1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/csrmbcs.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,529 @@
1.4 +/*
1.5 + **********************************************************************
1.6 + * Copyright (C) 2005-2012, International Business Machines
1.7 + * Corporation and others. All Rights Reserved.
1.8 + **********************************************************************
1.9 + */
1.10 +
1.11 +#include "unicode/utypes.h"
1.12 +
1.13 +#if !UCONFIG_NO_CONVERSION
1.14 +
1.15 +#include "csmatch.h"
1.16 +#include "csrmbcs.h"
1.17 +
1.18 +#include <math.h>
1.19 +
1.20 +U_NAMESPACE_BEGIN
1.21 +
1.22 +#define ARRAY_SIZE(array) (sizeof array / sizeof array[0])
1.23 +
1.24 +#define min(x,y) (((x)<(y))?(x):(y))
1.25 +
1.26 +static const uint16_t commonChars_sjis [] = {
1.27 +// TODO: This set of data comes from the character frequency-
1.28 +// of-occurence analysis tool. The data needs to be moved
1.29 +// into a resource and loaded from there.
1.30 +0x8140, 0x8141, 0x8142, 0x8145, 0x815b, 0x8169, 0x816a, 0x8175, 0x8176, 0x82a0,
1.31 +0x82a2, 0x82a4, 0x82a9, 0x82aa, 0x82ab, 0x82ad, 0x82af, 0x82b1, 0x82b3, 0x82b5,
1.32 +0x82b7, 0x82bd, 0x82be, 0x82c1, 0x82c4, 0x82c5, 0x82c6, 0x82c8, 0x82c9, 0x82cc,
1.33 +0x82cd, 0x82dc, 0x82e0, 0x82e7, 0x82e8, 0x82e9, 0x82ea, 0x82f0, 0x82f1, 0x8341,
1.34 +0x8343, 0x834e, 0x834f, 0x8358, 0x835e, 0x8362, 0x8367, 0x8375, 0x8376, 0x8389,
1.35 +0x838a, 0x838b, 0x838d, 0x8393, 0x8e96, 0x93fa, 0x95aa};
1.36 +
1.37 +static const uint16_t commonChars_euc_jp[] = {
1.38 +// TODO: This set of data comes from the character frequency-
1.39 +// of-occurence analysis tool. The data needs to be moved
1.40 +// into a resource and loaded from there.
1.41 +0xa1a1, 0xa1a2, 0xa1a3, 0xa1a6, 0xa1bc, 0xa1ca, 0xa1cb, 0xa1d6, 0xa1d7, 0xa4a2,
1.42 +0xa4a4, 0xa4a6, 0xa4a8, 0xa4aa, 0xa4ab, 0xa4ac, 0xa4ad, 0xa4af, 0xa4b1, 0xa4b3,
1.43 +0xa4b5, 0xa4b7, 0xa4b9, 0xa4bb, 0xa4bd, 0xa4bf, 0xa4c0, 0xa4c1, 0xa4c3, 0xa4c4,
1.44 +0xa4c6, 0xa4c7, 0xa4c8, 0xa4c9, 0xa4ca, 0xa4cb, 0xa4ce, 0xa4cf, 0xa4d0, 0xa4de,
1.45 +0xa4df, 0xa4e1, 0xa4e2, 0xa4e4, 0xa4e8, 0xa4e9, 0xa4ea, 0xa4eb, 0xa4ec, 0xa4ef,
1.46 +0xa4f2, 0xa4f3, 0xa5a2, 0xa5a3, 0xa5a4, 0xa5a6, 0xa5a7, 0xa5aa, 0xa5ad, 0xa5af,
1.47 +0xa5b0, 0xa5b3, 0xa5b5, 0xa5b7, 0xa5b8, 0xa5b9, 0xa5bf, 0xa5c3, 0xa5c6, 0xa5c7,
1.48 +0xa5c8, 0xa5c9, 0xa5cb, 0xa5d0, 0xa5d5, 0xa5d6, 0xa5d7, 0xa5de, 0xa5e0, 0xa5e1,
1.49 +0xa5e5, 0xa5e9, 0xa5ea, 0xa5eb, 0xa5ec, 0xa5ed, 0xa5f3, 0xb8a9, 0xb9d4, 0xbaee,
1.50 +0xbbc8, 0xbef0, 0xbfb7, 0xc4ea, 0xc6fc, 0xc7bd, 0xcab8, 0xcaf3, 0xcbdc, 0xcdd1};
1.51 +
1.52 +static const uint16_t commonChars_euc_kr[] = {
1.53 +// TODO: This set of data comes from the character frequency-
1.54 +// of-occurence analysis tool. The data needs to be moved
1.55 +// into a resource and loaded from there.
1.56 +0xb0a1, 0xb0b3, 0xb0c5, 0xb0cd, 0xb0d4, 0xb0e6, 0xb0ed, 0xb0f8, 0xb0fa, 0xb0fc,
1.57 +0xb1b8, 0xb1b9, 0xb1c7, 0xb1d7, 0xb1e2, 0xb3aa, 0xb3bb, 0xb4c2, 0xb4cf, 0xb4d9,
1.58 +0xb4eb, 0xb5a5, 0xb5b5, 0xb5bf, 0xb5c7, 0xb5e9, 0xb6f3, 0xb7af, 0xb7c2, 0xb7ce,
1.59 +0xb8a6, 0xb8ae, 0xb8b6, 0xb8b8, 0xb8bb, 0xb8e9, 0xb9ab, 0xb9ae, 0xb9cc, 0xb9ce,
1.60 +0xb9fd, 0xbab8, 0xbace, 0xbad0, 0xbaf1, 0xbbe7, 0xbbf3, 0xbbfd, 0xbcad, 0xbcba,
1.61 +0xbcd2, 0xbcf6, 0xbdba, 0xbdc0, 0xbdc3, 0xbdc5, 0xbec6, 0xbec8, 0xbedf, 0xbeee,
1.62 +0xbef8, 0xbefa, 0xbfa1, 0xbfa9, 0xbfc0, 0xbfe4, 0xbfeb, 0xbfec, 0xbff8, 0xc0a7,
1.63 +0xc0af, 0xc0b8, 0xc0ba, 0xc0bb, 0xc0bd, 0xc0c7, 0xc0cc, 0xc0ce, 0xc0cf, 0xc0d6,
1.64 +0xc0da, 0xc0e5, 0xc0fb, 0xc0fc, 0xc1a4, 0xc1a6, 0xc1b6, 0xc1d6, 0xc1df, 0xc1f6,
1.65 +0xc1f8, 0xc4a1, 0xc5cd, 0xc6ae, 0xc7cf, 0xc7d1, 0xc7d2, 0xc7d8, 0xc7e5, 0xc8ad};
1.66 +
1.67 +static const uint16_t commonChars_big5[] = {
1.68 +// TODO: This set of data comes from the character frequency-
1.69 +// of-occurence analysis tool. The data needs to be moved
1.70 +// into a resource and loaded from there.
1.71 +0xa140, 0xa141, 0xa142, 0xa143, 0xa147, 0xa149, 0xa175, 0xa176, 0xa440, 0xa446,
1.72 +0xa447, 0xa448, 0xa451, 0xa454, 0xa457, 0xa464, 0xa46a, 0xa46c, 0xa477, 0xa4a3,
1.73 +0xa4a4, 0xa4a7, 0xa4c1, 0xa4ce, 0xa4d1, 0xa4df, 0xa4e8, 0xa4fd, 0xa540, 0xa548,
1.74 +0xa558, 0xa569, 0xa5cd, 0xa5e7, 0xa657, 0xa661, 0xa662, 0xa668, 0xa670, 0xa6a8,
1.75 +0xa6b3, 0xa6b9, 0xa6d3, 0xa6db, 0xa6e6, 0xa6f2, 0xa740, 0xa751, 0xa759, 0xa7da,
1.76 +0xa8a3, 0xa8a5, 0xa8ad, 0xa8d1, 0xa8d3, 0xa8e4, 0xa8fc, 0xa9c0, 0xa9d2, 0xa9f3,
1.77 +0xaa6b, 0xaaba, 0xaabe, 0xaacc, 0xaafc, 0xac47, 0xac4f, 0xacb0, 0xacd2, 0xad59,
1.78 +0xaec9, 0xafe0, 0xb0ea, 0xb16f, 0xb2b3, 0xb2c4, 0xb36f, 0xb44c, 0xb44e, 0xb54c,
1.79 +0xb5a5, 0xb5bd, 0xb5d0, 0xb5d8, 0xb671, 0xb7ed, 0xb867, 0xb944, 0xbad8, 0xbb44,
1.80 +0xbba1, 0xbdd1, 0xc2c4, 0xc3b9, 0xc440, 0xc45f};
1.81 +
1.82 +static const uint16_t commonChars_gb_18030[] = {
1.83 +// TODO: This set of data comes from the character frequency-
1.84 +// of-occurence analysis tool. The data needs to be moved
1.85 +// into a resource and loaded from there.
1.86 +0xa1a1, 0xa1a2, 0xa1a3, 0xa1a4, 0xa1b0, 0xa1b1, 0xa1f1, 0xa1f3, 0xa3a1, 0xa3ac,
1.87 +0xa3ba, 0xb1a8, 0xb1b8, 0xb1be, 0xb2bb, 0xb3c9, 0xb3f6, 0xb4f3, 0xb5bd, 0xb5c4,
1.88 +0xb5e3, 0xb6af, 0xb6d4, 0xb6e0, 0xb7a2, 0xb7a8, 0xb7bd, 0xb7d6, 0xb7dd, 0xb8b4,
1.89 +0xb8df, 0xb8f6, 0xb9ab, 0xb9c9, 0xb9d8, 0xb9fa, 0xb9fd, 0xbacd, 0xbba7, 0xbbd6,
1.90 +0xbbe1, 0xbbfa, 0xbcbc, 0xbcdb, 0xbcfe, 0xbdcc, 0xbecd, 0xbedd, 0xbfb4, 0xbfc6,
1.91 +0xbfc9, 0xc0b4, 0xc0ed, 0xc1cb, 0xc2db, 0xc3c7, 0xc4dc, 0xc4ea, 0xc5cc, 0xc6f7,
1.92 +0xc7f8, 0xc8ab, 0xc8cb, 0xc8d5, 0xc8e7, 0xc9cf, 0xc9fa, 0xcab1, 0xcab5, 0xcac7,
1.93 +0xcad0, 0xcad6, 0xcaf5, 0xcafd, 0xccec, 0xcdf8, 0xceaa, 0xcec4, 0xced2, 0xcee5,
1.94 +0xcfb5, 0xcfc2, 0xcfd6, 0xd0c2, 0xd0c5, 0xd0d0, 0xd0d4, 0xd1a7, 0xd2aa, 0xd2b2,
1.95 +0xd2b5, 0xd2bb, 0xd2d4, 0xd3c3, 0xd3d0, 0xd3fd, 0xd4c2, 0xd4da, 0xd5e2, 0xd6d0};
1.96 +
1.97 +static int32_t binarySearch(const uint16_t *array, int32_t len, uint16_t value)
1.98 +{
1.99 + int32_t start = 0, end = len-1;
1.100 + int32_t mid = (start+end)/2;
1.101 +
1.102 + while(start <= end) {
1.103 + if(array[mid] == value) {
1.104 + return mid;
1.105 + }
1.106 +
1.107 + if(array[mid] < value){
1.108 + start = mid+1;
1.109 + } else {
1.110 + end = mid-1;
1.111 + }
1.112 +
1.113 + mid = (start+end)/2;
1.114 + }
1.115 +
1.116 + return -1;
1.117 +}
1.118 +
1.119 +IteratedChar::IteratedChar() :
1.120 +charValue(0), index(-1), nextIndex(0), error(FALSE), done(FALSE)
1.121 +{
1.122 + // nothing else to do.
1.123 +}
1.124 +
1.125 +/*void IteratedChar::reset()
1.126 +{
1.127 + charValue = 0;
1.128 + index = -1;
1.129 + nextIndex = 0;
1.130 + error = FALSE;
1.131 + done = FALSE;
1.132 +}*/
1.133 +
1.134 +int32_t IteratedChar::nextByte(InputText *det)
1.135 +{
1.136 + if (nextIndex >= det->fRawLength) {
1.137 + done = TRUE;
1.138 +
1.139 + return -1;
1.140 + }
1.141 +
1.142 + return det->fRawInput[nextIndex++];
1.143 +}
1.144 +
1.145 +CharsetRecog_mbcs::~CharsetRecog_mbcs()
1.146 +{
1.147 + // nothing to do.
1.148 +}
1.149 +
1.150 +int32_t CharsetRecog_mbcs::match_mbcs(InputText *det, const uint16_t commonChars[], int32_t commonCharsLen) const {
1.151 + int32_t singleByteCharCount = 0;
1.152 + int32_t doubleByteCharCount = 0;
1.153 + int32_t commonCharCount = 0;
1.154 + int32_t badCharCount = 0;
1.155 + int32_t totalCharCount = 0;
1.156 + int32_t confidence = 0;
1.157 + IteratedChar iter;
1.158 +
1.159 + while (nextChar(&iter, det)) {
1.160 + totalCharCount++;
1.161 +
1.162 + if (iter.error) {
1.163 + badCharCount++;
1.164 + } else {
1.165 + if (iter.charValue <= 0xFF) {
1.166 + singleByteCharCount++;
1.167 + } else {
1.168 + doubleByteCharCount++;
1.169 +
1.170 + if (commonChars != 0) {
1.171 + if (binarySearch(commonChars, commonCharsLen, iter.charValue) >= 0){
1.172 + commonCharCount += 1;
1.173 + }
1.174 + }
1.175 + }
1.176 + }
1.177 +
1.178 +
1.179 + if (badCharCount >= 2 && badCharCount*5 >= doubleByteCharCount) {
1.180 + // Bail out early if the byte data is not matching the encoding scheme.
1.181 + // break detectBlock;
1.182 + return confidence;
1.183 + }
1.184 + }
1.185 +
1.186 + if (doubleByteCharCount <= 10 && badCharCount == 0) {
1.187 + // Not many multi-byte chars.
1.188 + if (doubleByteCharCount == 0 && totalCharCount < 10) {
1.189 + // There weren't any multibyte sequences, and there was a low density of non-ASCII single bytes.
1.190 + // We don't have enough data to have any confidence.
1.191 + // Statistical analysis of single byte non-ASCII charcters would probably help here.
1.192 + confidence = 0;
1.193 + }
1.194 + else {
1.195 + // ASCII or ISO file? It's probably not our encoding,
1.196 + // but is not incompatible with our encoding, so don't give it a zero.
1.197 + confidence = 10;
1.198 + }
1.199 +
1.200 + return confidence;
1.201 + }
1.202 +
1.203 + //
1.204 + // No match if there are too many characters that don't fit the encoding scheme.
1.205 + // (should we have zero tolerance for these?)
1.206 + //
1.207 + if (doubleByteCharCount < 20*badCharCount) {
1.208 + confidence = 0;
1.209 +
1.210 + return confidence;
1.211 + }
1.212 +
1.213 + if (commonChars == 0) {
1.214 + // We have no statistics on frequently occuring characters.
1.215 + // Assess confidence purely on having a reasonable number of
1.216 + // multi-byte characters (the more the better)
1.217 + confidence = 30 + doubleByteCharCount - 20*badCharCount;
1.218 +
1.219 + if (confidence > 100) {
1.220 + confidence = 100;
1.221 + }
1.222 + } else {
1.223 + //
1.224 + // Frequency of occurence statistics exist.
1.225 + //
1.226 +
1.227 + double maxVal = log((double)doubleByteCharCount / 4); /*(float)?*/
1.228 + double scaleFactor = 90.0 / maxVal;
1.229 + confidence = (int32_t)(log((double)commonCharCount+1) * scaleFactor + 10.0);
1.230 +
1.231 + confidence = min(confidence, 100);
1.232 + }
1.233 +
1.234 + if (confidence < 0) {
1.235 + confidence = 0;
1.236 + }
1.237 +
1.238 + return confidence;
1.239 +}
1.240 +
1.241 +CharsetRecog_sjis::~CharsetRecog_sjis()
1.242 +{
1.243 + // nothing to do
1.244 +}
1.245 +
1.246 +UBool CharsetRecog_sjis::nextChar(IteratedChar* it, InputText* det) const {
1.247 + it->index = it->nextIndex;
1.248 + it->error = FALSE;
1.249 +
1.250 + int32_t firstByte = it->charValue = it->nextByte(det);
1.251 +
1.252 + if (firstByte < 0) {
1.253 + return FALSE;
1.254 + }
1.255 +
1.256 + if (firstByte <= 0x7F || (firstByte > 0xA0 && firstByte <= 0xDF)) {
1.257 + return TRUE;
1.258 + }
1.259 +
1.260 + int32_t secondByte = it->nextByte(det);
1.261 + if (secondByte >= 0) {
1.262 + it->charValue = (firstByte << 8) | secondByte;
1.263 + }
1.264 + // else we'll handle the error later.
1.265 +
1.266 + if (! ((secondByte >= 0x40 && secondByte <= 0x7F) || (secondByte >= 0x80 && secondByte <= 0xFE))) {
1.267 + // Illegal second byte value.
1.268 + it->error = TRUE;
1.269 + }
1.270 +
1.271 + return TRUE;
1.272 +}
1.273 +
1.274 +UBool CharsetRecog_sjis::match(InputText* det, CharsetMatch *results) const {
1.275 + int32_t confidence = match_mbcs(det, commonChars_sjis, ARRAY_SIZE(commonChars_sjis));
1.276 + results->set(det, this, confidence);
1.277 + return (confidence > 0);
1.278 +}
1.279 +
1.280 +const char *CharsetRecog_sjis::getName() const
1.281 +{
1.282 + return "Shift_JIS";
1.283 +}
1.284 +
1.285 +const char *CharsetRecog_sjis::getLanguage() const
1.286 +{
1.287 + return "ja";
1.288 +}
1.289 +
1.290 +CharsetRecog_euc::~CharsetRecog_euc()
1.291 +{
1.292 + // nothing to do
1.293 +}
1.294 +
1.295 +UBool CharsetRecog_euc::nextChar(IteratedChar* it, InputText* det) const {
1.296 + int32_t firstByte = 0;
1.297 + int32_t secondByte = 0;
1.298 + int32_t thirdByte = 0;
1.299 +
1.300 + it->index = it->nextIndex;
1.301 + it->error = FALSE;
1.302 + firstByte = it->charValue = it->nextByte(det);
1.303 +
1.304 + if (firstByte < 0) {
1.305 + // Ran off the end of the input data
1.306 + return FALSE;
1.307 + }
1.308 +
1.309 + if (firstByte <= 0x8D) {
1.310 + // single byte char
1.311 + return TRUE;
1.312 + }
1.313 +
1.314 + secondByte = it->nextByte(det);
1.315 + if (secondByte >= 0) {
1.316 + it->charValue = (it->charValue << 8) | secondByte;
1.317 + }
1.318 + // else we'll handle the error later.
1.319 +
1.320 + if (firstByte >= 0xA1 && firstByte <= 0xFE) {
1.321 + // Two byte Char
1.322 + if (secondByte < 0xA1) {
1.323 + it->error = TRUE;
1.324 + }
1.325 +
1.326 + return TRUE;
1.327 + }
1.328 +
1.329 + if (firstByte == 0x8E) {
1.330 + // Code Set 2.
1.331 + // In EUC-JP, total char size is 2 bytes, only one byte of actual char value.
1.332 + // In EUC-TW, total char size is 4 bytes, three bytes contribute to char value.
1.333 + // We don't know which we've got.
1.334 + // Treat it like EUC-JP. If the data really was EUC-TW, the following two
1.335 + // bytes will look like a well formed 2 byte char.
1.336 + if (secondByte < 0xA1) {
1.337 + it->error = TRUE;
1.338 + }
1.339 +
1.340 + return TRUE;
1.341 + }
1.342 +
1.343 + if (firstByte == 0x8F) {
1.344 + // Code set 3.
1.345 + // Three byte total char size, two bytes of actual char value.
1.346 + thirdByte = it->nextByte(det);
1.347 + it->charValue = (it->charValue << 8) | thirdByte;
1.348 +
1.349 + if (thirdByte < 0xa1) {
1.350 + // Bad second byte or ran off the end of the input data with a non-ASCII first byte.
1.351 + it->error = TRUE;
1.352 + }
1.353 + }
1.354 +
1.355 + return TRUE;
1.356 +
1.357 +}
1.358 +
1.359 +CharsetRecog_euc_jp::~CharsetRecog_euc_jp()
1.360 +{
1.361 + // nothing to do
1.362 +}
1.363 +
1.364 +const char *CharsetRecog_euc_jp::getName() const
1.365 +{
1.366 + return "EUC-JP";
1.367 +}
1.368 +
1.369 +const char *CharsetRecog_euc_jp::getLanguage() const
1.370 +{
1.371 + return "ja";
1.372 +}
1.373 +
1.374 +UBool CharsetRecog_euc_jp::match(InputText *det, CharsetMatch *results) const
1.375 +{
1.376 + int32_t confidence = match_mbcs(det, commonChars_euc_jp, ARRAY_SIZE(commonChars_euc_jp));
1.377 + results->set(det, this, confidence);
1.378 + return (confidence > 0);
1.379 +}
1.380 +
1.381 +CharsetRecog_euc_kr::~CharsetRecog_euc_kr()
1.382 +{
1.383 + // nothing to do
1.384 +}
1.385 +
1.386 +const char *CharsetRecog_euc_kr::getName() const
1.387 +{
1.388 + return "EUC-KR";
1.389 +}
1.390 +
1.391 +const char *CharsetRecog_euc_kr::getLanguage() const
1.392 +{
1.393 + return "ko";
1.394 +}
1.395 +
1.396 +UBool CharsetRecog_euc_kr::match(InputText *det, CharsetMatch *results) const
1.397 +{
1.398 + int32_t confidence = match_mbcs(det, commonChars_euc_kr, ARRAY_SIZE(commonChars_euc_kr));
1.399 + results->set(det, this, confidence);
1.400 + return (confidence > 0);
1.401 +}
1.402 +
1.403 +CharsetRecog_big5::~CharsetRecog_big5()
1.404 +{
1.405 + // nothing to do
1.406 +}
1.407 +
1.408 +UBool CharsetRecog_big5::nextChar(IteratedChar* it, InputText* det) const
1.409 +{
1.410 + int32_t firstByte;
1.411 +
1.412 + it->index = it->nextIndex;
1.413 + it->error = FALSE;
1.414 + firstByte = it->charValue = it->nextByte(det);
1.415 +
1.416 + if (firstByte < 0) {
1.417 + return FALSE;
1.418 + }
1.419 +
1.420 + if (firstByte <= 0x7F || firstByte == 0xFF) {
1.421 + // single byte character.
1.422 + return TRUE;
1.423 + }
1.424 +
1.425 + int32_t secondByte = it->nextByte(det);
1.426 + if (secondByte >= 0) {
1.427 + it->charValue = (it->charValue << 8) | secondByte;
1.428 + }
1.429 + // else we'll handle the error later.
1.430 +
1.431 + if (secondByte < 0x40 || secondByte == 0x7F || secondByte == 0xFF) {
1.432 + it->error = TRUE;
1.433 + }
1.434 +
1.435 + return TRUE;
1.436 +}
1.437 +
1.438 +const char *CharsetRecog_big5::getName() const
1.439 +{
1.440 + return "Big5";
1.441 +}
1.442 +
1.443 +const char *CharsetRecog_big5::getLanguage() const
1.444 +{
1.445 + return "zh";
1.446 +}
1.447 +
1.448 +UBool CharsetRecog_big5::match(InputText *det, CharsetMatch *results) const
1.449 +{
1.450 + int32_t confidence = match_mbcs(det, commonChars_big5, ARRAY_SIZE(commonChars_big5));
1.451 + results->set(det, this, confidence);
1.452 + return (confidence > 0);
1.453 +}
1.454 +
1.455 +CharsetRecog_gb_18030::~CharsetRecog_gb_18030()
1.456 +{
1.457 + // nothing to do
1.458 +}
1.459 +
1.460 +UBool CharsetRecog_gb_18030::nextChar(IteratedChar* it, InputText* det) const {
1.461 + int32_t firstByte = 0;
1.462 + int32_t secondByte = 0;
1.463 + int32_t thirdByte = 0;
1.464 + int32_t fourthByte = 0;
1.465 +
1.466 + it->index = it->nextIndex;
1.467 + it->error = FALSE;
1.468 + firstByte = it->charValue = it->nextByte(det);
1.469 +
1.470 + if (firstByte < 0) {
1.471 + // Ran off the end of the input data
1.472 + return FALSE;
1.473 + }
1.474 +
1.475 + if (firstByte <= 0x80) {
1.476 + // single byte char
1.477 + return TRUE;
1.478 + }
1.479 +
1.480 + secondByte = it->nextByte(det);
1.481 + if (secondByte >= 0) {
1.482 + it->charValue = (it->charValue << 8) | secondByte;
1.483 + }
1.484 + // else we'll handle the error later.
1.485 +
1.486 + if (firstByte >= 0x81 && firstByte <= 0xFE) {
1.487 + // Two byte Char
1.488 + if ((secondByte >= 0x40 && secondByte <= 0x7E) || (secondByte >=80 && secondByte <= 0xFE)) {
1.489 + return TRUE;
1.490 + }
1.491 +
1.492 + // Four byte char
1.493 + if (secondByte >= 0x30 && secondByte <= 0x39) {
1.494 + thirdByte = it->nextByte(det);
1.495 +
1.496 + if (thirdByte >= 0x81 && thirdByte <= 0xFE) {
1.497 + fourthByte = it->nextByte(det);
1.498 +
1.499 + if (fourthByte >= 0x30 && fourthByte <= 0x39) {
1.500 + it->charValue = (it->charValue << 16) | (thirdByte << 8) | fourthByte;
1.501 +
1.502 + return TRUE;
1.503 + }
1.504 + }
1.505 + }
1.506 +
1.507 + // Something wasn't valid, or we ran out of data (-1).
1.508 + it->error = TRUE;
1.509 + }
1.510 +
1.511 + return TRUE;
1.512 +}
1.513 +
1.514 +const char *CharsetRecog_gb_18030::getName() const
1.515 +{
1.516 + return "GB18030";
1.517 +}
1.518 +
1.519 +const char *CharsetRecog_gb_18030::getLanguage() const
1.520 +{
1.521 + return "zh";
1.522 +}
1.523 +
1.524 +UBool CharsetRecog_gb_18030::match(InputText *det, CharsetMatch *results) const
1.525 +{
1.526 + int32_t confidence = match_mbcs(det, commonChars_gb_18030, ARRAY_SIZE(commonChars_gb_18030));
1.527 + results->set(det, this, confidence);
1.528 + return (confidence > 0);
1.529 +}
1.530 +
1.531 +U_NAMESPACE_END
1.532 +#endif
