* Copyright (C) 1995-2007, Index Data ApS
* See the file LICENSE for details.
*
- * $Id: icu_I18N.c,v 1.6 2007-10-25 08:32:50 marc Exp $
+ * $Id: icu_I18N.c,v 1.13 2007-11-08 08:17:17 adam Exp $
*/
#if HAVE_CONFIG_H
#endif
-#if HAVE_ICU
+#if YAZ_HAVE_ICU
#include <yaz/icu_I18N.h>
#include <yaz/log.h>
size_t src8cstr_len = 0;
int32_t utf16_len = 0;
+ *status = U_ZERO_ERROR;
src8cstr_len = strlen(src8cstr);
u_strFromUTF8(dest16->utf16, dest16->utf16_cap,
switch(step->type) {
case ICU_chain_step_type_display:
break;
-#ifdef ICU_CHAIN_SORTKEY
- case ICU_chain_step_type_index:
- break;
- case ICU_chain_step_type_sortkey:
- break;
-#endif
case ICU_chain_step_type_casemap:
step->u.casemap = icu_casemap_create((char *) chain->locale,
(char) rule[0], status);
switch(step->type) {
case ICU_chain_step_type_display:
break;
-#ifdef ICU_CHAIN_SORTKEY
- case ICU_chain_step_type_index:
- break;
- case ICU_chain_step_type_sortkey:
- break;
-#endif
case ICU_chain_step_type_casemap:
icu_casemap_destroy(step->u.casemap);
icu_buf_utf16_destroy(step->buf16);
-struct icu_chain * icu_chain_create(const uint8_t * locale,
+struct icu_chain * icu_chain_create(const char *locale,
int sort,
UErrorCode * status)
{
struct icu_chain * chain
= (struct icu_chain *) malloc(sizeof(struct icu_chain));
+ *status = U_ZERO_ERROR;
+
strncpy((char *) chain->locale, (const char *) locale, 16);
chain->locale[16 - 1] = '\0';
chain->token_count = 0;
+ chain->src8cstr = 0;
+
chain->display8 = icu_buf_utf8_create(0);
chain->norm8 = icu_buf_utf8_create(0);
chain->sort8 = icu_buf_utf8_create(0);
chain->steps = 0;
-
return chain;
}
-struct icu_chain * icu_chain_xml_config(xmlNode *xml_node,
- const uint8_t * locale,
+struct icu_chain * icu_chain_xml_config(const xmlNode *xml_node,
+ const char *locale,
int sort,
UErrorCode * status){
xmlNode *node = 0;
struct icu_chain * chain = 0;
+ *status = U_ZERO_ERROR;
+
if (!xml_node
||xml_node->type != XML_ELEMENT_NODE
// || strcmp((const char *) xml_node->name, "icu_chain")
step = icu_chain_insert_step(chain, ICU_chain_step_type_display,
(const uint8_t *) "", status);
}
-#ifdef ICU_CHAIN_SORTKEY
- else if (!strcmp((const char *) node->name,
- (const char *) "index")){
- step = icu_chain_insert_step(chain, ICU_chain_step_type_index,
- (const uint8_t *) "", status);
- }
- else if (!strcmp((const char *) node->name,
- (const char *) "sortkey")){
- step = icu_chain_insert_step(chain, ICU_chain_step_type_sortkey,
- (const uint8_t *) "", status);
- }
-#endif
xmlFree(xml_rule);
if (!step || U_FAILURE(*status)){
icu_chain_destroy(chain);
case ICU_chain_step_type_display:
buf16 = src16;
break;
-#ifdef ICU_CHAIN_SORTKEY
- case ICU_chain_step_type_index:
- buf16 = src16;
- break;
- case ICU_chain_step_type_sortkey:
- buf16 = src16;
- break;
-#endif
case ICU_chain_step_type_casemap:
buf16 = icu_buf_utf16_create(0);
break;
if (step->previous){
src16 = step->previous->buf16;
/* tokens might be killed in previous steps, therefore looping */
+
while (step->need_new_token
&& step->previous->more_tokens
&& !got_new_token)
case ICU_chain_step_type_display:
icu_utf16_to_utf8(chain->display8, src16, status);
break;
-#ifdef ICU_CHAIN_SORTKEY
- // TODO
- case ICU_chain_step_type_index:
- icu_utf16_to_utf8(chain->norm8, src16, status);
- break;
- case ICU_chain_step_type_sortkey:
- icu_utf16_to_utf8(chain->sort8, src16, status);
- //UErrorCode icu_sortkey8_from_utf16(UCollator *coll,
- // struct icu_buf_utf8 * dest8,
- // struct icu_buf_utf16 * src16,
- // UErrorCode * status);
- break;
-#endif
case ICU_chain_step_type_casemap:
icu_casemap_casemap(step->u.casemap,
step->buf16, src16, status);
/* make sure to get new previous token if this one had been used up
by recursive call to _same_ step */
- if (!step->more_tokens)
+ if (!step->more_tokens){
step->more_tokens = icu_chain_step_next_token(chain, step, status);
+ return step->more_tokens; // avoid one token count too much!
+ }
- //if (0 == step->more_tokens)
- //return 0;
break;
default:
return 0;
return 0;
/* if token disappered into thin air, tell caller */
- if (!step->buf16->utf16_len)
- return 0;
-
- if (U_FAILURE(*status))
- return 0;
+ /* if (!step->buf16->utf16_len && !step->more_tokens) */
+ /* return 0; */
return 1;
}
-#if 0 /* backup */
-int icu_chain_step_next_token_BAK(struct icu_chain * chain,
- struct icu_chain_step * step,
- UErrorCode *status)
-{
- struct icu_buf_utf16 * src16 = 0;
-
- if (!chain || !chain->src16 || !step || !step->more_tokens)
- return 0;
-
- /* assign utf16 src buffers as neeed, advance in previous steps
- tokens until non-zero token met, and setting stop condition
- */
- if (step->previous){
- src16 = step->previous->buf16;
- if (step->need_new_token)
- step->more_tokens
- = icu_chain_step_next_token(chain, step->previous, status);
- }
- else { /* first step can only work once on chain->src16 input buffer */
- src16 = chain->src16;
- step->more_tokens = 1;
- }
-
- /* stop if nothing to process
- i.e new token source was not properly assigned
- or I did not get any tokens from previous step
- */
- if (!step->more_tokens || !src16)
- return 0;
-
- /* perform the work, eventually put this steps output in
- step->buf16 or the chains UTF8 output buffers */
- switch(step->type) {
- case ICU_chain_step_type_display:
- icu_utf16_to_utf8(chain->display8, src16, status);
- break;
- case ICU_chain_step_type_index:
- icu_utf16_to_utf8(chain->norm8, src16, status);
- break;
- case ICU_chain_step_type_sortkey:
- icu_utf16_to_utf8(chain->sort8, src16, status);
- break;
- case ICU_chain_step_type_casemap:
- icu_casemap_casemap(step->u.casemap,
- step->buf16, src16, status);
- break;
- case ICU_chain_step_type_normalize:
- icu_normalizer_normalize(step->u.normalizer,
- step->buf16, src16, status);
- break;
- case ICU_chain_step_type_tokenize:
- /* attach to new src16 token only first time during splitting */
- if (step->need_new_token){
- icu_tokenizer_attach(step->u.tokenizer, src16, status);
- step->need_new_token = 0;
- }
- /* splitting one src16 token into multiple buf16 tokens */
- step->more_tokens
- = icu_tokenizer_next_token(step->u.tokenizer,
- step->buf16, status);
- /* make sure to get new previous token if this one had been used up */
- if (step->previous && !step->more_tokens){
- if (icu_chain_step_next_token(chain, step->previous, status)){
- icu_tokenizer_attach(step->u.tokenizer, src16, status);
- step->need_new_token = 0;
- step->more_tokens
- = icu_tokenizer_next_token(step->u.tokenizer,
- step->buf16, status);
- }
- }
- if (0 == step->more_tokens)
- return 0;
- break;
- default:
- return 0;
- break;
- }
-
- /* stop further token processing if last step and
- new tokens can ot be obtained from last non-existing step
- */
- if (!step->previous)
- step->more_tokens = 0;
-
- if (U_FAILURE(*status))
- return 0;
-
- // result buf empty, got no token TODO: make while loop around all token
- // fetching operations
- //if (!step->buf16->utf16_len)
- // return 0;
-
- return 1;
-}
-
-#endif /* backup */
-
int icu_chain_assign_cstr(struct icu_chain * chain,
const char * src8cstr,
UErrorCode *status)
if (!chain || !src8cstr)
return 0;
+ chain->src8cstr = src8cstr;
+
stp = chain->steps;
/* clear token count */
stp = stp->previous;
}
- /* finally convert UTF8 to UTF16 string */
- icu_utf16_from_utf8_cstr(chain->src16, src8cstr, status);
+ /* finally convert UTF8 to UTF16 string if needed */
+ if (chain->steps || chain->sort)
+ icu_utf16_from_utf8_cstr(chain->src16, chain->src8cstr, status);
if (U_FAILURE(*status))
return 0;
{
int got_token = 0;
- if (!chain || !chain->steps)
+ *status = U_ZERO_ERROR;
+
+ if (!chain)
return 0;
- while(!got_token && chain->steps->more_tokens)
- got_token = icu_chain_step_next_token(chain, chain->steps, status);
-
- if (got_token){
- chain->token_count++;
+ /* special case with no steps - same as index type binary */
+ if (!chain->steps){
+ if (chain->token_count)
+ return 0;
+ else {
+ chain->token_count++;
+
+ if (chain->sort)
+ icu_sortkey8_from_utf16(chain->coll,
+ chain->sort8, chain->steps->buf16,
+ status);
+ return chain->token_count;
+ }
+ }
+ /* usual case, one or more icu chain steps existing */
+ else {
- icu_utf16_to_utf8(chain->norm8, chain->steps->buf16, status);
+ while(!got_token && chain->steps && chain->steps->more_tokens)
+ got_token = icu_chain_step_next_token(chain, chain->steps, status);
- icu_sortkey8_from_utf16(chain->coll,
- chain->sort8, chain->steps->buf16, status);
+ if (got_token){
+ chain->token_count++;
- return chain->token_count;
+ icu_utf16_to_utf8(chain->norm8, chain->steps->buf16, status);
+
+ if (chain->sort)
+ icu_sortkey8_from_utf16(chain->coll,
+ chain->sort8, chain->steps->buf16,
+ status);
+
+ return chain->token_count;
+ }
}
-
+
return 0;
}
-int icu_chain_get_token_count(struct icu_chain * chain)
+int icu_chain_token_number(struct icu_chain * chain)
{
if (!chain)
return 0;
}
-
-const char * icu_chain_get_display(struct icu_chain * chain)
+const char * icu_chain_token_display(struct icu_chain * chain)
{
if (chain->display8)
return icu_buf_utf8_to_cstr(chain->display8);
return 0;
}
-const char * icu_chain_get_norm(struct icu_chain * chain)
+const char * icu_chain_token_norm(struct icu_chain * chain)
{
+ if (!chain->steps)
+ return chain->src8cstr;
+
if (chain->norm8)
return icu_buf_utf8_to_cstr(chain->norm8);
return 0;
}
-const char * icu_chain_get_sort(struct icu_chain * chain)
+const char * icu_chain_token_sortkey(struct icu_chain * chain)
{
if (chain->sort8)
return icu_buf_utf8_to_cstr(chain->sort8);
}
-
-
-#endif /* HAVE_ICU */
+#endif /* YAZ_HAVE_ICU */