-/* $Id: recgrs.c,v 1.102 2005-06-23 06:45:47 adam Exp $
+/* $Id: recgrs.c,v 1.105 2005-08-22 08:19:51 adam Exp $
Copyright (C) 1995-2005
Index Data ApS
}
else if (sp->len > 0 && isdigit(*(unsigned char *)sp->tok))
{
- wrd->term_buf = nmem_malloc(sp->nmem, sp->len);
- memcpy(wrd->term_buf, sp->tok, sp->len);
+ char *b;
wrd->term_len = sp->len;
+ b = nmem_malloc(sp->nmem, sp->len);
+ memcpy(b, sp->tok, sp->len);
+ wrd->term_buf = b;
sp_lex(sp);
}
else if (sp->len > 2 && sp->tok[0] == '\'' && sp->tok[sp->len-1] == '\'')
{
+ char *b;
wrd->term_len = sp->len - 2;
- wrd->term_buf = nmem_malloc(sp->nmem, wrd->term_len);
- memcpy(wrd->term_buf, sp->tok+1, wrd->term_len);
+ b = nmem_malloc(sp->nmem, wrd->term_len);
+ memcpy(b, sp->tok+1, wrd->term_len);
+ wrd->term_buf = b;
sp_lex(sp);
}
else
}
+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
#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);
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);
- }
- if (ok)
- yaz_log(YLOG_DEBUG, " xpath match %s",xpe->xpath_expr);
- else
- yaz_log(YLOG_DEBUG, " xpath no match %s",xpe->xpath_expr);
+ int i;
+ ok = dfa_match_first(xpe->dfa->states, pexpr);
- pexpr--;
if (ok) {
#ifdef ENHANCED_XELM
/* we have to check the perdicates up to the root node */
}
+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;
+}
+
+
static void index_xpath(struct source_parser *sp, data1_node *n,
struct recExtractCtrl *p,
int level, RecWord *wrd,
{
int i;
char tag_path_full[1024];
- size_t flen = 0;
- data1_node *nn;
int termlist_only = 1;
data1_termlist *tl;
int xpdone = 0;
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 &&
}
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
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
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);
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]))
{