2 * Copyright (c) 1995, the EUROPAGATE consortium (see below).
4 * The EUROPAGATE consortium members are:
6 * University College Dublin
7 * Danmarks Teknologiske Videnscenter
8 * An Chomhairle Leabharlanna
9 * Consejo Superior de Investigaciones Cientificas
11 * Permission to use, copy, modify, distribute, and sell this software and
12 * its documentation, in whole or in part, for any purpose, is hereby granted,
15 * 1. This copyright and permission notice appear in all copies of the
16 * software and its documentation. Notices of copyright or attribution
17 * which appear at the beginning of any file must remain unchanged.
19 * 2. The names of EUROPAGATE or the project partners may not be used to
20 * endorse or promote products derived from this software without specific
21 * prior written permission.
23 * 3. Users of this software (implementors and gateway operators) agree to
24 * inform the EUROPAGATE consortium of their use of the software. This
25 * information will be used to evaluate the EUROPAGATE project and the
26 * software, and to plan further developments. The consortium may use
27 * the information in later publications.
29 * 4. Users of this software agree to make their best efforts, when
30 * documenting their use of the software, to acknowledge the EUROPAGATE
31 * consortium, and the role played by the software in their work.
33 * THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT WARRANTY OF ANY KIND,
34 * EXPRESS, IMPLIED, OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY
35 * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
36 * IN NO EVENT SHALL THE EUROPAGATE CONSORTIUM OR ITS MEMBERS BE LIABLE
37 * FOR ANY SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF
38 * ANY KIND, OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA
39 * OR PROFITS, WHETHER OR NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND
40 * ON ANY THEORY OF LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE
41 * USE OR PERFORMANCE OF THIS SOFTWARE.
44 /* CCL - lexical analysis
47 * $Id: ccltoken.c,v 1.19 2002-06-06 12:54:24 adam Exp $
51 * Revision 1.10 1995/07/11 12:28:31 adam
52 * New function: ccl_token_simple (split into simple tokens) and
53 * ccl_token_del (delete tokens).
55 * Revision 1.9 1995/05/16 09:39:28 adam
58 * Revision 1.8 1995/05/11 14:03:57 adam
59 * Changes in the reading of qualifier(s). New function: ccl_qual_fitem.
60 * New variable ccl_case_sensitive, which controls whether reserved
61 * words and field names are case sensitive or not.
63 * Revision 1.7 1995/04/19 12:11:24 adam
66 * Revision 1.6 1995/04/17 09:31:48 adam
67 * Improved handling of qualifiers. Aliases or reserved words.
69 * Revision 1.5 1995/02/23 08:32:00 adam
72 * Revision 1.3 1995/02/15 17:42:16 adam
73 * Minor changes of the api of this module. FILE* argument added
76 * Revision 1.2 1995/02/14 19:55:13 adam
77 * Header files ccl.h/cclp.h are gone! They have been merged an
78 * moved to ../include/ccl.h.
79 * Node kind(s) in ccl_rpn_node have changed names.
81 * Revision 1.1 1995/02/13 12:35:21 adam
82 * First version of CCL. Qualifiers aren't handled yet.
92 * token_cmp: Compare token with keyword(s)
93 * kw: Keyword list. Each keyword is separated by space.
95 * return: 1 if token string matches one of the keywords in list;
98 static int token_cmp (CCL_parser cclp, const char *kw, struct ccl_token *token)
100 const char *cp1 = kw;
103 int case_sensitive = cclp->ccl_case_sensitive;
105 aliases = ccl_qual_search_special(cclp->bibset, "case");
107 case_sensitive = atoi(aliases);
110 while ((cp2 = strchr (cp1, ' ')))
112 if (token->len == (size_t) (cp2-cp1))
116 if (!memcmp (cp1, token->name, token->len))
121 if (!ccl_memicmp (cp1, token->name, token->len))
128 return token->len == strlen(cp1)
129 && !memcmp (cp1, token->name, token->len);
130 return token->len == strlen(cp1) &&
131 !ccl_memicmp (cp1, token->name, token->len);
135 * ccl_token_simple: tokenize CCL raw tokens
137 struct ccl_token *ccl_token_simple (const char *command)
139 const char *cp = command;
140 struct ccl_token *first = NULL;
141 struct ccl_token *last = NULL;
145 while (*cp && strchr (" \t\r\n", *cp))
152 first = last = (struct ccl_token *)xmalloc (sizeof (*first));
158 last->next = (struct ccl_token *)xmalloc (sizeof(*first));
159 ccl_assert (last->next);
160 last->next->prev = last;
169 last->kind = CCL_TOK_EOL;
172 last->kind = CCL_TOK_TERM;
175 while (*cp && *cp != '\"')
184 while (*cp && !strchr (" \t\n\r", *cp))
189 last->kind = CCL_TOK_TERM;
197 * ccl_tokenize: tokenize CCL command string.
198 * return: CCL token list.
200 struct ccl_token *ccl_parser_tokenize (CCL_parser cclp, const char *command)
203 const char *cp = command;
204 struct ccl_token *first = NULL;
205 struct ccl_token *last = NULL;
209 while (*cp && strchr (" \t\r\n", *cp))
216 first = last = (struct ccl_token *)xmalloc (sizeof (*first));
222 last->next = (struct ccl_token *)xmalloc (sizeof(*first));
223 ccl_assert (last->next);
224 last->next->prev = last;
233 last->kind = CCL_TOK_EOL;
236 last->kind = CCL_TOK_LP;
239 last->kind = CCL_TOK_RP;
242 last->kind = CCL_TOK_COMMA;
246 last->kind = CCL_TOK_PROX;
247 while (*cp == '%' || *cp == '!')
256 if (*cp == '=' || *cp == '<' || *cp == '>')
259 last->kind = CCL_TOK_REL;
262 else if (cp[-1] == '=')
263 last->kind = CCL_TOK_EQ;
265 last->kind = CCL_TOK_REL;
268 last->kind = CCL_TOK_TERM;
271 while (*cp && *cp != '\"')
280 if (!strchr ("(),%!><=- \t\n\r", cp[-1]))
282 while (*cp && !strchr ("(),%!><=- \t\n\r", *cp))
288 last->kind = CCL_TOK_TERM;
290 aliases = ccl_qual_search_special(cclp->bibset, "and");
292 aliases = cclp->ccl_token_and;
293 if (token_cmp (cclp, aliases, last))
294 last->kind = CCL_TOK_AND;
296 aliases = ccl_qual_search_special(cclp->bibset, "or");
298 aliases = cclp->ccl_token_or;
299 if (token_cmp (cclp, aliases, last))
300 last->kind = CCL_TOK_OR;
302 aliases = ccl_qual_search_special(cclp->bibset, "not");
304 aliases = cclp->ccl_token_not;
305 if (token_cmp (cclp, aliases, last))
306 last->kind = CCL_TOK_NOT;
308 aliases = ccl_qual_search_special(cclp->bibset, "set");
310 aliases = cclp->ccl_token_set;
312 if (token_cmp (cclp, aliases, last))
313 last->kind = CCL_TOK_SET;
319 struct ccl_token *ccl_tokenize (const char *command)
321 CCL_parser cclp = ccl_parser_create ();
322 struct ccl_token *list;
324 list = ccl_parser_tokenize (cclp, command);
326 ccl_parser_destroy (cclp);
331 * ccl_token_del: delete CCL tokens
333 void ccl_token_del (struct ccl_token *list)
335 struct ccl_token *list1;
345 char *ccl_strdup (const char *str)
347 int len = strlen(str);
348 char *p = (char*) xmalloc (len+1);
353 CCL_parser ccl_parser_create (void)
355 CCL_parser p = (CCL_parser)xmalloc (sizeof(*p));
358 p->look_token = NULL;
363 p->ccl_token_and = ccl_strdup("and");
364 p->ccl_token_or = ccl_strdup("or");
365 p->ccl_token_not = ccl_strdup("not andnot");
366 p->ccl_token_set = ccl_strdup("set");
367 p->ccl_case_sensitive = 1;
372 void ccl_parser_destroy (CCL_parser p)
376 xfree (p->ccl_token_and);
377 xfree (p->ccl_token_or);
378 xfree (p->ccl_token_not);
379 xfree (p->ccl_token_set);
383 void ccl_parser_set_op_and (CCL_parser p, const char *op)
387 if (p->ccl_token_and)
388 xfree (p->ccl_token_and);
389 p->ccl_token_and = ccl_strdup (op);
393 void ccl_parser_set_op_or (CCL_parser p, const char *op)
398 xfree (p->ccl_token_or);
399 p->ccl_token_or = ccl_strdup (op);
402 void ccl_parser_set_op_not (CCL_parser p, const char *op)
406 if (p->ccl_token_not)
407 xfree (p->ccl_token_not);
408 p->ccl_token_not = ccl_strdup (op);
411 void ccl_parser_set_op_set (CCL_parser p, const char *op)
415 if (p->ccl_token_set)
416 xfree (p->ccl_token_set);
417 p->ccl_token_set = ccl_strdup (op);
421 void ccl_parser_set_case (CCL_parser p, int case_sensitivity_flag)
424 p->ccl_case_sensitive = case_sensitivity_flag;