1 /* $Id: rsbetween.c,v 1.11 2004-01-30 13:07:14 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_count_between (RSET ct);
47 static int r_read_between (RSFD rfd, void *buf, int *term_index);
48 static int r_write_between (RSFD rfd, const void *buf);
50 static const struct rset_control control_between =
65 const struct rset_control *rset_kind_between = &control_between;
67 struct rset_between_info {
74 int (*cmp)(const void *p1, const void *p2);
75 char *(*printer)(const void *p1, char *buf);
76 struct rset_between_rfd *rfd_list;
79 struct rset_between_rfd {
96 struct rset_between_rfd *next;
97 struct rset_between_info *info;
100 static void *r_create_between (RSET ct, const struct rset_control *sel,
103 rset_between_parms *between_parms = (rset_between_parms *) parms;
104 struct rset_between_info *info;
106 info = (struct rset_between_info *) xmalloc (sizeof(*info));
107 info->key_size = between_parms->key_size;
108 info->rset_l = between_parms->rset_l;
109 info->rset_m = between_parms->rset_m;
110 info->rset_r = between_parms->rset_r;
111 info->rset_attr = between_parms->rset_attr;
112 if (rset_is_volatile(info->rset_l) ||
113 rset_is_volatile(info->rset_m) ||
114 rset_is_volatile(info->rset_r))
115 ct->flags |= RSET_FLAG_VOLATILE;
116 info->cmp = between_parms->cmp;
117 info->printer = between_parms->printer;
118 info->rfd_list = NULL;
120 info->term_index_s = info->rset_l->no_rset_terms;
124 info->rset_l->no_rset_terms +
125 info->rset_m->no_rset_terms +
126 info->rset_r->no_rset_terms;
127 ct->rset_terms = (RSET_TERM *)
128 xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms);
129 memcpy (ct->rset_terms, info->rset_l->rset_terms,
130 info->rset_l->no_rset_terms * sizeof(*ct->rset_terms));
131 memcpy (ct->rset_terms + info->rset_l->no_rset_terms,
132 info->rset_m->rset_terms,
133 info->rset_m->no_rset_terms * sizeof(*ct->rset_terms));
134 memcpy (ct->rset_terms + info->rset_l->no_rset_terms +
135 info->rset_m->no_rset_terms,
136 info->rset_r->rset_terms,
137 info->rset_r->no_rset_terms * sizeof(*ct->rset_terms));
142 info->rset_l->no_rset_terms +
143 info->rset_r->no_rset_terms;
144 ct->rset_terms = (RSET_TERM *)
145 xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms);
146 memcpy (ct->rset_terms, info->rset_l->rset_terms,
147 info->rset_l->no_rset_terms * sizeof(*ct->rset_terms));
148 memcpy (ct->rset_terms + info->rset_l->no_rset_terms,
149 info->rset_r->rset_terms,
150 info->rset_r->no_rset_terms * sizeof(*ct->rset_terms));
156 static RSFD r_open_between (RSET ct, int flag)
158 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
159 struct rset_between_rfd *rfd;
161 if (flag & RSETF_WRITE)
163 logf (LOG_FATAL, "between set type is read-only");
166 rfd = (struct rset_between_rfd *) xmalloc (sizeof(*rfd));
167 rfd->next = info->rfd_list;
168 info->rfd_list = rfd;
171 rfd->buf_l = xmalloc (info->key_size);
172 rfd->buf_m = xmalloc (info->key_size);
173 rfd->buf_r = xmalloc (info->key_size);
174 rfd->buf_attr = xmalloc (info->key_size);
176 rfd->rfd_l = rset_open (info->rset_l, RSETF_READ);
177 rfd->rfd_m = rset_open (info->rset_m, RSETF_READ);
178 rfd->rfd_r = rset_open (info->rset_r, RSETF_READ);
180 rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l,
182 rfd->more_m = rset_read (info->rset_m, rfd->rfd_m, rfd->buf_m,
184 rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r,
189 rfd->rfd_attr = rset_open (info->rset_attr, RSETF_READ);
190 rfd->more_attr = rset_read (info->rset_attr, rfd->rfd_attr,
191 rfd->buf_attr, &dummy);
197 static void r_close_between (RSFD rfd)
199 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
200 struct rset_between_rfd **rfdp;
202 for (rfdp = &info->rfd_list; *rfdp; rfdp = &(*rfdp)->next)
205 xfree ((*rfdp)->buf_l);
206 xfree ((*rfdp)->buf_m);
207 xfree ((*rfdp)->buf_r);
208 xfree ((*rfdp)->buf_attr);
209 rset_close (info->rset_l, (*rfdp)->rfd_l);
210 rset_close (info->rset_m, (*rfdp)->rfd_m);
211 rset_close (info->rset_r, (*rfdp)->rfd_r);
213 rset_close (info->rset_attr, (*rfdp)->rfd_attr);
215 *rfdp = (*rfdp)->next;
219 logf (LOG_FATAL, "r_close_between but no rfd match!");
223 static void r_delete_between (RSET ct)
225 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
227 assert (info->rfd_list == NULL);
228 xfree (ct->rset_terms);
229 rset_delete (info->rset_l);
230 rset_delete (info->rset_m);
231 rset_delete (info->rset_r);
233 rset_delete (info->rset_attr);
237 static void r_rewind_between (RSFD rfd)
239 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
240 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
242 logf (LOG_DEBUG, "rsbetween_rewind");
243 rset_rewind (info->rset_l, p->rfd_l);
244 rset_rewind (info->rset_m, p->rfd_m);
245 rset_rewind (info->rset_r, p->rfd_r);
246 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, &p->term_index_l);
247 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m, &p->term_index_m);
248 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, &p->term_index_r);
252 rset_rewind (info->rset_attr, p->rfd_attr);
253 p->more_attr = rset_read (info->rset_attr, p->rfd_attr, p->buf_attr,
259 static int r_count_between (RSET ct)
265 static void log2 (struct rset_between_rfd *p, char *msg, int cmp_l, int cmp_r)
270 logf(LOG_DEBUG,"btw: %s l=%s(%d/%d) m=%s(%d) r=%s(%d/%d), lev=%d",
272 (*p->info->printer)(p->buf_l, buf_l), p->more_l, cmp_l,
273 (*p->info->printer)(p->buf_m, buf_m), p->more_m,
274 (*p->info->printer)(p->buf_r, buf_r), p->more_r, cmp_r,
278 static int r_read_between (RSFD rfd, void *buf, int *term_index)
280 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
281 struct rset_between_info *info = p->info;
288 log2( p, "start of loop", cmp_l, cmp_r);
290 /* forward L until past m, count levels, note rec boundaries */
292 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
296 cmp_l=2; /* past this record */
298 log2( p, "after first L", cmp_l, cmp_r);
300 while (cmp_l < 0) /* l before m */
303 p->level=0; /* earlier record */
306 p->level++; /* relevant start tag */
308 if (!info->rset_attr)
317 cmp_attr = (*info->cmp)(p->buf_attr, p->buf_l);
323 else if (cmp_attr > 0)
325 p->more_attr = rset_read (info->rset_attr, p->rfd_attr,
326 p->buf_attr, &dummy_term);
330 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l,
334 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
338 log2( p, "end of L loop", cmp_l, cmp_r);
342 /* forward R until past m, count levels */
343 log2( p, "Before moving R", cmp_l, cmp_r);
345 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
348 log2( p, "after first R", cmp_l, cmp_r);
349 while (cmp_r < 0) /* r before m */
351 /* -2, earlier record, doesn't matter */
353 p->level--; /* relevant end tag */
356 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r,
358 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
362 log2( p, "End of R loop", cmp_l, cmp_r);
365 if ( ( p->level <= 0 ) && ! p->more_l)
366 return 0; /* no more start tags, nothing more to find */
368 if ( attr_match && p->level > 0) /* within a tag pair (or deeper) */
370 memcpy (buf, p->buf_m, info->key_size);
371 *term_index = p->term_index_m;
372 log2( p, "Returning a hit (and forwarding m)", cmp_l, cmp_r);
373 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
379 else if ( ! p->more_l ) /* not in data, no more starts */
381 log2( p, "no more starts, exiting without a hit", cmp_l, cmp_r);
382 return 0; /* ergo, nothing can be found. stop scanning */
386 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m,
388 log2( p, "End of M loop", cmp_l, cmp_r);
391 log2( p, "Exiting, nothing more in m", cmp_l, cmp_r);
392 return 0; /* no more data possible */
398 static int r_write_between (RSFD rfd, const void *buf)
400 logf (LOG_FATAL, "between set type is read-only");