-/* $Id: recgrs.c,v 1.73 2003-02-18 11:57:58 adam Exp $
- Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002
+/* $Id: recgrs.c,v 1.79 2003-05-05 20:14:42 adam Exp $
+ Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003
Index Data Aps
This file is part of the Zebra server.
xfree (h);
}
-int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) {
- int res = 1;
- char *attname;
- data1_xattr *attr;
-
- if (!p) {
- return (1);
- } else {
- if (p->which == XPATH_PREDICATE_RELATION) {
- if (p->u.relation.name[0]) {
- if (*p->u.relation.name != '@') {
- logf(LOG_WARN,
- " Only attributes (@) are supported in xelm xpath predicates");
- logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
- return (1);
- }
- attname = p->u.relation.name + 1;
- res = 0;
- /* looking for the attribute with a specified name */
- for (attr = n->u.tag.attributes; attr; attr = attr->next) {
- logf(LOG_DEBUG," - attribute %s <-> %s", attname, attr->name );
-
- if (!strcmp(attr->name, attname)) {
- if (p->u.relation.op[0]) {
- if (*p->u.relation.op != '=') {
- logf(LOG_WARN,
- "Only '=' relation is supported (%s)",p->u.relation.op);
- logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
- res = 1; break;
- } else {
- logf(LOG_DEBUG," - value %s <-> %s",
- p->u.relation.value, attr->value );
- if (!strcmp(attr->value, p->u.relation.value)) {
- res = 1; break;
- }
- }
- } else {
- /* attribute exists, no value specified */
- res = 1; break;
- }
- }
- }
- return (res);
- } else {
- return (1);
- }
- }
- else if (p->which == XPATH_PREDICATE_BOOLEAN) {
- if (!strcmp(p->u.boolean.op,"and")) {
- return (d1_check_xpath_predicate(n, p->u.boolean.left)
- && d1_check_xpath_predicate(n, p->u.boolean.right));
- }
- else if (!strcmp(p->u.boolean.op,"or")) {
- return (d1_check_xpath_predicate(n, p->u.boolean.left)
- || d1_check_xpath_predicate(n, p->u.boolean.right));
- } else {
- logf(LOG_WARN, "Unknown boolean relation %s, ignored",p->u.boolean.op);
- return (1);
- }
+int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p)
+{
+ int res = 1;
+ char *attname;
+ data1_xattr *attr;
+
+ if (!p) {
+ return 1;
+ } else {
+ if (p->which == XPATH_PREDICATE_RELATION) {
+ if (p->u.relation.name[0]) {
+ if (*p->u.relation.name != '@') {
+ logf(LOG_WARN,
+ " Only attributes (@) are supported in xelm xpath predicates");
+ logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
+ return (1);
+ }
+ attname = p->u.relation.name + 1;
+ res = 0;
+ /* looking for the attribute with a specified name */
+ for (attr = n->u.tag.attributes; attr; attr = attr->next) {
+ logf(LOG_DEBUG," - attribute %s <-> %s", attname, attr->name );
+
+ if (!strcmp(attr->name, attname)) {
+ if (p->u.relation.op[0]) {
+ if (*p->u.relation.op != '=') {
+ logf(LOG_WARN,
+ "Only '=' relation is supported (%s)",p->u.relation.op);
+ logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
+ res = 1; break;
+ } else {
+ logf(LOG_DEBUG," - value %s <-> %s",
+ p->u.relation.value, attr->value );
+ if (!strcmp(attr->value, p->u.relation.value)) {
+ res = 1; break;
+ }
+ }
+ } else {
+ /* attribute exists, no value specified */
+ res = 1; break;
+ }
+ }
+ }
+ return res;
+ } else {
+ return 1;
+ }
+ }
+ else if (p->which == XPATH_PREDICATE_BOOLEAN) {
+ if (!strcmp(p->u.boolean.op,"and")) {
+ return d1_check_xpath_predicate(n, p->u.boolean.left)
+ && d1_check_xpath_predicate(n, p->u.boolean.right);
+ }
+ else if (!strcmp(p->u.boolean.op,"or")) {
+ return (d1_check_xpath_predicate(n, p->u.boolean.left)
+ || d1_check_xpath_predicate(n, p->u.boolean.right));
+ } else {
+ logf(LOG_WARN, "Unknown boolean relation %s, ignored",p->u.boolean.op);
+ return 1;
+ }
+ }
}
- }
- return 0;
+ return 0;
}
/* *ostrich*
- New function, looking for xpath "element" definitions in abs, by
- tagpath, using a kind of ugly regxp search.The DFA was built while
- parsing abs, so here we just go trough them and try to match
- against the given tagpath. The first matching entry is returned.
+New function, looking for xpath "element" definitions in abs, by
+tagpath, using a kind of ugly regxp search.The DFA was built while
+parsing abs, so here we just go trough them and try to match
+against the given tagpath. The first matching entry is returned.
- pop, 2002-12-13
+pop, 2002-12-13
- Added support for enhanced xelm. Now [] predicates are considered
- as well, when selecting indexing rules... (why the hell it's called
- termlist???)
+Added support for enhanced xelm. Now [] predicates are considered
+as well, when selecting indexing rules... (why the hell it's called
+termlist???)
- pop, 2003-01-17
+pop, 2003-01-17
- */
+*/
data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
{
struct xpath_location_step *xp;
#endif
- char *pexpr = malloc(strlen(tagpath)+2);
+ char *pexpr = xmalloc(strlen(tagpath)+2);
int ok = 0;
sprintf (pexpr, "%s\n", tagpath);
+ logf(LOG_DEBUG,"Checking tagpath %s",tagpath);
while (xpe)
{
struct DFA_state **dfaar = xpe->dfa->states;
int i;
unsigned char c;
int start_line = 1;
-
+
+ logf(LOG_DEBUG," - xpath %s",xpe->xpath_expr);
+
c = *pexpr++; t = s->trans; i = s->tran_no;
- if (c >= t->ch[0] && c <= t->ch[1]) {
+ if ((c >= t->ch[0] && c <= t->ch[1]) || (!t->ch[0])) {
p = pexpr;
do {
if ((s = dfaar[t->to])->rule_no &&
}
p++;
} while (i >= 0);
- }
+ }
pexpr--;
if (ok) {
#ifdef ENHANCED_XELM
- /* we have to check the perdicates up to the root node */
- xp = xpe->xpath;
-
- /* find the first tag up in the node structure */
- nn = n; while (nn && nn->which != DATA1N_tag) {
- nn = nn->parent;
- }
-
- /* go from inside out in the node structure, while going
- backwards trough xpath location steps ... */
- for (i=xpe->xpath_len - 1; i>0; i--) {
-
- logf(LOG_DEBUG,"Checking step %d: %s on tag %s",
+ /* we have to check the perdicates up to the root node */
+ xp = xpe->xpath;
+
+ /* find the first tag up in the node structure */
+ nn = n; while (nn && nn->which != DATA1N_tag) {
+ nn = nn->parent;
+ }
+
+ /* go from inside out in the node structure, while going
+ backwards trough xpath location steps ... */
+ for (i=xpe->xpath_len - 1; i>0; i--) {
+
+ logf(LOG_DEBUG,"Checking step %d: %s on tag %s",
i,xp[i].part,nn->u.tag.tag);
-
- if (!d1_check_xpath_predicate(nn, xp[i].predicate)) {
- logf(LOG_DEBUG," Predicates didn't match");
- ok = 0;
- break;
- }
-
- if (nn->which == DATA1N_tag) {
- nn = nn->parent;
- }
- }
+
+ if (!d1_check_xpath_predicate(nn, xp[i].predicate)) {
+ logf(LOG_DEBUG," Predicates didn't match");
+ ok = 0;
+ break;
+ }
+
+ if (nn->which == DATA1N_tag) {
+ nn = nn->parent;
+ }
+ }
#endif
- if (ok) {
- break;
- }
+ if (ok) {
+ break;
+ }
}
xpe = xpe->next;
}
+ xfree(pexpr);
+
if (ok) {
logf(LOG_DEBUG,"Got it");
return xpe->termlists;
modified by pop, 2002-12-13
*/
+/* add xpath index for an attribute */
+static void index_xpath_attr (char *tag_path, char *name, char *value,
+ char *structure, struct recExtractCtrl *p,
+ RecWord *wrd)
+{
+ wrd->attrSet = VAL_IDXPATH;
+ wrd->attrUse = 1;
+ wrd->reg_type = '0';
+ wrd->string = tag_path;
+ wrd->length = strlen(tag_path);
+ (*p->tokenAdd)(wrd);
+
+ if (value) {
+ wrd->attrUse = 1015;
+ wrd->reg_type = 'w';
+ wrd->string = value;
+ wrd->length = strlen(value);
+ (*p->tokenAdd)(wrd);
+ }
+
+ wrd->attrUse = 2;
+ wrd->reg_type = '0';
+ wrd->string = tag_path;
+ wrd->length = strlen(tag_path);
+ (*p->tokenAdd)(wrd);
+}
+
+
static void index_xpath (data1_node *n, struct recExtractCtrl *p,
int level, RecWord *wrd, int use)
{
char tag_path_full[1024];
size_t flen = 0;
data1_node *nn;
+ int termlist_only = 1;
+
+ if ((!n->root->u.root.absyn) ||
+ (n->root->u.root.absyn->enable_xpath_indexing)) {
+ termlist_only = 0;
+ }
switch (n->which)
{
int xpdone = 0;
flen = 0;
+ // logf (LOG_LOG,"Index value, %s",wrd->string);
+
/* we have to fetch the whole path to the data tag */
for (nn = n; nn; nn = nn->parent) {
if (nn->which == DATA1N_tag) {
}
}
/* xpath indexing is done, if there was no termlist given,
- or no ! attribute... */
- if (!xpdone) {
+ or no ! in the termlist, and default indexing is enabled... */
+ if ((!xpdone) && (!termlist_only)) {
wrd->attrSet = VAL_IDXPATH;
wrd->attrUse = use;
wrd->reg_type = 'w';
(*p->tokenAdd)(wrd);
}
- }
+ }
break;
case DATA1N_tag:
flen = 0;
else
{
data1_xattr *xp;
- (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */
- if (use == 1)
+ 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 */
+ if (!(do_xpindex = 1 - termlist_only)) {
+ if ((tl = xpath_termlist_by_tagpath(tag_path_full, n))) {
+ for (; tl; tl = tl->next) { if (!tl->att) {do_xpindex = 1;} }
+ }
+ }
+ if (do_xpindex) {
+ (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */
+ }
+
+ if (use == 1) /* only for the starting tag... */
{
- for (xp = n->u.tag.attributes; xp; xp = xp->next)
- {
+
+#define MAX_ATTR_COUNT 50
+
+ data1_termlist *tll[MAX_ATTR_COUNT];
+
+ int i = 0;
+
+ /* get termlists for attributes, and find out, if we have to do xpath indexing */
+ for (xp = n->u.tag.attributes; xp; xp = xp->next) {
+ i++;
+ }
+
+ i = 0;
+ for (xp = n->u.tag.attributes; xp; xp = xp->next) {
char comb[512];
- /* attribute (no value) */
- wrd->reg_type = '0';
- wrd->attrUse = 3;
- wrd->string = xp->name;
- wrd->length = strlen(xp->name);
+ int do_xpindex = 1 - termlist_only;
+ data1_termlist *tl;
+ char attr_tag_path_full[1024];
+ int int_len = flen;
- wrd->seqno--;
- (*p->tokenAdd)(wrd);
-
- if (xp->value &&
- strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2)
+ /* this could be cached as well */
+ sprintf (attr_tag_path_full, "@%s/%.*s",
+ xp->name, int_len, tag_path_full);
+
+ tll[i] = xpath_termlist_by_tagpath(attr_tag_path_full,n);
+
+ /* if there is a ! in the xelm termlist, or default indexing is on,
+ proceed with xpath idx */
+ if ((tl = tll[i]))
{
- /* attribute value exact */
- strcpy (comb, xp->name);
- strcat (comb, "=");
- strcat (comb, xp->value);
+ for (; tl; tl = tl->next)
+ {
+ if (!tl->att)
+ do_xpindex = 1;
+ }
+ }
+
+ if (do_xpindex) {
- wrd->attrUse = 3;
+ /* attribute (no value) */
wrd->reg_type = '0';
- wrd->string = comb;
- wrd->length = strlen(comb);
- wrd->seqno--;
+ wrd->attrUse = 3;
+ wrd->string = xp->name;
+ wrd->length = strlen(xp->name);
+ wrd->seqno--;
(*p->tokenAdd)(wrd);
- }
- }
- for (xp = n->u.tag.attributes; xp; xp = xp->next)
- {
+
+ if (xp->value &&
+ strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2) {
+
+ /* attribute value exact */
+ strcpy (comb, xp->name);
+ strcat (comb, "=");
+ strcat (comb, xp->value);
+
+ wrd->attrUse = 3;
+ wrd->reg_type = '0';
+ wrd->string = comb;
+ wrd->length = strlen(comb);
+ wrd->seqno--;
+
+ (*p->tokenAdd)(wrd);
+ }
+ }
+ i++;
+ }
+
+ i = 0;
+ 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);
- wrd->reg_type = '0';
- wrd->attrUse = 1;
- wrd->string = attr_tag_path_full;
- wrd->length = strlen(attr_tag_path_full);
- (*p->tokenAdd)(wrd);
- if (xp->value)
- {
- /* the same jokes, as with the data nodes ... */
- data1_termlist *tl;
- int xpdone = 0;
-
- wrd->string = xp->value;
- wrd->length = strlen(xp->value);
- wrd->reg_type = 'w';
-
- if ((tl = xpath_termlist_by_tagpath(attr_tag_path_full,
- n))) {
- for (; tl; tl = tl->next) {
- wrd->reg_type = *tl->structure;
- if (!tl->att) {
- wrd->attrSet = VAL_IDXPATH;
- wrd->attrUse = 1015;
- (*p->tokenAdd)(wrd);
- xpdone = 1;
- } else {
- wrd->attrSet = (int) (tl->att->parent->reference);
+ if ((tl = tll[i]))
+ {
+ /* If there is a termlist given (=xelm directive) */
+ for (; tl; tl = tl->next)
+ {
+ if (!tl->att) {
+ /* add xpath index for the attribute */
+ index_xpath_attr (attr_tag_path_full, xp->name,
+ xp->value, tl->structure,
+ p, wrd);
+ xpdone = 1;
+ } else {
+ /* add attribute based index for the attribute */
+ if (xp->value) {
+ wrd->attrSet = (int)
+ (tl->att->parent->reference);
wrd->attrUse = tl->att->locals->local;
+ wrd->reg_type = *tl->structure;
+ wrd->string = xp->value;
+ wrd->length = strlen(xp->value);
(*p->tokenAdd)(wrd);
}
}
-
- }
- if (!xpdone) {
- wrd->attrSet = VAL_IDXPATH;
- wrd->attrUse = 1015;
- wrd->reg_type = 'w';
- (*p->tokenAdd)(wrd);
}
}
-
- wrd->attrSet = VAL_IDXPATH;
- wrd->reg_type = '0';
- wrd->attrUse = 2;
- wrd->string = attr_tag_path_full;
- wrd->length = strlen(attr_tag_path_full);
- (*p->tokenAdd)(wrd);
+ /* if there was no termlist for the given path,
+ or the termlist didn't have a ! element, index
+ the attribute as "w" */
+ if ((!xpdone) && (!termlist_only))
+ {
+ index_xpath_attr (attr_tag_path_full, xp->name,
+ xp->value, "w", p, wrd);
+ }
+ i++;
}
}
- }
+ }
}
}
tlist->att->name, tlist->att->value,
tlist->source);
printf (" data=\"");
- for (i = 0; i<wrd->length && i < 8; i++)
+ for (i = 0; i<wrd->length && i < 40; i++)
fputc (wrd->string[i], stdout);
fputc ('"', stdout);
- if (wrd->length > 8)
+ if (wrd->length > 40)
printf (" ...");
fputc ('\n', stdout);
}
index_termlist (n, n, p, level, wrd);
/* index start tag */
assert (n->root->u.root.absyn);
-
- if (!n->root->u.root.absyn)
- index_xpath (n, p, level, wrd, 1);
- else if (n->root->u.root.absyn->enable_xpath_indexing)
- index_xpath (n, p, level, wrd, 1);
- }
+ index_xpath (n, p, level, wrd, 1);
+ }
if (n->child)
if (dumpkeys(n->child, p, level + 1, wrd) < 0)
if (par)
index_termlist (par, n, p, level, wrd);
- if (!n->root->u.root.absyn)
- index_xpath (n, p, level, wrd, 1016);
- else if (n->root->u.root.absyn->enable_xpath_indexing)
- index_xpath (n, p, level, wrd, 1016);
+
+ index_xpath (n, p, level, wrd, 1016);
}
if (n->which == DATA1N_tag)
{
/* index end tag */
- if (!n->root->u.root.absyn)
- index_xpath (n, p, level, wrd, 2);
- else if (n->root->u.root.absyn->enable_xpath_indexing)
- index_xpath (n, p, level, wrd, 2);
+ index_xpath (n, p, level, wrd, 2);
}
if (p->flagShowRecords && n->which == DATA1N_root)
if ((oid_ent_to_oid (&oe, oidtmp)))
(*p->schemaAdd)(p, oidtmp);
}
+ data1_concat_text(p->dh, mem, n);
/* ensure our data1 tree is UTF-8 */
data1_iconv (p->dh, mem, n, "UTF-8", data1_get_encoding(p->dh, n));
nmem_destroy (mem);
return 0;
}
+ data1_concat_text(p->dh, mem, node);
+
/* ensure our data1 tree is UTF-8 */
data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node));
projects / idzebra-moved-to-github.git / blobdiff