1 /* $Id: rsbetween.c,v 1.12 2004-02-12 15:15:54 heikki Exp $
2 Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002
5 This file is part of the Zebra server.
7 Zebra is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 2, or (at your option) any later
12 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with Zebra; see the file LICENSE.zebra. If not, write to the
19 Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
24 /* rsbetween is (mostly) used for xml searches. It returns the hits of the
25 * "middle" rset, that are in between the "left" and "right" rsets. For
26 * example "Shakespeare" in between "<title>" and </title>. The thing is
27 * complicated by the inclusion of attributes (from their own rset). If attrs
28 * specified, they must match the "left" rset (start tag). "Hamlet" between
29 * "<title lang=eng>" and "</title>". (This assumes that the attributes are
30 * indexed to the same seqno as the tags).
38 #include <rsbetween.h>
41 static void *r_create_between(RSET ct, const struct rset_control *sel, void *parms);
42 static RSFD r_open_between (RSET ct, int flag);
43 static void r_close_between (RSFD rfd);
44 static void r_delete_between (RSET ct);
45 static void r_rewind_between (RSFD rfd);
46 static int r_forward_between(RSET ct, RSFD rfd, void *buf, int *term_index,
47 int (*cmpfunc)(const void *p1, const void *p2),
48 const void *untilbuf);
49 static int r_count_between (RSET ct);
50 static int r_read_between (RSFD rfd, void *buf, int *term_index);
51 static int r_write_between (RSFD rfd, const void *buf);
53 static const struct rset_control control_between =
61 r_forward_between, /* rset_default_forward, */
68 const struct rset_control *rset_kind_between = &control_between;
70 struct rset_between_info {
77 int (*cmp)(const void *p1, const void *p2);
78 char *(*printer)(const void *p1, char *buf);
79 struct rset_between_rfd *rfd_list;
82 struct rset_between_rfd {
99 struct rset_between_rfd *next;
100 struct rset_between_info *info;
103 static void log2 (struct rset_between_rfd *p, char *msg, int cmp_l, int cmp_r)
108 logf(LOG_DEBUG,"btw: %s l=%s(%d/%d) m=%s(%d) r=%s(%d/%d), lev=%d",
110 (*p->info->printer)(p->buf_l, buf_l), p->more_l, cmp_l,
111 (*p->info->printer)(p->buf_m, buf_m), p->more_m,
112 (*p->info->printer)(p->buf_r, buf_r), p->more_r, cmp_r,
116 static void *r_create_between (RSET ct, const struct rset_control *sel,
119 rset_between_parms *between_parms = (rset_between_parms *) parms;
120 struct rset_between_info *info;
122 info = (struct rset_between_info *) xmalloc (sizeof(*info));
123 info->key_size = between_parms->key_size;
124 info->rset_l = between_parms->rset_l;
125 info->rset_m = between_parms->rset_m;
126 info->rset_r = between_parms->rset_r;
127 info->rset_attr = between_parms->rset_attr;
128 if (rset_is_volatile(info->rset_l) ||
129 rset_is_volatile(info->rset_m) ||
130 rset_is_volatile(info->rset_r))
131 ct->flags |= RSET_FLAG_VOLATILE;
132 info->cmp = between_parms->cmp;
133 info->printer = between_parms->printer;
134 info->rfd_list = NULL;
136 info->term_index_s = info->rset_l->no_rset_terms;
140 info->rset_l->no_rset_terms +
141 info->rset_m->no_rset_terms +
142 info->rset_r->no_rset_terms;
143 ct->rset_terms = (RSET_TERM *)
144 xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms);
145 memcpy (ct->rset_terms, info->rset_l->rset_terms,
146 info->rset_l->no_rset_terms * sizeof(*ct->rset_terms));
147 memcpy (ct->rset_terms + info->rset_l->no_rset_terms,
148 info->rset_m->rset_terms,
149 info->rset_m->no_rset_terms * sizeof(*ct->rset_terms));
150 memcpy (ct->rset_terms + info->rset_l->no_rset_terms +
151 info->rset_m->no_rset_terms,
152 info->rset_r->rset_terms,
153 info->rset_r->no_rset_terms * sizeof(*ct->rset_terms));
158 info->rset_l->no_rset_terms +
159 info->rset_r->no_rset_terms;
160 ct->rset_terms = (RSET_TERM *)
161 xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms);
162 memcpy (ct->rset_terms, info->rset_l->rset_terms,
163 info->rset_l->no_rset_terms * sizeof(*ct->rset_terms));
164 memcpy (ct->rset_terms + info->rset_l->no_rset_terms,
165 info->rset_r->rset_terms,
166 info->rset_r->no_rset_terms * sizeof(*ct->rset_terms));
172 static RSFD r_open_between (RSET ct, int flag)
174 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
175 struct rset_between_rfd *rfd;
177 if (flag & RSETF_WRITE)
179 logf (LOG_FATAL, "between set type is read-only");
182 rfd = (struct rset_between_rfd *) xmalloc (sizeof(*rfd));
183 rfd->next = info->rfd_list;
184 info->rfd_list = rfd;
187 rfd->buf_l = xmalloc (info->key_size);
188 rfd->buf_m = xmalloc (info->key_size);
189 rfd->buf_r = xmalloc (info->key_size);
190 rfd->buf_attr = xmalloc (info->key_size);
192 rfd->rfd_l = rset_open (info->rset_l, RSETF_READ);
193 rfd->rfd_m = rset_open (info->rset_m, RSETF_READ);
194 rfd->rfd_r = rset_open (info->rset_r, RSETF_READ);
196 rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l,
198 rfd->more_m = rset_read (info->rset_m, rfd->rfd_m, rfd->buf_m,
200 rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r,
205 rfd->rfd_attr = rset_open (info->rset_attr, RSETF_READ);
206 rfd->more_attr = rset_read (info->rset_attr, rfd->rfd_attr,
207 rfd->buf_attr, &dummy);
213 static void r_close_between (RSFD rfd)
215 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
216 struct rset_between_rfd **rfdp;
218 for (rfdp = &info->rfd_list; *rfdp; rfdp = &(*rfdp)->next)
221 xfree ((*rfdp)->buf_l);
222 xfree ((*rfdp)->buf_m);
223 xfree ((*rfdp)->buf_r);
224 xfree ((*rfdp)->buf_attr);
225 rset_close (info->rset_l, (*rfdp)->rfd_l);
226 rset_close (info->rset_m, (*rfdp)->rfd_m);
227 rset_close (info->rset_r, (*rfdp)->rfd_r);
229 rset_close (info->rset_attr, (*rfdp)->rfd_attr);
231 *rfdp = (*rfdp)->next;
235 logf (LOG_FATAL, "r_close_between but no rfd match!");
239 static void r_delete_between (RSET ct)
241 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
243 assert (info->rfd_list == NULL);
244 xfree (ct->rset_terms);
245 rset_delete (info->rset_l);
246 rset_delete (info->rset_m);
247 rset_delete (info->rset_r);
249 rset_delete (info->rset_attr);
253 static void r_rewind_between (RSFD rfd)
255 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
256 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
258 logf (LOG_DEBUG, "rsbetween_rewind");
259 rset_rewind (info->rset_l, p->rfd_l);
260 rset_rewind (info->rset_m, p->rfd_m);
261 rset_rewind (info->rset_r, p->rfd_r);
262 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, &p->term_index_l);
263 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m, &p->term_index_m);
264 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, &p->term_index_r);
268 rset_rewind (info->rset_attr, p->rfd_attr);
269 p->more_attr = rset_read (info->rset_attr, p->rfd_attr, p->buf_attr,
277 static int r_forward_between(RSET ct, RSFD rfd, void *buf, int *term_index,
278 int (*cmpfunc)(const void *p1, const void *p2),
279 const void *untilbuf)
281 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
282 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
284 log2( p, "fwd: before forward", 0,0);
285 /* It is enough to forward the m pointer here, the read will */
286 /* naturally forward the l, m, and attr pointers */
288 p->more_m=rset_forward(info->rset_m,p->rfd_m, p->buf_m,
289 &p->term_index_m, info->cmp,untilbuf);
290 log2( p, "fwd: after forward M", 0,0);
291 rc = r_read_between(rfd, buf, term_index);
292 log2( p, "fwd: after forward", 0,0);
296 static int r_count_between (RSET ct)
303 static int r_read_between (RSFD rfd, void *buf, int *term_index)
305 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
306 struct rset_between_info *info = p->info;
313 log2( p, "start of loop", cmp_l, cmp_r);
315 /* forward L until past m, count levels, note rec boundaries */
317 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
321 cmp_l=2; /* past this record */
323 log2( p, "after first L", cmp_l, cmp_r);
325 while (cmp_l < 0) /* l before m */
328 p->level=0; /* earlier record */
331 p->level++; /* relevant start tag */
333 if (!info->rset_attr)
342 cmp_attr = (*info->cmp)(p->buf_attr, p->buf_l);
348 else if (cmp_attr > 0)
350 else if (cmp_attr==-1)
351 p->more_attr = rset_read (info->rset_attr, p->rfd_attr,
352 p->buf_attr, &dummy_term);
353 /* if we had a forward that went all the way to
354 * the seqno, we could use that. But fwd only goes
356 else if (cmp_attr==-2)
358 p->more_attr = rset_forward(
359 info->rset_attr, p->rfd_attr,
360 p->buf_attr, &dummy_term,
361 info->cmp, p->buf_l);
362 logf(LOG_DEBUG, "btw: after frowarding attr m=%d",p->more_attr);
364 } /* while more_attr */
373 p->more_l=rset_forward(
374 info->rset_l, p->rfd_l,
375 p->buf_l, &p->term_index_l,
376 info->cmp, p->buf_m);
378 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
381 log2( p, "after forwarding L", cmp_l, cmp_r);
385 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
389 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
394 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
398 log2( p, "end of L loop", cmp_l, cmp_r);
402 /* forward R until past m, count levels */
403 log2( p, "Before moving R", cmp_l, cmp_r);
405 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
408 log2( p, "after first R", cmp_l, cmp_r);
409 while (cmp_r < 0) /* r before m */
411 /* -2, earlier record, don't count level */
413 p->level--; /* relevant end tag */
419 p->more_r=rset_forward(
420 info->rset_r, p->rfd_r,
421 p->buf_r, &p->term_index_r,
422 info->cmp, p->buf_m);
425 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
429 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
432 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
434 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
439 log2( p, "End of R loop", cmp_l, cmp_r);
442 if ( ( p->level <= 0 ) && ! p->more_l)
443 return 0; /* no more start tags, nothing more to find */
445 if ( attr_match && p->level > 0) /* within a tag pair (or deeper) */
447 memcpy (buf, p->buf_m, info->key_size);
448 *term_index = p->term_index_m;
449 log2( p, "Returning a hit (and forwarding m)", cmp_l, cmp_r);
450 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
456 else if ( ! p->more_l ) /* not in data, no more starts */
458 log2( p, "no more starts, exiting without a hit", cmp_l, cmp_r);
459 return 0; /* ergo, nothing can be found. stop scanning */
465 p->more_m=rset_forward(
466 info->rset_m, p->rfd_m,
467 p->buf_m, &p->term_index_m,
468 info->cmp, p->buf_l);
471 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
477 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
480 log2( p, "End of M loop", cmp_l, cmp_r);
483 log2( p, "Exiting, nothing more in m", cmp_l, cmp_r);
484 return 0; /* no more data possible */
490 static int r_write_between (RSFD rfd, const void *buf)
492 logf (LOG_FATAL, "between set type is read-only");