &&  // char before is a number

   353 				 (aSample[increment1] >= 0x30) && (aSample[increment1] <= 0x39) )   // char after is a number

   354 				{

   355 				aConfidenceLevel+=5;

   356 				}

   357 			}

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

   359 		// currency symbol is numeric

   360 		if((aSample[i]>=0xa2) && (aSample[i] <= 0xa5) && increment1<sampleLength)

   361 			{

   362 			if ((aSample[increment1] >= 0x30) && (aSample[increment1] <= 0x39))

   363 				{

   364 				aConfidenceLevel+=5;

   365 				}

   366 			}

   367 		} // for loop

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

   369 	}

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