intl/icu/source/i18n/ucsdet.cpp

Wed, 31 Dec 2014 06:09:35 +0100

author
Michael Schloh von Bennewitz <michael@schloh.com>
date
Wed, 31 Dec 2014 06:09:35 +0100
changeset 0
6474c204b198
permissions
-rw-r--r--

Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.

michael@0 1 /*
michael@0 2 ********************************************************************************
michael@0 3 * Copyright (C) 2005-2013, International Business Machines
michael@0 4 * Corporation and others. All Rights Reserved.
michael@0 5 ********************************************************************************
michael@0 6 */
michael@0 7
michael@0 8 #include "unicode/utypes.h"
michael@0 9
michael@0 10 #if !UCONFIG_NO_CONVERSION
michael@0 11 #include "unicode/ucsdet.h"
michael@0 12 #include "csdetect.h"
michael@0 13 #include "csmatch.h"
michael@0 14 #include "csrsbcs.h"
michael@0 15 #include "csrmbcs.h"
michael@0 16 #include "csrutf8.h"
michael@0 17 #include "csrucode.h"
michael@0 18 #include "csr2022.h"
michael@0 19
michael@0 20 #include "cmemory.h"
michael@0 21
michael@0 22 U_NAMESPACE_USE
michael@0 23
michael@0 24 #define ARRAY_SIZE(array) (sizeof array / sizeof array[0])
michael@0 25
michael@0 26 #define NEW_ARRAY(type,count) (type *) uprv_malloc((count) * sizeof(type))
michael@0 27 #define DELETE_ARRAY(array) uprv_free((void *) (array))
michael@0 28
michael@0 29 U_CDECL_BEGIN
michael@0 30
michael@0 31 U_CAPI UCharsetDetector * U_EXPORT2
michael@0 32 ucsdet_open(UErrorCode *status)
michael@0 33 {
michael@0 34 if(U_FAILURE(*status)) {
michael@0 35 return 0;
michael@0 36 }
michael@0 37
michael@0 38 CharsetDetector* csd = new CharsetDetector(*status);
michael@0 39
michael@0 40 if (U_FAILURE(*status)) {
michael@0 41 delete csd;
michael@0 42 csd = NULL;
michael@0 43 }
michael@0 44
michael@0 45 return (UCharsetDetector *) csd;
michael@0 46 }
michael@0 47
michael@0 48 U_CAPI void U_EXPORT2
michael@0 49 ucsdet_close(UCharsetDetector *ucsd)
michael@0 50 {
michael@0 51 CharsetDetector *csd = (CharsetDetector *) ucsd;
michael@0 52 delete csd;
michael@0 53 }
michael@0 54
michael@0 55 U_CAPI void U_EXPORT2
michael@0 56 ucsdet_setText(UCharsetDetector *ucsd, const char *textIn, int32_t len, UErrorCode *status)
michael@0 57 {
michael@0 58 if(U_FAILURE(*status)) {
michael@0 59 return;
michael@0 60 }
michael@0 61
michael@0 62 ((CharsetDetector *) ucsd)->setText(textIn, len);
michael@0 63 }
michael@0 64
michael@0 65 U_CAPI const char * U_EXPORT2
michael@0 66 ucsdet_getName(const UCharsetMatch *ucsm, UErrorCode *status)
michael@0 67 {
michael@0 68 if(U_FAILURE(*status)) {
michael@0 69 return NULL;
michael@0 70 }
michael@0 71
michael@0 72 return ((CharsetMatch *) ucsm)->getName();
michael@0 73 }
michael@0 74
michael@0 75 U_CAPI int32_t U_EXPORT2
michael@0 76 ucsdet_getConfidence(const UCharsetMatch *ucsm, UErrorCode *status)
michael@0 77 {
michael@0 78 if(U_FAILURE(*status)) {
michael@0 79 return 0;
michael@0 80 }
michael@0 81
michael@0 82 return ((CharsetMatch *) ucsm)->getConfidence();
michael@0 83 }
michael@0 84
michael@0 85 U_CAPI const char * U_EXPORT2
michael@0 86 ucsdet_getLanguage(const UCharsetMatch *ucsm, UErrorCode *status)
michael@0 87 {
michael@0 88 if(U_FAILURE(*status)) {
michael@0 89 return NULL;
michael@0 90 }
michael@0 91
michael@0 92 return ((CharsetMatch *) ucsm)->getLanguage();
michael@0 93 }
michael@0 94
michael@0 95 U_CAPI const UCharsetMatch * U_EXPORT2
michael@0 96 ucsdet_detect(UCharsetDetector *ucsd, UErrorCode *status)
michael@0 97 {
michael@0 98 if(U_FAILURE(*status)) {
michael@0 99 return NULL;
michael@0 100 }
michael@0 101
michael@0 102 return (const UCharsetMatch *) ((CharsetDetector *) ucsd)->detect(*status);
michael@0 103 }
michael@0 104
michael@0 105 U_CAPI void U_EXPORT2
michael@0 106 ucsdet_setDeclaredEncoding(UCharsetDetector *ucsd, const char *encoding, int32_t length, UErrorCode *status)
michael@0 107 {
michael@0 108 if(U_FAILURE(*status)) {
michael@0 109 return;
michael@0 110 }
michael@0 111
michael@0 112 ((CharsetDetector *) ucsd)->setDeclaredEncoding(encoding,length);
michael@0 113 }
michael@0 114
michael@0 115 U_CAPI const UCharsetMatch**
michael@0 116 ucsdet_detectAll(UCharsetDetector *ucsd,
michael@0 117 int32_t *maxMatchesFound, UErrorCode *status)
michael@0 118 {
michael@0 119 if(U_FAILURE(*status)) {
michael@0 120 return NULL;
michael@0 121 }
michael@0 122
michael@0 123 CharsetDetector *csd = (CharsetDetector *) ucsd;
michael@0 124
michael@0 125 return (const UCharsetMatch**)csd->detectAll(*maxMatchesFound,*status);
michael@0 126 }
michael@0 127
michael@0 128 // U_CAPI const char * U_EXPORT2
michael@0 129 // ucsdet_getDetectableCharsetName(const UCharsetDetector *csd, int32_t index, UErrorCode *status)
michael@0 130 // {
michael@0 131 // if(U_FAILURE(*status)) {
michael@0 132 // return 0;
michael@0 133 // }
michael@0 134 // return csd->getCharsetName(index,*status);
michael@0 135 // }
michael@0 136
michael@0 137 // U_CAPI int32_t U_EXPORT2
michael@0 138 // ucsdet_getDetectableCharsetsCount(const UCharsetDetector *csd, UErrorCode *status)
michael@0 139 // {
michael@0 140 // if(U_FAILURE(*status)) {
michael@0 141 // return -1;
michael@0 142 // }
michael@0 143 // return UCharsetDetector::getDetectableCount();
michael@0 144 // }
michael@0 145
michael@0 146 U_CAPI UBool U_EXPORT2
michael@0 147 ucsdet_isInputFilterEnabled(const UCharsetDetector *ucsd)
michael@0 148 {
michael@0 149 // todo: could use an error return...
michael@0 150 if (ucsd == NULL) {
michael@0 151 return FALSE;
michael@0 152 }
michael@0 153
michael@0 154 return ((CharsetDetector *) ucsd)->getStripTagsFlag();
michael@0 155 }
michael@0 156
michael@0 157 U_CAPI UBool U_EXPORT2
michael@0 158 ucsdet_enableInputFilter(UCharsetDetector *ucsd, UBool filter)
michael@0 159 {
michael@0 160 // todo: could use an error return...
michael@0 161 if (ucsd == NULL) {
michael@0 162 return FALSE;
michael@0 163 }
michael@0 164
michael@0 165 CharsetDetector *csd = (CharsetDetector *) ucsd;
michael@0 166 UBool prev = csd->getStripTagsFlag();
michael@0 167
michael@0 168 csd->setStripTagsFlag(filter);
michael@0 169
michael@0 170 return prev;
michael@0 171 }
michael@0 172
michael@0 173 U_CAPI int32_t U_EXPORT2
michael@0 174 ucsdet_getUChars(const UCharsetMatch *ucsm,
michael@0 175 UChar *buf, int32_t cap, UErrorCode *status)
michael@0 176 {
michael@0 177 if(U_FAILURE(*status)) {
michael@0 178 return 0;
michael@0 179 }
michael@0 180
michael@0 181 return ((CharsetMatch *) ucsm)->getUChars(buf, cap, status);
michael@0 182 }
michael@0 183
michael@0 184 U_CAPI void U_EXPORT2
michael@0 185 ucsdet_setDetectableCharset(UCharsetDetector *ucsd, const char *encoding, UBool enabled, UErrorCode *status)
michael@0 186 {
michael@0 187 ((CharsetDetector *)ucsd)->setDetectableCharset(encoding, enabled, *status);
michael@0 188 }
michael@0 189
michael@0 190 U_CAPI UEnumeration * U_EXPORT2
michael@0 191 ucsdet_getAllDetectableCharsets(const UCharsetDetector * /*ucsd*/, UErrorCode *status)
michael@0 192 {
michael@0 193 return CharsetDetector::getAllDetectableCharsets(*status);
michael@0 194 }
michael@0 195
michael@0 196 U_DRAFT UEnumeration * U_EXPORT2
michael@0 197 ucsdet_getDetectableCharsets(const UCharsetDetector *ucsd, UErrorCode *status)
michael@0 198 {
michael@0 199 return ((CharsetDetector *)ucsd)->getDetectableCharsets(*status);
michael@0 200 }
michael@0 201
michael@0 202 U_CDECL_END
michael@0 203
michael@0 204
michael@0 205 #endif

mercurial