X-Git-Url: http://lists.indexdata.dk/cgi-bin?a=blobdiff_plain;f=src%2Ficu_I18N.c;h=230d59f02e6ece70697460a28979e78ca255c30e;hb=e005fd551460aab9ccefb561e42c746cc58fd302;hp=9e6cdd9640e758004d6f1b549cfec43c31cc0fd4;hpb=c168c38b0beb83bca23c72d9203cc796a374bfac;p=yaz-moved-to-github.git

diff --git a/src/icu_I18N.c b/src/icu_I18N.c
index 9e6cdd9..230d59f 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 1.2 2007-10-22 17:32:07 adam Exp $
+ * $Id: icu_I18N.c,v 1.5 2007-10-24 14:48:17 marc Exp $
  */
 
 #if HAVE_CONFIG_H
@@ -61,27 +61,38 @@ struct icu_buf_utf16 * icu_buf_utf16_create(size_t capacity)
     return buf16;
 }
 
-struct icu_buf_utf16 * icu_buf_utf16_resize(struct icu_buf_utf16 * buf16,
-                                            size_t capacity)
+struct icu_buf_utf16 * icu_buf_utf16_clear(struct icu_buf_utf16 * buf16)
 {
     if (buf16){
-        if (capacity >  0){
-            if (0 == buf16->utf16)
-                buf16->utf16 = (UChar *) malloc(sizeof(UChar) * capacity);
-            else
-                buf16->utf16 
-                    = (UChar *) realloc(buf16->utf16, sizeof(UChar) * capacity);
+        if (buf16->utf16)
             buf16->utf16[0] = (UChar) 0;
-            buf16->utf16_len = 0;
-            buf16->utf16_cap = capacity;
-        } 
-        else { 
-            if (buf16->utf16)
-                free(buf16->utf16);
-            buf16->utf16 = 0;
-            buf16->utf16_len = 0;
-            buf16->utf16_cap = 0;
-        }
+        buf16->utf16_len = 0;
+    }
+    return buf16;
+}
+
+struct icu_buf_utf16 * icu_buf_utf16_resize(struct icu_buf_utf16 * buf16,
+                                            size_t capacity)
+{
+    if (!buf16)
+        return 0;
+    
+    if (capacity >  0){
+        if (0 == buf16->utf16)
+            buf16->utf16 = (UChar *) malloc(sizeof(UChar) * capacity);
+        else
+            buf16->utf16 
+                = (UChar *) realloc(buf16->utf16, sizeof(UChar) * capacity);
+
+        icu_buf_utf16_clear(buf16);
+        buf16->utf16_cap = capacity;
+    } 
+    else { 
+        if (buf16->utf16)
+            free(buf16->utf16);
+        buf16->utf16 = 0;
+        buf16->utf16_len = 0;
+        buf16->utf16_cap = 0;
     }
 
     return buf16;
@@ -137,29 +148,41 @@ struct icu_buf_utf8 * icu_buf_utf8_create(size_t capacity)
 }
 
 
+struct icu_buf_utf8 * icu_buf_utf8_clear(struct icu_buf_utf8 * buf8)
+{
+    if (buf8){
+        if (buf8->utf8)
+            buf8->utf8[0] = (uint8_t) 0;
+        buf8->utf8_len = 0;
+    }
+    return buf8;
+}
+
 
 struct icu_buf_utf8 * icu_buf_utf8_resize(struct icu_buf_utf8 * buf8,
                                           size_t capacity)
 {
-    if (buf8){
-        if (capacity >  0){
-            if (0 == buf8->utf8)
-                buf8->utf8 = (uint8_t *) malloc(sizeof(uint8_t) * capacity);
-            else
-                buf8->utf8 
-                    = (uint8_t *) realloc(buf8->utf8, 
-                                          sizeof(uint8_t) * capacity);
-            buf8->utf8_cap = capacity;
-        } 
-        else { 
-            if (buf8->utf8)
-                free(buf8->utf8);
-            buf8->utf8 = 0;
-            buf8->utf8_len = 0;
-            buf8->utf8_cap = 0;
-        }
-    }
+    if (!buf8)
+        return 0;
 
+    if (capacity >  0){
+        if (0 == buf8->utf8)
+            buf8->utf8 = (uint8_t *) malloc(sizeof(uint8_t) * capacity);
+        else
+            buf8->utf8 
+                = (uint8_t *) realloc(buf8->utf8, sizeof(uint8_t) * capacity);
+
+        icu_buf_utf8_clear(buf8);
+        buf8->utf8_cap = capacity;
+    } 
+    else { 
+        if (buf8->utf8)
+            free(buf8->utf8);
+        buf8->utf8 = 0;
+        buf8->utf8_len = 0;
+        buf8->utf8_cap = 0;
+    }
+    
     return buf8;
 }
 
@@ -185,9 +208,12 @@ const char *icu_buf_utf8_to_cstr(struct icu_buf_utf8 *src8)
 {
     if (!src8 || src8->utf8_len == 0)
         return "";
+
     if (src8->utf8_len == src8->utf8_cap)
         src8 = icu_buf_utf8_resize(src8, src8->utf8_len * 2 + 1);
+
     src8->utf8[src8->utf8_len] = '\0';
+
     return (const char *) src8->utf8;
 }
 
@@ -226,10 +252,8 @@ UErrorCode icu_utf16_from_utf8(struct icu_buf_utf16 * dest16,
     if (U_SUCCESS(*status)  
         && utf16_len <= dest16->utf16_cap)
         dest16->utf16_len = utf16_len;
-    else {
-        dest16->utf16[0] = (UChar) 0;
-        dest16->utf16_len = 0;
-    }
+    else 
+        icu_buf_utf16_clear(dest16);
   
     return *status;
 }
@@ -262,10 +286,8 @@ UErrorCode icu_utf16_from_utf8_cstr(struct icu_buf_utf16 * dest16,
     if (U_SUCCESS(*status)  
         && utf16_len <= dest16->utf16_cap)
         dest16->utf16_len = utf16_len;
-    else {
-        dest16->utf16[0] = (UChar) 0;
-        dest16->utf16_len = 0;
-    }
+    else 
+        icu_buf_utf16_clear(dest16);
   
     return *status;
 }
@@ -297,10 +319,8 @@ UErrorCode icu_utf16_to_utf8(struct icu_buf_utf8 * dest8,
     if (U_SUCCESS(*status)  
         && utf8_len <= dest8->utf8_cap)
         dest8->utf8_len = utf8_len;
-    else {
-        dest8->utf8[0] = (uint8_t) 0;
-        dest8->utf8_len = 0;
-    }
+    else 
+        icu_buf_utf8_clear(dest8);
   
     return *status;
 }
@@ -317,12 +337,13 @@ struct icu_casemap * icu_casemap_create(const char *locale, char 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);
@@ -358,24 +379,37 @@ int icu_utf16_casemap(struct icu_buf_utf16 * dest16,
                       UErrorCode *status)
 {
     int32_t dest16_len = 0;
+
+
+    if (!src16->utf16_len){           //guarding for empty source string
+        if (dest16->utf16)
+            dest16->utf16[0] = (UChar) 0;
+        dest16->utf16_len = 0;
+        return U_ZERO_ERROR;
+    }
+
     
     switch(action) {    
     case 'l':    
+    case 'L':    
         dest16_len = u_strToLower(dest16->utf16, dest16->utf16_cap,
                                   src16->utf16, src16->utf16_len, 
                                   locale, status);
         break;
     case 'u':    
+    case 'U':    
         dest16_len = u_strToUpper(dest16->utf16, dest16->utf16_cap,
                                   src16->utf16, src16->utf16_len, 
                                   locale, status);
         break;
     case 't':    
+    case 'T':    
         dest16_len = u_strToTitle(dest16->utf16, dest16->utf16_cap,
                                   src16->utf16, src16->utf16_len,
                                   0, locale, status);
         break;
     case 'f':    
+    case 'F':    
         dest16_len = u_strFoldCase(dest16->utf16, dest16->utf16_cap,
                                    src16->utf16, src16->utf16_len,
                                    U_FOLD_CASE_DEFAULT, status);
@@ -396,21 +430,25 @@ int icu_utf16_casemap(struct icu_buf_utf16 * dest16,
     
         switch(action) {    
         case 'l':    
+        case 'L':    
             dest16_len = u_strToLower(dest16->utf16, dest16->utf16_cap,
                                       src16->utf16, src16->utf16_len, 
                                       locale, status);
             break;
         case 'u':    
+        case 'U':    
             dest16_len = u_strToUpper(dest16->utf16, dest16->utf16_cap,
                                       src16->utf16, src16->utf16_len, 
                                       locale, status);
             break;
         case 't':    
+        case 'T':    
             dest16_len = u_strToTitle(dest16->utf16, dest16->utf16_cap,
                                       src16->utf16, src16->utf16_len,
                                       0, locale, status);
             break;
         case 'f':    
+        case 'F':    
             dest16_len = u_strFoldCase(dest16->utf16, dest16->utf16_cap,
                                        src16->utf16, src16->utf16_len,
                                        U_FOLD_CASE_DEFAULT, status);
@@ -426,7 +464,8 @@ int icu_utf16_casemap(struct icu_buf_utf16 * dest16,
         && dest16_len <= dest16->utf16_cap)
         dest16->utf16_len = dest16_len;
     else {
-        dest16->utf16[0] = (UChar) 0;
+        if (dest16->utf16)
+            dest16->utf16[0] = (UChar) 0;
         dest16->utf16_len = 0;
     }
   
@@ -456,11 +495,9 @@ UErrorCode icu_sortkey8_from_utf16(UCollator *coll,
     if (U_SUCCESS(*status)
         && sortkey_len > 0)
         dest8->utf8_len = sortkey_len;
-    else {
-        dest8->utf8[0] = (UChar) 0;
-        dest8->utf8_len = 0;
-    }
-
+    else 
+        icu_buf_utf8_clear(dest8);
+    
     return sortkey_len;
 }
 
@@ -484,26 +521,31 @@ struct icu_tokenizer * icu_tokenizer_create(const char *locale, char action,
 
     switch(tokenizer->action) {    
     case 'l':
+    case 'L':
         tokenizer->bi
             = ubrk_open(UBRK_LINE, tokenizer->locale,
                         0, 0, status);
         break;
     case 's':
+    case 'S':
         tokenizer->bi
             = ubrk_open(UBRK_SENTENCE, tokenizer->locale,
                         0, 0, status);
         break;
     case 'w':
+    case 'W':
         tokenizer->bi 
             = ubrk_open(UBRK_WORD, tokenizer->locale,
                         0, 0, status);
         break;
     case 'c':
+    case 'C':
         tokenizer->bi 
             = ubrk_open(UBRK_CHARACTER, tokenizer->locale,
                         0, 0, status);
         break;
     case 't':
+    case 'T':
         tokenizer->bi 
             = ubrk_open(UBRK_TITLE, tokenizer->locale,
                         0, 0, status);
@@ -661,6 +703,7 @@ struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
      
     switch(normalizer->action) {    
     case 'f':
+    case 'F':
         normalizer->trans
             = utrans_openU(normalizer->rules16->utf16, 
                            normalizer->rules16->utf16_len,
@@ -669,6 +712,7 @@ struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
                            normalizer->parse_error, status);
         break;
     case 'r':
+    case 'R':
         normalizer->trans
             = utrans_openU(normalizer->rules16->utf16,
                            normalizer->rules16->utf16_len,
@@ -696,9 +740,7 @@ void icu_normalizer_destroy(struct icu_normalizer * normalizer){
         if (normalizer->rules16) 
             icu_buf_utf16_destroy(normalizer->rules16);
         if (normalizer->trans)
-        {
             utrans_close(normalizer->trans);
-        }
         free(normalizer);
     }
 }
@@ -710,21 +752,27 @@ int icu_normalizer_normalize(struct icu_normalizer * normalizer,
                              struct icu_buf_utf16 * src16,
                              UErrorCode *status)
 {
-    if (!normalizer || !normalizer->trans || !src16 || !dest16)
+    if (!normalizer || !normalizer->trans 
+        || !src16
+        || !dest16)
         return 0;
 
+    if (!src16->utf16_len){           //guarding for empty source string
+        icu_buf_utf16_clear(dest16);
+        return 0;
+    }
+
     if (!icu_buf_utf16_copy(dest16, src16))
         return 0;
 
+   
     utrans_transUChars (normalizer->trans, 
                         dest16->utf16, &(dest16->utf16_len),
                         dest16->utf16_cap,
                         0, &(src16->utf16_len), status);
 
-    if (U_FAILURE(*status)){
-        dest16->utf16[0] = (UChar) 0;
-        dest16->utf16_len = 0;
-    }
+    if (U_FAILURE(*status))
+        icu_buf_utf16_clear(dest16);
     
     return dest16->utf16_len;
 }
@@ -810,15 +858,15 @@ void icu_chain_step_destroy(struct icu_chain_step * step){
 
 
 
-struct icu_chain * icu_chain_create(const uint8_t * identifier,
+struct icu_chain * icu_chain_create( //const uint8_t * identifier,
                                     const uint8_t * locale)
 {
 
     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';
+    //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';
 
@@ -853,6 +901,7 @@ void icu_chain_destroy(struct icu_chain * chain)
 
 
 struct icu_chain * icu_chain_xml_config(xmlNode *xml_node, 
+                                        const uint8_t * locale, 
                                         UErrorCode * status){
 
     xmlNode *node = 0;
@@ -860,26 +909,17 @@ struct icu_chain * icu_chain_xml_config(xmlNode *xml_node,
    
     if (!xml_node 
         ||xml_node->type != XML_ELEMENT_NODE 
-        || strcmp((const char *) xml_node->name, "icu_chain"))
+        // || strcmp((const char *) xml_node->name, "icu_chain")
+        )
 
         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;
-        
-        chain = icu_chain_create((const uint8_t *) xml_id, 
-                                 (const uint8_t *) xml_locale);
-        
-        xmlFree(xml_id);
-        xmlFree(xml_locale);
-    }
+        chain = icu_chain_create( // (const uint8_t *) xml_id, 
+                                 locale);
+
     if (!chain)
         return 0;
+
         
     for (node = xml_node->children; node; node = node->next)
     {
@@ -996,6 +1036,104 @@ 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 */
+
+    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 */
+        src16 = chain->src16;
+        step->more_tokens = 0;
+        got_new_token = 1;
+    }
+
+    if (!src16)
+        return 0;
+
+    /* 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);
+        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
+           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;
@@ -1016,7 +1154,8 @@ int icu_chain_step_next_token(struct icu_chain * chain,
 
     /* 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;
 
@@ -1068,21 +1207,24 @@ int icu_chain_step_next_token(struct icu_chain * chain,
         break;
     }
 
-
-
     /* stop further token processing if last step and 
-       new tokens are needed from previous (non-existing) step 
+       new tokens can ot be obtained from last non-existing step 
     */
-    if (!step->previous && step->need_new_token)
+    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, 
@@ -1119,14 +1261,14 @@ int icu_chain_assign_cstr(struct icu_chain * chain,
 int icu_chain_next_token(struct icu_chain * chain,
                          UErrorCode *status)
 {
-    int success = 0;
+    int got_token = 0;
     
     if (!chain || !chain->steps)
         return 0;
 
-    success = icu_chain_step_next_token(chain, chain->steps, status);
+    got_token = icu_chain_step_next_token(chain, chain->steps, status);
     
-    if (success){
+    if (got_token){
         chain->token_count++;
         return chain->token_count;
     }