1 // Copyright (C) 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
4 *******************************************************************************
6 * Copyright (C) 1999-2014, International Business Machines
7 * Corporation and others. All Rights Reserved.
9 *******************************************************************************
10 * file name: unistr_cnv.cpp
12 * tab size: 8 (not used)
15 * created on: 2004aug19
16 * created by: Markus W. Scherer
18 * Character conversion functions moved here from unistr.cpp
21 #include "unicode/utypes.h"
23 #if !UCONFIG_NO_CONVERSION
25 #include "unicode/putil.h"
28 #include "unicode/ustring.h"
29 #include "unicode/unistr.h"
30 #include "unicode/ucnv.h"
38 //========================================
40 //========================================
42 #if !U_CHARSET_IS_UTF8
44 UnicodeString::UnicodeString(const char *codepageData) {
45 fUnion.fFields.fLengthAndFlags = kShortString;
46 if(codepageData != 0) {
47 doCodepageCreate(codepageData, (int32_t)uprv_strlen(codepageData), 0);
51 UnicodeString::UnicodeString(const char *codepageData,
53 fUnion.fFields.fLengthAndFlags = kShortString;
54 if(codepageData != 0) {
55 doCodepageCreate(codepageData, dataLength, 0);
59 // else see unistr.cpp
62 UnicodeString::UnicodeString(const char *codepageData,
63 const char *codepage) {
64 fUnion.fFields.fLengthAndFlags = kShortString;
65 if(codepageData != 0) {
66 doCodepageCreate(codepageData, (int32_t)uprv_strlen(codepageData), codepage);
70 UnicodeString::UnicodeString(const char *codepageData,
72 const char *codepage) {
73 fUnion.fFields.fLengthAndFlags = kShortString;
74 if(codepageData != 0) {
75 doCodepageCreate(codepageData, dataLength, codepage);
79 UnicodeString::UnicodeString(const char *src, int32_t srcLength,
81 UErrorCode &errorCode) {
82 fUnion.fFields.fLengthAndFlags = kShortString;
83 if(U_SUCCESS(errorCode)) {
86 // treat as an empty string, do nothing more
87 } else if(srcLength<-1) {
88 errorCode=U_ILLEGAL_ARGUMENT_ERROR;
92 srcLength=(int32_t)uprv_strlen(src);
96 // use the provided converter
97 ucnv_resetToUnicode(cnv);
98 doCodepageCreate(src, srcLength, cnv, errorCode);
100 // use the default converter
101 cnv=u_getDefaultConverter(&errorCode);
102 doCodepageCreate(src, srcLength, cnv, errorCode);
103 u_releaseDefaultConverter(cnv);
108 if(U_FAILURE(errorCode)) {
114 //========================================
115 // Codeset conversion
116 //========================================
118 #if !U_CHARSET_IS_UTF8
121 UnicodeString::extract(int32_t start,
124 uint32_t dstSize) const {
125 return extract(start, length, target, dstSize, 0);
128 // else see unistr.cpp
132 UnicodeString::extract(int32_t start,
136 const char *codepage) const
138 // if the arguments are illegal, then do nothing
139 if(/*dstSize < 0 || */(dstSize > 0 && target == 0)) {
143 // pin the indices to legal values
144 pinIndices(start, length);
146 // We need to cast dstSize to int32_t for all subsequent code.
147 // I don't know why the API was defined with uint32_t but we are stuck with it.
148 // Also, dstSize==0xffffffff means "unlimited" but if we use target+dstSize
149 // as a limit in some functions, it may wrap around and yield a pointer
150 // that compares less-than target.
152 if(dstSize < 0x7fffffff) {
153 // Assume that the capacity is real and a limit pointer won't wrap around.
154 capacity = (int32_t)dstSize;
156 // Pin the capacity so that a limit pointer does not wrap around.
157 char *targetLimit = (char *)U_MAX_PTR(target);
158 // U_MAX_PTR(target) returns a targetLimit that is at most 0x7fffffff
159 // greater than target and does not wrap around the top of the address space.
160 capacity = (int32_t)(targetLimit - target);
163 // create the converter
164 UConverter *converter;
165 UErrorCode status = U_ZERO_ERROR;
167 // just write the NUL if the string length is 0
169 return u_terminateChars(target, capacity, 0, &status);
172 // if the codepage is the default, use our cache
173 // if it is an empty string, then use the "invariant character" conversion
175 const char *defaultName = ucnv_getDefaultName();
176 if(UCNV_FAST_IS_UTF8(defaultName)) {
177 return toUTF8(start, length, target, capacity);
179 converter = u_getDefaultConverter(&status);
180 } else if (*codepage == 0) {
181 // use the "invariant characters" conversion
183 if(length <= capacity) {
186 destLength = capacity;
188 u_UCharsToChars(getArrayStart() + start, target, destLength);
189 return u_terminateChars(target, capacity, length, &status);
191 converter = ucnv_open(codepage, &status);
194 length = doExtract(start, length, target, capacity, converter, status);
196 // close the converter
198 u_releaseDefaultConverter(converter);
200 ucnv_close(converter);
207 UnicodeString::extract(char *dest, int32_t destCapacity,
209 UErrorCode &errorCode) const
211 if(U_FAILURE(errorCode)) {
215 if(isBogus() || destCapacity<0 || (destCapacity>0 && dest==0)) {
216 errorCode=U_ILLEGAL_ARGUMENT_ERROR;
222 return u_terminateChars(dest, destCapacity, 0, &errorCode);
226 UBool isDefaultConverter;
228 isDefaultConverter=TRUE;
229 cnv=u_getDefaultConverter(&errorCode);
230 if(U_FAILURE(errorCode)) {
234 isDefaultConverter=FALSE;
235 ucnv_resetFromUnicode(cnv);
239 int32_t len=doExtract(0, length(), dest, destCapacity, cnv, errorCode);
241 // release the converter
242 if(isDefaultConverter) {
243 u_releaseDefaultConverter(cnv);
250 UnicodeString::doExtract(int32_t start, int32_t length,
251 char *dest, int32_t destCapacity,
253 UErrorCode &errorCode) const
255 if(U_FAILURE(errorCode)) {
256 if(destCapacity!=0) {
262 const UChar *src=getArrayStart()+start, *srcLimit=src+length;
263 char *originalDest=dest;
264 const char *destLimit;
266 if(destCapacity==0) {
268 } else if(destCapacity==-1) {
269 // Pin the limit to U_MAX_PTR if the "magic" destCapacity is used.
270 destLimit=(char*)U_MAX_PTR(dest);
271 // for NUL-termination, translate into highest int32_t
272 destCapacity=0x7fffffff;
274 destLimit=dest+destCapacity;
277 // perform the conversion
278 ucnv_fromUnicode(cnv, &dest, destLimit, &src, srcLimit, 0, TRUE, &errorCode);
279 length=(int32_t)(dest-originalDest);
281 // if an overflow occurs, then get the preflighting length
282 if(errorCode==U_BUFFER_OVERFLOW_ERROR) {
285 destLimit=buffer+sizeof(buffer);
288 errorCode=U_ZERO_ERROR;
289 ucnv_fromUnicode(cnv, &dest, destLimit, &src, srcLimit, 0, TRUE, &errorCode);
290 length+=(int32_t)(dest-buffer);
291 } while(errorCode==U_BUFFER_OVERFLOW_ERROR);
294 return u_terminateChars(originalDest, destCapacity, length, &errorCode);
298 UnicodeString::doCodepageCreate(const char *codepageData,
300 const char *codepage)
302 // if there's nothing to convert, do nothing
303 if(codepageData == 0 || dataLength == 0 || dataLength < -1) {
306 if(dataLength == -1) {
307 dataLength = (int32_t)uprv_strlen(codepageData);
310 UErrorCode status = U_ZERO_ERROR;
312 // create the converter
313 // if the codepage is the default, use our cache
314 // if it is an empty string, then use the "invariant character" conversion
315 UConverter *converter;
317 const char *defaultName = ucnv_getDefaultName();
318 if(UCNV_FAST_IS_UTF8(defaultName)) {
319 setToUTF8(StringPiece(codepageData, dataLength));
322 converter = u_getDefaultConverter(&status);
323 } else if(*codepage == 0) {
324 // use the "invariant characters" conversion
325 if(cloneArrayIfNeeded(dataLength, dataLength, FALSE)) {
326 u_charsToUChars(codepageData, getArrayStart(), dataLength);
327 setLength(dataLength);
333 converter = ucnv_open(codepage, &status);
336 // if we failed, set the appropriate flags and return
337 if(U_FAILURE(status)) {
342 // perform the conversion
343 doCodepageCreate(codepageData, dataLength, converter, status);
344 if(U_FAILURE(status)) {
348 // close the converter
350 u_releaseDefaultConverter(converter);
352 ucnv_close(converter);
357 UnicodeString::doCodepageCreate(const char *codepageData,
359 UConverter *converter,
362 if(U_FAILURE(status)) {
366 // set up the conversion parameters
367 const char *mySource = codepageData;
368 const char *mySourceEnd = mySource + dataLength;
369 UChar *array, *myTarget;
371 // estimate the size needed:
373 if(dataLength <= US_STACKBUF_SIZE) {
374 // try to use the stack buffer
375 arraySize = US_STACKBUF_SIZE;
377 // 1.25 UChar's per source byte should cover most cases
378 arraySize = dataLength + (dataLength >> 2);
381 // we do not care about the current contents
382 UBool doCopyArray = FALSE;
384 if(!cloneArrayIfNeeded(arraySize, arraySize, doCopyArray)) {
389 // perform the conversion
390 array = getArrayStart();
391 myTarget = array + length();
392 ucnv_toUnicode(converter, &myTarget, array + getCapacity(),
393 &mySource, mySourceEnd, 0, TRUE, &status);
395 // update the conversion parameters
396 setLength((int32_t)(myTarget - array));
398 // allocate more space and copy data, if needed
399 if(status == U_BUFFER_OVERFLOW_ERROR) {
400 // reset the error code
401 status = U_ZERO_ERROR;
403 // keep the previous conversion results
406 // estimate the new size needed, larger than before
407 // try 2 UChar's per remaining source byte
408 arraySize = (int32_t)(length() + 2 * (mySourceEnd - mySource));