X-Git-Url: https://git.saurik.com/apple/icu.git/blobdiff_plain/46f4442e9a5a4f3b98b7c1083586332f6a8a99a4..b331163bffd790ced0e88b73f44f86d49ccc48a5:/icuSources/common/utext.cpp?ds=sidebyside diff --git a/icuSources/common/utext.cpp b/icuSources/common/utext.cpp index 40266ffb..6dc31a04 100644 --- a/icuSources/common/utext.cpp +++ b/icuSources/common/utext.cpp @@ -1,7 +1,7 @@ /* ******************************************************************************* * -* Copyright (C) 2005-2009, International Business Machines +* Copyright (C) 2005-2014, International Business Machines * Corporation and others. All Rights Reserved. * ******************************************************************************* @@ -19,10 +19,14 @@ #include "unicode/unistr.h" #include "unicode/chariter.h" #include "unicode/utext.h" +#include "unicode/utf.h" +#include "unicode/utf8.h" +#include "unicode/utf16.h" #include "ustr_imp.h" #include "cmemory.h" #include "cstring.h" #include "uassert.h" +#include "putilimp.h" U_NAMESPACE_USE @@ -117,13 +121,13 @@ utext_setNativeIndex(UText *ut, int64_t index) { // Adjust the index position if it is in the middle of a surrogate pair. if (ut->chunkOffsetchunkLength) { UChar c= ut->chunkContents[ut->chunkOffset]; - if (UTF16_IS_TRAIL(c)) { + if (U16_IS_TRAIL(c)) { if (ut->chunkOffset==0) { ut->pFuncs->access(ut, ut->chunkNativeStart, FALSE); } if (ut->chunkOffset>0) { UChar lead = ut->chunkContents[ut->chunkOffset-1]; - if (UTF16_IS_LEAD(lead)) { + if (U16_IS_LEAD(lead)) { ut->chunkOffset--; } } @@ -512,8 +516,17 @@ utext_copy(UText *ut, U_CAPI UText * U_EXPORT2 utext_clone(UText *dest, const UText *src, UBool deep, UBool readOnly, UErrorCode *status) { - UText *result; - result = src->pFuncs->clone(dest, src, deep, status); + if (U_FAILURE(*status)) { + return dest; + } + UText *result = src->pFuncs->clone(dest, src, deep, status); + if (U_FAILURE(*status)) { + return result; + } + if (result == NULL) { + *status = U_MEMORY_ALLOCATION_ERROR; + return result; + } if (readOnly) { utext_freeze(result); } @@ -800,7 +813,12 @@ shallowTextClone(UText * dest, const UText * src, UErrorCode * status) { adjustPointer(dest, &dest->p, src); adjustPointer(dest, &dest->q, src); adjustPointer(dest, &dest->r, src); - adjustPointer(dest, (const void **)&dest->chunkContents, src); + adjustPointer(dest, (const void **)&dest->chunkContents, src); + + // The newly shallow-cloned UText does _not_ own the underlying storage for the text. + // (The source for the clone may or may not have owned the text.) + + dest->providerProperties &= ~I32_FLAG(UTEXT_PROVIDER_OWNS_TEXT); return dest; } @@ -932,7 +950,7 @@ utf8TextAccess(UText *ut, int64_t index, UBool forward) { if (ix>length) { if (length>=0) { ix=length; - } else if (ix>ut->c) { + } else if (ix>=ut->c) { // Zero terminated string, and requested index is beyond // the region that has already been scanned. // Scan up to either the end of the string or to the @@ -1190,7 +1208,7 @@ fillForward: int32_t destIx = 0; int32_t srcIx = ix; UBool seenNonAscii = FALSE; - UChar32 c; + UChar32 c = 0; // Fill the chunk buffer and mapping arrays. while (destIx0 && c<0x80) { // Special case ASCII range for speed. // zero is excluded to simplify bounds checking. - buf[destIx] = c; - mapToNative[destIx] = srcIx - ix; - mapToUChars[srcIx-ix] = destIx; + buf[destIx] = (UChar)c; + mapToNative[destIx] = (uint8_t)(srcIx - ix); + mapToUChars[srcIx-ix] = (uint8_t)destIx; srcIx++; destIx++; } else { @@ -1213,23 +1231,19 @@ fillForward: int32_t cIx = srcIx; int32_t dIx = destIx; int32_t dIxSaved = destIx; - U8_NEXT(s8, srcIx, strLen, c); + U8_NEXT_OR_FFFD(s8, srcIx, strLen, c); if (c==0 && nulTerminated) { srcIx--; break; } - if (c<0) { - // Illegal UTF-8. Replace with sub character. - c = 0x0fffd; - } U16_APPEND_UNSAFE(buf, destIx, c); do { - mapToNative[dIx++] = cIx - ix; + mapToNative[dIx++] = (uint8_t)(cIx - ix); } while (dIx < destIx); do { - mapToUChars[cIx++ - ix] = dIxSaved; + mapToUChars[cIx++ - ix] = (uint8_t)dIxSaved; } while (cIx < srcIx); } if (srcIx>=strLen) { @@ -1240,8 +1254,8 @@ fillForward: // store Native <--> Chunk Map entries for the end of the buffer. // There is no actual character here, but the index position is valid. - mapToNative[destIx] = srcIx - ix; - mapToUChars[srcIx - ix] = destIx; + mapToNative[destIx] = (uint8_t)(srcIx - ix); + mapToUChars[srcIx - ix] = (uint8_t)destIx; // fill in Buffer descriptor u8b->bufNativeStart = ix; @@ -1307,8 +1321,8 @@ fillReverse: // Map to/from Native Indexes, fill in for the position at the end of // the buffer. // - mapToNative[destIx] = srcIx - toUCharsMapStart; - mapToUChars[srcIx - toUCharsMapStart] = destIx; + mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart); + mapToUChars[srcIx - toUCharsMapStart] = (uint8_t)destIx; // Fill the chunk buffer // Work backwards, filling from the end of the buffer towards the front. @@ -1321,39 +1335,35 @@ fillReverse: c = s8[srcIx]; if (c<0x80) { // Special case ASCII range for speed. - buf[destIx] = c; - mapToUChars[srcIx - toUCharsMapStart] = destIx; - mapToNative[destIx] = srcIx - toUCharsMapStart; + buf[destIx] = (UChar)c; + mapToUChars[srcIx - toUCharsMapStart] = (uint8_t)destIx; + mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart); } else { // General case, handle everything non-ASCII. int32_t sIx = srcIx; // ix of last byte of multi-byte u8 char // Get the full character from the UTF8 string. - // use code derived from tbe macros in utf.8 + // use code derived from tbe macros in utf8.h // Leaves srcIx pointing at the first byte of the UTF-8 char. // - if (c<=0xbf) { - c=utf8_prevCharSafeBody(s8, 0, &srcIx, c, -1); - // leaves srcIx at first byte of the multi-byte char. - } else { - c=0x0fffd; - } + c=utf8_prevCharSafeBody(s8, 0, &srcIx, c, -3); + // leaves srcIx at first byte of the multi-byte char. // Store the character in UTF-16 buffer. if (c<0x10000) { - buf[destIx] = c; - mapToNative[destIx] = srcIx - toUCharsMapStart; + buf[destIx] = (UChar)c; + mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart); } else { buf[destIx] = U16_TRAIL(c); - mapToNative[destIx] = srcIx - toUCharsMapStart; + mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart); buf[--destIx] = U16_LEAD(c); - mapToNative[destIx] = srcIx - toUCharsMapStart; + mapToNative[destIx] = (uint8_t)(srcIx - toUCharsMapStart); } // Fill in the map from native indexes to UChars buf index. do { - mapToUChars[sIx-- - toUCharsMapStart] = destIx; + mapToUChars[sIx-- - toUCharsMapStart] = (uint8_t)destIx; } while (sIx >= srcIx); // Set native indexing limit to be the current position. @@ -1399,7 +1409,7 @@ utext_strFromUTF8(UChar *dest, { UChar *pDest = dest; - UChar *pDestLimit = dest+destCapacity; + UChar *pDestLimit = (dest!=NULL)?(dest+destCapacity):NULL; UChar32 ch=0; int32_t index = 0; int32_t reqLength = 0; @@ -1411,16 +1421,13 @@ utext_strFromUTF8(UChar *dest, if(ch <=0x7f){ *pDest++=(UChar)ch; }else{ - ch=utf8_nextCharSafeBody(pSrc, &index, srcLength, ch, -1); - if(ch<0){ - ch = 0xfffd; - } - if(ch<=0xFFFF){ + ch=utf8_nextCharSafeBody(pSrc, &index, srcLength, ch, -3); + if(U_IS_BMP(ch)){ *(pDest++)=(UChar)ch; }else{ - *(pDest++)=UTF16_LEAD(ch); + *(pDest++)=U16_LEAD(ch); if(pDestchunkNativeLimit) { for (i=0; i<3; i++) { - if (U8_IS_LEAD(buf[start32]) || start32==0) { + if (U8_IS_SINGLE(buf[start32]) || U8_IS_LEAD(buf[start32]) || start32==0) { break; } start32--; @@ -1494,7 +1498,7 @@ utf8TextExtract(UText *ut, if (limit32 < ut->chunkNativeLimit) { for (i=0; i<3; i++) { - if (U8_IS_LEAD(buf[limit32]) || limit32==0) { + if (U8_IS_SINGLE(buf[limit32]) || U8_IS_LEAD(buf[limit32]) || limit32==0) { break; } limit32--; @@ -1506,6 +1510,7 @@ utf8TextExtract(UText *ut, utext_strFromUTF8(dest, destCapacity, &destLength, (const char *)ut->context+start32, limit32-start32, pErrorCode); + utf8TextAccess(ut, limit32, TRUE); return destLength; } @@ -1584,7 +1589,7 @@ utf8TextClose(UText *ut) { U_CDECL_END -static const struct UTextFuncs utf8Funcs = +static const struct UTextFuncs utf8Funcs = { sizeof(UTextFuncs), 0, 0, 0, // Reserved alignment padding @@ -1603,11 +1608,17 @@ static const struct UTextFuncs utf8Funcs = }; +static const char gEmptyString[] = {0}; + U_CAPI UText * U_EXPORT2 utext_openUTF8(UText *ut, const char *s, int64_t length, UErrorCode *status) { if(U_FAILURE(*status)) { return NULL; } + if(s==NULL && length==0) { + s = gEmptyString; + } + if(s==NULL || length<-1 || length>INT32_MAX) { *status=U_ILLEGAL_ARGUMENT_ERROR; return NULL; @@ -1864,6 +1875,8 @@ repTextExtract(UText *ut, } UnicodeString buffer(dest, 0, destCapacity); // writable alias rep->extractBetween(start32, limit32, buffer); + repTextAccess(ut, limit32, TRUE); + return u_terminateUChars(dest, destCapacity, length, status); } @@ -1985,7 +1998,7 @@ repTextCopy(UText *ut, repTextAccess(ut, nativeIterIndex, TRUE); } -static const struct UTextFuncs repFuncs = +static const struct UTextFuncs repFuncs = { sizeof(UTextFuncs), 0, 0, 0, // Reserved alignment padding @@ -1993,8 +2006,8 @@ static const struct UTextFuncs repFuncs = repTextLength, repTextAccess, repTextExtract, - repTextReplace, - repTextCopy, + repTextReplace, + repTextCopy, NULL, // MapOffsetToNative, NULL, // MapIndexToUTF16, repTextClose, @@ -2132,6 +2145,9 @@ unistrTextExtract(UText *t, trimmedLength=destCapacity; } us->extract(start32, trimmedLength, dest); + t->chunkOffset = start32+trimmedLength; + } else { + t->chunkOffset = start32; } u_terminateUChars(dest, destCapacity, length, pErrorCode); return length; @@ -2233,7 +2249,7 @@ unistrTextCopy(UText *ut, } -static const struct UTextFuncs unistrFuncs = +static const struct UTextFuncs unistrFuncs = { sizeof(UTextFuncs), 0, 0, 0, // Reserved alignment padding @@ -2241,8 +2257,8 @@ static const struct UTextFuncs unistrFuncs = unistrTextLength, unistrTextAccess, unistrTextExtract, - unistrTextReplace, - unistrTextCopy, + unistrTextReplace, + unistrTextCopy, NULL, // MapOffsetToNative, NULL, // MapIndexToUTF16, unistrTextClose, @@ -2258,20 +2274,9 @@ U_CDECL_END U_CAPI UText * U_EXPORT2 utext_openUnicodeString(UText *ut, UnicodeString *s, UErrorCode *status) { - // TODO: use openConstUnicodeString, then add in the differences. - // - ut = utext_setup(ut, 0, status); + ut = utext_openConstUnicodeString(ut, s, status); if (U_SUCCESS(*status)) { - ut->pFuncs = &unistrFuncs; - ut->context = s; - ut->providerProperties = I32_FLAG(UTEXT_PROVIDER_STABLE_CHUNKS)| - I32_FLAG(UTEXT_PROVIDER_WRITABLE); - - ut->chunkContents = s->getBuffer(); - ut->chunkLength = s->length(); - ut->chunkNativeStart = 0; - ut->chunkNativeLimit = ut->chunkLength; - ut->nativeIndexingLimit = ut->chunkLength; + ut->providerProperties |= I32_FLAG(UTEXT_PROVIDER_WRITABLE); } return ut; } @@ -2280,6 +2285,13 @@ utext_openUnicodeString(UText *ut, UnicodeString *s, UErrorCode *status) { U_CAPI UText * U_EXPORT2 utext_openConstUnicodeString(UText *ut, const UnicodeString *s, UErrorCode *status) { + if (U_SUCCESS(*status) && s->isBogus()) { + // The UnicodeString is bogus, but we still need to detach the UText + // from whatever it was hooked to before, if anything. + utext_openUChars(ut, NULL, 0, status); + *status = U_ILLEGAL_ARGUMENT_ERROR; + return ut; + } ut = utext_setup(ut, 0, status); // note: use the standard (writable) function table for UnicodeString. // The flag settings disable writing, so having the functions in @@ -2484,7 +2496,7 @@ ucstrTextExtract(UText *ut, return 0; } - const UChar *s=(const UChar *)ut->context; + //const UChar *s=(const UChar *)ut->context; int32_t si, di; int32_t start32; @@ -2494,8 +2506,8 @@ ucstrTextExtract(UText *ut, // Pins 'start' to the length of the string, if it came in out-of-bounds. // Snaps 'start' to the beginning of a code point. ucstrTextAccess(ut, start, TRUE); - U_ASSERT(start <= INT32_MAX); - start32 = (int32_t)start; + const UChar *s=ut->chunkContents; + start32 = ut->chunkOffset; int32_t strLength=(int32_t)ut->a; if (strLength >= 0) { @@ -2503,7 +2515,6 @@ ucstrTextExtract(UText *ut, } else { limit32 = pinIndex(limit, INT32_MAX); } - di = 0; for (si=start32; si=0); /* to ensure di never exceeds INT32_MAX, which must not happen logically */ if (di=0) { // We have filled the destination buffer, and the string length is known. // Cut the loop short. There is no need to scan string termination. - di = strLength; + di = limit32 - start32; si = limit32; break; } @@ -2542,7 +2554,7 @@ ucstrTextExtract(UText *ut, } // Put iteration position at the point just following the extracted text - ut->chunkOffset = si; + ut->chunkOffset = uprv_min(strLength, start32 + destCapacity); // Add a terminating NUL if space in the buffer permits, // and set the error status as required. @@ -2550,7 +2562,7 @@ ucstrTextExtract(UText *ut, return di; } -static const struct UTextFuncs ucstrFuncs = +static const struct UTextFuncs ucstrFuncs = { sizeof(UTextFuncs), 0, 0, 0, // Reserved alignment padding @@ -2570,13 +2582,17 @@ static const struct UTextFuncs ucstrFuncs = U_CDECL_END +static const UChar gEmptyUString[] = {0}; U_CAPI UText * U_EXPORT2 utext_openUChars(UText *ut, const UChar *s, int64_t length, UErrorCode *status) { if (U_FAILURE(*status)) { return NULL; } - if (length < -1 || length>INT32_MAX) { + if(s==NULL && length==0) { + s = gEmptyUString; + } + if (s==NULL || length < -1 || length>INT32_MAX) { *status = U_ILLEGAL_ARGUMENT_ERROR; return NULL; } @@ -2744,15 +2760,19 @@ charIterTextExtract(UText *ut, int32_t limit32 = pinIndex(limit, length); int32_t desti = 0; int32_t srci; + int32_t copyLimit; CharacterIterator *ci = (CharacterIterator *)ut->context; ci->setIndex32(start32); // Moves ix to lead of surrogate pair, if needed. srci = ci->getIndex(); + copyLimit = srci; while (srcinext32PostInc(); int32_t len = U16_LENGTH(c); + U_ASSERT(desti+len>0); /* to ensure desti+len never exceeds MAX_INT32, which must not happen logically */ if (desti+len <= destCapacity) { U16_APPEND_UNSAFE(dest, desti, c); + copyLimit = srci+len; } else { desti += len; *status = U_BUFFER_OVERFLOW_ERROR; @@ -2760,11 +2780,13 @@ charIterTextExtract(UText *ut, srci += len; } + charIterTextAccess(ut, copyLimit, TRUE); + u_terminateUChars(dest, destCapacity, desti, status); return desti; } -static const struct UTextFuncs charIterFuncs = +static const struct UTextFuncs charIterFuncs = { sizeof(UTextFuncs), 0, 0, 0, // Reserved alignment padding @@ -2824,6 +2846,3 @@ utext_openCharacterIterator(UText *ut, CharacterIterator *ci, UErrorCode *status } return ut; } - - -