Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(698)

Unified Diff: icu4c/source/common/unistr_case.cpp

Issue 311670043: ICU ticket #12410: CaseMap with Edits (Closed) Base URL: http://source.icu-project.org/repos/icu/branches/markus/caseix/
Patch Set: changes for review feedback (for real now) Created 7 years, 1 month ago
Use n/p to move between diff chunks; N/P to move between comments. Please Sign in to add in-line comments.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « icu4c/source/common/uniset_props.cpp ('k') | icu4c/source/common/unistr_case_locale.cpp » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: icu4c/source/common/unistr_case.cpp
===================================================================
--- icu4c/source/common/unistr_case.cpp (revision 39545)
+++ icu4c/source/common/unistr_case.cpp (revision 39683)
@@ -19,6 +19,8 @@
*/
#include "unicode/utypes.h"
+#include "unicode/casemap.h"
+#include "unicode/edits.h"
#include "unicode/putil.h"
#include "cstring.h"
#include "cmemory.h"
@@ -25,8 +27,9 @@
#include "unicode/ustring.h"
#include "unicode/unistr.h"
#include "unicode/uchar.h"
+#include "uassert.h"
+#include "ucasemap_imp.h"
#include "uelement.h"
-#include "ustr_imp.h"
U_NAMESPACE_BEGIN
@@ -87,7 +90,7 @@
//========================================
UnicodeString &
-UnicodeString::caseMap(const UCaseMap *csm,
+UnicodeString::caseMap(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_PARAM
UStringCaseMapper *stringCaseMapper) {
if(isEmpty() || !isWritable()) {
// nothing to do
@@ -94,49 +97,97 @@
return *this;
}
- // We need to allocate a new buffer for the internal string case mapping function.
- // This is very similar to how doReplace() keeps the old array pointer
- // and deletes the old array itself after it is done.
- // In addition, we are forcing cloneArrayIfNeeded() to always allocate a new array.
- UChar oldStackBuffer[US_STACKBUF_SIZE];
+ UChar oldBuffer[2 * US_STACKBUF_SIZE];
UChar *oldArray;
- int32_t oldLength;
+ int32_t oldLength = length();
+ int32_t newLength;
+ UBool writable = isBufferWritable();
+ UErrorCode errorCode = U_ZERO_ERROR;
- if(fUnion.fFields.fLengthAndFlags&kUsingStackBuffer) {
- // copy the stack buffer contents because it will be overwritten
- oldArray = oldStackBuffer;
- oldLength = getShortLength();
- u_memcpy(oldStackBuffer, fUnion.fStackFields.fBuffer, oldLength);
+ // Try to avoid heap-allocating a new character array for this string.
+ if (writable ? oldLength <= UPRV_LENGTHOF(oldBuffer) : oldLength < US_STACKBUF_SIZE) {
+ // Short string: Copy the contents into a temporary buffer and
+ // case-map back into the current array, or into the stack buffer.
+ UChar *buffer = getArrayStart();
+ int32_t capacity;
+ oldArray = oldBuffer;
+ u_memcpy(oldBuffer, buffer, oldLength);
+ if (writable) {
+ capacity = getCapacity();
+ } else {
+ // Switch from the read-only alias or shared heap buffer to the stack buffer.
+ if (!cloneArrayIfNeeded(US_STACKBUF_SIZE, US_STACKBUF_SIZE, /* doCopyArray= */ FALSE)) {
+ return *this;
+ }
+ U_ASSERT(fUnion.fFields.fLengthAndFlags & kUsingStackBuffer);
+ buffer = fUnion.fStackFields.fBuffer;
+ capacity = US_STACKBUF_SIZE;
+ }
+ newLength = stringCaseMapper(caseLocale, options, UCASEMAP_BREAK_ITERATOR
+ buffer, capacity,
+ oldArray, oldLength, NULL, errorCode);
+ if (U_SUCCESS(errorCode)) {
+ setLength(newLength);
+ return *this;
+ } else if (errorCode == U_BUFFER_OVERFLOW_ERROR) {
+ // common overflow handling below
+ } else {
+ setToBogus();
+ return *this;
+ }
} else {
+ // Longer string or read-only buffer:
+ // Collect only changes and then apply them to this string.
+ // Case mapping often changes only small parts of a string,
+ // and often does not change its length.
oldArray = getArrayStart();
- oldLength = length();
+ Edits edits;
+ UChar replacementChars[200];
+ stringCaseMapper(caseLocale, options | UCASEMAP_OMIT_UNCHANGED_TEXT, UCASEMAP_BREAK_ITERATOR
+ replacementChars, UPRV_LENGTHOF(replacementChars),
+ oldArray, oldLength, &edits, errorCode);
+ if (U_SUCCESS(errorCode)) {
+ // Grow the buffer at most once, not for multiple doReplace() calls.
+ newLength = oldLength + edits.lengthDelta();
+ if (newLength > oldLength && !cloneArrayIfNeeded(newLength, newLength)) {
+ return *this;
+ }
+ for (Edits::Iterator ei = edits.getCoarseChangesIterator(); ei.next(errorCode);) {
+ doReplace(ei.destinationIndex(), ei.oldLength(),
+ replacementChars, ei.replacementIndex(), ei.newLength());
+ }
+ if (U_FAILURE(errorCode)) {
+ setToBogus();
+ }
+ return *this;
+ } else if (errorCode == U_BUFFER_OVERFLOW_ERROR) {
+ // common overflow handling below
+ newLength = oldLength + edits.lengthDelta();
+ } else {
+ setToBogus();
+ return *this;
+ }
}
- int32_t capacity;
- if(oldLength <= US_STACKBUF_SIZE) {
- capacity = US_STACKBUF_SIZE;
- } else {
- capacity = oldLength + 20;
- }
+ // Handle buffer overflow, newLength is known.
+ // We need to allocate a new buffer for the internal string case mapping function.
+ // This is very similar to how doReplace() keeps the old array pointer
+ // and deletes the old array itself after it is done.
+ // In addition, we are forcing cloneArrayIfNeeded() to always allocate a new array.
int32_t *bufferToDelete = 0;
- if(!cloneArrayIfNeeded(capacity, capacity, FALSE, &bufferToDelete, TRUE)) {
+ if (!cloneArrayIfNeeded(newLength, newLength, FALSE, &bufferToDelete, TRUE)) {
return *this;
}
-
- // Case-map, and if the result is too long, then reallocate and repeat.
- UErrorCode errorCode;
- int32_t newLength;
- do {
- errorCode = U_ZERO_ERROR;
- newLength = stringCaseMapper(csm, getArrayStart(), getCapacity(),
- oldArray, oldLength, &errorCode);
- setLength(newLength);
- } while(errorCode==U_BUFFER_OVERFLOW_ERROR && cloneArrayIfNeeded(newLength, newLength, FALSE));
-
+ errorCode = U_ZERO_ERROR;
+ newLength = stringCaseMapper(caseLocale, options, UCASEMAP_BREAK_ITERATOR
+ getArrayStart(), getCapacity(),
+ oldArray, oldLength, NULL, errorCode);
if (bufferToDelete) {
uprv_free(bufferToDelete);
}
- if(U_FAILURE(errorCode)) {
+ if (U_SUCCESS(errorCode)) {
+ setLength(newLength);
+ } else {
setToBogus();
}
return *this;
@@ -144,10 +195,7 @@
UnicodeString &
UnicodeString::foldCase(uint32_t options) {
- UCaseMap csm=UCASEMAP_INITIALIZER;
- csm.csp=ucase_getSingleton();
- csm.options=options;
- return caseMap(&csm, ustrcase_internalFold);
+ return caseMap(UCASE_LOC_ROOT, options, UCASEMAP_BREAK_ITERATOR_NULL ustrcase_internalFold);
}
U_NAMESPACE_END
« no previous file with comments | « icu4c/source/common/uniset_props.cpp ('k') | icu4c/source/common/unistr_case_locale.cpp » ('j') | no next file with comments »

Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld f62528b