Bug Summary

File:root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp
Warning:line 281, column 8
Excessive padding in 'struct ShortStringOptions' (11 padding bytes, where 3 is optimal). Optimal fields order: action, attr, optionStart, consider reordering the fields or adding explicit padding members

Annotated Source Code

Press '?' to see keyboard shortcuts

clang -cc1 -cc1 -triple x86_64-pc-linux-gnu -analyze -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name ucol_sit.cpp -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=cplusplus -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -setup-static-analyzer -analyzer-config-compatibility-mode=true -mrelocation-model pic -pic-level 2 -fhalf-no-semantic-interposition -mframe-pointer=all -relaxed-aliasing -ffp-contract=off -fno-rounding-math -mconstructor-aliases -funwind-tables=2 -target-cpu x86-64 -tune-cpu generic -debugger-tuning=gdb -fdebug-compilation-dir=/root/firefox-clang/obj-x86_64-pc-linux-gnu/config/external/icu/i18n -fcoverage-compilation-dir=/root/firefox-clang/obj-x86_64-pc-linux-gnu/config/external/icu/i18n -resource-dir /usr/lib/llvm-21/lib/clang/21 -include /root/firefox-clang/config/gcc_hidden.h -include /root/firefox-clang/obj-x86_64-pc-linux-gnu/mozilla-config.h -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/system_wrappers -U _FORTIFY_SOURCE -D _FORTIFY_SOURCE=2 -D _GLIBCXX_ASSERTIONS -D DEBUG=1 -D U_I18N_IMPLEMENTATION -D _LIBCPP_DISABLE_DEPRECATION_WARNINGS -D U_USING_ICU_NAMESPACE=0 -D U_NO_DEFAULT_INCLUDE_UTF_HEADERS=1 -D U_HIDE_OBSOLETE_UTF_OLD_H=1 -D UCONFIG_NO_LEGACY_CONVERSION -D UCONFIG_NO_TRANSLITERATION -D UCONFIG_NO_REGULAR_EXPRESSIONS -D UCONFIG_NO_BREAK_ITERATION -D UCONFIG_NO_IDNA -D UCONFIG_NO_MF2 -D U_CHARSET_IS_UTF8 -D UNISTR_FROM_CHAR_EXPLICIT=explicit -D UNISTR_FROM_STRING_EXPLICIT=explicit -D U_ENABLE_DYLOAD=0 -D U_DEBUG=1 -I /root/firefox-clang/config/external/icu/i18n -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/config/external/icu/i18n -I /root/firefox-clang/intl/icu/source/common -I /root/firefox-clang/mfbt/double-conversion -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/include -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/include/nspr -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/include/nss -D MOZILLA_CLIENT -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../include/c++/14 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../include/x86_64-linux-gnu/c++/14 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../include/c++/14/backward -internal-isystem /usr/lib/llvm-21/lib/clang/21/include -internal-isystem /usr/local/include -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../x86_64-linux-gnu/include -internal-externc-isystem /usr/include/x86_64-linux-gnu -internal-externc-isystem /include -internal-externc-isystem /usr/include -O2 -Wno-error=pessimizing-move -Wno-error=large-by-value-copy=128 -Wno-error=implicit-int-float-conversion -Wno-error=thread-safety-analysis -Wno-error=tautological-type-limit-compare -Wno-invalid-offsetof -Wno-range-loop-analysis -Wno-deprecated-anon-enum-enum-conversion -Wno-deprecated-enum-enum-conversion -Wno-deprecated-this-capture -Wno-inline-new-delete -Wno-error=deprecated-declarations -Wno-error=array-bounds -Wno-error=free-nonheap-object -Wno-error=atomic-alignment -Wno-error=deprecated-builtins -Wno-psabi -Wno-error=builtin-macro-redefined -Wno-vla-cxx-extension -Wno-unknown-warning-option -Wno-comma -Wno-implicit-const-int-float-conversion -Wno-macro-redefined -Wno-microsoft-include -Wno-tautological-unsigned-enum-zero-compare -Wno-unreachable-code-loop-increment -Wno-unreachable-code-return -fdeprecated-macro -ferror-limit 19 -fstrict-flex-arrays=1 -stack-protector 2 -fstack-clash-protection -ftrivial-auto-var-init=pattern -fgnuc-version=4.2.1 -fskip-odr-check-in-gmf -fno-sized-deallocation -fno-aligned-allocation -vectorize-loops -vectorize-slp -analyzer-checker optin.performance.Padding -analyzer-output=html -analyzer-config stable-report-filename=true -faddrsig -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /tmp/scan-build-2025-06-27-100320-3286336-1 -x c++ /root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp
1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4*******************************************************************************
5* Copyright (C) 2004-2016, International Business Machines
6* Corporation and others. All Rights Reserved.
7*******************************************************************************
8* file name: ucol_sit.cpp
9* encoding: UTF-8
10* tab size: 8 (not used)
11* indentation:4
12*
13* Modification history
14* Date Name Comments
15* 03/12/2004 weiv Creation
16*/
17
18#include "unicode/ustring.h"
19#include "unicode/udata.h"
20#include "unicode/utf16.h"
21#include "utracimp.h"
22#include "ucol_imp.h"
23#include "ulocimp.h"
24#include "cmemory.h"
25#include "cstring.h"
26#include "uresimp.h"
27#include "unicode/coll.h"
28#include "unicode/stringpiece.h"
29#include "charstr.h"
30
31U_NAMESPACE_USEusing namespace icu_77;
32
33#ifdef UCOL_TRACE_SIT
34# include <stdio.h>
35#endif
36
37#if !UCONFIG_NO_COLLATION0
38
39#include "unicode/tblcoll.h"
40
41enum OptionsList {
42 UCOL_SIT_LANGUAGE = 0,
43 UCOL_SIT_SCRIPT = 1,
44 UCOL_SIT_REGION = 2,
45 UCOL_SIT_VARIANT = 3,
46 UCOL_SIT_KEYWORD = 4,
47 UCOL_SIT_PROVIDER = 5,
48 UCOL_SIT_LOCELEMENT_MAX = UCOL_SIT_PROVIDER, /* the last element that's part of LocElements */
49
50 UCOL_SIT_BCP47,
51 UCOL_SIT_STRENGTH,
52 UCOL_SIT_CASE_LEVEL,
53 UCOL_SIT_CASE_FIRST,
54 UCOL_SIT_NUMERIC_COLLATION,
55 UCOL_SIT_ALTERNATE_HANDLING,
56 UCOL_SIT_NORMALIZATION_MODE,
57 UCOL_SIT_FRENCH_COLLATION,
58 UCOL_SIT_HIRAGANA_QUATERNARY,
59 UCOL_SIT_VARIABLE_TOP,
60 UCOL_SIT_VARIABLE_TOP_VALUE,
61 UCOL_SIT_ITEMS_COUNT
62};
63
64/* option starters chars. */
65static const char alternateHArg = 'A';
66static const char variableTopValArg = 'B';
67static const char caseFirstArg = 'C';
68static const char numericCollArg = 'D';
69static const char caseLevelArg = 'E';
70static const char frenchCollArg = 'F';
71static const char hiraganaQArg = 'H';
72static const char keywordArg = 'K';
73static const char languageArg = 'L';
74static const char normArg = 'N';
75static const char providerArg = 'P';
76static const char regionArg = 'R';
77static const char strengthArg = 'S';
78static const char variableTopArg = 'T';
79static const char variantArg = 'V';
80static const char RFC3066Arg = 'X';
81static const char scriptArg = 'Z';
82
83static const char collationKeyword[] = "@collation=";
84static const char providerKeyword[] = "@sp=";
85
86
87static const int32_t locElementCount = UCOL_SIT_LOCELEMENT_MAX+1;
88static const int32_t locElementCapacity = 32;
89static const int32_t loc3066Capacity = 256;
90
91/* structure containing specification of a collator. Initialized
92 * from a short string. Also used to construct a short string from a
93 * collator instance
94 */
95struct CollatorSpec {
96 inline CollatorSpec();
97
98 CharString locElements[locElementCount];
99 CharString locale;
100 UColAttributeValue options[UCOL_ATTRIBUTE_COUNT];
101 uint32_t variableTopValue;
102 char16_t variableTopString[locElementCapacity];
103 int32_t variableTopStringLen;
104 UBool variableTopSet;
105 CharString entries[UCOL_SIT_ITEMS_COUNT];
106};
107
108CollatorSpec::CollatorSpec() :
109locale(),
110variableTopValue(0),
111variableTopString(),
112variableTopSet(false)
113 {
114 // set collation options to default
115 for(int32_t i = 0; i < UCOL_ATTRIBUTE_COUNT; i++) {
116 options[i] = UCOL_DEFAULT;
117 }
118}
119
120
121/* structure for converting between character attribute
122 * representation and real collation attribute value.
123 */
124struct AttributeConversion {
125 char letter;
126 UColAttributeValue value;
127};
128
129static const AttributeConversion conversions[12] = {
130 { '1', UCOL_PRIMARY },
131 { '2', UCOL_SECONDARY },
132 { '3', UCOL_TERTIARY },
133 { '4', UCOL_QUATERNARY },
134 { 'D', UCOL_DEFAULT },
135 { 'I', UCOL_IDENTICAL },
136 { 'L', UCOL_LOWER_FIRST },
137 { 'N', UCOL_NON_IGNORABLE },
138 { 'O', UCOL_ON },
139 { 'S', UCOL_SHIFTED },
140 { 'U', UCOL_UPPER_FIRST },
141 { 'X', UCOL_OFF }
142};
143
144
145static UColAttributeValue
146ucol_sit_letterToAttributeValue(char letter, UErrorCode *status) {
147 uint32_t i = 0;
148 for(i = 0; i < UPRV_LENGTHOF(conversions)(int32_t)(sizeof(conversions)/sizeof((conversions)[0])); i++) {
149 if(conversions[i].letter == letter) {
150 return conversions[i].value;
151 }
152 }
153 *status = U_ILLEGAL_ARGUMENT_ERROR;
154#ifdef UCOL_TRACE_SIT
155 fprintf(stderrstderr, "%s:%d: unknown letter %c: %s\n", __FILE__"/root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp", __LINE__155, letter, u_errorNameu_errorName_77(*status));
156#endif
157 return UCOL_DEFAULT;
158}
159
160/* function prototype for functions used to parse a short string */
161U_CDECL_BEGINextern "C" {
162typedef const char* U_CALLCONV
163ActionFunction(CollatorSpec *spec, uint32_t value1, const char* string,
164 UErrorCode *status);
165U_CDECL_END}
166
167U_CDECL_BEGINextern "C" {
168static const char* U_CALLCONV
169_processLocaleElement(CollatorSpec *spec, uint32_t value, const char* string,
170 UErrorCode *status)
171{
172 do {
173 if(value == UCOL_SIT_LANGUAGE || value == UCOL_SIT_KEYWORD || value == UCOL_SIT_PROVIDER) {
174 spec->locElements[value].append(uprv_toloweruprv_asciitolower_77(*string), *status);
175 } else {
176 spec->locElements[value].append(*string, *status);
177 }
178 } while(*(++string) != '_' && *string && U_SUCCESS(*status));
179 // don't skip the underscore at the end
180 return string;
181}
182U_CDECL_END}
183
184U_CDECL_BEGINextern "C" {
185static const char* U_CALLCONV
186_processRFC3066Locale(CollatorSpec *spec, uint32_t, const char* string,
187 UErrorCode *status)
188{
189 char terminator = *string;
190 string++;
191 const char *end = uprv_strchr(string+1, terminator):: strchr(string+1, terminator);
192 if(end == nullptr || end - string >= loc3066Capacity) {
193 *status = U_BUFFER_OVERFLOW_ERROR;
194 return string;
195 } else {
196 spec->locale.copyFrom(CharString(string, static_cast<int32_t>(end-string), *status), *status);
197 return end+1;
198 }
199}
200
201U_CDECL_END}
202
203U_CDECL_BEGINextern "C" {
204static const char* U_CALLCONV
205_processCollatorOption(CollatorSpec *spec, uint32_t option, const char* string,
206 UErrorCode *status)
207{
208 spec->options[option] = ucol_sit_letterToAttributeValue(*string, status);
209 if((*(++string) != '_' && *string) || U_FAILURE(*status)) {
210#ifdef UCOL_TRACE_SIT
211 fprintf(stderrstderr, "%s:%d: unknown collator option at '%s': %s\n", __FILE__"/root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp", __LINE__211, string, u_errorNameu_errorName_77(*status));
212#endif
213 *status = U_ILLEGAL_ARGUMENT_ERROR;
214 }
215 return string;
216}
217U_CDECL_END}
218
219
220static char16_t
221readHexCodeUnit(const char **string, UErrorCode *status)
222{
223 char16_t result = 0;
224 int32_t value = 0;
225 char c;
226 int32_t noDigits = 0;
227 while((c = **string) != 0 && noDigits < 4) {
228 if( c >= '0' && c <= '9') {
229 value = c - '0';
230 } else if ( c >= 'a' && c <= 'f') {
231 value = c - 'a' + 10;
232 } else if ( c >= 'A' && c <= 'F') {
233 value = c - 'A' + 10;
234 } else {
235 *status = U_ILLEGAL_ARGUMENT_ERROR;
236#ifdef UCOL_TRACE_SIT
237 fprintf(stderrstderr, "%s:%d: Bad hex char at '%s': %s\n", __FILE__"/root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp", __LINE__237, *string, u_errorNameu_errorName_77(*status));
238#endif
239 return 0;
240 }
241 result = (result << 4) | static_cast<char16_t>(value);
242 noDigits++;
243 (*string)++;
244 }
245 // if the string was terminated before we read 4 digits, set an error
246 if(noDigits < 4) {
247 *status = U_ILLEGAL_ARGUMENT_ERROR;
248#ifdef UCOL_TRACE_SIT
249 fprintf(stderrstderr, "%s:%d: Short (only %d digits, wanted 4) at '%s': %s\n", __FILE__"/root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp", __LINE__249, noDigits,*string, u_errorNameu_errorName_77(*status));
250#endif
251 }
252 return result;
253}
254
255U_CDECL_BEGINextern "C" {
256static const char* U_CALLCONV
257_processVariableTop(CollatorSpec *spec, uint32_t value1, const char* string, UErrorCode *status)
258{
259 // get four digits
260 int32_t i = 0;
261 if(!value1) {
262 while(U_SUCCESS(*status) && i < locElementCapacity && *string != 0 && *string != '_') {
263 spec->variableTopString[i++] = readHexCodeUnit(&string, status);
264 }
265 spec->variableTopStringLen = i;
266 if(i == locElementCapacity && *string != 0 && *string != '_') {
267 *status = U_BUFFER_OVERFLOW_ERROR;
268 }
269 } else {
270 spec->variableTopValue = readHexCodeUnit(&string, status);
271 }
272 if(U_SUCCESS(*status)) {
273 spec->variableTopSet = true;
274 }
275 return string;
276}
277U_CDECL_END}
278
279
280/* Table for parsing short strings */
281struct ShortStringOptions {
Excessive padding in 'struct ShortStringOptions' (11 padding bytes, where 3 is optimal). Optimal fields order: action, attr, optionStart, consider reordering the fields or adding explicit padding members
282 char optionStart;
283 ActionFunction *action;
284 uint32_t attr;
285};
286
287static const ShortStringOptions options[UCOL_SIT_ITEMS_COUNT] =
288{
289/* 10 ALTERNATE_HANDLING */ {alternateHArg, _processCollatorOption, UCOL_ALTERNATE_HANDLING }, // alternate N, S, D
290/* 15 VARIABLE_TOP_VALUE */ {variableTopValArg, _processVariableTop, 1 },
291/* 08 CASE_FIRST */ {caseFirstArg, _processCollatorOption, UCOL_CASE_FIRST }, // case first L, U, X, D
292/* 09 NUMERIC_COLLATION */ {numericCollArg, _processCollatorOption, UCOL_NUMERIC_COLLATION }, // codan O, X, D
293/* 07 CASE_LEVEL */ {caseLevelArg, _processCollatorOption, UCOL_CASE_LEVEL }, // case level O, X, D
294/* 12 FRENCH_COLLATION */ {frenchCollArg, _processCollatorOption, UCOL_FRENCH_COLLATION }, // french O, X, D
295/* 13 HIRAGANA_QUATERNARY] */ {hiraganaQArg, _processCollatorOption, UCOL_HIRAGANA_QUATERNARY_MODE }, // hiragana O, X, D
296/* 04 KEYWORD */ {keywordArg, _processLocaleElement, UCOL_SIT_KEYWORD }, // keyword
297/* 00 LANGUAGE */ {languageArg, _processLocaleElement, UCOL_SIT_LANGUAGE }, // language
298/* 11 NORMALIZATION_MODE */ {normArg, _processCollatorOption, UCOL_NORMALIZATION_MODE }, // norm O, X, D
299/* 02 REGION */ {regionArg, _processLocaleElement, UCOL_SIT_REGION }, // region
300/* 06 STRENGTH */ {strengthArg, _processCollatorOption, UCOL_STRENGTH }, // strength 1, 2, 3, 4, I, D
301/* 14 VARIABLE_TOP */ {variableTopArg, _processVariableTop, 0 },
302/* 03 VARIANT */ {variantArg, _processLocaleElement, UCOL_SIT_VARIANT }, // variant
303/* 05 RFC3066BIS */ {RFC3066Arg, _processRFC3066Locale, 0 }, // rfc3066bis locale name
304/* 01 SCRIPT */ {scriptArg, _processLocaleElement, UCOL_SIT_SCRIPT }, // script
305/* PROVIDER */ {providerArg, _processLocaleElement, UCOL_SIT_PROVIDER }
306};
307
308
309static
310const char* ucol_sit_readOption(const char *start, CollatorSpec *spec,
311 UErrorCode *status)
312{
313 int32_t i = 0;
314
315 for(i = 0; i < UCOL_SIT_ITEMS_COUNT; i++) {
316 if(*start == options[i].optionStart) {
317 const char* end = options[i].action(spec, options[i].attr, start+1, status);
318#ifdef UCOL_TRACE_SIT
319 fprintf(stderrstderr, "***Set %d to %s...\n", i, start);
320#endif
321 // assume 'start' does not go away through all this
322 spec->entries[i].copyFrom(CharString(start, static_cast<int32_t>(end - start), *status), *status);
323 return end;
324 }
325 }
326 *status = U_ILLEGAL_ARGUMENT_ERROR;
327#ifdef UCOL_TRACE_SIT
328 fprintf(stderrstderr, "%s:%d: Unknown option at '%s': %s\n", __FILE__"/root/firefox-clang/intl/icu/source/i18n/ucol_sit.cpp", __LINE__328, start, u_errorNameu_errorName_77(*status));
329#endif
330 return start;
331}
332
333static const char*
334ucol_sit_readSpecs(CollatorSpec *s, const char *string,
335 UParseError *parseError, UErrorCode *status)
336{
337 const char *definition = string;
338 while(U_SUCCESS(*status) && *string) {
339 string = ucol_sit_readOption(string, s, status);
340 // advance over '_'
341 while(*string && *string == '_') {
342 string++;
343 }
344 }
345 if(U_FAILURE(*status)) {
346 parseError->offset = static_cast<int32_t>(string - definition);
347 }
348 return string;
349}
350
351static
352int32_t ucol_sit_dumpSpecs(CollatorSpec *s, char *destination, int32_t capacity, UErrorCode *status)
353{
354 int32_t i = 0, j = 0;
355 int32_t len = 0;
356 char optName;
357 if(U_SUCCESS(*status)) {
358 for(i = 0; i < UCOL_SIT_ITEMS_COUNT; i++) {
359 if(!s->entries[i].isEmpty()) {
360 if(len) {
361 if(len < capacity) {
362 uprv_strcat(destination, "_"):: strcat(destination, "_");
363 }
364 len++;
365 }
366 optName = s->entries[i][0];
367 if(optName == languageArg || optName == regionArg || optName == variantArg || optName == keywordArg) {
368 for(j = 0; j < s->entries[i].length(); j++) {
369 if(len + j < capacity) {
370 destination[len+j] = uprv_toupperuprv_toupper_77(s->entries[i][j]);
371 }
372 }
373 len += s->entries[i].length();
374 } else {
375 len += s->entries[i].extract(destination + len, capacity - len, *status);
376 }
377 }
378 }
379 return len;
380 } else {
381 return 0;
382 }
383}
384
385static void
386ucol_sit_calculateWholeLocale(CollatorSpec *s, UErrorCode &status) {
387 // put the locale together, unless we have a done
388 // locale
389 if(s->locale.isEmpty()) {
390 // first the language
391 s->locale.append(s->locElements[UCOL_SIT_LANGUAGE], status);
392 // then the script, if present
393 if(!s->locElements[UCOL_SIT_SCRIPT].isEmpty()) {
394 s->locale.append("_", status);
395 s->locale.append(s->locElements[UCOL_SIT_SCRIPT], status);
396 }
397 // then the region, if present
398 if(!s->locElements[UCOL_SIT_REGION].isEmpty()) {
399 s->locale.append("_", status);
400 s->locale.append(s->locElements[UCOL_SIT_REGION], status);
401 } else if(!s->locElements[UCOL_SIT_VARIANT].isEmpty()) { // if there is a variant, we need an underscore
402 s->locale.append("_", status);
403 }
404 // add variant, if there
405 if(!s->locElements[UCOL_SIT_VARIANT].isEmpty()) {
406 s->locale.append("_", status);
407 s->locale.append(s->locElements[UCOL_SIT_VARIANT], status);
408 }
409
410 // if there is a collation keyword, add that too
411 if(!s->locElements[UCOL_SIT_KEYWORD].isEmpty()) {
412 s->locale.append(collationKeyword, status);
413 s->locale.append(s->locElements[UCOL_SIT_KEYWORD], status);
414 }
415
416 // if there is a provider keyword, add that too
417 if(!s->locElements[UCOL_SIT_PROVIDER].isEmpty()) {
418 s->locale.append(providerKeyword, status);
419 s->locale.append(s->locElements[UCOL_SIT_PROVIDER], status);
420 }
421 }
422}
423
424
425U_CAPIextern "C" void U_EXPORT2
426ucol_prepareShortStringOpenucol_prepareShortStringOpen_77( const char *definition,
427 UBool,
428 UParseError *parseError,
429 UErrorCode *status)
430{
431 if(U_FAILURE(*status)) return;
432
433 UParseError internalParseError;
434
435 if(!parseError) {
436 parseError = &internalParseError;
437 }
438 parseError->line = 0;
439 parseError->offset = 0;
440 parseError->preContext[0] = 0;
441 parseError->postContext[0] = 0;
442
443
444 // first we want to pick stuff out of short string.
445 // we'll end up with an UCA version, locale and a bunch of
446 // settings
447
448 // analyse the string in order to get everything we need.
449 CollatorSpec s;
450 ucol_sit_readSpecs(&s, definition, parseError, status);
451 ucol_sit_calculateWholeLocale(&s, *status);
452
453 CharString buffer = ulocimp_canonicalizeulocimp_canonicalize_77(s.locale.toStringPiece(), *status);
454
455 UResourceBundle *b = ures_openures_open_77(U_ICUDATA_COLL"icudt" "77" "l" "-" "coll", buffer.data(), status);
456 /* we try to find stuff from keyword */
457 UResourceBundle *collations = ures_getByKeyures_getByKey_77(b, "collations", nullptr, status);
458 UResourceBundle *collElem = nullptr;
459 // if there is a keyword, we pick it up and try to get elements
460 CharString keyBuffer = ulocimp_getKeywordValueulocimp_getKeywordValue_77(buffer.data(), "collation", *status);
461 if(keyBuffer.isEmpty()) {
462 // no keyword
463 // we try to find the default setting, which will give us the keyword value
464 UResourceBundle *defaultColl = ures_getByKeyWithFallbackures_getByKeyWithFallback_77(collations, "default", nullptr, status);
465 if(U_SUCCESS(*status)) {
466 int32_t defaultKeyLen = 0;
467 const char16_t *defaultKey = ures_getStringures_getString_77(defaultColl, &defaultKeyLen, status);
468 keyBuffer.appendInvariantChars(defaultKey, defaultKeyLen, *status);
469 } else {
470 *status = U_INTERNAL_PROGRAM_ERROR;
471 return;
472 }
473 ures_closeures_close_77(defaultColl);
474 }
475 collElem = ures_getByKeyWithFallbackures_getByKeyWithFallback_77(collations, keyBuffer.data(), collElem, status);
476 ures_closeures_close_77(collElem);
477 ures_closeures_close_77(collations);
478 ures_closeures_close_77(b);
479}
480
481
482U_CAPIextern "C" UCollator* U_EXPORT2
483ucol_openFromShortStringucol_openFromShortString_77( const char *definition,
484 UBool forceDefaults,
485 UParseError *parseError,
486 UErrorCode *status)
487{
488 UTRACE_ENTRY_OC(UTRACE_UCOL_OPEN_FROM_SHORT_STRING);
489 UTRACE_DATA1(UTRACE_INFO, "short string = \"%s\"", definition);
490
491 if (U_FAILURE(*status)) return nullptr;
492
493 UParseError internalParseError;
494
495 if(!parseError) {
496 parseError = &internalParseError;
497 }
498 parseError->line = 0;
499 parseError->offset = 0;
500 parseError->preContext[0] = 0;
501 parseError->postContext[0] = 0;
502
503
504 // first we want to pick stuff out of short string.
505 // we'll end up with an UCA version, locale and a bunch of
506 // settings
507
508 // analyse the string in order to get everything we need.
509 const char *string = definition;
510 CollatorSpec s;
511 string = ucol_sit_readSpecs(&s, definition, parseError, status);
512 ucol_sit_calculateWholeLocale(&s, *status);
513
514#ifdef UCOL_TRACE_SIT
515 fprintf(stderrstderr, "DEF %s, DATA %s, ERR %s\n", definition, s.locale.data(), u_errorNameu_errorName_77(*status));
516#endif
517 CharString buffer = ulocimp_canonicalizeulocimp_canonicalize_77(s.locale.toStringPiece(), *status);
518
519 UCollator *result = ucol_openucol_open_77(buffer.data(), status);
520 int32_t i = 0;
521
522 for(i = 0; i < UCOL_ATTRIBUTE_COUNT; i++) {
523 if(s.options[i] != UCOL_DEFAULT) {
524 if(forceDefaults || ucol_getAttributeucol_getAttribute_77(result, (UColAttribute)i, status) != s.options[i]) {
525 ucol_setAttributeucol_setAttribute_77(result, (UColAttribute)i, s.options[i], status);
526 }
527
528 if(U_FAILURE(*status)) {
529 parseError->offset = (int32_t)(string - definition);
530 ucol_closeucol_close_77(result);
531 return nullptr;
532 }
533
534 }
535 }
536 if(s.variableTopSet) {
537 if(s.variableTopString[0]) {
538 ucol_setVariableTopucol_setVariableTop_77(result, s.variableTopString, s.variableTopStringLen, status);
539 } else { // we set by value, using 'B'
540 ucol_restoreVariableTopucol_restoreVariableTop_77(result, s.variableTopValue, status);
541 }
542 }
543
544
545 if(U_FAILURE(*status)) { // here it can only be a bogus value
546 ucol_closeucol_close_77(result);
547 result = nullptr;
548 }
549
550 UTRACE_EXIT_PTR_STATUS(result, *status);
551 return result;
552}
553
554
555U_CAPIextern "C" int32_t U_EXPORT2
556ucol_getShortDefinitionStringucol_getShortDefinitionString_77(const UCollator *coll,
557 const char *locale,
558 char *dst,
559 int32_t capacity,
560 UErrorCode *status)
561{
562 if(U_FAILURE(*status)) return 0;
563 if(coll == nullptr) {
564 *status = U_ILLEGAL_ARGUMENT_ERROR;
565 return 0;
566 }
567 return ((icu::Collator*)coll)->internalGetShortDefinitionString(locale,dst,capacity,*status);
568}
569
570U_CAPIextern "C" int32_t U_EXPORT2
571ucol_normalizeShortDefinitionStringucol_normalizeShortDefinitionString_77(const char *definition,
572 char *destination,
573 int32_t capacity,
574 UParseError *parseError,
575 UErrorCode *status)
576{
577
578 if(U_FAILURE(*status)) {
579 return 0;
580 }
581
582 if(destination) {
583 uprv_memset(destination, 0, capacity*sizeof(char)):: memset(destination, 0, capacity*sizeof(char));
584 }
585
586 UParseError pe;
587 if(!parseError) {
588 parseError = &pe;
589 }
590
591 // validate
592 CollatorSpec s;
593 ucol_sit_readSpecs(&s, definition, parseError, status);
594 return ucol_sit_dumpSpecs(&s, destination, capacity, status);
595}
596
597/**
598 * Get a set containing the contractions defined by the collator. The set includes
599 * both the UCA contractions and the contractions defined by the collator
600 * @param coll collator
601 * @param conts the set to hold the result
602 * @param status to hold the error code
603 * @return the size of the contraction set
604 */
605U_CAPIextern "C" int32_t U_EXPORT2
606ucol_getContractionsucol_getContractions_77( const UCollator *coll,
607 USet *contractions,
608 UErrorCode *status)
609{
610 ucol_getContractionsAndExpansionsucol_getContractionsAndExpansions_77(coll, contractions, nullptr, false, status);
611 return uset_getItemCountuset_getItemCount_77(contractions);
612}
613
614/**
615 * Get a set containing the expansions defined by the collator. The set includes
616 * both the UCA expansions and the expansions defined by the tailoring
617 * @param coll collator
618 * @param conts the set to hold the result
619 * @param addPrefixes add the prefix contextual elements to contractions
620 * @param status to hold the error code
621 *
622 * @draft ICU 3.4
623 */
624U_CAPIextern "C" void U_EXPORT2
625ucol_getContractionsAndExpansionsucol_getContractionsAndExpansions_77( const UCollator *coll,
626 USet *contractions,
627 USet *expansions,
628 UBool addPrefixes,
629 UErrorCode *status)
630{
631 if(U_FAILURE(*status)) {
632 return;
633 }
634 if(coll == nullptr) {
635 *status = U_ILLEGAL_ARGUMENT_ERROR;
636 return;
637 }
638 const icu::RuleBasedCollator *rbc = icu::RuleBasedCollator::rbcFromUCollator(coll);
639 if(rbc == nullptr) {
640 *status = U_UNSUPPORTED_ERROR;
641 return;
642 }
643 rbc->internalGetContractionsAndExpansions(
644 icu::UnicodeSet::fromUSet(contractions),
645 icu::UnicodeSet::fromUSet(expansions),
646 addPrefixes, *status);
647}
648#endif