-/*
- * Copyright (C) 1995-2007, Index Data ApS
+/* This file is part of the YAZ toolkit.
+ * Copyright (C) 1995-2009 Index Data
* See the file LICENSE for details.
- *
- * $Id: icu_I18N.c,v 1.4 2007-10-24 13:23:34 marc Exp $
+ */
+
+/**
+ * \file icu_I18N.c
+ * \brief ICU utilities
*/
#if HAVE_CONFIG_H
#include <yaz/timing.h>
#endif
+#if YAZ_HAVE_ICU
+#include <yaz/xmalloc.h>
-#if HAVE_ICU
#include <yaz/icu_I18N.h>
#include <yaz/log.h>
int icu_check_status (UErrorCode status)
{
- if(U_FAILURE(status)){
- yaz_log(YLOG_WARN,
- "ICU: %d %s\n", status, u_errorName(status));
+ if (U_FAILURE(status))
+ {
+ yaz_log(YLOG_WARN, "ICU: %d %s\n", status, u_errorName(status));
return 0;
}
return 1;
struct icu_buf_utf16 * icu_buf_utf16_create(size_t capacity)
{
struct icu_buf_utf16 * buf16
- = (struct icu_buf_utf16 *) malloc(sizeof(struct icu_buf_utf16));
+ = (struct icu_buf_utf16 *) xmalloc(sizeof(struct icu_buf_utf16));
buf16->utf16 = 0;
buf16->utf16_len = 0;
buf16->utf16_cap = 0;
if (capacity > 0){
- buf16->utf16 = (UChar *) malloc(sizeof(UChar) * capacity);
+ buf16->utf16 = (UChar *) xmalloc(sizeof(UChar) * capacity);
buf16->utf16[0] = (UChar) 0;
buf16->utf16_cap = capacity;
}
if (capacity > 0){
if (0 == buf16->utf16)
- buf16->utf16 = (UChar *) malloc(sizeof(UChar) * capacity);
+ buf16->utf16 = (UChar *) xmalloc(sizeof(UChar) * capacity);
else
buf16->utf16
- = (UChar *) realloc(buf16->utf16, sizeof(UChar) * capacity);
+ = (UChar *) xrealloc(buf16->utf16, sizeof(UChar) * capacity);
icu_buf_utf16_clear(buf16);
buf16->utf16_cap = capacity;
}
else {
- if (buf16->utf16)
- free(buf16->utf16);
+ xfree(buf16->utf16);
buf16->utf16 = 0;
buf16->utf16_len = 0;
buf16->utf16_cap = 0;
void icu_buf_utf16_destroy(struct icu_buf_utf16 * buf16)
{
- if (buf16){
- if (buf16->utf16)
- free(buf16->utf16);
- free(buf16);
- }
+ if (buf16)
+ xfree(buf16->utf16);
+ xfree(buf16);
}
-
-
-
struct icu_buf_utf8 * icu_buf_utf8_create(size_t capacity)
{
struct icu_buf_utf8 * buf8
- = (struct icu_buf_utf8 *) malloc(sizeof(struct icu_buf_utf8));
+ = (struct icu_buf_utf8 *) xmalloc(sizeof(struct icu_buf_utf8));
buf8->utf8 = 0;
buf8->utf8_len = 0;
buf8->utf8_cap = 0;
if (capacity > 0){
- buf8->utf8 = (uint8_t *) malloc(sizeof(uint8_t) * capacity);
+ buf8->utf8 = (uint8_t *) xmalloc(sizeof(uint8_t) * capacity);
buf8->utf8[0] = (uint8_t) 0;
buf8->utf8_cap = capacity;
}
if (capacity > 0){
if (0 == buf8->utf8)
- buf8->utf8 = (uint8_t *) malloc(sizeof(uint8_t) * capacity);
+ buf8->utf8 = (uint8_t *) xmalloc(sizeof(uint8_t) * capacity);
else
buf8->utf8
- = (uint8_t *) realloc(buf8->utf8, sizeof(uint8_t) * capacity);
-
- icu_buf_utf8_clear(buf8);
+ = (uint8_t *) xrealloc(buf8->utf8, sizeof(uint8_t) * capacity);
+
buf8->utf8_cap = capacity;
}
else {
- if (buf8->utf8)
- free(buf8->utf8);
+ xfree(buf8->utf8);
buf8->utf8 = 0;
buf8->utf8_len = 0;
buf8->utf8_cap = 0;
}
-struct icu_buf_utf8 * icu_buf_utf8_copy(struct icu_buf_utf8 * dest8,
- struct icu_buf_utf8 * src8)
-{
- if(!dest8 || !src8
- || dest8 == src8)
- return 0;
-
-
- if (dest8->utf8_cap < src8->utf8_len)
- icu_buf_utf8_resize(dest8, src8->utf8_len * 2);
-
- strncpy((char*) dest8->utf8, (char*) src8->utf8, src8->utf8_len);
-
- return dest8;
-}
-
-
const char *icu_buf_utf8_to_cstr(struct icu_buf_utf8 *src8)
{
if (!src8 || src8->utf8_len == 0)
void icu_buf_utf8_destroy(struct icu_buf_utf8 * buf8)
{
- if (buf8){
- if (buf8->utf8)
- free(buf8->utf8);
- free(buf8);
- }
+ if (buf8)
+ xfree(buf8->utf8);
+ xfree(buf8);
}
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,
-struct icu_casemap * icu_casemap_create(const char *locale, char action,
- UErrorCode *status)
+struct icu_casemap * icu_casemap_create(char action, UErrorCode *status)
{
struct icu_casemap * casemap
- = (struct icu_casemap *) malloc(sizeof(struct icu_casemap));
- strcpy(casemap->locale, locale);
+ = (struct icu_casemap *) xmalloc(sizeof(struct icu_casemap));
casemap->action = action;
switch(casemap->action) {
void icu_casemap_destroy(struct icu_casemap * casemap)
{
- if (casemap)
- free(casemap);
+ xfree(casemap);
}
int icu_casemap_casemap(struct icu_casemap * casemap,
struct icu_buf_utf16 * dest16,
struct icu_buf_utf16 * src16,
- UErrorCode *status)
+ UErrorCode *status,
+ const char *locale)
{
if(!casemap)
return 0;
- return icu_utf16_casemap(dest16, src16,
- casemap->locale, casemap->action, status);
+ return icu_utf16_casemap(dest16, src16, locale,
+ casemap->action, status);
}
int32_t dest16_len = 0;
- if (!src16->utf16_len){ //guarding for empty source string
+ if (!src16->utf16_len){ /* guarding for empty source string */
if (dest16->utf16)
dest16->utf16[0] = (UChar) 0;
dest16->utf16_len = 0;
-UErrorCode icu_sortkey8_from_utf16(UCollator *coll,
- struct icu_buf_utf8 * dest8,
- struct icu_buf_utf16 * src16,
- UErrorCode * status)
+void icu_sortkey8_from_utf16(UCollator *coll,
+ struct icu_buf_utf8 * dest8,
+ struct icu_buf_utf16 * src16,
+ UErrorCode * status)
{
int32_t sortkey_len = 0;
dest8->utf8_len = sortkey_len;
else
icu_buf_utf8_clear(dest8);
-
- return sortkey_len;
}
UErrorCode *status)
{
struct icu_tokenizer * tokenizer
- = (struct icu_tokenizer *) malloc(sizeof(struct icu_tokenizer));
+ = (struct icu_tokenizer *) xmalloc(sizeof(struct icu_tokenizer));
- strcpy(tokenizer->locale, locale);
tokenizer->action = action;
tokenizer->bi = 0;
tokenizer->buf16 = 0;
switch(tokenizer->action) {
case 'l':
case 'L':
- tokenizer->bi
- = ubrk_open(UBRK_LINE, tokenizer->locale,
- 0, 0, status);
+ tokenizer->bi = ubrk_open(UBRK_LINE, locale, 0, 0, status);
break;
case 's':
case 'S':
- tokenizer->bi
- = ubrk_open(UBRK_SENTENCE, tokenizer->locale,
- 0, 0, status);
+ tokenizer->bi = ubrk_open(UBRK_SENTENCE, locale, 0, 0, status);
break;
case 'w':
case 'W':
- tokenizer->bi
- = ubrk_open(UBRK_WORD, tokenizer->locale,
- 0, 0, status);
+ tokenizer->bi = ubrk_open(UBRK_WORD, locale, 0, 0, status);
break;
case 'c':
case 'C':
- tokenizer->bi
- = ubrk_open(UBRK_CHARACTER, tokenizer->locale,
- 0, 0, status);
+ tokenizer->bi = ubrk_open(UBRK_CHARACTER, locale, 0, 0, status);
break;
case 't':
case 'T':
- tokenizer->bi
- = ubrk_open(UBRK_TITLE, tokenizer->locale,
- 0, 0, status);
+ tokenizer->bi = ubrk_open(UBRK_TITLE, locale, 0, 0, status);
break;
default:
*status = U_UNSUPPORTED_ERROR;
if (tokenizer) {
if (tokenizer->bi)
ubrk_close(tokenizer->bi);
- free(tokenizer);
+ xfree(tokenizer);
}
}
-struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
- UErrorCode *status)
+struct icu_transform * icu_transform_create(const char *id, char action,
+ const char *rules,
+ UErrorCode *status)
{
+ struct icu_buf_utf16 *id16 = icu_buf_utf16_create(0);
+ struct icu_buf_utf16 *rules16 = icu_buf_utf16_create(0);
+
+ struct icu_transform * transform
+ = (struct icu_transform *) xmalloc(sizeof(struct icu_transform));
- struct icu_normalizer * normalizer
- = (struct icu_normalizer *) malloc(sizeof(struct icu_normalizer));
+ transform->action = action;
+ transform->trans = 0;
- normalizer->action = action;
- normalizer->trans = 0;
- normalizer->rules16 = icu_buf_utf16_create(0);
- icu_utf16_from_utf8_cstr(normalizer->rules16, rules, status);
-
- switch(normalizer->action) {
+ if (id)
+ icu_utf16_from_utf8_cstr(id16, id, status);
+ if (rules)
+ icu_utf16_from_utf8_cstr(rules16, rules, status);
+
+ switch(transform->action)
+ {
case 'f':
case 'F':
- normalizer->trans
- = utrans_openU(normalizer->rules16->utf16,
- normalizer->rules16->utf16_len,
+ transform->trans
+ = utrans_openU(id16->utf16,
+ id16->utf16_len,
UTRANS_FORWARD,
- 0, 0,
- normalizer->parse_error, status);
+ rules16->utf16,
+ rules16->utf16_len,
+ &transform->parse_error, status);
break;
case 'r':
case 'R':
- normalizer->trans
- = utrans_openU(normalizer->rules16->utf16,
- normalizer->rules16->utf16_len,
+ transform->trans
+ = utrans_openU(id16->utf16,
+ id16->utf16_len,
UTRANS_REVERSE ,
- 0, 0,
- normalizer->parse_error, status);
+ rules16->utf16,
+ rules16->utf16_len,
+ &transform->parse_error, status);
break;
default:
*status = U_UNSUPPORTED_ERROR;
- return 0;
break;
}
+ icu_buf_utf16_destroy(rules16);
+ icu_buf_utf16_destroy(id16);
if (U_SUCCESS(*status))
- return normalizer;
+ return transform;
/* freeing if failed */
- icu_normalizer_destroy(normalizer);
+ icu_transform_destroy(transform);
return 0;
}
-void icu_normalizer_destroy(struct icu_normalizer * normalizer){
- if (normalizer) {
- if (normalizer->rules16)
- icu_buf_utf16_destroy(normalizer->rules16);
- if (normalizer->trans)
- utrans_close(normalizer->trans);
- free(normalizer);
+void icu_transform_destroy(struct icu_transform * transform){
+ if (transform) {
+ if (transform->trans)
+ utrans_close(transform->trans);
+ xfree(transform);
}
}
-int icu_normalizer_normalize(struct icu_normalizer * normalizer,
- struct icu_buf_utf16 * dest16,
- struct icu_buf_utf16 * src16,
- UErrorCode *status)
+int icu_transform_trans(struct icu_transform * transform,
+ struct icu_buf_utf16 * dest16,
+ struct icu_buf_utf16 * src16,
+ UErrorCode *status)
{
- if (!normalizer || !normalizer->trans
+ if (!transform || !transform->trans
|| !src16
|| !dest16)
return 0;
- if (!src16->utf16_len){ //guarding for empty source string
+ if (!src16->utf16_len){ /* guarding for empty source string */
icu_buf_utf16_clear(dest16);
return 0;
}
return 0;
- utrans_transUChars (normalizer->trans,
+ utrans_transUChars (transform->trans,
dest16->utf16, &(dest16->utf16_len),
dest16->utf16_cap,
0, &(src16->utf16_len), status);
if(!chain || !type || !rule)
return 0;
- step = (struct icu_chain_step *) malloc(sizeof(struct icu_chain_step));
+ step = (struct icu_chain_step *) xmalloc(sizeof(struct icu_chain_step));
step->type = type;
switch(step->type) {
case ICU_chain_step_type_display:
break;
- case ICU_chain_step_type_index:
- break;
- case ICU_chain_step_type_sortkey:
- break;
case ICU_chain_step_type_casemap:
- step->u.casemap = icu_casemap_create((char *) chain->locale,
- (char) rule[0], status);
+ step->u.casemap = icu_casemap_create(rule[0], status);
break;
- case ICU_chain_step_type_normalize:
- step->u.normalizer = icu_normalizer_create((char *) rule, 'f', status);
+ case ICU_chain_step_type_transform:
+ /* rule omitted. Only ID used */
+ step->u.transform = icu_transform_create((const char *) rule, 'f',
+ 0, status);
break;
case ICU_chain_step_type_tokenize:
step->u.tokenizer = icu_tokenizer_create((char *) chain->locale,
(char) rule[0], status);
break;
+ case ICU_chain_step_type_transliterate:
+ /* we pass a dummy ID to utrans_openU.. */
+ step->u.transform = icu_transform_create("custom", 'f',
+ (const char *) rule, status);
+ break;
default:
break;
}
switch(step->type) {
case ICU_chain_step_type_display:
break;
- case ICU_chain_step_type_index:
- break;
- case ICU_chain_step_type_sortkey:
- break;
case ICU_chain_step_type_casemap:
icu_casemap_destroy(step->u.casemap);
icu_buf_utf16_destroy(step->buf16);
break;
- case ICU_chain_step_type_normalize:
- icu_normalizer_destroy(step->u.normalizer);
+ case ICU_chain_step_type_transform:
+ case ICU_chain_step_type_transliterate:
+ icu_transform_destroy(step->u.transform);
icu_buf_utf16_destroy(step->buf16);
break;
case ICU_chain_step_type_tokenize:
default:
break;
}
- free(step);
+ xfree(step);
}
-struct icu_chain * icu_chain_create(const uint8_t * identifier,
- 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));
+ = (struct icu_chain *) xmalloc(sizeof(struct icu_chain));
- strncpy((char *) chain->identifier, (const char *) identifier, 128);
- chain->identifier[128 - 1] = '\0';
- strncpy((char *) chain->locale, (const char *) locale, 16);
- chain->locale[16 - 1] = '\0';
+ *status = U_ZERO_ERROR;
+
+ chain->locale = xstrdup(locale);
+
+ chain->sort = sort;
+
+ chain->coll = ucol_open((const char *) chain->locale, status);
+
+ if (U_FAILURE(*status))
+ return 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);
void icu_chain_destroy(struct icu_chain * chain)
{
- if (chain){
+ if (chain)
+ {
+ if (chain->coll)
+ ucol_close(chain->coll);
+
icu_buf_utf8_destroy(chain->display8);
icu_buf_utf8_destroy(chain->norm8);
icu_buf_utf8_destroy(chain->sort8);
icu_buf_utf16_destroy(chain->src16);
icu_chain_step_destroy(chain->steps);
- free(chain);
+ xfree(chain->locale);
+ xfree(chain);
}
}
-struct icu_chain * icu_chain_xml_config(xmlNode *xml_node,
- UErrorCode * status){
-
+struct icu_chain * icu_chain_xml_config(const xmlNode *xml_node,
+ int sort,
+ UErrorCode * status)
+{
xmlNode *node = 0;
struct icu_chain * chain = 0;
- if (!xml_node
- ||xml_node->type != XML_ELEMENT_NODE
- || strcmp((const char *) xml_node->name, "icu_chain"))
+ *status = U_ZERO_ERROR;
+ if (!xml_node ||xml_node->type != XML_ELEMENT_NODE)
return 0;
-
- {
- xmlChar *xml_id = xmlGetProp(xml_node, (xmlChar *) "id");
- xmlChar *xml_locale = xmlGetProp(xml_node, (xmlChar *) "locale");
-
- if (!xml_id || !strlen((const char *) xml_id)
- || !xml_locale || !strlen((const char *) xml_locale))
- return 0;
+
+ {
+ xmlChar * xml_locale = xmlGetProp((xmlNode *) xml_node,
+ (xmlChar *) "locale");
- chain = icu_chain_create((const uint8_t *) xml_id,
- (const uint8_t *) xml_locale);
+ if (xml_locale)
+ {
+ chain = icu_chain_create((const char *) xml_locale, sort, status);
+ xmlFree(xml_locale);
+ }
- xmlFree(xml_id);
- xmlFree(xml_locale);
}
if (!chain)
return 0;
-
+
for (node = xml_node->children; node; node = node->next)
{
xmlChar *xml_rule;
xml_rule = xmlGetProp(node, (xmlChar *) "rule");
- if (!strcmp((const char *) node->name,
- (const char *) "casemap")){
+ if (!strcmp((const char *) node->name, "casemap"))
step = icu_chain_insert_step(chain, ICU_chain_step_type_casemap,
(const uint8_t *) xml_rule, status);
- }
- else if (!strcmp((const char *) node->name,
- (const char *) "normalize")){
- step = icu_chain_insert_step(chain, ICU_chain_step_type_normalize,
+ else if (!strcmp((const char *) node->name, "transform"))
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_transform,
(const uint8_t *) xml_rule, status);
- }
- else if (!strcmp((const char *) node->name,
- (const char *) "tokenize")){
+ else if (!strcmp((const char *) node->name, "transliterate"))
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_transliterate,
+ (const uint8_t *) xml_rule, status);
+ else if (!strcmp((const char *) node->name, "tokenize"))
step = icu_chain_insert_step(chain, ICU_chain_step_type_tokenize,
(const uint8_t *) xml_rule, status);
- }
- else if (!strcmp((const char *) node->name,
- (const char *) "display")){
+ else if (!strcmp((const char *) node->name, "display"))
step = icu_chain_insert_step(chain, ICU_chain_step_type_display,
(const uint8_t *) "", status);
+ else if (!strcmp((const char *) node->name, "normalize"))
+ {
+ yaz_log(YLOG_WARN, "Element %s is deprecated. "
+ "Use transform instead", node->name);
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_transform,
+ (const uint8_t *) xml_rule, status);
}
- 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, "index")
+ || !strcmp((const char *) node->name, "sortkey"))
+ {
+ yaz_log(YLOG_WARN, "Element %s is no longer needed. "
+ "Remove it from the configuration", node->name);
}
- 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);
+ else
+ {
+ yaz_log(YLOG_WARN, "Unknown element %s", node->name);
+ icu_chain_destroy(chain);
+ return 0;
}
-
xmlFree(xml_rule);
- if (!step || U_FAILURE(*status)){
+ if (step && U_FAILURE(*status))
+ {
icu_chain_destroy(chain);
return 0;
}
-
-
}
-
return chain;
}
/* create utf16 destination buffers as needed, or */
- switch(type) {
+ switch(type)
+ {
case ICU_chain_step_type_display:
buf16 = src16;
break;
- case ICU_chain_step_type_index:
- buf16 = src16;
- break;
- case ICU_chain_step_type_sortkey:
- buf16 = src16;
- break;
case ICU_chain_step_type_casemap:
buf16 = icu_buf_utf16_create(0);
break;
- case ICU_chain_step_type_normalize:
+ case ICU_chain_step_type_transform:
+ case ICU_chain_step_type_transliterate:
buf16 = icu_buf_utf16_create(0);
break;
case ICU_chain_step_type_tokenize:
buf16 = icu_buf_utf16_create(0);
break;
+ break;
default:
break;
}
/* assign utf16 src buffers as neeed, advance in previous steps
tokens until non-zero token met, and setting stop condition */
- if (step->previous){
+ 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)
got_new_token
= icu_chain_step_next_token(chain, step->previous, status);
}
- else { /* first step can only work once on chain->src16 input buffer */
+ else
+ { /* first step can only work once on chain->src16 input buffer */
src16 = chain->src16;
step->more_tokens = 0;
got_new_token = 1;
return 0;
/* stop if nothing to process */
- if (step->need_new_token && !got_new_token){
+ if (step->need_new_token && !got_new_token)
+ {
step->more_tokens = 0;
return 0;
}
perform the work, eventually put this steps output in
step->buf16 or the chains UTF8 output buffers */
- switch(step->type) {
+ 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);
+ step->buf16, src16, status,
+ chain->locale);
break;
- case ICU_chain_step_type_normalize:
- icu_normalizer_normalize(step->u.normalizer,
- step->buf16, src16, status);
+ case ICU_chain_step_type_transform:
+ case ICU_chain_step_type_transliterate:
+ icu_transform_trans(step->u.transform,
+ 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){
+ 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,
by recursive call to _same_ step */
if (!step->more_tokens)
+ {
step->more_tokens = icu_chain_step_next_token(chain, step, status);
-
- //if (0 == step->more_tokens)
- //return 0;
- break;
- default:
- return 0;
- break;
- }
-
- if (U_FAILURE(*status))
- return 0;
-
- /* if token disappered into thin air, tell caller */
- if (!step->buf16->utf16_len)
- 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;
+ return step->more_tokens; /* avoid one token count too much! */
}
- /* 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;
+ /* if token disappered into thin air, tell caller */
+ /* if (!step->buf16->utf16_len && !step->more_tokens) */
+ /* return 0; */
return 1;
}
-#endif /* backup */
int icu_chain_assign_cstr(struct icu_chain * chain,
const char * src8cstr,
if (!chain || !src8cstr)
return 0;
+ chain->src8cstr = src8cstr;
+
stp = chain->steps;
/* clear token count */
chain->token_count = 0;
/* clear all steps stop states */
- while (stp){
+ while (stp)
+ {
stp->more_tokens = 1;
stp->need_new_token = 1;
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;
- got_token = icu_chain_step_next_token(chain, chain->steps, status);
-
- if (got_token){
- chain->token_count++;
- return 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
+ {
+ while(!got_token && chain->steps && chain->steps->more_tokens)
+ got_token = icu_chain_step_next_token(chain, chain->steps, status);
+
+ if (got_token)
+ {
+ 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);
return 0;
}
+const UCollator * icu_chain_get_coll(struct icu_chain * chain)
+{
+ return chain->coll;
+}
-#endif /* HAVE_ICU */
-
-
-
+#endif /* YAZ_HAVE_ICU */
/*
* Local variables:
* c-basic-offset: 4
+ * c-file-style: "Stroustrup"
* indent-tabs-mode: nil
* End:
* vim: shiftwidth=4 tabstop=8 expandtab
*/
+