1 /* $Id: rsbetween.c,v 1.14 2004-08-03 12:15:45 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 #define RSBETWEEN_DEBUG 0
43 static void *r_create_between(RSET ct, const struct rset_control *sel, void *parms);
44 static RSFD r_open_between (RSET ct, int flag);
45 static void r_close_between (RSFD rfd);
46 static void r_delete_between (RSET ct);
47 static void r_rewind_between (RSFD rfd);
48 static int r_forward_between(RSET ct, RSFD rfd, void *buf, int *term_index,
49 int (*cmpfunc)(const void *p1, const void *p2),
50 const void *untilbuf);
51 /* static int r_count_between (RSET ct); */
52 static int r_read_between (RSFD rfd, void *buf, int *term_index);
53 static int r_write_between (RSFD rfd, const void *buf);
55 static const struct rset_control control_between =
63 r_forward_between, /* rset_default_forward, */
64 /* r_count_between, */
70 const struct rset_control *rset_kind_between = &control_between;
72 struct rset_between_info {
79 int (*cmp)(const void *p1, const void *p2);
80 char *(*printer)(const void *p1, char *buf);
81 struct rset_between_rfd *rfd_list;
84 struct rset_between_rfd {
101 struct rset_between_rfd *next;
102 struct rset_between_info *info;
106 static void log2 (struct rset_between_rfd *p, char *msg, int cmp_l, int cmp_r)
111 logf(LOG_DEBUG,"btw: %s l=%s(%d/%d) m=%s(%d) r=%s(%d/%d), lev=%d",
113 (*p->info->printer)(p->buf_l, buf_l), p->more_l, cmp_l,
114 (*p->info->printer)(p->buf_m, buf_m), p->more_m,
115 (*p->info->printer)(p->buf_r, buf_r), p->more_r, cmp_r,
120 static void *r_create_between (RSET ct, const struct rset_control *sel,
123 rset_between_parms *between_parms = (rset_between_parms *) parms;
124 struct rset_between_info *info;
126 info = (struct rset_between_info *) xmalloc (sizeof(*info));
127 info->key_size = between_parms->key_size;
128 info->rset_l = between_parms->rset_l;
129 info->rset_m = between_parms->rset_m;
130 info->rset_r = between_parms->rset_r;
131 info->rset_attr = between_parms->rset_attr;
132 if (rset_is_volatile(info->rset_l) ||
133 rset_is_volatile(info->rset_m) ||
134 rset_is_volatile(info->rset_r))
135 ct->flags |= RSET_FLAG_VOLATILE;
136 info->cmp = between_parms->cmp;
137 info->printer = between_parms->printer;
138 info->rfd_list = NULL;
140 info->term_index_s = info->rset_l->no_rset_terms;
144 info->rset_l->no_rset_terms +
145 info->rset_m->no_rset_terms +
146 info->rset_r->no_rset_terms;
147 ct->rset_terms = (RSET_TERM *)
148 xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms);
149 memcpy (ct->rset_terms, info->rset_l->rset_terms,
150 info->rset_l->no_rset_terms * sizeof(*ct->rset_terms));
151 memcpy (ct->rset_terms + info->rset_l->no_rset_terms,
152 info->rset_m->rset_terms,
153 info->rset_m->no_rset_terms * sizeof(*ct->rset_terms));
154 memcpy (ct->rset_terms + info->rset_l->no_rset_terms +
155 info->rset_m->no_rset_terms,
156 info->rset_r->rset_terms,
157 info->rset_r->no_rset_terms * sizeof(*ct->rset_terms));
162 info->rset_l->no_rset_terms +
163 info->rset_r->no_rset_terms;
164 ct->rset_terms = (RSET_TERM *)
165 xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms);
166 memcpy (ct->rset_terms, info->rset_l->rset_terms,
167 info->rset_l->no_rset_terms * sizeof(*ct->rset_terms));
168 memcpy (ct->rset_terms + info->rset_l->no_rset_terms,
169 info->rset_r->rset_terms,
170 info->rset_r->no_rset_terms * sizeof(*ct->rset_terms));
176 static RSFD r_open_between (RSET ct, int flag)
178 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
179 struct rset_between_rfd *rfd;
181 if (flag & RSETF_WRITE)
183 logf (LOG_FATAL, "between set type is read-only");
186 rfd = (struct rset_between_rfd *) xmalloc (sizeof(*rfd));
187 rfd->next = info->rfd_list;
188 info->rfd_list = rfd;
191 rfd->buf_l = xmalloc (info->key_size);
192 rfd->buf_m = xmalloc (info->key_size);
193 rfd->buf_r = xmalloc (info->key_size);
194 rfd->buf_attr = xmalloc (info->key_size);
196 rfd->rfd_l = rset_open (info->rset_l, RSETF_READ);
197 rfd->rfd_m = rset_open (info->rset_m, RSETF_READ);
198 rfd->rfd_r = rset_open (info->rset_r, RSETF_READ);
200 rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l,
202 rfd->more_m = rset_read (info->rset_m, rfd->rfd_m, rfd->buf_m,
204 rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r,
209 rfd->rfd_attr = rset_open (info->rset_attr, RSETF_READ);
210 rfd->more_attr = rset_read (info->rset_attr, rfd->rfd_attr,
211 rfd->buf_attr, &dummy);
217 static void r_close_between (RSFD rfd)
219 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
220 struct rset_between_rfd **rfdp;
222 for (rfdp = &info->rfd_list; *rfdp; rfdp = &(*rfdp)->next)
225 xfree ((*rfdp)->buf_l);
226 xfree ((*rfdp)->buf_m);
227 xfree ((*rfdp)->buf_r);
228 xfree ((*rfdp)->buf_attr);
229 rset_close (info->rset_l, (*rfdp)->rfd_l);
230 rset_close (info->rset_m, (*rfdp)->rfd_m);
231 rset_close (info->rset_r, (*rfdp)->rfd_r);
233 rset_close (info->rset_attr, (*rfdp)->rfd_attr);
235 *rfdp = (*rfdp)->next;
239 logf (LOG_FATAL, "r_close_between but no rfd match!");
243 static void r_delete_between (RSET ct)
245 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
247 assert (info->rfd_list == NULL);
248 xfree (ct->rset_terms);
249 rset_delete (info->rset_l);
250 rset_delete (info->rset_m);
251 rset_delete (info->rset_r);
253 rset_delete (info->rset_attr);
257 static void r_rewind_between (RSFD rfd)
259 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
260 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
263 logf (LOG_DEBUG, "rsbetween_rewind");
265 rset_rewind (info->rset_l, p->rfd_l);
266 rset_rewind (info->rset_m, p->rfd_m);
267 rset_rewind (info->rset_r, p->rfd_r);
268 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, &p->term_index_l);
269 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m, &p->term_index_m);
270 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, &p->term_index_r);
274 rset_rewind (info->rset_attr, p->rfd_attr);
275 p->more_attr = rset_read (info->rset_attr, p->rfd_attr, p->buf_attr,
283 static int r_forward_between(RSET ct, RSFD rfd, void *buf, int *term_index,
284 int (*cmpfunc)(const void *p1, const void *p2),
285 const void *untilbuf)
287 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
288 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
291 log2( p, "fwd: before forward", 0,0);
293 /* It is enough to forward the m pointer here, the read will */
294 /* naturally forward the l, m, and attr pointers */
296 p->more_m=rset_forward(info->rset_m,p->rfd_m, p->buf_m,
297 &p->term_index_m, info->cmp,untilbuf);
299 log2( p, "fwd: after forward M", 0,0);
301 rc = r_read_between(rfd, buf, term_index);
303 log2( p, "fwd: after forward", 0,0);
309 static int r_count_between (RSET ct)
317 static int r_read_between (RSFD rfd, void *buf, int *term_index)
319 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
320 struct rset_between_info *info = p->info;
328 log2( p, "start of loop", cmp_l, cmp_r);
331 /* forward L until past m, count levels, note rec boundaries */
333 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
337 cmp_l=2; /* past this record */
340 log2( p, "after first L", cmp_l, cmp_r);
343 while (cmp_l < 0) /* l before m */
346 p->level=0; /* earlier record */
349 p->level++; /* relevant start tag */
351 if (!info->rset_attr)
360 cmp_attr = (*info->cmp)(p->buf_attr, p->buf_l);
366 else if (cmp_attr > 0)
368 else if (cmp_attr==-1)
369 p->more_attr = rset_read (info->rset_attr, p->rfd_attr,
370 p->buf_attr, &dummy_term);
371 /* if we had a forward that went all the way to
372 * the seqno, we could use that. But fwd only goes
374 else if (cmp_attr==-2)
376 p->more_attr = rset_forward(
377 info->rset_attr, p->rfd_attr,
378 p->buf_attr, &dummy_term,
379 info->cmp, p->buf_l);
381 logf(LOG_DEBUG, "btw: after frowarding attr m=%d",p->more_attr);
384 } /* while more_attr */
393 p->more_l=rset_forward(
394 info->rset_l, p->rfd_l,
395 p->buf_l, &p->term_index_l,
396 info->cmp, p->buf_m);
398 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
402 log2( p, "after forwarding L", cmp_l, cmp_r);
407 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
411 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
416 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
421 log2( p, "end of L loop", cmp_l, cmp_r);
426 /* forward R until past m, count levels */
428 log2( p, "Before moving R", cmp_l, cmp_r);
431 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
435 log2( p, "after first R", cmp_l, cmp_r);
437 while (cmp_r < 0) /* r before m */
439 /* -2, earlier record, don't count level */
441 p->level--; /* relevant end tag */
447 p->more_r=rset_forward(
448 info->rset_r, p->rfd_r,
449 p->buf_r, &p->term_index_r,
450 info->cmp, p->buf_m);
453 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
457 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
460 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
462 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
468 log2( p, "End of R loop", cmp_l, cmp_r);
472 if ( ( p->level <= 0 ) && ! p->more_l)
473 return 0; /* no more start tags, nothing more to find */
475 if ( attr_match && p->level > 0) /* within a tag pair (or deeper) */
477 memcpy (buf, p->buf_m, info->key_size);
478 *term_index = p->term_index_m;
480 log2( p, "Returning a hit (and forwarding m)", cmp_l, cmp_r);
482 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
488 else if ( ! p->more_l ) /* not in data, no more starts */
491 log2( p, "no more starts, exiting without a hit", cmp_l, cmp_r);
493 return 0; /* ergo, nothing can be found. stop scanning */
499 p->more_m=rset_forward(
500 info->rset_m, p->rfd_m,
501 p->buf_m, &p->term_index_m,
502 info->cmp, p->buf_l);
505 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
511 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
515 log2( p, "End of M loop", cmp_l, cmp_r);
520 log2( p, "Exiting, nothing more in m", cmp_l, cmp_r);
522 return 0; /* no more data possible */
528 static int r_write_between (RSFD rfd, const void *buf)
530 logf (LOG_FATAL, "between set type is read-only");