X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Ficu_I18N.c;h=4e9dcf74ef7327d23291de566f40cfa8bac400d5;hb=e089d89f38b4fc61fce16eb03150776844b053f8;hp=5e0a0cb884c8046659dc11730dc0c56991561cc6;hpb=9edd0f41a0323104c2632c2291eae336ab60874c;p=yaz-moved-to-github.git

diff --git a/src/icu_I18N.c b/src/icu_I18N.c
index 5e0a0cb..4e9dcf7 100644
--- a/src/icu_I18N.c
+++ b/src/icu_I18N.c
@@ -2,7 +2,7 @@
  * Copyright (C) 1995-2007, Index Data ApS
  * See the file LICENSE for details.
  *
- * $Id: icu_I18N.c,v 1.3 2007-10-24 07:41:48 marc Exp $
+ * $Id: icu_I18N.c,v 1.14 2007-11-08 13:35:13 adam Exp $
  */
 
 #if HAVE_CONFIG_H
@@ -15,7 +15,7 @@
 #endif
 
 
-#if HAVE_ICU
+#if YAZ_HAVE_ICU
 #include <yaz/icu_I18N.h>
 
 #include <yaz/log.h>
@@ -267,6 +267,7 @@ UErrorCode icu_utf16_from_utf8_cstr(struct icu_buf_utf16 * dest16,
     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,
@@ -332,17 +333,17 @@ struct icu_casemap * icu_casemap_create(const char *locale, char action,
 {    
     struct icu_casemap * casemap
         = (struct icu_casemap *) malloc(sizeof(struct icu_casemap));
-    strcpy(casemap->locale, locale);
     casemap->action = action;
 
     switch(casemap->action) {    
     case 'l':   
-        break;
+    case 'L':   
     case 'u':   
-        break;
+    case 'U':   
     case 't':  
-        break;
+    case 'T':  
     case 'f':  
+    case 'F':  
         break;
     default:
         icu_casemap_destroy(casemap);
@@ -354,7 +355,7 @@ struct icu_casemap * icu_casemap_create(const char *locale, char action,
 
 void icu_casemap_destroy(struct icu_casemap * casemap)
 {
-    if (casemap) 
+    if (casemap)
         free(casemap);
 }
 
@@ -362,13 +363,14 @@ void icu_casemap_destroy(struct icu_casemap * 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);
 }
 
 
@@ -800,10 +802,6 @@ struct icu_chain_step * icu_chain_step_create(struct icu_chain * chain,
     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);
@@ -833,10 +831,6 @@ void icu_chain_step_destroy(struct icu_chain_step * step){
     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);
@@ -857,20 +851,31 @@ void icu_chain_step_destroy(struct icu_chain_step * 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));
 
-    strncpy((char *) chain->identifier, (const char *) identifier, 128);
-    chain->identifier[128 - 1] = '\0';
+    *status = U_ZERO_ERROR;
+
     strncpy((char *) chain->locale, (const char *) locale, 16);    
     chain->locale[16 - 1] = '\0';
 
+    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);
@@ -886,6 +891,10 @@ struct icu_chain * icu_chain_create(const uint8_t * identifier,
 void icu_chain_destroy(struct icu_chain * 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);
@@ -899,35 +908,32 @@ void icu_chain_destroy(struct icu_chain * 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;
@@ -958,17 +964,6 @@ struct icu_chain * icu_chain_xml_config(xmlNode *xml_node,
             step = icu_chain_insert_step(chain, ICU_chain_step_type_display, 
                                          (const uint8_t *) "", 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,
-                         (const char *) "sortkey")){
-            step = icu_chain_insert_step(chain, ICU_chain_step_type_sortkey, 
-                                         (const uint8_t *) "", status);
-        }
-
         xmlFree(xml_rule);
         if (!step || U_FAILURE(*status)){
             icu_chain_destroy(chain);
@@ -1009,12 +1004,6 @@ struct icu_chain_step * icu_chain_insert_step(struct icu_chain * chain,
     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;
@@ -1043,46 +1032,51 @@ int icu_chain_step_next_token(struct icu_chain * chain,
                               UErrorCode *status)
 {
     struct icu_buf_utf16 * src16 = 0;
-    
+    int got_new_token = 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
-    */
+       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 
+        /* 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 */
         src16 = chain->src16;
-        step->more_tokens = 1;
+        step->more_tokens = 0;
+        got_new_token = 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)
+    if (!src16)
         return 0;
 
-    /* perform the work, eventually put this steps output in 
+    /* stop if nothing to process */
+    if (step->need_new_token && !got_new_token){
+        step->more_tokens = 0;
+        return 0;
+    }
+
+    /* either an old token not finished yet, or a new token, thus
+       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);
+                            step->buf16, src16, status,
+                            chain->locale);
         break;
     case ICU_chain_step_type_normalize:
         icu_normalizer_normalize(step->u.normalizer,
@@ -1094,47 +1088,38 @@ int icu_chain_step_next_token(struct icu_chain * chain,
             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);
-            }
+
+        /* make sure to get new previous token if this one had been used up
+           by recursive call to _same_ step */
+
+        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;
         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;
 }
 
 
-
 int icu_chain_assign_cstr(struct icu_chain * chain,
                           const char * src8cstr, 
                           UErrorCode *status)
@@ -1144,6 +1129,8 @@ int icu_chain_assign_cstr(struct icu_chain * chain,
     if (!chain || !src8cstr)
         return 0;
 
+    chain->src8cstr = src8cstr;
+
     stp = chain->steps;
     
     /* clear token count */
@@ -1156,8 +1143,9 @@ int icu_chain_assign_cstr(struct icu_chain * chain,
         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;
@@ -1172,20 +1160,49 @@ int icu_chain_next_token(struct icu_chain * chain,
 {
     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;
@@ -1194,8 +1211,7 @@ int icu_chain_get_token_count(struct icu_chain * chain)
 }
 
 
-
-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);
@@ -1203,15 +1219,18 @@ const char * icu_chain_get_display(struct icu_chain * chain)
     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);
@@ -1219,8 +1238,13 @@ const char * icu_chain_get_sort(struct icu_chain * chain)
     return 0;
 }
 
+const UCollator * icu_chain_get_coll(struct icu_chain * chain)
+{
+    return chain->coll;
+}
+
 
-#endif /* HAVE_ICU */
+#endif /* YAZ_HAVE_ICU */