X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=ccl%2Fcclfind.c;h=ba12079846fb8b2047839f74bddd24c9df630336;hb=45f8f517ebe10930067907c19a56557afd779826;hp=efd807bcd813546e8c8e7ee60f22bfc7f1abe08a;hpb=3888e88cde70f36851544ab7acc04e0943470ec6;p=yaz-moved-to-github.git diff --git a/ccl/cclfind.c b/ccl/cclfind.c index efd807b..ba12079 100644 --- a/ccl/cclfind.c +++ b/ccl/cclfind.c @@ -45,7 +45,49 @@ * Europagate, 1995 * * $Log: cclfind.c,v $ - * Revision 1.5 1996-10-11 15:00:24 adam + * Revision 1.17 2000-05-01 09:36:50 adam + * Range operator only treated in ordered ranges so that minus (-) can be + * used for, say, the and-not operator. + * + * Revision 1.16 2000/03/14 09:06:11 adam + * Added POSIX threads support for frontend server. + * + * Revision 1.15 2000/02/24 23:49:13 adam + * Fixed memory allocation problem. + * + * Revision 1.14 2000/01/31 13:15:21 adam + * Removed uses of assert(3). Cleanup of ODR. CCL parser update so + * that some characters are not surrounded by spaces in resulting term. + * ILL-code updates. + * + * Revision 1.13 1999/12/22 13:13:32 adam + * Search terms may include "operators" without causing error. + * + * Revision 1.12 1999/11/30 13:47:11 adam + * Improved installation. Moved header files to include/yaz. + * + * Revision 1.11 1999/03/31 11:15:37 adam + * Fixed memory leaks in ccl_find_str and ccl_qual_rm. + * + * Revision 1.10 1998/02/11 11:53:33 adam + * Changed code so that it compiles as C++. + * + * Revision 1.9 1997/09/29 08:56:37 adam + * Changed CCL parser to be thread safe. New type, CCL_parser, declared + * and a create/destructers ccl_parser_create/ccl_parser/destory has + * been added. + * + * Revision 1.8 1997/09/01 08:48:11 adam + * New windows NT/95 port using MSV5.0. Only a few changes made + * to avoid warnings. + * + * Revision 1.7 1997/05/14 06:53:26 adam + * C++ support. + * + * Revision 1.6 1997/04/30 08:52:06 quinn + * Null + * + * Revision 1.5 1996/10/11 15:00:24 adam * CCL parser from Europagate Email gateway 1.0. * * Revision 1.16 1996/01/08 08:41:13 adam @@ -98,27 +140,16 @@ * */ -#include #include -#include #include -#include - -/* current lookahead token */ -static struct ccl_token *look_token; - -/* holds error no if error occur */ -static int ccl_error; - -/* current bibset */ -static CCL_bibset bibset; +#include /* returns type of current lookahead */ -#define KIND (look_token->kind) +#define KIND (cclp->look_token->kind) /* move one token forward */ -#define ADVANCE look_token = look_token->next +#define ADVANCE cclp->look_token = cclp->look_token->next /* * qual_val_type: test for existance of attribute type/value pair. @@ -166,8 +197,8 @@ static void strxcat (char *n, const char *src, int len) */ static char *copy_token_name (struct ccl_token *tp) { - char *str = malloc (tp->len + 1); - assert (str); + char *str = (char *)malloc (tp->len + 1); + ccl_assert (str); memcpy (str, tp->name, tp->len); str[tp->len] = '\0'; return str; @@ -178,11 +209,11 @@ static char *copy_token_name (struct ccl_token *tp) * kind: Type of node. * return: pointer to allocated node. */ -static struct ccl_rpn_node *mk_node (enum rpn_node_kind kind) +static struct ccl_rpn_node *mk_node (int kind) { struct ccl_rpn_node *p; - p = malloc (sizeof(*p)); - assert (p); + p = (struct ccl_rpn_node *)malloc (sizeof(*p)); + ccl_assert (p); p->kind = kind; return p; } @@ -223,8 +254,19 @@ void ccl_rpn_delete (struct ccl_rpn_node *rpn) free (rpn); } -static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa); -static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa); +static struct ccl_rpn_node *find_spec (CCL_parser cclp, + struct ccl_rpn_attr **qa); + +static int is_term_ok (int look, int *list) +{ + for (;*list >= 0; list++) + if (look == *list) + return 1; + return 0; +} + +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. @@ -236,8 +278,8 @@ static void add_attr (struct ccl_rpn_node *p, int type, int value) { struct ccl_rpn_attr *n; - n = malloc (sizeof(*n)); - assert (n); + n = (struct ccl_rpn_attr *)malloc (sizeof(*n)); + ccl_assert (n); n->type = type; n->value = value; n->next = p->u.t.attr_list; @@ -246,15 +288,19 @@ static void add_attr (struct ccl_rpn_node *p, int type, int value) /* * search_term: Parse CCL search term. + * cclp: CCL Parser * qa: Qualifier attributes already applied. * return: pointer to node(s); NULL on error. */ -static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) +static struct ccl_rpn_node *search_term_x (CCL_parser cclp, + struct ccl_rpn_attr **qa, + int *term_list) { + struct ccl_rpn_attr *qa_tmp[2]; struct ccl_rpn_node *p; - struct ccl_token *lookahead = look_token; + struct ccl_token *lookahead = cclp->look_token; int len = 0; - int no, i; + size_t no, i; int left_trunc = 0; int right_trunc = 0; int mid_trunc = 0; @@ -264,9 +310,9 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) int truncation_value = -1; int completeness_value = -1; - if (KIND != CCL_TOK_TERM) + if (!is_term_ok(KIND, term_list)) { - ccl_error = CCL_ERR_TERM_EXPECTED; + cclp->error_code = CCL_ERR_TERM_EXPECTED; return NULL; } /* create the term node, but wait a moment before adding the term */ @@ -278,9 +324,9 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) { /* no qualifier(s) applied. Use 'term' if it is defined */ - qa = malloc (2*sizeof(*qa)); - assert (qa); - qa[0] = ccl_qual_search (bibset, "term", 4); + qa = qa_tmp; + ccl_assert (qa); + qa[0] = ccl_qual_search (cclp, "term", 4); qa[1] = NULL; } @@ -326,14 +372,14 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) /* go through each TERM token. If no truncation attribute is yet met, then look for left/right truncation markers (?) and set left_trunc/right_trunc/mid_trunc accordingly */ - for (no = 0; lookahead->kind == CCL_TOK_TERM; no++) + for (no = 0; is_term_ok(lookahead->kind, term_list); no++) { for (i = 0; ilen; i++) if (truncation_value == -1 && lookahead->name[i] == '?') { if (no == 0 && i == 0 && lookahead->len >= 1) left_trunc = 1; - else if (lookahead->next->kind != CCL_TOK_TERM && + else if (!is_term_ok(lookahead->next->kind, term_list) && i == lookahead->len-1 && i >= 1) right_trunc = 1; else @@ -356,13 +402,13 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) } /* make the RPN token */ - p->u.t.term = malloc (len); - assert (p->u.t.term); + p->u.t.term = (char *)malloc (len); + ccl_assert (p->u.t.term); p->u.t.term[0] = '\0'; for (i = 0; iname; - int src_len = look_token->len; + const char *src_str = cclp->look_token->name; + int src_len = cclp->look_token->len; if (i == 0 && left_trunc) { @@ -371,8 +417,16 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) } else if (i == no-1 && right_trunc) src_len--; - if (i) - strcat (p->u.t.term, " "); + if (src_len) + { + int len = strlen(p->u.t.term); + if (len && + !strchr("-+", *src_str) && + !strchr("-+", p->u.t.term[len-1])) + { + strcat (p->u.t.term, " "); + } + } strxcat (p->u.t.term, src_str, src_len); ADVANCE; } @@ -380,8 +434,7 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) { if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH)) { - ccl_error = CCL_ERR_TRUNC_NOT_BOTH; - free (qa); + cclp->error_code = CCL_ERR_TRUNC_NOT_BOTH; ccl_rpn_delete (p); return NULL; } @@ -391,8 +444,7 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) { if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT)) { - ccl_error = CCL_ERR_TRUNC_NOT_RIGHT; - free (qa); + cclp->error_code = CCL_ERR_TRUNC_NOT_RIGHT; ccl_rpn_delete (p); return NULL; } @@ -402,8 +454,7 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) { if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_LEFT)) { - ccl_error = CCL_ERR_TRUNC_NOT_LEFT; - free (qa); + cclp->error_code = CCL_ERR_TRUNC_NOT_LEFT; ccl_rpn_delete (p); return NULL; } @@ -417,39 +468,49 @@ static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa) return p; } +static struct ccl_rpn_node *search_term (CCL_parser cclp, + struct ccl_rpn_attr **qa) +{ + static int list[] = {CCL_TOK_TERM, CCL_TOK_COMMA, -1}; + return search_term_x(cclp, qa, list); +} + /* * qualifiers: 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 (struct ccl_token *la, +static struct ccl_rpn_node *qualifiers (CCL_parser cclp, struct ccl_token *la, struct ccl_rpn_attr **qa) { - struct ccl_token *lookahead = look_token; + struct ccl_token *lookahead = cclp->look_token; struct ccl_rpn_attr **ap; int no = 0; int i, rel; #if 0 if (qa) { - ccl_error = CCL_ERR_DOUBLE_QUAL; + cclp->error_code = CCL_ERR_DOUBLE_QUAL; return NULL; } #endif - for (lookahead = look_token; lookahead != la; lookahead=lookahead->next) + for (lookahead = cclp->look_token; lookahead != la; + lookahead=lookahead->next) no++; if (qa) for (i=0; qa[i]; i++) no++; - ap = malloc ((no+1) * sizeof(*ap)); - assert (ap); - for (i = 0; look_token != la; i++) + ap = (struct ccl_rpn_attr **)malloc ((no+1) * sizeof(*ap)); + ccl_assert (ap); + for (i = 0; cclp->look_token != la; i++) { - ap[i] = ccl_qual_search (bibset, look_token->name, look_token->len); + ap[i] = ccl_qual_search (cclp, cclp->look_token->name, + cclp->look_token->len); if (!ap[i]) { - ccl_error = CCL_ERR_UNKNOWN_QUAL; + cclp->error_code = CCL_ERR_UNKNOWN_QUAL; free (ap); return NULL; } @@ -467,7 +528,7 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, struct ccl_rpn_node *p; if (KIND != CCL_TOK_EQ) { - ccl_error = CCL_ERR_EQ_EXPECTED; + cclp->error_code = CCL_ERR_EQ_EXPECTED; free (ap); return NULL; } @@ -475,14 +536,14 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, if (KIND == CCL_TOK_LP) { ADVANCE; - if (!(p = find_spec (ap))) + if (!(p = find_spec (cclp, ap))) { free (ap); return NULL; } if (KIND != CCL_TOK_RP) { - ccl_error = CCL_ERR_RP_EXPECTED; + cclp->error_code = CCL_ERR_RP_EXPECTED; ccl_rpn_delete (p); free (ap); return NULL; @@ -490,40 +551,43 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, ADVANCE; } else - p = search_terms (ap); + p = search_terms (cclp, ap); free (ap); return p; } + /* ordered relation ... */ rel = 0; - if (look_token->len == 1) + if (cclp->look_token->len == 1) { - if (look_token->name[0] == '<') + if (cclp->look_token->name[0] == '<') rel = 1; - else if (look_token->name[0] == '=') + else if (cclp->look_token->name[0] == '=') rel = 3; - else if (look_token->name[0] == '>') + else if (cclp->look_token->name[0] == '>') rel = 5; } - else if (look_token->len == 2) + else if (cclp->look_token->len == 2) { - if (!memcmp (look_token->name, "<=", 2)) + if (!memcmp (cclp->look_token->name, "<=", 2)) rel = 2; - else if (!memcmp (look_token->name, ">=", 2)) + else if (!memcmp (cclp->look_token->name, ">=", 2)) rel = 4; - else if (!memcmp (look_token->name, "<>", 2)) + else if (!memcmp (cclp->look_token->name, "<>", 2)) rel = 6; } if (!rel) - ccl_error = CCL_ERR_BAD_RELATION; + cclp->error_code = CCL_ERR_BAD_RELATION; else { struct ccl_rpn_node *p; ADVANCE; /* skip relation */ - if (KIND == CCL_TOK_TERM && look_token->next->kind == CCL_TOK_MINUS) + if (KIND == CCL_TOK_TERM && + cclp->look_token->next->len == 1 && + cclp->look_token->next->name[0] == '-') { struct ccl_rpn_node *p1; - if (!(p1 = search_term (ap))) + if (!(p1 = search_term (cclp, ap))) { free (ap); return NULL; @@ -533,7 +597,7 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, { struct ccl_rpn_node *p2; - if (!(p2 = search_term (ap))) + if (!(p2 = search_term (cclp, ap))) { ccl_rpn_delete (p1); free (ap); @@ -554,10 +618,11 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, return p1; } } - else if (KIND == CCL_TOK_MINUS) /* = - term ? */ + else if (cclp->look_token->len == 1 && + cclp->look_token->name[0] == '"') /* = - term ? */ { ADVANCE; - if (!(p = search_term (ap))) + if (!(p = search_term (cclp, ap))) { free (ap); return NULL; @@ -569,14 +634,14 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, else if (KIND == CCL_TOK_LP) { ADVANCE; - if (!(p = find_spec (ap))) + if (!(p = find_spec (cclp, ap))) { free (ap); return NULL; } if (KIND != CCL_TOK_RP) { - ccl_error = CCL_ERR_RP_EXPECTED; + cclp->error_code = CCL_ERR_RP_EXPECTED; ccl_rpn_delete (p); free (ap); return NULL; @@ -587,7 +652,7 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, } else { - if (!(p = search_terms (ap))) + if (!(p = search_terms (cclp, ap))) { free (ap); return NULL; @@ -596,7 +661,7 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, free (ap); return p; } - ccl_error = CCL_ERR_TERM_EXPECTED; + cclp->error_code = CCL_ERR_TERM_EXPECTED; } free (ap); return NULL; @@ -604,13 +669,17 @@ static struct ccl_rpn_node *qualifiers (struct ccl_token *la, /* * search_terms: Parse CCL search terms - including proximity. + * cclp: CCL Parser * qa: Qualifier attributes already applied. * return: pointer to node(s); NULL on error. */ -static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa) +static struct ccl_rpn_node *search_terms (CCL_parser cclp, + struct ccl_rpn_attr **qa) { + static int list[] = { + CCL_TOK_TERM, CCL_TOK_COMMA,CCL_TOK_EQ, CCL_TOK_REL, -1}; struct ccl_rpn_node *p1, *p2, *pn; - p1 = search_term (qa); + p1 = search_term_x (cclp, qa, list); if (!p1) return NULL; while (1) @@ -618,7 +687,7 @@ static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa) if (KIND == CCL_TOK_PROX) { ADVANCE; - p2 = search_term (qa); + p2 = search_term_x (cclp, qa, list); if (!p2) { ccl_rpn_delete (p1); @@ -629,9 +698,9 @@ static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa) pn->u.p[1] = p2; p1 = pn; } - else if (KIND == CCL_TOK_TERM) + else if (is_term_ok(KIND, list)) { - p2 = search_term (qa); + p2 = search_term_x (cclp, qa, list); if (!p2) { ccl_rpn_delete (p1); @@ -650,22 +719,24 @@ static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa) /* * search_elements: Parse CCL search elements + * cclp: CCL Parser * qa: Qualifier attributes already applied. * return: pointer to node(s); NULL on error. */ -static struct ccl_rpn_node *search_elements (struct ccl_rpn_attr **qa) +static struct ccl_rpn_node *search_elements (CCL_parser cclp, + struct ccl_rpn_attr **qa) { struct ccl_rpn_node *p1; struct ccl_token *lookahead; if (KIND == CCL_TOK_LP) { ADVANCE; - p1 = find_spec (qa); + p1 = find_spec (cclp, qa); if (!p1) return NULL; if (KIND != CCL_TOK_RP) { - ccl_error = CCL_ERR_RP_EXPECTED; + cclp->error_code = CCL_ERR_RP_EXPECTED; ccl_rpn_delete (p1); return NULL; } @@ -679,37 +750,39 @@ static struct ccl_rpn_node *search_elements (struct ccl_rpn_attr **qa) ADVANCE; if (KIND != CCL_TOK_TERM) { - ccl_error = CCL_ERR_SETNAME_EXPECTED; + cclp->error_code = CCL_ERR_SETNAME_EXPECTED; return NULL; } p1 = mk_node (CCL_RPN_SET); - p1->u.setname = copy_token_name (look_token); + p1->u.setname = copy_token_name (cclp->look_token); ADVANCE; return p1; } - lookahead = look_token; + lookahead = cclp->look_token; while (lookahead->kind==CCL_TOK_TERM) { lookahead = lookahead->next; if (lookahead->kind == CCL_TOK_REL || lookahead->kind == CCL_TOK_EQ) - return qualifiers (lookahead, qa); + return qualifiers (cclp, lookahead, qa); if (lookahead->kind != CCL_TOK_COMMA) break; lookahead = lookahead->next; } - return search_terms (qa); + return search_terms (cclp, qa); } /* * find_spec: Parse CCL find specification + * cclp: CCL Parser * qa: Qualifier attributes already applied. * return: pointer to node(s); NULL on error. */ -static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa) +static struct ccl_rpn_node *find_spec (CCL_parser cclp, + struct ccl_rpn_attr **qa) { struct ccl_rpn_node *p1, *p2, *pn; - if (!(p1 = search_elements (qa))) + if (!(p1 = search_elements (cclp, qa))) return NULL; while (1) { @@ -717,7 +790,7 @@ static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa) { case CCL_TOK_AND: ADVANCE; - p2 = search_elements (qa); + p2 = search_elements (cclp, qa); if (!p2) { ccl_rpn_delete (p1); @@ -730,7 +803,7 @@ static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa) continue; case CCL_TOK_OR: ADVANCE; - p2 = search_elements (qa); + p2 = search_elements (cclp, qa); if (!p2) { ccl_rpn_delete (p1); @@ -743,7 +816,7 @@ static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa) continue; case CCL_TOK_NOT: ADVANCE; - p2 = search_elements (qa); + p2 = search_elements (cclp, qa); if (!p2) { ccl_rpn_delete (p1); @@ -760,36 +833,52 @@ static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa) return p1; } -/* - * ccl_find: Parse CCL find - token representation - * abibset: Bibset to be used for the parsing - * list: List of tokens - * error: Pointer to integer. Holds error no. on completion. - * pos: Pointer to char position. Holds approximate error position. - * return: RPN tree on successful completion; NULL otherwise. - */ -struct ccl_rpn_node *ccl_find (CCL_bibset abibset, struct ccl_token *list, - int *error, const char **pos) +struct ccl_rpn_node *ccl_parser_find (CCL_parser cclp, struct ccl_token *list) { struct ccl_rpn_node *p; - look_token = list; - bibset = abibset; - p = find_spec (NULL); + cclp->look_token = list; + p = find_spec (cclp, NULL); if (p && KIND != CCL_TOK_EOL) { if (KIND == CCL_TOK_RP) - ccl_error = CCL_ERR_BAD_RP; + cclp->error_code = CCL_ERR_BAD_RP; else - ccl_error = CCL_ERR_OP_EXPECTED; + cclp->error_code = CCL_ERR_OP_EXPECTED; ccl_rpn_delete (p); p = NULL; } - *pos = look_token->name; + cclp->error_pos = cclp->look_token->name; if (p) - *error = CCL_ERR_OK; + cclp->error_code = CCL_ERR_OK; else - *error = ccl_error; + cclp->error_code = cclp->error_code; + return p; +} + +/* + * ccl_find: Parse CCL find - token representation + * bibset: Bibset to be used for the parsing + * list: List of tokens + * error: Pointer to integer. Holds error no. on completion. + * pos: Pointer to char position. Holds approximate error position. + * return: RPN tree on successful completion; NULL otherwise. + */ +struct ccl_rpn_node *ccl_find (CCL_bibset bibset, struct ccl_token *list, + int *error, const char **pos) +{ + struct ccl_rpn_node *p; + CCL_parser cclp = ccl_parser_create (); + + cclp->bibset = bibset; + + p = ccl_parser_find (cclp, list); + + *error = cclp->error_code; + *pos = cclp->error_pos; + + ccl_parser_destroy (cclp); + return p; } @@ -804,13 +893,19 @@ struct ccl_rpn_node *ccl_find (CCL_bibset abibset, struct ccl_token *list, struct ccl_rpn_node *ccl_find_str (CCL_bibset bibset, const char *str, int *error, int *pos) { + CCL_parser cclp = ccl_parser_create (); struct ccl_token *list; - struct ccl_rpn_node *rpn; - const char *char_pos; + struct ccl_rpn_node *p; - list = ccl_tokenize (str); - rpn = ccl_find (bibset, list, error, &char_pos); + cclp->bibset = bibset; + + list = ccl_parser_tokenize (cclp, str); + p = ccl_parser_find (cclp, list); + + *error = cclp->error_code; if (*error) - *pos = char_pos - str; - return rpn; + *pos = cclp->error_pos - str; + ccl_parser_destroy (cclp); + ccl_token_del (list); + return p; }