Fixed bug #317: xelm only indexes cdata at matching node (not children).

diff --git a/data1/d1_absyn.c b/data1/d1_absyn.c
index 45724a9..34c4dd1 100644 (file)
--- a/data1/d1_absyn.c
+++ b/data1/d1_absyn.c
@@ -1,4 +1,4 @@
-/* $Id: d1_absyn.c,v 1.20 2005-06-23 06:45:46 adam Exp $
+/* $Id: d1_absyn.c,v 1.21 2005-06-29 16:52:26 adam Exp $
    Copyright (C) 1995-2005
    Index Data ApS
 
@@ -439,65 +439,73 @@ void fix_element_ref (data1_handle dh, data1_absyn *absyn, data1_element *e)
 
  */
 
-const char * mk_xpath_regexp (data1_handle dh, char *expr) 
+static const char * mk_xpath_regexp (data1_handle dh, const char *expr) 
 {
-    char *p = expr;
-    char *pp;
-    char *s;
+    const char *p = expr;
     int abs = 1;
     int i;
-    int j;
-    int e=0;
-    int is_predicate = 0;
+    int e = 0;
+    char *stack[32];
+    char *res_p, *res = 0;
+    size_t res_size = 1;
     
-    static char *stack[32];
-    static char res[1024];
-    char *r = "";
-    
-    if (*p != '/') { return (""); }
+    if (*p != '/')
+       return ("");
     p++;
-    if (*p == '/') { abs=0; p++; }
-    
-    while (*p) {
-        i=0;
-        while (*p && !strchr("/",*p)) { 
-         i++; p++; 
-       }
-        stack[e] = (char *) nmem_malloc (data1_nmem_get (dh), i+1);
+    if (*p == '/') 
+    { 
+       abs =0;
+       p++;
+    }
+    while (*p)
+    {
+       int is_predicate = 0;
+       char *s;
+       int j;
+        for (i = 0; *p && !strchr("/",*p); i++, p++)
+           ;
+       res_size += (i+3); /* we'll add / between later .. */
+        stack[e] = (char *) nmem_malloc(data1_nmem_get(dh), i+1);
        s = stack[e];
-       for (j=0; j< i; j++) {
-         pp = p-i+j;
-         if (*pp == '[') {
-           is_predicate=1;
-         }
-         else if (*pp == ']') {
-           is_predicate=0;
-         }
-         else {
-           if (!is_predicate) {
-             if (*pp == '*') 
-                *s++ = '.';
-             *s++ = *pp;
+       for (j = 0; j < i; j++)
+       {
+           const char *pp = p-i+j;
+           if (*pp == '[')
+               is_predicate=1;
+           else if (*pp == ']')
+               is_predicate=0;
+           else 
+           {
+               if (!is_predicate) {
+                   if (*pp == '*') 
+                       *s++ = '.';
+                   *s++ = *pp;
+               }
            }
-         }
        }
        *s = 0;
         e++;
-        if (*p) {p++;}
+        if (*p)
+           p++;
     }
-    e--;  p = &res[0]; i=0;
-    sprintf (p, "^"); p++;
-    while (e >= 0) {
-        /* !!! res size is not checked !!! */
-        sprintf (p, "%s/",stack[e]);
-        p += strlen(stack[e]) + 1;
-        e--;
+    res_p = res = nmem_malloc(data1_nmem_get(dh), res_size + 10);
+
+    i = 0;
+    sprintf(res_p, ".*/");
+    res_p = res_p + strlen(res_p);
+    while (--e >= 0) {
+       sprintf(res_p, "%s/", stack[e]);
+       res_p += strlen(stack[e]) + 1;
     }
-    if (!abs) { sprintf (p, ".*"); p+=2; }
-    sprintf (p, "$"); p++;
-    r = nmem_strdup (data1_nmem_get (dh), res);
-    yaz_log(YLOG_DEBUG,"Got regexp: %s",r);
-    return (r);
+    if (!abs)
+    {
+       sprintf(res_p, ".*"); 
+       res_p += 2;
+    }
+    sprintf (res_p, "$");
+    res_p++;
+    yaz_log(YLOG_DEBUG, "Got regexp: %s", res);
+    return res;
 }
 
 /* *ostrich*

diff --git a/include/d1_absyn.h b/include/d1_absyn.h
index 771911a..80556ed 100644 (file)
--- a/include/d1_absyn.h
+++ b/include/d1_absyn.h
@@ -1,3 +1,24 @@
+/* $Id: d1_absyn.h,v 1.2 2005-06-29 16:52:27 adam Exp $
+   Copyright (C) 1995-2005
+   Index Data ApS
+
+This file is part of the Zebra server.
+
+Zebra 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
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with Zebra; see the file LICENSE.zebra.  If not, write to the
+Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA.
+*/
 
 #ifndef D1_ABSYN_H
 #define D1_ABSYN_H 1

diff --git a/recctrl/recgrs.c b/recctrl/recgrs.c
index 99eb954..eb80884 100644 (file)
--- a/recctrl/recgrs.c
+++ b/recctrl/recgrs.c
@@ -1,4 +1,4 @@
-/* $Id: recgrs.c,v 1.102 2005-06-23 06:45:47 adam Exp $
+/* $Id: recgrs.c,v 1.103 2005-06-29 16:52:27 adam Exp $
    Copyright (C) 1995-2005
    Index Data ApS
 
@@ -337,6 +337,33 @@ int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p)
 }
 
 
+static int dfa_match_first(struct DFA_state **dfaar, const char *text)
+{
+    struct DFA_state *s = dfaar[0]; /* start state */
+    struct DFA_tran *t;
+    int i;
+    const char *p = text;
+    unsigned char c;
+    
+    for (c = *p++, t = s->trans, i = s->tran_no; --i >= 0; t++)
+       if (c >= t->ch[0] && c <= t->ch[1])
+       {
+           while (i >= 0)
+           {
+               /* move to next state and return if we get a match */
+               s = dfaar[t->to];
+               if (s->rule_no)
+                   return 1;
+               /* next char */
+               c = *p++;
+               for (t = s->trans, i = s->tran_no; --i >= 0; t++)
+                   if (c >= t->ch[0] && c <= t->ch[1])
+                       break;
+           }
+       }
+    return 0;
+}
+
 /* *ostrich*
    
 New function, looking for xpath "element" definitions in abs, by
@@ -362,44 +389,20 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
 #ifdef ENHANCED_XELM 
     struct xpath_location_step *xp;
 #endif
-    char *pexpr = xmalloc(strlen(tagpath)+2);
+    char *pexpr = xmalloc(strlen(tagpath)+5);
     int ok = 0;
     
-    sprintf (pexpr, "%s\n", tagpath);
-    yaz_log(YLOG_DEBUG, "Checking tagpath %s",tagpath);
+    sprintf (pexpr, "/%s\n", tagpath);
+    yaz_log(YLOG_LOG, "Checking tagpath %s", pexpr);
     for (; xpe; xpe = xpe->next)
     {
-        struct DFA_state **dfaar = xpe->dfa->states;
-        struct DFA_state *s = dfaar[0];
-        struct DFA_tran *t = s->trans;
-        int i = s->tran_no;
-        unsigned char c = *pexpr++;
-        int start_line = 1;
-
-       if ((c >= t->ch[0] && c <= t->ch[1]) || (!t->ch[0]))
-       {
-            const char *p = pexpr;
-            do 
-           {
-                if ((s = dfaar[t->to])->rule_no && 
-                    (start_line || s->rule_nno))
-               {
-                    ok = 1;
-                    break;
-                }
-                for (t=s->trans, i=s->tran_no; --i >= 0; t++)
-                    if ((unsigned) *p >= t->ch[0] && (unsigned) *p <= t->ch[1])
-                        break;
-                p++;
-            } 
-           while (i >= 0);
-       }
+       int i;
+       ok = dfa_match_first(xpe->dfa->states, pexpr);
        if (ok)
-           yaz_log(YLOG_DEBUG, " xpath match %s",xpe->xpath_expr);
+           yaz_log(YLOG_LOG, " xpath got match %s",xpe->xpath_expr);
        else
-           yaz_log(YLOG_DEBUG, " xpath no match %s",xpe->xpath_expr);
+           yaz_log(YLOG_LOG, " xpath no match %s",xpe->xpath_expr);
 
-        pexpr--;
         if (ok) {
 #ifdef ENHANCED_XELM 
             /* we have to check the perdicates up to the root node */
@@ -499,6 +502,32 @@ static void index_xpath_attr (char *tag_path, char *name, char *value,
 }
 
 
+static void mk_tag_path_full(char *tag_path_full, size_t max, data1_node *n)
+{
+    size_t flen = 0;
+    data1_node *nn;
+
+    /* we have to fetch the whole path to the data tag */
+    for (nn = n; nn; nn = nn->parent)
+    {
+       if (nn->which == DATA1N_tag)
+       {
+           size_t tlen = strlen(nn->u.tag.tag);
+           if (tlen + flen > (max - 2))
+               break;
+           memcpy (tag_path_full + flen, nn->u.tag.tag, tlen);
+           flen += tlen;
+           tag_path_full[flen++] = '/';
+       }
+       else
+           if (nn->which == DATA1N_root)
+               break;
+    }
+    tag_path_full[flen] = 0;
+    yaz_log(YLOG_LOG, "mk_tag_path_full=%s", tag_path_full);
+}
+       
+
 static void index_xpath(struct source_parser *sp, data1_node *n,
                        struct recExtractCtrl *p,
                        int level, RecWord *wrd,
@@ -512,8 +541,6 @@ static void index_xpath(struct source_parser *sp, data1_node *n,
 {
     int i;
     char tag_path_full[1024];
-    size_t flen = 0;
-    data1_node *nn;
     int termlist_only = 1;
     data1_termlist *tl;
     int xpdone = 0;
@@ -541,26 +568,8 @@ static void index_xpath(struct source_parser *sp, data1_node *n,
         wrd->term_buf = n->u.data.data;
         wrd->term_len = n->u.data.len;
         xpdone = 0;
-        flen = 0;
-            
-       /* we have to fetch the whole path to the data tag */
-       for (nn = n; nn; nn = nn->parent)
-       {
-           if (nn->which == DATA1N_tag)
-           {
-               size_t tlen = strlen(nn->u.tag.tag);
-               if (tlen + flen > (sizeof(tag_path_full)-2))
-                   break;
-               memcpy (tag_path_full + flen, nn->u.tag.tag, tlen);
-               flen += tlen;
-               tag_path_full[flen++] = '/';
-           }
-           else
-               if (nn->which == DATA1N_root)
-                   break;
-       }
-       
-       tag_path_full[flen] = 0;
+
+       mk_tag_path_full(tag_path_full, sizeof(tag_path_full), n);
        
        /* If we have a matching termlist... */
        if (n->root->u.root.absyn && 
@@ -651,26 +660,11 @@ static void index_xpath(struct source_parser *sp, data1_node *n,
        }
         break;
     case DATA1N_tag:
-        flen = 0;
-        for (nn = n; nn; nn = nn->parent)
-        {
-            if (nn->which == DATA1N_tag)
-            {
-                size_t tlen = strlen(nn->u.tag.tag);
-                if (tlen + flen > (sizeof(tag_path_full)-2))
-                   break;
-                memcpy (tag_path_full + flen, nn->u.tag.tag, tlen);
-                flen += tlen;
-                tag_path_full[flen++] = '/';
-            }
-            else if (nn->which == DATA1N_root)
-                break;
-        }
-
+       mk_tag_path_full(tag_path_full, sizeof(tag_path_full), n);
 
         wrd->index_type = '0';
         wrd->term_buf = tag_path_full;
-        wrd->term_len = flen;
+        wrd->term_len = strlen(tag_path_full);
 #if NATTR
        wrd->index_name = xpath_index;
 #else
@@ -692,8 +686,6 @@ static void index_xpath(struct source_parser *sp, data1_node *n,
             data1_termlist *tl;
            int do_xpindex;
             
-           tag_path_full[flen] = 0;
-            
             /* Add tag start/end xpath index, only when there is a ! in
               the apropriate xelm directive, or default xpath indexing
               is enabled 
@@ -735,11 +727,10 @@ static void index_xpath(struct source_parser *sp, data1_node *n,
                     int do_xpindex = 1 - termlist_only;
                     data1_termlist *tl;
                     char attr_tag_path_full[1024]; 
-                    int int_len = flen;
                     
                     /* this could be cached as well */
-                    sprintf (attr_tag_path_full, "@%s/%.*s",
-                             xp->name, int_len, tag_path_full);
+                    sprintf (attr_tag_path_full, "@%s/%s",
+                             xp->name, tag_path_full);
                     
                     tll[i] = xpath_termlist_by_tagpath(attr_tag_path_full,n);
                     
@@ -802,11 +793,10 @@ static void index_xpath(struct source_parser *sp, data1_node *n,
                 for (xp = n->u.tag.attributes; xp; xp = xp->next) {
                     data1_termlist *tl;
                     char attr_tag_path_full[1024];
-                    int int_len = flen;
                     int xpdone = 0;
                     
-                    sprintf (attr_tag_path_full, "@%s/%.*s",
-                             xp->name, int_len, tag_path_full);
+                    sprintf (attr_tag_path_full, "@%s/%s",
+                             xp->name, tag_path_full);
                     
                     if ((tl = tll[i]))
                     {

diff --git a/test/xpath/assembled.abs b/test/xpath/assembled.abs
index 7daa026..2c06889 100644 (file)
--- a/test/xpath/assembled.abs
+++ b/test/xpath/assembled.abs
@@ -4,3 +4,4 @@ xpath enable
 
 xelm /assembled/basic/names/CASno !:w,!:p,!:s,NAL-call-number:p,Title-key:s
 xelm /*/orgs  body-of-text:w
+xelm //something Title:w

diff --git a/test/xpath/xpath6.c b/test/xpath/xpath6.c
index c0ef3c3..9828f8d 100644 (file)
--- a/test/xpath/xpath6.c
+++ b/test/xpath/xpath6.c
@@ -1,4 +1,4 @@
-/* $Id: xpath6.c,v 1.3 2005-05-04 12:53:05 adam Exp $
+/* $Id: xpath6.c,v 1.4 2005-06-29 16:52:27 adam Exp $
    Copyright (C) 1995-2005
    Index Data ApS
 
@@ -46,6 +46,7 @@ int main(int argc, char **argv)
     zebra_end_trans(zh);
     zebra_commit(zh);
     
+
     do_query(__LINE__, zh, "@attr 5=1 @attr 6=3  @attr 4=1 @attr 1=/assembled/basic/names/CASno \"367-93-1\"", 2);
 
     do_query(__LINE__, zh, "@attr 5=1 @attr 6=3  @attr 4=1 @attr 1=18 \"367-93-1\"", 2);
@@ -60,10 +61,8 @@ int main(int argc, char **argv)
             "@and @attr 1=/assembled/orgs/org 1 @attr 5=1 @attr 6=3  @attr 4=1 "
             "@attr 1=/assembled/basic/names/CASno 367-93-1", 2);
 
-#if 0
     /* bug #317 */
-    do_query(__LINE__, zh, "@attr 1=1010 46", 1);
-#endif
+    do_query(__LINE__, zh, "@attr 1=1010 46", 2);
 
     return close_down(zh, zs, 0);
 }