X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=retrieval%2Fd1_grs.c;h=873549ca7f18df43ad2f39b2d4822bfc74087239;hb=8f3f8256752d09f66cd2c7a412cc8c30a0fc90ea;hp=746413f31e01622d1836a4a13ccf5e16f0a66e95;hpb=751fc8b0c5b555a41feb9d8c0bf8ac71d5707d25;p=yaz-moved-to-github.git

diff --git a/retrieval/d1_grs.c b/retrieval/d1_grs.c
index 746413f..873549c 100644
--- a/retrieval/d1_grs.c
+++ b/retrieval/d1_grs.c
@@ -3,7 +3,7 @@
  * See the file LICENSE for details.
  * Sebastian Hammer, Adam Dickmeiss
  *
- * $Id: d1_grs.c,v 1.22 2002-07-03 14:09:34 adam Exp $
+ * $Id: d1_grs.c,v 1.25 2002-08-23 14:27:18 adam Exp $
  *
  */
 
@@ -114,6 +114,45 @@ static int traverse_triples(data1_node *n, int level, Z_ElementMetaData *m,
     return 0;
 }
 
+/*
+ * Locate some data under this node. This routine should handle variants
+ * prettily.
+ */
+static char *get_data(data1_node *n, int *len)
+{
+    char *r;
+    data1_node *np = 0;
+
+    while (n)
+    {
+        if (n->which == DATA1N_data)
+        {
+            int i;
+            *len = n->u.data.len;
+
+            for (i = 0; i<*len; i++)
+                if (!d1_isspace(n->u.data.data[i]))
+                    break;
+            while (*len && d1_isspace(n->u.data.data[*len - 1]))
+                (*len)--;
+            *len = *len - i;
+            if (*len > 0)
+                return n->u.data.data + i;
+        }
+        if (n->which == DATA1N_tag)
+            np = n->child;
+        n = n->next;
+        if (!n)
+        {
+            n = np;
+            np = 0;
+        }
+    }
+    r = "";
+    *len = strlen(r);
+    return r;
+}
+
 static Z_ElementData *nodetoelementdata(data1_handle dh, data1_node *n,
 					int select, int leaf,
 					ODR o, int *len)
@@ -125,39 +164,37 @@ static Z_ElementData *nodetoelementdata(data1_handle dh, data1_node *n,
 	res->which = Z_ElementData_elementNotThere;
 	res->u.elementNotThere = odr_nullval();
     }
-    else if (n->which == DATA1N_data && (leaf || n->next == NULL))
+    else if (n->which == DATA1N_data && leaf)
     {
-	char str[512];
-	int toget;
-	data1_node *p;
+	char str[64], *cp;
+	int toget = n->u.data.len;
 
-	for (p = n->parent; p && p->which != DATA1N_tag; p = p->parent)
-	    ;
+        cp = get_data (n, &toget);
 
 	switch (n->u.data.what)
 	{
         case DATA1I_num:
             res->which = Z_ElementData_numeric;
             res->u.numeric = (int *)odr_malloc(o, sizeof(int));
-            *res->u.numeric = atoi(n->u.data.data);
+            *res->u.numeric = atoi_n (cp, toget);
             *len += 4;
             break;
         case DATA1I_text:
-            toget = n->u.data.len;
-            if (p && p->u.tag.get_bytes > 0 && p->u.tag.get_bytes < toget)
-                toget = p->u.tag.get_bytes;
             res->which = Z_ElementData_string;
             res->u.string = (char *)odr_malloc(o, toget+1);
-            memcpy(res->u.string, n->u.data.data, toget);
+            if (toget)
+                memcpy(res->u.string, cp, toget);
             res->u.string[toget] = '\0';
             *len += toget;
             break;
         case DATA1I_oid:
             res->which = Z_ElementData_oid;
-            strncpy(str, n->u.data.data, n->u.data.len);
-            str[n->u.data.len] = '\0';
+            if (toget > 63)
+                toget = 63;
+            memcpy (str, cp, toget);
+            str[toget] = '\0';
             res->u.oid = odr_getoidbystr(o, str);
-            *len += n->u.data.len;
+            *len += oid_oidlen(res->u.oid) * sizeof(int);
             break;
         default:
             yaz_log(LOG_WARN, "Can't handle datatype.");
@@ -173,6 +210,21 @@ static Z_ElementData *nodetoelementdata(data1_handle dh, data1_node *n,
     return res;
 }
 
+static int is_empty_data (data1_node *n)
+{
+    if (n && n->which == DATA1N_data && n->u.data.what == DATA1I_text)
+    {
+        int i = n->u.data.len;
+        
+        while (i > 0 && n->u.data.data[i-1] == '\n')
+            i--;
+        if (i == 0)
+            return 1;
+    }
+    return 0;
+}
+
+
 static Z_TaggedElement *nodetotaggedelement(data1_handle dh, data1_node *n,
 					    int select, ODR o,
 					    int *len)
@@ -187,7 +239,20 @@ static Z_TaggedElement *nodetotaggedelement(data1_handle dh, data1_node *n,
 	if (n->u.tag.element)
 	    tag = n->u.tag.element->tag;
 	data = n->child;
-	leaf = 0;
+
+        /* skip empty data children */
+        while (is_empty_data(data))
+            data = data->next;
+        if (!data)
+            data = n->child;
+        else
+        {   /* got one. see if this is the only non-empty one */
+            data1_node *sub = data->next;
+            while (sub && is_empty_data(sub))
+                sub = sub->next;
+            if (!sub)
+                leaf = 1;  /* all empty. 'data' is the only child */
+        }
     }
     /*
      * If we're a data element at this point, we need to insert a
@@ -204,6 +269,8 @@ static Z_TaggedElement *nodetotaggedelement(data1_handle dh, data1_node *n,
 	}
 	data = n;
 	leaf = 1;
+        if (is_empty_data(data))
+            return 0;
     }
     else
     {
@@ -229,7 +296,7 @@ static Z_TaggedElement *nodetotaggedelement(data1_handle dh, data1_node *n,
 	else if (tag)                    
 	    tagstr = tag->value.string;  /* no take from well-known */
 	else
-	    tagstr = "???";                /* no tag at all! */
+            return 0;
 	res->tagValue->which = Z_StringOrNumeric_string;
 	res->tagValue->u.string = odr_strdup(o, tagstr);
     }
@@ -286,9 +353,9 @@ Z_GenericRecord *data1_nodetogr(data1_handle dh, data1_node *n,
     {
 	if (c->which == DATA1N_tag && select && !c->u.tag.node_selected)
 	    continue;
-	if (!(res->elements[res->num_elements++] =
-	      nodetotaggedelement (dh, c, select, o, len)))
-	    return 0;
+	if ((res->elements[res->num_elements] =
+             nodetotaggedelement (dh, c, select, o, len)))
+	    res->num_elements++;
     }
     return res;
 }