/* CCL find (to rpn conversion)
* Europagate, 1995
*
- * $Id: cclfind.c,v 1.27 2001-11-27 22:38:50 adam Exp $
+ * $Id: cclfind.c,v 1.34 2003-06-23 10:22:21 adam Exp $
*
* Old Europagate log:
*
for (i = 0; (q=qa[i]); i++)
while (q)
{
- if (q->type == type && q->value == value)
+ if (q->type == type && q->kind == CCL_RPN_ATTR_NUMERIC &&
+ q->value.numeric == value)
{
if (attset)
*attset = q->set;
*/
static char *copy_token_name (struct ccl_token *tp)
{
- char *str = (char *)malloc (tp->len + 1);
+ char *str = (char *)xmalloc (tp->len + 1);
ccl_assert (str);
memcpy (str, tp->name, tp->len);
str[tp->len] = '\0';
static struct ccl_rpn_node *mk_node (int kind)
{
struct ccl_rpn_node *p;
- p = (struct ccl_rpn_node *)malloc (sizeof(*p));
+ p = (struct ccl_rpn_node *)xmalloc (sizeof(*p));
ccl_assert (p);
p->kind = kind;
return p;
ccl_rpn_delete (rpn->u.p[1]);
break;
case CCL_RPN_TERM:
- free (rpn->u.t.term);
+ xfree (rpn->u.t.term);
for (attr = rpn->u.t.attr_list; attr; attr = attr1)
{
attr1 = attr->next;
+ if (attr->kind == CCL_RPN_ATTR_STRING)
+ xfree(attr->value.str);
if (attr->set)
- free (attr->set);
- free (attr);
+ xfree (attr->set);
+ xfree (attr);
}
break;
case CCL_RPN_SET:
- free (rpn->u.setname);
+ xfree (rpn->u.setname);
break;
case CCL_RPN_PROX:
ccl_rpn_delete (rpn->u.p[0]);
ccl_rpn_delete (rpn->u.p[1]);
break;
}
- free (rpn);
+ xfree (rpn);
}
static struct ccl_rpn_node *find_spec (CCL_parser cclp,
static struct ccl_rpn_node *search_terms (CCL_parser cclp,
struct ccl_rpn_attr **qa);
-/*
- * add_attr: Add attribute (type/value) to RPN term node.
- * p: RPN node of type term.
- * type: Type of attribute
- * value: Value of attribute
- * set: Attribute set name
- */
-static void add_attr (struct ccl_rpn_node *p, const char *set,
- int type, int value)
+static struct ccl_rpn_attr *add_attr_node (struct ccl_rpn_node *p,
+ const char *set, int type)
{
struct ccl_rpn_attr *n;
-
- n = (struct ccl_rpn_attr *)malloc (sizeof(*n));
+
+ n = (struct ccl_rpn_attr *)xmalloc (sizeof(*n));
ccl_assert (n);
if (set)
{
- n->set = (char*) malloc (strlen(set)+1);
+ n->set = (char*) xmalloc (strlen(set)+1);
strcpy (n->set, set);
}
else
n->set = 0;
n->type = type;
- n->value = value;
n->next = p->u.t.attr_list;
p->u.t.attr_list = n;
+
+ n->kind = CCL_RPN_ATTR_NUMERIC;
+ n->value.numeric = 0;
+ return n;
}
/*
+ * add_attr_numeric: Add attribute (type/value) to RPN term node.
+ * p: RPN node of type term.
+ * type: Type of attribute
+ * value: Value of attribute
+ * set: Attribute set name
+ */
+static void add_attr_numeric (struct ccl_rpn_node *p, const char *set,
+ int type, int value)
+{
+ struct ccl_rpn_attr *n;
+
+ n = add_attr_node(p, set, type);
+ n->kind = CCL_RPN_ATTR_NUMERIC;
+ n->value.numeric = value;
+}
+
+static void add_attr_string (struct ccl_rpn_node *p, const char *set,
+ int type, char *value)
+{
+ struct ccl_rpn_attr *n;
+
+ n = add_attr_node(p, set, type);
+ n->kind = CCL_RPN_ATTR_STRING;
+ n->value.str = xstrdup(value);
+}
+
+
+/*
* search_term: Parse CCL search term.
* cclp: CCL Parser
* qa: Qualifier attributes already applied.
struct ccl_rpn_attr *attr;
for (attr = qa[i]; attr; attr = attr->next)
- if (attr->value > 0)
- { /* deal only with REAL attributes (positive) */
- switch (attr->type)
- {
- case CCL_BIB1_REL:
- if (relation_value != -1)
- continue;
- relation_value = attr->value;
- break;
- case CCL_BIB1_POS:
- if (position_value != -1)
- continue;
- position_value = attr->value;
- break;
- case CCL_BIB1_STR:
- if (structure_value != -1)
- continue;
- structure_value = attr->value;
- break;
- case CCL_BIB1_TRU:
- if (truncation_value != -1)
- continue;
- truncation_value = attr->value;
- left_trunc = right_trunc = mid_trunc = 0;
- break;
- case CCL_BIB1_COM:
- if (completeness_value != -1)
- continue;
- completeness_value = attr->value;
- break;
- }
- add_attr (p, attr->set, attr->type, attr->value);
- }
+ switch(attr->kind)
+ {
+ case CCL_RPN_ATTR_STRING:
+ add_attr_string(p, attr->set, attr->type,
+ attr->value.str);
+ break;
+ case CCL_RPN_ATTR_NUMERIC:
+ if (attr->value.numeric > 0)
+ { /* deal only with REAL attributes (positive) */
+ switch (attr->type)
+ {
+ case CCL_BIB1_REL:
+ if (relation_value != -1)
+ continue;
+ relation_value = attr->value.numeric;
+ break;
+ case CCL_BIB1_POS:
+ if (position_value != -1)
+ continue;
+ position_value = attr->value.numeric;
+ break;
+ case CCL_BIB1_STR:
+ if (structure_value != -1)
+ continue;
+ structure_value = attr->value.numeric;
+ break;
+ case CCL_BIB1_TRU:
+ if (truncation_value != -1)
+ continue;
+ truncation_value = attr->value.numeric;
+ left_trunc = right_trunc = mid_trunc = 0;
+ break;
+ case CCL_BIB1_COM:
+ if (completeness_value != -1)
+ continue;
+ completeness_value = attr->value.numeric;
+ break;
+ }
+ add_attr_numeric(p, attr->set, attr->type,
+ attr->value.numeric);
+ }
+ }
}
/* len now holds the number of characters in the RPN term */
/* no holds the number of CCL tokens (1 or more) */
{ /* no structure attribute met. Apply either structure attribute
WORD or PHRASE depending on number of CCL tokens */
if (no == 1 && no_spaces == 0)
- add_attr (p, attset, CCL_BIB1_STR, 2);
+ add_attr_numeric (p, attset, CCL_BIB1_STR, 2);
else
- add_attr (p, attset, CCL_BIB1_STR, 1);
+ add_attr_numeric (p, attset, CCL_BIB1_STR, 1);
}
/* make the RPN token */
- p->u.t.term = (char *)malloc (len);
+ p->u.t.term = (char *)xmalloc (len);
ccl_assert (p->u.t.term);
p->u.t.term[0] = '\0';
for (i = 0; i<no; i++)
ccl_rpn_delete (p);
return NULL;
}
- add_attr (p, attset, CCL_BIB1_TRU, 3);
+ add_attr_numeric (p, attset, CCL_BIB1_TRU, 3);
}
else if (right_trunc)
{
ccl_rpn_delete (p);
return NULL;
}
- add_attr (p, attset, CCL_BIB1_TRU, 1);
+ add_attr_numeric (p, attset, CCL_BIB1_TRU, 1);
}
else if (left_trunc)
{
ccl_rpn_delete (p);
return NULL;
}
- add_attr (p, attset, CCL_BIB1_TRU, 2);
+ add_attr_numeric (p, attset, CCL_BIB1_TRU, 2);
}
else
{
if (qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE,
&attset))
- add_attr (p, attset, CCL_BIB1_TRU, 100);
+ add_attr_numeric (p, attset, CCL_BIB1_TRU, 100);
}
if (!multi)
break;
}
p = mk_node (CCL_RPN_AND);
p->u.p[0] = p1;
- add_attr (p1, attset, CCL_BIB1_REL, 4);
+ add_attr_numeric (p1, attset, CCL_BIB1_REL, 4);
p->u.p[1] = p2;
- add_attr (p2, attset, CCL_BIB1_REL, 2);
+ add_attr_numeric (p2, attset, CCL_BIB1_REL, 2);
return p;
}
else /* = term - */
{
- add_attr (p1, attset, CCL_BIB1_REL, 4);
+ add_attr_numeric (p1, attset, CCL_BIB1_REL, 4);
return p1;
}
}
else if (cclp->look_token->len == 1 &&
- cclp->look_token->name[0] == '"') /* = - term ? */
+ cclp->look_token->name[0] == '-') /* = - term ? */
{
ADVANCE;
if (!(p = search_term (cclp, ap)))
return NULL;
- add_attr (p, attset, CCL_BIB1_REL, 2);
+ add_attr_numeric (p, attset, CCL_BIB1_REL, 2);
return p;
}
else if (KIND == CCL_TOK_LP)
{
if (!(p = search_terms (cclp, ap)))
return NULL;
- add_attr (p, attset, CCL_BIB1_REL, rel);
+ add_attr_numeric (p, attset, CCL_BIB1_REL, rel);
return p;
}
cclp->error_code = CCL_ERR_TERM_EXPECTED;
}
/*
- * qualifiers: Parse CCL qualifiers and search terms.
+ * qualifiers1: Parse CCL qualifiers and search terms.
* cclp: CCL Parser
* la: Token pointer to RELATION token.
* qa: Qualifier attributes already applied.
* return: pointer to node(s); NULL on error.
*/
-static struct ccl_rpn_node *qualifiers (CCL_parser cclp, struct ccl_token *la,
- struct ccl_rpn_attr **qa)
+static struct ccl_rpn_node *qualifiers1 (CCL_parser cclp, struct ccl_token *la,
+ struct ccl_rpn_attr **qa)
{
struct ccl_token *lookahead = cclp->look_token;
struct ccl_token *look_start = cclp->look_token;
struct ccl_rpn_attr **ap;
struct ccl_rpn_node *node = 0;
+ const char *field_str;
int no = 0;
int seq = 0;
int i;
+ int mode_merge = 1;
#if 0
if (qa)
{
if (qa)
for (i=0; qa[i]; i++)
no++;
- ap = (struct ccl_rpn_attr **)malloc ((no+1) * sizeof(*ap));
+ ap = (struct ccl_rpn_attr **)xmalloc ((no ? (no+1) : 2) * sizeof(*ap));
ccl_assert (ap);
- while (1)
+ field_str = ccl_qual_search_special(cclp->bibset, "field");
+ if (field_str)
+ {
+ if (!strcmp (field_str, "or"))
+ mode_merge = 0;
+ else if (!strcmp (field_str, "merge"))
+ mode_merge = 1;
+ }
+ if (!mode_merge)
{
- struct ccl_rpn_node *node_sub;
- int found = 0;
+ /* consider each field separately and OR */
lookahead = look_start;
- for (i = 0; lookahead != la; i++)
+ while (lookahead != la)
{
- ap[i] = ccl_qual_search (cclp, lookahead->name,
- lookahead->len, seq);
- if (ap[i])
- found++;
- if (!ap[i] && seq > 0)
- ap[i] = ccl_qual_search (cclp, lookahead->name,
- lookahead->len, 0);
- if (!ap[i])
+ ap[1] = 0;
+ seq = 0;
+ while ((ap[0] = ccl_qual_search (cclp, lookahead->name,
+ lookahead->len, seq)) != 0)
+ {
+ struct ccl_rpn_node *node_sub;
+ cclp->look_token = la;
+
+ node_sub = qualifiers2(cclp, ap);
+ if (!node_sub)
+ {
+ ccl_rpn_delete (node);
+ xfree (ap);
+ return 0;
+ }
+ if (node)
+ {
+ struct ccl_rpn_node *node_this = mk_node(CCL_RPN_OR);
+ node_this->u.p[0] = node;
+ node_this->u.p[1] = node_sub;
+ node = node_this;
+ }
+ else
+ node = node_sub;
+ seq++;
+ }
+ if (seq == 0)
{
cclp->look_token = lookahead;
cclp->error_code = CCL_ERR_UNKNOWN_QUAL;
- free (ap);
+ xfree (ap);
return NULL;
}
lookahead = lookahead->next;
if (lookahead->kind == CCL_TOK_COMMA)
lookahead = lookahead->next;
}
- if (qa)
- while (*qa)
- ap[i++] = *qa++;
- ap[i] = NULL;
-
- if (!found)
- break;
-
- cclp->look_token = lookahead;
-
- node_sub = qualifiers2(cclp, ap);
- if (!node_sub)
- {
- ccl_rpn_delete (node);
- break;
- }
- if (node)
+ }
+ else
+ {
+ /* merge attributes from ALL fields - including inherited ones */
+ while (1)
{
- struct ccl_rpn_node *node_this = mk_node(CCL_RPN_OR);
- node_this->u.p[0] = node;
- node_this->u.p[1] = node_sub;
- node = node_this;
+ struct ccl_rpn_node *node_sub;
+ int found = 0;
+ lookahead = look_start;
+ for (i = 0; lookahead != la; i++)
+ {
+ ap[i] = ccl_qual_search (cclp, lookahead->name,
+ lookahead->len, seq);
+ if (ap[i])
+ found++;
+ if (!ap[i] && seq > 0)
+ ap[i] = ccl_qual_search (cclp, lookahead->name,
+ lookahead->len, 0);
+ if (!ap[i])
+ {
+ cclp->look_token = lookahead;
+ cclp->error_code = CCL_ERR_UNKNOWN_QUAL;
+ xfree (ap);
+ return NULL;
+ }
+ lookahead = lookahead->next;
+ if (lookahead->kind == CCL_TOK_COMMA)
+ lookahead = lookahead->next;
+ }
+ if (qa)
+ {
+ struct ccl_rpn_attr **qa0 = qa;
+
+ while (*qa0)
+ ap[i++] = *qa0++;
+ }
+ ap[i] = NULL;
+
+ if (!found)
+ break;
+
+ cclp->look_token = lookahead;
+
+ node_sub = qualifiers2(cclp, ap);
+ if (!node_sub)
+ {
+ ccl_rpn_delete (node);
+ break;
+ }
+ if (node)
+ {
+ struct ccl_rpn_node *node_this = mk_node(CCL_RPN_OR);
+ node_this->u.p[0] = node;
+ node_this->u.p[1] = node_sub;
+ node = node_this;
+ }
+ else
+ node = node_sub;
+ seq++;
}
- else
- node = node_sub;
- seq++;
}
- free (ap);
+ xfree (ap);
return node;
}
struct ccl_rpn_attr **qa)
{
static int list[] = {
- CCL_TOK_TERM, CCL_TOK_COMMA,CCL_TOK_EQ, CCL_TOK_REL, -1};
+ CCL_TOK_TERM, CCL_TOK_COMMA,CCL_TOK_EQ, CCL_TOK_REL, CCL_TOK_SET, -1};
struct ccl_rpn_node *p1, *p2, *pn;
p1 = search_term_x (cclp, qa, list, 1);
if (!p1)
{
if (KIND == CCL_TOK_PROX)
{
+ struct ccl_rpn_node *p_prox = 0;
+ /* ! word order specified */
+ /* % word order not specified */
+ p_prox = mk_node(CCL_RPN_TERM);
+ p_prox->u.t.term = (char *) xmalloc(cclp->look_token->len);
+ memcpy(p_prox->u.t.term, cclp->look_token->name,
+ cclp->look_token->len);
+ p_prox->u.t.term[cclp->look_token->len] = 0;
+ p_prox->u.t.attr_list = 0;
+
ADVANCE;
p2 = search_term_x (cclp, qa, list, 1);
if (!p2)
pn = mk_node (CCL_RPN_PROX);
pn->u.p[0] = p1;
pn->u.p[1] = p2;
+ pn->u.p[2] = p_prox;
p1 = pn;
}
else if (is_term_ok(KIND, list))
pn = mk_node (CCL_RPN_PROX);
pn->u.p[0] = p1;
pn->u.p[1] = p2;
+ pn->u.p[2] = 0;
p1 = pn;
}
else
{
lookahead = lookahead->next;
if (lookahead->kind == CCL_TOK_REL || lookahead->kind == CCL_TOK_EQ)
- return qualifiers (cclp, lookahead, qa);
+ return qualifiers1 (cclp, lookahead, qa);
if (lookahead->kind != CCL_TOK_COMMA)
break;
lookahead = lookahead->next;
struct ccl_rpn_node *node_this = mk_node(CCL_RPN_OR);
node_this->u.p[0] = node;
node_this->u.p[1] = node_sub;
+ node_this->u.p[2] = 0;
node = node_this;
}
else
pn = mk_node (CCL_RPN_AND);
pn->u.p[0] = p1;
pn->u.p[1] = p2;
+ pn->u.p[2] = 0;
p1 = pn;
continue;
case CCL_TOK_OR:
pn = mk_node (CCL_RPN_OR);
pn->u.p[0] = p1;
pn->u.p[1] = p2;
+ pn->u.p[2] = 0;
p1 = pn;
continue;
case CCL_TOK_NOT:
pn = mk_node (CCL_RPN_NOT);
pn->u.p[0] = p1;
pn->u.p[1] = p2;
+ pn->u.p[2] = 0;
p1 = pn;
continue;
}