2 * Copyright (C) 1995-2007, Index Data ApS
3 * See the file LICENSE for details.
5 * $Id: icu_I18N.c,v 1.15 2007-11-08 17:08:11 adam Exp $
14 #include <yaz/timing.h>
19 #include <yaz/icu_I18N.h>
27 #include <unicode/ustring.h> /* some more string fcns*/
28 #include <unicode/uchar.h> /* char names */
31 #include <unicode/ucol.h>
34 int icu_check_status (UErrorCode status)
36 if(U_FAILURE(status)){
38 "ICU: %d %s\n", status, u_errorName(status));
47 struct icu_buf_utf16 * icu_buf_utf16_create(size_t capacity)
49 struct icu_buf_utf16 * buf16
50 = (struct icu_buf_utf16 *) malloc(sizeof(struct icu_buf_utf16));
57 buf16->utf16 = (UChar *) malloc(sizeof(UChar) * capacity);
58 buf16->utf16[0] = (UChar) 0;
59 buf16->utf16_cap = capacity;
64 struct icu_buf_utf16 * icu_buf_utf16_clear(struct icu_buf_utf16 * buf16)
68 buf16->utf16[0] = (UChar) 0;
74 struct icu_buf_utf16 * icu_buf_utf16_resize(struct icu_buf_utf16 * buf16,
81 if (0 == buf16->utf16)
82 buf16->utf16 = (UChar *) malloc(sizeof(UChar) * capacity);
85 = (UChar *) realloc(buf16->utf16, sizeof(UChar) * capacity);
87 icu_buf_utf16_clear(buf16);
88 buf16->utf16_cap = capacity;
102 struct icu_buf_utf16 * icu_buf_utf16_copy(struct icu_buf_utf16 * dest16,
103 struct icu_buf_utf16 * src16)
109 if (dest16->utf16_cap < src16->utf16_len)
110 icu_buf_utf16_resize(dest16, src16->utf16_len * 2);
112 u_strncpy(dest16->utf16, src16->utf16, src16->utf16_len);
113 dest16->utf16_len = src16->utf16_len;
119 void icu_buf_utf16_destroy(struct icu_buf_utf16 * buf16)
133 struct icu_buf_utf8 * icu_buf_utf8_create(size_t capacity)
135 struct icu_buf_utf8 * buf8
136 = (struct icu_buf_utf8 *) malloc(sizeof(struct icu_buf_utf8));
143 buf8->utf8 = (uint8_t *) malloc(sizeof(uint8_t) * capacity);
144 buf8->utf8[0] = (uint8_t) 0;
145 buf8->utf8_cap = capacity;
151 struct icu_buf_utf8 * icu_buf_utf8_clear(struct icu_buf_utf8 * buf8)
155 buf8->utf8[0] = (uint8_t) 0;
162 struct icu_buf_utf8 * icu_buf_utf8_resize(struct icu_buf_utf8 * buf8,
170 buf8->utf8 = (uint8_t *) malloc(sizeof(uint8_t) * capacity);
173 = (uint8_t *) realloc(buf8->utf8, sizeof(uint8_t) * capacity);
175 icu_buf_utf8_clear(buf8);
176 buf8->utf8_cap = capacity;
190 struct icu_buf_utf8 * icu_buf_utf8_copy(struct icu_buf_utf8 * dest8,
191 struct icu_buf_utf8 * src8)
198 if (dest8->utf8_cap < src8->utf8_len)
199 icu_buf_utf8_resize(dest8, src8->utf8_len * 2);
201 strncpy((char*) dest8->utf8, (char*) src8->utf8, src8->utf8_len);
207 const char *icu_buf_utf8_to_cstr(struct icu_buf_utf8 *src8)
209 if (!src8 || src8->utf8_len == 0)
212 if (src8->utf8_len == src8->utf8_cap)
213 src8 = icu_buf_utf8_resize(src8, src8->utf8_len * 2 + 1);
215 src8->utf8[src8->utf8_len] = '\0';
217 return (const char *) src8->utf8;
221 void icu_buf_utf8_destroy(struct icu_buf_utf8 * buf8)
232 UErrorCode icu_utf16_from_utf8(struct icu_buf_utf16 * dest16,
233 struct icu_buf_utf8 * src8,
236 int32_t utf16_len = 0;
238 u_strFromUTF8(dest16->utf16, dest16->utf16_cap,
240 (const char *) src8->utf8, src8->utf8_len, status);
242 /* check for buffer overflow, resize and retry */
243 if (*status == U_BUFFER_OVERFLOW_ERROR)
245 icu_buf_utf16_resize(dest16, utf16_len * 2);
246 *status = U_ZERO_ERROR;
247 u_strFromUTF8(dest16->utf16, dest16->utf16_cap,
249 (const char *) src8->utf8, src8->utf8_len, status);
252 if (U_SUCCESS(*status)
253 && utf16_len <= dest16->utf16_cap)
254 dest16->utf16_len = utf16_len;
256 icu_buf_utf16_clear(dest16);
263 UErrorCode icu_utf16_from_utf8_cstr(struct icu_buf_utf16 * dest16,
264 const char * src8cstr,
267 size_t src8cstr_len = 0;
268 int32_t utf16_len = 0;
270 *status = U_ZERO_ERROR;
271 src8cstr_len = strlen(src8cstr);
273 u_strFromUTF8(dest16->utf16, dest16->utf16_cap,
275 src8cstr, src8cstr_len, status);
277 /* check for buffer overflow, resize and retry */
278 if (*status == U_BUFFER_OVERFLOW_ERROR)
280 icu_buf_utf16_resize(dest16, utf16_len * 2);
281 *status = U_ZERO_ERROR;
282 u_strFromUTF8(dest16->utf16, dest16->utf16_cap,
284 src8cstr, src8cstr_len, status);
287 if (U_SUCCESS(*status)
288 && utf16_len <= dest16->utf16_cap)
289 dest16->utf16_len = utf16_len;
291 icu_buf_utf16_clear(dest16);
299 UErrorCode icu_utf16_to_utf8(struct icu_buf_utf8 * dest8,
300 struct icu_buf_utf16 * src16,
303 int32_t utf8_len = 0;
305 u_strToUTF8((char *) dest8->utf8, dest8->utf8_cap,
307 src16->utf16, src16->utf16_len, status);
309 /* check for buffer overflow, resize and retry */
310 if (*status == U_BUFFER_OVERFLOW_ERROR)
312 icu_buf_utf8_resize(dest8, utf8_len * 2);
313 *status = U_ZERO_ERROR;
314 u_strToUTF8((char *) dest8->utf8, dest8->utf8_cap,
316 src16->utf16, src16->utf16_len, status);
320 if (U_SUCCESS(*status)
321 && utf8_len <= dest8->utf8_cap)
322 dest8->utf8_len = utf8_len;
324 icu_buf_utf8_clear(dest8);
331 struct icu_casemap * icu_casemap_create(char action, UErrorCode *status)
333 struct icu_casemap * casemap
334 = (struct icu_casemap *) malloc(sizeof(struct icu_casemap));
335 casemap->action = action;
337 switch(casemap->action) {
348 icu_casemap_destroy(casemap);
355 void icu_casemap_destroy(struct icu_casemap * casemap)
362 int icu_casemap_casemap(struct icu_casemap * casemap,
363 struct icu_buf_utf16 * dest16,
364 struct icu_buf_utf16 * src16,
371 return icu_utf16_casemap(dest16, src16, locale,
372 casemap->action, status);
376 int icu_utf16_casemap(struct icu_buf_utf16 * dest16,
377 struct icu_buf_utf16 * src16,
378 const char *locale, char action,
381 int32_t dest16_len = 0;
384 if (!src16->utf16_len){ //guarding for empty source string
386 dest16->utf16[0] = (UChar) 0;
387 dest16->utf16_len = 0;
395 dest16_len = u_strToLower(dest16->utf16, dest16->utf16_cap,
396 src16->utf16, src16->utf16_len,
401 dest16_len = u_strToUpper(dest16->utf16, dest16->utf16_cap,
402 src16->utf16, src16->utf16_len,
407 dest16_len = u_strToTitle(dest16->utf16, dest16->utf16_cap,
408 src16->utf16, src16->utf16_len,
413 dest16_len = u_strFoldCase(dest16->utf16, dest16->utf16_cap,
414 src16->utf16, src16->utf16_len,
415 U_FOLD_CASE_DEFAULT, status);
419 return U_UNSUPPORTED_ERROR;
423 /* check for buffer overflow, resize and retry */
424 if (*status == U_BUFFER_OVERFLOW_ERROR
425 && dest16 != src16 /* do not resize if in-place conversion */
427 icu_buf_utf16_resize(dest16, dest16_len * 2);
428 *status = U_ZERO_ERROR;
434 dest16_len = u_strToLower(dest16->utf16, dest16->utf16_cap,
435 src16->utf16, src16->utf16_len,
440 dest16_len = u_strToUpper(dest16->utf16, dest16->utf16_cap,
441 src16->utf16, src16->utf16_len,
446 dest16_len = u_strToTitle(dest16->utf16, dest16->utf16_cap,
447 src16->utf16, src16->utf16_len,
452 dest16_len = u_strFoldCase(dest16->utf16, dest16->utf16_cap,
453 src16->utf16, src16->utf16_len,
454 U_FOLD_CASE_DEFAULT, status);
458 return U_UNSUPPORTED_ERROR;
463 if (U_SUCCESS(*status)
464 && dest16_len <= dest16->utf16_cap)
465 dest16->utf16_len = dest16_len;
468 dest16->utf16[0] = (UChar) 0;
469 dest16->utf16_len = 0;
477 UErrorCode icu_sortkey8_from_utf16(UCollator *coll,
478 struct icu_buf_utf8 * dest8,
479 struct icu_buf_utf16 * src16,
483 int32_t sortkey_len = 0;
485 sortkey_len = ucol_getSortKey(coll, src16->utf16, src16->utf16_len,
486 dest8->utf8, dest8->utf8_cap);
488 /* check for buffer overflow, resize and retry */
489 if (sortkey_len > dest8->utf8_cap) {
490 icu_buf_utf8_resize(dest8, sortkey_len * 2);
491 sortkey_len = ucol_getSortKey(coll, src16->utf16, src16->utf16_len,
492 dest8->utf8, dest8->utf8_cap);
495 if (U_SUCCESS(*status)
497 dest8->utf8_len = sortkey_len;
499 icu_buf_utf8_clear(dest8);
506 struct icu_tokenizer * icu_tokenizer_create(const char *locale, char action,
509 struct icu_tokenizer * tokenizer
510 = (struct icu_tokenizer *) malloc(sizeof(struct icu_tokenizer));
512 strcpy(tokenizer->locale, locale);
513 tokenizer->action = action;
515 tokenizer->buf16 = 0;
516 tokenizer->token_count = 0;
517 tokenizer->token_id = 0;
518 tokenizer->token_start = 0;
519 tokenizer->token_end = 0;
522 switch(tokenizer->action) {
526 = ubrk_open(UBRK_LINE, tokenizer->locale,
532 = ubrk_open(UBRK_SENTENCE, tokenizer->locale,
538 = ubrk_open(UBRK_WORD, tokenizer->locale,
544 = ubrk_open(UBRK_CHARACTER, tokenizer->locale,
550 = ubrk_open(UBRK_TITLE, tokenizer->locale,
554 *status = U_UNSUPPORTED_ERROR;
559 /* ICU error stuff is a very funny business */
560 if (U_SUCCESS(*status))
563 /* freeing if failed */
564 icu_tokenizer_destroy(tokenizer);
568 void icu_tokenizer_destroy(struct icu_tokenizer * tokenizer)
572 ubrk_close(tokenizer->bi);
577 int icu_tokenizer_attach(struct icu_tokenizer * tokenizer,
578 struct icu_buf_utf16 * src16,
581 if (!tokenizer || !tokenizer->bi || !src16)
585 tokenizer->buf16 = src16;
586 tokenizer->token_count = 0;
587 tokenizer->token_id = 0;
588 tokenizer->token_start = 0;
589 tokenizer->token_end = 0;
591 ubrk_setText(tokenizer->bi, src16->utf16, src16->utf16_len, status);
594 if (U_FAILURE(*status))
600 int32_t icu_tokenizer_next_token(struct icu_tokenizer * tokenizer,
601 struct icu_buf_utf16 * tkn16,
604 int32_t tkn_start = 0;
609 if (!tokenizer || !tokenizer->bi
610 || !tokenizer->buf16 || !tokenizer->buf16->utf16_len)
614 never change tokenizer->buf16 and keep always invariant
615 0 <= tokenizer->token_start
616 <= tokenizer->token_end
617 <= tokenizer->buf16->utf16_len
618 returns length of token
621 if (0 == tokenizer->token_end) /* first call */
622 tkn_start = ubrk_first(tokenizer->bi);
623 else /* successive calls */
624 tkn_start = tokenizer->token_end;
626 /* get next position */
627 tkn_end = ubrk_next(tokenizer->bi);
629 /* repairing invariant at end of ubrk, which is UBRK_DONE = -1 */
630 if (UBRK_DONE == tkn_end)
631 tkn_end = tokenizer->buf16->utf16_len;
633 /* copy out if everything is well */
634 if(U_FAILURE(*status))
637 /* everything OK, now update internal state */
638 tkn_len = tkn_end - tkn_start;
641 tokenizer->token_count++;
642 tokenizer->token_id++;
644 tokenizer->token_id = 0;
646 tokenizer->token_start = tkn_start;
647 tokenizer->token_end = tkn_end;
650 /* copying into token buffer if it exists */
652 if (tkn16->utf16_cap < tkn_len)
653 icu_buf_utf16_resize(tkn16, (size_t) tkn_len * 2);
655 u_strncpy(tkn16->utf16, &(tokenizer->buf16->utf16)[tkn_start],
658 tkn16->utf16_len = tkn_len;
665 int32_t icu_tokenizer_token_id(struct icu_tokenizer * tokenizer)
667 return tokenizer->token_id;
670 int32_t icu_tokenizer_token_start(struct icu_tokenizer * tokenizer)
672 return tokenizer->token_start;
675 int32_t icu_tokenizer_token_end(struct icu_tokenizer * tokenizer)
677 return tokenizer->token_end;
680 int32_t icu_tokenizer_token_length(struct icu_tokenizer * tokenizer)
682 return (tokenizer->token_end - tokenizer->token_start);
685 int32_t icu_tokenizer_token_count(struct icu_tokenizer * tokenizer)
687 return tokenizer->token_count;
692 struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
696 struct icu_normalizer * normalizer
697 = (struct icu_normalizer *) malloc(sizeof(struct icu_normalizer));
699 normalizer->action = action;
700 normalizer->trans = 0;
701 normalizer->rules16 = icu_buf_utf16_create(0);
702 icu_utf16_from_utf8_cstr(normalizer->rules16, rules, status);
704 switch(normalizer->action) {
708 = utrans_openU(normalizer->rules16->utf16,
709 normalizer->rules16->utf16_len,
712 normalizer->parse_error, status);
717 = utrans_openU(normalizer->rules16->utf16,
718 normalizer->rules16->utf16_len,
721 normalizer->parse_error, status);
724 *status = U_UNSUPPORTED_ERROR;
729 if (U_SUCCESS(*status))
732 /* freeing if failed */
733 icu_normalizer_destroy(normalizer);
738 void icu_normalizer_destroy(struct icu_normalizer * normalizer){
740 if (normalizer->rules16)
741 icu_buf_utf16_destroy(normalizer->rules16);
742 if (normalizer->trans)
743 utrans_close(normalizer->trans);
750 int icu_normalizer_normalize(struct icu_normalizer * normalizer,
751 struct icu_buf_utf16 * dest16,
752 struct icu_buf_utf16 * src16,
755 if (!normalizer || !normalizer->trans
760 if (!src16->utf16_len){ //guarding for empty source string
761 icu_buf_utf16_clear(dest16);
765 if (!icu_buf_utf16_copy(dest16, src16))
769 utrans_transUChars (normalizer->trans,
770 dest16->utf16, &(dest16->utf16_len),
772 0, &(src16->utf16_len), status);
774 if (U_FAILURE(*status))
775 icu_buf_utf16_clear(dest16);
777 return dest16->utf16_len;
783 struct icu_chain_step * icu_chain_step_create(struct icu_chain * chain,
784 enum icu_chain_step_type type,
785 const uint8_t * rule,
786 struct icu_buf_utf16 * buf16,
789 struct icu_chain_step * step = 0;
791 if(!chain || !type || !rule)
794 step = (struct icu_chain_step *) malloc(sizeof(struct icu_chain_step));
800 /* create auxilary objects */
802 case ICU_chain_step_type_display:
804 case ICU_chain_step_type_casemap:
805 step->u.casemap = icu_casemap_create(rule[0], status);
807 case ICU_chain_step_type_normalize:
808 step->u.normalizer = icu_normalizer_create((char *) rule, 'f', status);
810 case ICU_chain_step_type_tokenize:
811 step->u.tokenizer = icu_tokenizer_create((char *) chain->locale,
812 (char) rule[0], status);
822 void icu_chain_step_destroy(struct icu_chain_step * step){
827 icu_chain_step_destroy(step->previous);
830 case ICU_chain_step_type_display:
832 case ICU_chain_step_type_casemap:
833 icu_casemap_destroy(step->u.casemap);
834 icu_buf_utf16_destroy(step->buf16);
836 case ICU_chain_step_type_normalize:
837 icu_normalizer_destroy(step->u.normalizer);
838 icu_buf_utf16_destroy(step->buf16);
840 case ICU_chain_step_type_tokenize:
841 icu_tokenizer_destroy(step->u.tokenizer);
842 icu_buf_utf16_destroy(step->buf16);
852 struct icu_chain * icu_chain_create(const char *locale,
857 struct icu_chain * chain
858 = (struct icu_chain *) malloc(sizeof(struct icu_chain));
860 *status = U_ZERO_ERROR;
862 strncpy((char *) chain->locale, (const char *) locale, 16);
863 chain->locale[16 - 1] = '\0';
867 chain->coll = ucol_open((const char *) chain->locale, status);
869 if (U_FAILURE(*status))
873 chain->token_count = 0;
877 chain->display8 = icu_buf_utf8_create(0);
878 chain->norm8 = icu_buf_utf8_create(0);
879 chain->sort8 = icu_buf_utf8_create(0);
881 chain->src16 = icu_buf_utf16_create(0);
889 void icu_chain_destroy(struct icu_chain * chain)
894 ucol_close(chain->coll);
896 icu_buf_utf8_destroy(chain->display8);
897 icu_buf_utf8_destroy(chain->norm8);
898 icu_buf_utf8_destroy(chain->sort8);
900 icu_buf_utf16_destroy(chain->src16);
902 icu_chain_step_destroy(chain->steps);
909 struct icu_chain * icu_chain_xml_config(const xmlNode *xml_node,
914 struct icu_chain * chain = 0;
916 *status = U_ZERO_ERROR;
918 if (!xml_node ||xml_node->type != XML_ELEMENT_NODE)
922 xmlChar * xml_locale = xmlGetProp((xmlNode *) xml_node,
923 (xmlChar *) "locale");
927 chain = icu_chain_create((const char *) xml_locale, sort, status);
935 for (node = xml_node->children; node; node = node->next)
938 struct icu_chain_step * step = 0;
940 if (node->type != XML_ELEMENT_NODE)
943 xml_rule = xmlGetProp(node, (xmlChar *) "rule");
945 if (!strcmp((const char *) node->name,
946 (const char *) "casemap")){
947 step = icu_chain_insert_step(chain, ICU_chain_step_type_casemap,
948 (const uint8_t *) xml_rule, status);
950 else if (!strcmp((const char *) node->name,
951 (const char *) "normalize")){
952 step = icu_chain_insert_step(chain, ICU_chain_step_type_normalize,
953 (const uint8_t *) xml_rule, status);
955 else if (!strcmp((const char *) node->name,
956 (const char *) "tokenize")){
957 step = icu_chain_insert_step(chain, ICU_chain_step_type_tokenize,
958 (const uint8_t *) xml_rule, status);
960 else if (!strcmp((const char *) node->name,
961 (const char *) "display")){
962 step = icu_chain_insert_step(chain, ICU_chain_step_type_display,
963 (const uint8_t *) "", status);
966 if (!step || U_FAILURE(*status)){
967 icu_chain_destroy(chain);
979 struct icu_chain_step * icu_chain_insert_step(struct icu_chain * chain,
980 enum icu_chain_step_type type,
981 const uint8_t * rule,
984 struct icu_chain_step * step = 0;
985 struct icu_buf_utf16 * src16 = 0;
986 struct icu_buf_utf16 * buf16 = 0;
988 if (!chain || !type || !rule)
991 /* assign utf16 src buffers as needed */
992 if (chain->steps && chain->steps->buf16)
993 src16 = chain->steps->buf16;
994 else if (chain->src16)
995 src16 = chain->src16;
1000 /* create utf16 destination buffers as needed, or */
1002 case ICU_chain_step_type_display:
1005 case ICU_chain_step_type_casemap:
1006 buf16 = icu_buf_utf16_create(0);
1008 case ICU_chain_step_type_normalize:
1009 buf16 = icu_buf_utf16_create(0);
1011 case ICU_chain_step_type_tokenize:
1012 buf16 = icu_buf_utf16_create(0);
1018 /* create actual chain step with this buffer */
1019 step = icu_chain_step_create(chain, type, rule, buf16, status);
1021 step->previous = chain->steps;
1022 chain->steps = step;
1028 int icu_chain_step_next_token(struct icu_chain * chain,
1029 struct icu_chain_step * step,
1032 struct icu_buf_utf16 * src16 = 0;
1033 int got_new_token = 0;
1035 if (!chain || !chain->src16 || !step || !step->more_tokens)
1038 /* assign utf16 src buffers as neeed, advance in previous steps
1039 tokens until non-zero token met, and setting stop condition */
1041 if (step->previous){
1042 src16 = step->previous->buf16;
1043 /* tokens might be killed in previous steps, therefore looping */
1045 while (step->need_new_token
1046 && step->previous->more_tokens
1049 = icu_chain_step_next_token(chain, step->previous, status);
1051 else { /* first step can only work once on chain->src16 input buffer */
1052 src16 = chain->src16;
1053 step->more_tokens = 0;
1060 /* stop if nothing to process */
1061 if (step->need_new_token && !got_new_token){
1062 step->more_tokens = 0;
1066 /* either an old token not finished yet, or a new token, thus
1067 perform the work, eventually put this steps output in
1068 step->buf16 or the chains UTF8 output buffers */
1070 switch(step->type) {
1071 case ICU_chain_step_type_display:
1072 icu_utf16_to_utf8(chain->display8, src16, status);
1074 case ICU_chain_step_type_casemap:
1075 icu_casemap_casemap(step->u.casemap,
1076 step->buf16, src16, status,
1079 case ICU_chain_step_type_normalize:
1080 icu_normalizer_normalize(step->u.normalizer,
1081 step->buf16, src16, status);
1083 case ICU_chain_step_type_tokenize:
1084 /* attach to new src16 token only first time during splitting */
1085 if (step->need_new_token){
1086 icu_tokenizer_attach(step->u.tokenizer, src16, status);
1087 step->need_new_token = 0;
1091 /* splitting one src16 token into multiple buf16 tokens */
1093 = icu_tokenizer_next_token(step->u.tokenizer,
1094 step->buf16, status);
1096 /* make sure to get new previous token if this one had been used up
1097 by recursive call to _same_ step */
1099 if (!step->more_tokens){
1100 step->more_tokens = icu_chain_step_next_token(chain, step, status);
1101 return step->more_tokens; // avoid one token count too much!
1110 if (U_FAILURE(*status))
1113 /* if token disappered into thin air, tell caller */
1114 /* if (!step->buf16->utf16_len && !step->more_tokens) */
1121 int icu_chain_assign_cstr(struct icu_chain * chain,
1122 const char * src8cstr,
1125 struct icu_chain_step * stp = 0;
1127 if (!chain || !src8cstr)
1130 chain->src8cstr = src8cstr;
1134 /* clear token count */
1135 chain->token_count = 0;
1137 /* clear all steps stop states */
1139 stp->more_tokens = 1;
1140 stp->need_new_token = 1;
1141 stp = stp->previous;
1144 /* finally convert UTF8 to UTF16 string if needed */
1145 if (chain->steps || chain->sort)
1146 icu_utf16_from_utf8_cstr(chain->src16, chain->src8cstr, status);
1148 if (U_FAILURE(*status))
1156 int icu_chain_next_token(struct icu_chain * chain,
1161 *status = U_ZERO_ERROR;
1166 /* special case with no steps - same as index type binary */
1168 if (chain->token_count)
1171 chain->token_count++;
1174 icu_sortkey8_from_utf16(chain->coll,
1175 chain->sort8, chain->steps->buf16,
1177 return chain->token_count;
1180 /* usual case, one or more icu chain steps existing */
1183 while(!got_token && chain->steps && chain->steps->more_tokens)
1184 got_token = icu_chain_step_next_token(chain, chain->steps, status);
1187 chain->token_count++;
1189 icu_utf16_to_utf8(chain->norm8, chain->steps->buf16, status);
1192 icu_sortkey8_from_utf16(chain->coll,
1193 chain->sort8, chain->steps->buf16,
1196 return chain->token_count;
1203 int icu_chain_token_number(struct icu_chain * chain)
1208 return chain->token_count;
1212 const char * icu_chain_token_display(struct icu_chain * chain)
1214 if (chain->display8)
1215 return icu_buf_utf8_to_cstr(chain->display8);
1220 const char * icu_chain_token_norm(struct icu_chain * chain)
1223 return chain->src8cstr;
1226 return icu_buf_utf8_to_cstr(chain->norm8);
1231 const char * icu_chain_token_sortkey(struct icu_chain * chain)
1234 return icu_buf_utf8_to_cstr(chain->sort8);
1239 const UCollator * icu_chain_get_coll(struct icu_chain * chain)
1245 #endif /* YAZ_HAVE_ICU */
1253 * indent-tabs-mode: nil
1255 * vim: shiftwidth=4 tabstop=8 expandtab