X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Fcharsets.c;h=f867b6f4a31aa98cf6b34a1b79e1b0393ee0f48e;hb=fc05a25c3100a114f64450604fc897c77c4f0db6;hp=0eefa079cc459ab5fa2233754f8d773b3dc9373d;hpb=ba5951a80fdd0da1d28c318852b69a1178cd6bfa;p=pazpar2-moved-to-github.git

diff --git a/src/charsets.c b/src/charsets.c
index 0eefa07..f867b6f 100644
--- a/src/charsets.c
+++ b/src/charsets.c
@@ -28,6 +28,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/xmalloc.h>
 #include <yaz/wrbuf.h>
 #include <yaz/log.h>
+#include <yaz/yaz-version.h>
 #include <ctype.h>
 #include <assert.h>
 #include <string.h>
@@ -39,7 +40,6 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/icu.h>
 #endif
 
-
 /* charset handle */
 struct pp2_charset_s {
     const char *(*token_next_handler)(pp2_relevance_token_t prt);
@@ -66,6 +66,9 @@ struct pp2_relevance_token_s {
     pp2_charset_t pct;  /* our main charset handle (type+config) */
     WRBUF norm_str;     /* normized string we return (temporarily) */
     WRBUF sort_str;     /* sort string we return (temporarily) */
+#if YAZ_HAVE_ICU
+    yaz_icu_iter_t iter;
+#endif
 };
 
 
@@ -142,14 +145,30 @@ void pp2_charset_destroy(pp2_charset_t pct)
     }
 }
 
-pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct,
-                                             const char *buf,
-                                             int skip_article)
+pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct)
 {
     pp2_relevance_token_t prt = xmalloc(sizeof(*prt));
 
     assert(pct);
 
+    prt->norm_str = wrbuf_alloc();
+    prt->sort_str = wrbuf_alloc();
+    prt->cp = 0;
+    prt->last_cp = 0;
+    prt->pct = pct;
+
+#if YAZ_HAVE_ICU
+    prt->iter = 0;
+    if (pct->icu_chn)
+        prt->iter = icu_iter_create(pct->icu_chn);
+#endif
+    return prt;
+}
+
+void pp2_relevance_first(pp2_relevance_token_t prt,
+                         const char *buf,
+                         int skip_article)
+{ 
     if (skip_article)
     {
         const char *p = buf;
@@ -167,29 +186,26 @@ pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct,
             buf = p;
     }
 
-    prt->norm_str = wrbuf_alloc();
-    prt->sort_str = wrbuf_alloc();
+    wrbuf_rewind(prt->norm_str);
+    wrbuf_rewind(prt->sort_str);
     prt->cp = buf;
     prt->last_cp = 0;
-    prt->pct = pct;
 
 #if YAZ_HAVE_ICU
-    if (pct->icu_chn)
+    if (prt->iter)
     {
-        int ok = 0;
-        pct->icu_sts = U_ZERO_ERROR;
-        ok = icu_chain_assign_cstr(pct->icu_chn, buf, &pct->icu_sts);
-        //printf("\nfield ok: %d '%s'\n", ok, buf);
-        prt->pct = pct;
+        icu_iter_first(prt->iter, buf);
     }
 #endif // YAZ_HAVE_ICU
-    return prt;
 }
 
-
 void pp2_relevance_token_destroy(pp2_relevance_token_t prt)
 {
     assert(prt);
+#if YAZ_HAVE_ICU
+    if (prt->iter)
+        icu_iter_destroy(prt->iter);
+#endif
     if(prt->norm_str) 
         wrbuf_destroy(prt->norm_str);
     if(prt->sort_str) 
@@ -260,20 +276,16 @@ static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt)
 #if YAZ_HAVE_ICU
 static const char *pp2_relevance_token_icu(pp2_relevance_token_t prt)
 {
-    if (icu_chain_next_token(prt->pct->icu_chn, &prt->pct->icu_sts))
+    if (icu_iter_next(prt->iter))
     {
-        if (U_FAILURE(prt->pct->icu_sts))
-        {
-            return 0;
-        }
-        return icu_chain_token_norm(prt->pct->icu_chn);
+        return icu_iter_get_norm(prt->iter);
     }
     return 0;
 }
 
 static const char *pp2_get_sort_icu(pp2_relevance_token_t prt)
 {
-    return icu_chain_token_sortkey(prt->pct->icu_chn);
+    return icu_iter_get_sortkey(prt->iter);
 }
 
 #endif // YAZ_HAVE_ICU