X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;ds=sidebyside;f=src%2Fcharsets.c;h=5fa6c39d16064daedb686e1fc4a45880fc581d27;hb=8961ed761e348e972f00d015284ce75c16b1648c;hp=397e73b8d337a85083b28a82592d2a2036ea5211;hpb=6a3018550c63b06a788bfeae83de16318965d43b;p=pazpar2-moved-to-github.git

diff --git a/src/charsets.c b/src/charsets.c
index 397e73b..5fa6c39 100644
--- a/src/charsets.c
+++ b/src/charsets.c
@@ -1,5 +1,5 @@
 /* This file is part of Pazpar2.
-   Copyright (C) 2006-2011 Index Data
+   Copyright (C) 2006-2012 Index Data
 
 Pazpar2 is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free
@@ -45,32 +45,32 @@ static pp2_charset_t pp2_charset_create_xml(xmlNode *xml_node);
 static pp2_charset_t pp2_charset_create(struct icu_chain * icu_chn);
 static pp2_charset_t pp2_charset_create_a_to_z(void);
 static void pp2_charset_destroy(pp2_charset_t pct);
-static pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct);
+static pp2_charset_token_t pp2_charset_tokenize(pp2_charset_t pct);
 
 /* charset handle */
 struct pp2_charset_s {
-    const char *(*token_next_handler)(pp2_relevance_token_t prt);
-    const char *(*get_sort_handler)(pp2_relevance_token_t prt);
-    const char *(*get_display_handler)(pp2_relevance_token_t prt);
+    const char *(*token_next_handler)(pp2_charset_token_t prt);
+    const char *(*get_sort_handler)(pp2_charset_token_t prt);
+    const char *(*get_display_handler)(pp2_charset_token_t prt);
 #if YAZ_HAVE_ICU
     struct icu_chain * icu_chn;
     UErrorCode icu_sts;
 #endif
 };
 
-static const char *pp2_relevance_token_null(pp2_relevance_token_t prt);
-static const char *pp2_relevance_token_a_to_z(pp2_relevance_token_t prt);
-static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt);
-static const char *pp2_get_display_ascii(pp2_relevance_token_t prt);
+static const char *pp2_charset_token_null(pp2_charset_token_t prt);
+static const char *pp2_charset_token_a_to_z(pp2_charset_token_t prt);
+static const char *pp2_get_sort_ascii(pp2_charset_token_t prt);
+static const char *pp2_get_display_ascii(pp2_charset_token_t prt);
 
 #if YAZ_HAVE_ICU
-static const char *pp2_relevance_token_icu(pp2_relevance_token_t prt);
-static const char *pp2_get_sort_icu(pp2_relevance_token_t prt);
-static const char *pp2_get_display_icu(pp2_relevance_token_t prt);
+static const char *pp2_charset_token_icu(pp2_charset_token_t prt);
+static const char *pp2_get_sort_icu(pp2_charset_token_t prt);
+static const char *pp2_get_display_icu(pp2_charset_token_t prt);
 #endif
 
 /* tokenzier handle */
-struct pp2_relevance_token_s {
+struct pp2_charset_token_s {
     const char *cp;     /* unnormalized buffer we're tokenizing */
     const char *last_cp;  /* pointer to last token we're dealing with */
     pp2_charset_t pct;  /* our main charset handle (type+config) */
@@ -160,23 +160,26 @@ int pp2_charset_fact_define(pp2_charset_fact_t pft,
 {
     int r;
     pp2_charset_t pct;
-    xmlChar *id;
+    xmlChar *id = 0;
 
     assert(xml_node);
     pct = pp2_charset_create_xml(xml_node);
     if (!pct)
         return -1;
-    id = xmlGetProp(xml_node, (xmlChar*) "id");
-    if (id)
-        default_id = (const char *) id;
     if (!default_id)
     {
-        yaz_log(YLOG_WARN, "Missing id for icu_chain");
-        pp2_charset_destroy(pct);
-        return -1;
+        id = xmlGetProp(xml_node, (xmlChar*) "id");
+        if (!id)
+        {
+            yaz_log(YLOG_WARN, "Missing id for icu_chain");
+            pp2_charset_destroy(pct);
+            return -1;
+        }
+        default_id = (const char *) id;
     }
     r = pp2_charset_fact_add(pft, pct, default_id);
-    xmlFree(id);
+    if (id)
+        xmlFree(id);
     return r;
 }
 
@@ -198,7 +201,7 @@ pp2_charset_t pp2_charset_create_xml(xmlNode *xml_node)
         //xmlChar *xmlstr = 0;
                 //int size = 0;
                 //xmlDocDumpMemory(icu_doc, size);
-        
+
         yaz_log(YLOG_FATAL, "Could not parse ICU chain config:\n"
                 "<%s>\n ... \n</%s>",
                 xml_node->name, xml_node->name);
@@ -209,7 +212,7 @@ pp2_charset_t pp2_charset_create_xml(xmlNode *xml_node)
     yaz_log(YLOG_FATAL, "Error: ICU support requested with element:\n"
             "<%s>\n ... \n</%s>",
             xml_node->name, xml_node->name);
-    yaz_log(YLOG_FATAL, 
+    yaz_log(YLOG_FATAL,
             "But no ICU support is compiled into the YAZ library.");
     return 0;
 #endif // YAZ_HAVE_ICU
@@ -218,7 +221,7 @@ pp2_charset_t pp2_charset_create_xml(xmlNode *xml_node)
 pp2_charset_t pp2_charset_create_a_to_z(void)
 {
     pp2_charset_t pct = pp2_charset_create(0);
-    pct->token_next_handler = pp2_relevance_token_a_to_z;
+    pct->token_next_handler = pp2_charset_token_a_to_z;
     return pct;
 }
 
@@ -226,7 +229,7 @@ pp2_charset_t pp2_charset_create(struct icu_chain *icu_chn)
 {
     pp2_charset_t pct = xmalloc(sizeof(*pct));
 
-    pct->token_next_handler = pp2_relevance_token_null;
+    pct->token_next_handler = pp2_charset_token_null;
     pct->get_sort_handler  = pp2_get_sort_ascii;
     pct->get_display_handler  = pp2_get_display_ascii;
 #if YAZ_HAVE_ICU
@@ -235,7 +238,7 @@ pp2_charset_t pp2_charset_create(struct icu_chain *icu_chn)
     {
         pct->icu_chn = icu_chn;
         pct->icu_sts = U_ZERO_ERROR;
-        pct->token_next_handler = pp2_relevance_token_icu;
+        pct->token_next_handler = pp2_charset_token_icu;
         pct->get_sort_handler = pp2_get_sort_icu;
         pct->get_display_handler = pp2_get_display_icu;
     }
@@ -251,19 +254,19 @@ void pp2_charset_destroy(pp2_charset_t pct)
     xfree(pct);
 }
 
-pp2_relevance_token_t pp2_relevance_create(pp2_charset_fact_t pft,
-                                           const char *id)
+pp2_charset_token_t pp2_charset_token_create(pp2_charset_fact_t pft,
+                                               const char *id)
 {
     struct pp2_charset_entry *pce;
     for (pce = pft->list; pce; pce = pce->next)
         if (!strcmp(id, pce->name))
-            return pp2_relevance_tokenize(pce->pct);
+            return pp2_charset_tokenize(pce->pct);
     return 0;
 }
 
-pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct)
+pp2_charset_token_t pp2_charset_tokenize(pp2_charset_t pct)
 {
-    pp2_relevance_token_t prt = xmalloc(sizeof(*prt));
+    pp2_charset_token_t prt = xmalloc(sizeof(*prt));
 
     assert(pct);
 
@@ -281,17 +284,16 @@ pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct)
     return prt;
 }
 
-void pp2_relevance_first(pp2_relevance_token_t prt,
-                         const char *buf,
-                         int skip_article)
-{ 
+void pp2_charset_token_first(pp2_charset_token_t prt,
+                             const char *buf, int skip_article)
+{
     if (skip_article)
     {
         const char *p = buf;
         char firstword[64];
         char *pout = firstword;
         char articles[] = "the den der die des an a "; // must end in space
-        
+
         for (; *p && *p != ' ' && pout - firstword < (sizeof(firstword)-2); p++)
             *pout++ = tolower(*(unsigned char *)p);
         *pout++ = ' ';
@@ -313,32 +315,32 @@ void pp2_relevance_first(pp2_relevance_token_t prt,
 #endif // YAZ_HAVE_ICU
 }
 
-void pp2_relevance_token_destroy(pp2_relevance_token_t prt)
+void pp2_charset_token_destroy(pp2_charset_token_t prt)
 {
     assert(prt);
 #if YAZ_HAVE_ICU
     if (prt->iter)
         icu_iter_destroy(prt->iter);
 #endif
-    if(prt->norm_str) 
+    if(prt->norm_str)
         wrbuf_destroy(prt->norm_str);
-    if(prt->sort_str) 
+    if(prt->sort_str)
         wrbuf_destroy(prt->sort_str);
     xfree(prt);
 }
 
-const char *pp2_relevance_token_next(pp2_relevance_token_t prt)
+const char *pp2_charset_token_next(pp2_charset_token_t prt)
 {
     assert(prt);
     return (prt->pct->token_next_handler)(prt);
 }
 
-const char *pp2_get_sort(pp2_relevance_token_t prt)
+const char *pp2_get_sort(pp2_charset_token_t prt)
 {
     return prt->pct->get_sort_handler(prt);
 }
 
-const char *pp2_get_display(pp2_relevance_token_t prt)
+const char *pp2_get_display(pp2_charset_token_t prt)
 {
     return prt->pct->get_display_handler(prt);
 }
@@ -347,7 +349,7 @@ const char *pp2_get_display(pp2_relevance_token_t prt)
 /* original tokenizer with our tokenize interface, but we
    add +1 to ensure no '\0' are in our string (except for EOF)
 */
-static const char *pp2_relevance_token_a_to_z(pp2_relevance_token_t prt)
+static const char *pp2_charset_token_a_to_z(pp2_charset_token_t prt)
 {
     const char *cp = prt->cp;
     int c;
@@ -374,7 +376,7 @@ static const char *pp2_relevance_token_a_to_z(pp2_relevance_token_t prt)
     return wrbuf_cstr(prt->norm_str);
 }
 
-static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt)
+static const char *pp2_get_sort_ascii(pp2_charset_token_t prt)
 {
     if (prt->last_cp == 0)
         return 0;
@@ -383,7 +385,7 @@ static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt)
         char *tmp = xstrdup(prt->last_cp);
         char *result = 0;
         result = normalize7bit_mergekey(tmp);
-        
+
         wrbuf_rewind(prt->sort_str);
         wrbuf_puts(prt->sort_str, result);
         xfree(tmp);
@@ -391,7 +393,7 @@ static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt)
     }
 }
 
-static const char *pp2_get_display_ascii(pp2_relevance_token_t prt)
+static const char *pp2_get_display_ascii(pp2_charset_token_t prt)
 {
     if (prt->last_cp == 0)
         return 0;
@@ -401,7 +403,7 @@ static const char *pp2_get_display_ascii(pp2_relevance_token_t prt)
     }
 }
 
-static const char *pp2_relevance_token_null(pp2_relevance_token_t prt)
+static const char *pp2_charset_token_null(pp2_charset_token_t prt)
 {
     const char *cp = prt->cp;
 
@@ -413,7 +415,7 @@ static const char *pp2_relevance_token_null(pp2_relevance_token_t prt)
 }
 
 #if YAZ_HAVE_ICU
-static const char *pp2_relevance_token_icu(pp2_relevance_token_t prt)
+static const char *pp2_charset_token_icu(pp2_charset_token_t prt)
 {
     if (icu_iter_next(prt->iter))
     {
@@ -422,12 +424,12 @@ static const char *pp2_relevance_token_icu(pp2_relevance_token_t prt)
     return 0;
 }
 
-static const char *pp2_get_sort_icu(pp2_relevance_token_t prt)
+static const char *pp2_get_sort_icu(pp2_charset_token_t prt)
 {
     return icu_iter_get_sortkey(prt->iter);
 }
 
-static const char *pp2_get_display_icu(pp2_relevance_token_t prt)
+static const char *pp2_get_display_icu(pp2_charset_token_t prt)
 {
     return icu_iter_get_display(prt->iter);
 }