1 /* This file is part of the Zebra server.
2 Copyright (C) 2004-2013 Index Data
3
4 Zebra is free software; you can redistribute it and/or modify it under
5 the terms of the GNU General Public License as published by the Free
6 Software Foundation; either version 2, or (at your option) any later
7 version.
8
9 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
10 WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
12 for more details.
13
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software
16 Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
17
18 */
19
20 #if HAVE_CONFIG_H
21 #include <config.h>
22 #endif
23 #include <assert.h>
24 #include <fcntl.h>
25 #include <stdlib.h>
26 #include <stdio.h>
27 #include <string.h>
28 #ifdef WIN32
29 #include <io.h>
30 #endif
31 #if HAVE_UNISTD_H
32 #include <unistd.h>
33 #endif
34 #include <sys/types.h>
35
36 #include <idzebra/util.h>
37 #include <rset.h>
38
39 static RSFD r_open(RSET ct, int flag);
40 static void r_close(RSFD rfd);
41 static void r_delete(RSET ct);
42 static int r_read(RSFD rfd, void *buf, TERMID *term);
43 static int r_write(RSFD rfd, const void *buf);
44 static void r_pos(RSFD rfd, double *current, double *total);
45 static void r_flush(RSFD rfd, int mk);
46 static void r_reread(RSFD rfd);
47
48 static const struct rset_control control =
49 {
50 "temp",
51 r_delete,
52 rset_get_one_term,
53 r_open,
54 r_close,
55 0, /* no forward */
56 r_pos,
57 r_read,
58 r_write,
59 };
60
61 struct rset_private {
62 int fd; /* file descriptor for temp file */
63 char *fname; /* name of temp file */
64 char *buf_mem; /* window buffer */
65 size_t buf_size; /* size of window */
66 size_t pos_end; /* last position in set */
67 size_t pos_buf; /* position of first byte in window */
68 size_t pos_border; /* position of last byte+1 in window */
69 int dirty; /* window is dirty */
70 zint hits; /* no of hits */
71 char *temp_path;
72 };
73
74 struct rfd_private {
75 void *buf;
76 size_t pos_cur; /* current position in set */
77 /* FIXME - term pos or what ?? */
78 zint cur; /* number of the current hit */
79 };
80
81 static int log_level = 0;
82 static int log_level_initialized = 0;
83
rset_create_temp(NMEM nmem,struct rset_key_control * kcontrol,int scope,const char * temp_path,TERMID term)84 RSET rset_create_temp(NMEM nmem, struct rset_key_control *kcontrol,
85 int scope, const char *temp_path, TERMID term)
86 {
87 RSET rnew = rset_create_base(&control, nmem, kcontrol, scope, term,
88 0, 0);
89 struct rset_private *info;
90 if (!log_level_initialized)
91 {
92 log_level = yaz_log_module_level("rstemp");
93 log_level_initialized = 1;
94 }
95 info = (struct rset_private *) nmem_malloc(rnew->nmem, sizeof(*info));
96 info->fd = -1;
97 info->fname = NULL;
98 info->buf_size = 4096;
99 info->buf_mem = (char *) nmem_malloc(rnew->nmem, info->buf_size);
100 info->pos_end = 0;
101 info->pos_buf = 0;
102 info->dirty = 0;
103 info->hits = 0;
104
105 if (!temp_path)
106 info->temp_path = NULL;
107 else
108 info->temp_path = nmem_strdup(rnew->nmem, temp_path);
109 rnew->priv = info;
110 return rnew;
111 } /* rstemp_create */
112
r_delete(RSET ct)113 static void r_delete(RSET ct)
114 {
115 struct rset_private *info = (struct rset_private*) ct->priv;
116
117 yaz_log(log_level, "r_delete: set size %ld", (long) info->pos_end);
118 if (info->fname)
119 {
120 yaz_log(log_level, "r_delete: unlink %s", info->fname);
121 unlink(info->fname);
122 }
123 }
124
r_open(RSET ct,int flag)125 static RSFD r_open(RSET ct, int flag)
126 {
127 struct rset_private *info = (struct rset_private *) ct->priv;
128 RSFD rfd;
129 struct rfd_private *prfd;
130
131 if (info->fd == -1 && info->fname)
132 {
133 if (flag & RSETF_WRITE)
134 info->fd = open(info->fname, O_BINARY|O_RDWR|O_CREAT, 0666);
135 else
136 info->fd = open(info->fname, O_BINARY|O_RDONLY);
137 if (info->fd == -1)
138 {
139 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: open failed %s", info->fname);
140 zebra_exit("r_open");
141 }
142 }
143 rfd = rfd_create_base(ct);
144 if (!rfd->priv)
145 {
146 prfd = (struct rfd_private *) nmem_malloc(ct->nmem, sizeof(*prfd));
147 rfd->priv = (void *)prfd;
148 prfd->buf = nmem_malloc(ct->nmem,ct->keycontrol->key_size);
149 }
150 else
151 prfd= rfd->priv;
152 r_flush(rfd, 0);
153 prfd->pos_cur = 0;
154 info->pos_buf = 0;
155 r_reread(rfd);
156 prfd->cur = 0;
157 return rfd;
158 }
159
160 /* r_flush:
161 flush current window to file if file is assocated with set
162 */
r_flush(RSFD rfd,int mk)163 static void r_flush(RSFD rfd, int mk)
164 {
165 struct rset_private *info = rfd->rset->priv;
166
167 if (!info->fname && mk)
168 {
169 #if HAVE_MKSTEMP
170 char template[1024];
171
172 *template = '\0';
173
174 if (info->temp_path)
175 sprintf(template, "%s/", info->temp_path);
176 strcat(template, "zrs_");
177 #if HAVE_UNISTD_H
178 sprintf(template + strlen(template), "%ld_", (long) getpid());
179 #endif
180 strcat(template, "XXXXXX");
181
182 info->fd = mkstemp(template);
183 if (info->fd == -1)
184 {
185 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: mkstemp %s", template);
186 zebra_exit("r_flush");
187 }
188 info->fname = nmem_strdup(rfd->rset->nmem, template);
189 #else
190 char *s = (char*) tempnam(info->temp_path, "zrs");
191 info->fname= nmem_strdup(rfd->rset->nmem, s);
192
193 yaz_log(log_level, "creating tempfile %s", info->fname);
194 info->fd = open(info->fname, O_BINARY|O_RDWR|O_CREAT, 0666);
195 if (info->fd == -1)
196 {
197 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: open %s", info->fname);
198 zebra_exit("r_flush");
199 }
200 #endif
201 }
202 if (info->fname && info->fd != -1 && info->dirty)
203 {
204 size_t count;
205 int r;
206
207 if (lseek(info->fd, info->pos_buf, SEEK_SET) == -1)
208 {
209 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: lseek (1) %s", info->fname);
210 zebra_exit("r_flusxh");
211 }
212 count = info->buf_size;
213 if (count > info->pos_end - info->pos_buf)
214 count = info->pos_end - info->pos_buf;
215 if ((r = write(info->fd, info->buf_mem, count)) < (int) count)
216 {
217 if (r == -1)
218 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: write %s", info->fname);
219 else
220 yaz_log(YLOG_FATAL, "rstemp: write of %ld but got %ld",
221 (long) count, (long) r);
222 zebra_exit("r_flush");
223 }
224 info->dirty = 0;
225 }
226 }
227
r_close(RSFD rfd)228 static void r_close(RSFD rfd)
229 {
230 struct rset_private *info = (struct rset_private *)rfd->rset->priv;
231 if (rfd_is_last(rfd))
232 {
233 r_flush(rfd, 0);
234 if (info->fname && info->fd != -1)
235 {
236 close(info->fd);
237 info->fd = -1;
238 }
239 }
240 }
241
242
243 /* r_reread:
244 read from file to window if file is assocated with set -
245 indicated by fname
246 */
r_reread(RSFD rfd)247 static void r_reread(RSFD rfd)
248 {
249 struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
250 struct rset_private *info = (struct rset_private *)rfd->rset->priv;
251
252 if (info->fname)
253 {
254 size_t count;
255 int r;
256
257 info->pos_border = mrfd->pos_cur +
258 info->buf_size;
259 if (info->pos_border > info->pos_end)
260 info->pos_border = info->pos_end;
261 count = info->pos_border - info->pos_buf;
262 if (count > 0)
263 {
264 if (lseek(info->fd, info->pos_buf, SEEK_SET) == -1)
265 {
266 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: lseek (2) %s fd=%d", info->fname, info->fd);
267 zebra_exit("r_reread");
268 }
269 if ((r = read(info->fd, info->buf_mem, count)) < (int) count)
270 {
271 if (r == -1)
272 yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: read %s", info->fname);
273 else
274 yaz_log(YLOG_FATAL, "read of %ld but got %ld",
275 (long) count, (long) r);
276 zebra_exit("r_reread");
277 }
278 }
279 }
280 else
281 info->pos_border = info->pos_end;
282 }
283
r_read(RSFD rfd,void * buf,TERMID * term)284 static int r_read(RSFD rfd, void *buf, TERMID *term)
285 {
286 struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
287 struct rset_private *info = (struct rset_private *)rfd->rset->priv;
288
289 size_t nc = mrfd->pos_cur + rfd->rset->keycontrol->key_size;
290
291 if (mrfd->pos_cur < info->pos_buf || nc > info->pos_border)
292 {
293 if (nc > info->pos_end)
294 return 0;
295 r_flush(rfd, 0);
296 info->pos_buf = mrfd->pos_cur;
297 r_reread(rfd);
298 }
299 memcpy(buf, info->buf_mem + (mrfd->pos_cur - info->pos_buf),
300 rfd->rset->keycontrol->key_size);
301 if (term)
302 *term = rfd->rset->term;
303 /* FIXME - should we store and return terms ?? */
304 mrfd->pos_cur = nc;
305 mrfd->cur++;
306 return 1;
307 }
308
r_write(RSFD rfd,const void * buf)309 static int r_write(RSFD rfd, const void *buf)
310 {
311 struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
312 struct rset_private *info = (struct rset_private *)rfd->rset->priv;
313
314 size_t nc = mrfd->pos_cur + rfd->rset->keycontrol->key_size;
315
316 if (nc > info->pos_buf + info->buf_size)
317 {
318 r_flush(rfd, 1);
319 info->pos_buf = mrfd->pos_cur;
320 if (info->pos_buf < info->pos_end)
321 r_reread(rfd);
322 }
323 info->dirty = 1;
324 memcpy(info->buf_mem + (mrfd->pos_cur - info->pos_buf), buf,
325 rfd->rset->keycontrol->key_size);
326 mrfd->pos_cur = nc;
327 if (nc > info->pos_end)
328 info->pos_border = info->pos_end = nc;
329 info->hits++;
330 return 1;
331 }
332
r_pos(RSFD rfd,double * current,double * total)333 static void r_pos(RSFD rfd, double *current, double *total)
334 {
335 struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
336 struct rset_private *info = (struct rset_private *)rfd->rset->priv;
337
338 *current = (double) mrfd->cur;
339 *total = (double) info->hits;
340 }
341 /*
342 * Local variables:
343 * c-basic-offset: 4
344 * c-file-style: "Stroustrup"
345 * indent-tabs-mode: nil
346 * End:
347 * vim: shiftwidth=4 tabstop=8 expandtab
348 */
349
350