-/*
- * Copyright (C) 1994-1995, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: rstemp.c,v $
- * Revision 1.12 1995-09-28 09:52:11 adam
- * xfree/xmalloc used everywhere.
- *
- * Revision 1.11 1995/09/18 14:17:56 adam
- * Bug fixes.
- *
- * Revision 1.10 1995/09/15 14:45:39 adam
- * Bug fixes.
- *
- * Revision 1.9 1995/09/15 09:20:42 adam
- * Bug fixes.
- *
- * Revision 1.8 1995/09/08 14:52:42 adam
- * Work on relevance feedback.
- *
- * Revision 1.7 1995/09/07 13:58:44 adam
- * New parameter: result-set file descriptor (RSFD) to support multiple
- * positions within the same result-set.
- * Boolean operators: and, or, not implemented.
- *
- * Revision 1.6 1995/09/06 16:11:56 adam
- * More work on boolean sets.
- *
- * Revision 1.5 1995/09/05 16:36:59 adam
- * Minor changes.
- *
- * Revision 1.4 1995/09/05 11:43:24 adam
- * Complete version of temporary sets. Not tested yet though.
- *
- * Revision 1.3 1995/09/04 15:20:40 adam
- * More work on temp sets. is_open member removed.
- *
- * Revision 1.2 1995/09/04 09:10:56 adam
- * Minor changes.
- *
- * Revision 1.1 1994/11/04 13:21:30 quinn
- * Working.
- *
- */
+/* $Id: rstemp.c,v 1.65 2005-06-14 20:28:54 adam Exp $
+ Copyright (C) 1995-2005
+ Index Data ApS
+
+This file is part of the Zebra server.
+
+Zebra is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with Zebra; see the file LICENSE.zebra. If not, write to the
+Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA.
+*/
-#include <fcntl.h>
#include <assert.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#ifdef WIN32
+#include <io.h>
+#endif
+#if HAVE_UNISTD_H
#include <unistd.h>
+#endif
#include <sys/types.h>
-#include <stdio.h>
-#include <alexutil.h>
-#include <rstemp.h>
+#include <idzebra/util.h>
+#include <rset.h>
-static rset_control *r_create(const struct rset_control *sel, void *parms);
-static RSFD r_open (rset_control *ct, int wflag);
-static void r_close (RSFD rfd);
-static void r_delete (rset_control *ct);
-static void r_rewind (RSFD rfd);
-static int r_count (rset_control *ct);
-static int r_read (RSFD rfd, void *buf);
-static int r_write (RSFD rfd, const void *buf);
+static RSFD r_open(RSET ct, int flag);
+static void r_close(RSFD rfd);
+static void r_delete(RSET ct);
+static int r_read(RSFD rfd, void *buf, TERMID *term);
+static int r_write(RSFD rfd, const void *buf);
+static void r_pos(RSFD rfd, double *current, double *total);
+static void r_flush(RSFD rfd, int mk);
+static void r_reread(RSFD rfd);
-static const rset_control control =
+static const struct rset_control control =
{
- "Temporary set",
- 0,
- r_create,
+ "temp",
+ r_delete,
+ rset_get_one_term,
r_open,
r_close,
- r_delete,
- r_rewind,
- r_count,
+ 0, /* no forward */
+ r_pos,
r_read,
- r_write
+ r_write,
};
-const rset_control *rset_kind_temp = &control;
-
-struct rset_temp_info {
- int fd;
- char *fname;
- size_t key_size; /* key size */
+struct rset_private {
+ int fd; /* file descriptor for temp file */
+ char *fname; /* name of temp file */
char *buf_mem; /* window buffer */
size_t buf_size; /* size of window */
size_t pos_end; /* last position in set */
- size_t pos_cur; /* current position in set */
size_t pos_buf; /* position of first byte in window */
size_t pos_border; /* position of last byte+1 in window */
int dirty; /* window is dirty */
+ zint hits; /* no of hits */
+ char *temp_path;
};
-struct rset_temp_rfd {
- struct rset_temp_info *info;
- struct rset_temp_rfd *next;
+struct rfd_private {
+ void *buf;
+ size_t pos_cur; /* current position in set */
+ /* FIXME - term pos or what ?? */
+ zint cur; /* number of the current hit */
};
-static struct rset_control *r_create(const struct rset_control *sel,
- void *parms)
-{
- rset_control *newct;
- rset_temp_parms *temp_parms = parms;
- struct rset_temp_info *info;
-
- newct = xmalloc(sizeof(*newct));
- memcpy(newct, sel, sizeof(*sel));
- newct->buf = xmalloc (sizeof(struct rset_temp_info));
- info = newct->buf;
+static int log_level = 0;
+static int log_level_initialized = 0;
+RSET rstemp_create(NMEM nmem, struct rset_key_control *kcontrol,
+ int scope, const char *temp_path, TERMID term)
+{
+ RSET rnew = rset_create_base(&control, nmem, kcontrol, scope, term,
+ 0, 0);
+ struct rset_private *info;
+ if (!log_level_initialized)
+ {
+ log_level = yaz_log_module_level("rstemp");
+ log_level_initialized = 1;
+ }
+ info = (struct rset_private *) nmem_malloc(rnew->nmem, sizeof(*info));
info->fd = -1;
info->fname = NULL;
- info->key_size = temp_parms->key_size;
info->buf_size = 4096;
- info->buf_mem = xmalloc (info->buf_size);
- info->pos_cur = 0;
+ info->buf_mem = (char *) nmem_malloc(rnew->nmem, info->buf_size);
info->pos_end = 0;
info->pos_buf = 0;
info->dirty = 0;
+ info->hits = 0;
- return newct;
-}
+ if (!temp_path)
+ info->temp_path = NULL;
+ else
+ info->temp_path = nmem_strdup(rnew->nmem, temp_path);
+ rnew->priv = info;
+ return rnew;
+} /* rstemp_create */
-static RSFD r_open (struct rset_control *ct, int wflag)
+static void r_delete(RSET ct)
{
- struct rset_temp_info *info = ct->buf;
- struct rset_temp_rfd *rfd;
+ struct rset_private *info = (struct rset_private*) ct->priv;
- assert (info->fd == -1);
+ yaz_log(log_level, "r_delete: set size %ld", (long) info->pos_end);
if (info->fname)
{
- if (wflag)
- info->fd = open (info->fname, O_RDWR|O_CREAT, 0666);
+ yaz_log(log_level, "r_delete: unlink %s", info->fname);
+ unlink(info->fname);
+ }
+}
+
+static RSFD r_open(RSET ct, int flag)
+{
+ struct rset_private *info = (struct rset_private *) ct->priv;
+ RSFD rfd;
+ struct rfd_private *prfd;
+
+ if (info->fd == -1 && info->fname)
+ {
+ if (flag & RSETF_WRITE)
+ info->fd = open(info->fname, O_BINARY|O_RDWR|O_CREAT, 0666);
else
- info->fd = open (info->fname, O_RDONLY);
+ info->fd = open(info->fname, O_BINARY|O_RDONLY);
if (info->fd == -1)
{
- logf (LOG_FATAL|LOG_ERRNO, "open %s", info->fname);
- exit (1);
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: open failed %s", info->fname);
+ exit(1);
}
}
- rfd = xmalloc (sizeof(*rfd));
- rfd->info = info;
- r_rewind (rfd);
+ rfd = rfd_create_base(ct);
+ if (!rfd->priv)
+ {
+ prfd = (struct rfd_private *) nmem_malloc(ct->nmem, sizeof(*prfd));
+ rfd->priv = (void *)prfd;
+ prfd->buf = nmem_malloc(ct->nmem,ct->keycontrol->key_size);
+ }
+ else
+ prfd= rfd->priv;
+ r_flush(rfd, 0);
+ prfd->pos_cur = 0;
+ info->pos_buf = 0;
+ r_reread(rfd);
+ prfd->cur = 0;
return rfd;
}
/* r_flush:
flush current window to file if file is assocated with set
*/
-static void r_flush (RSFD rfd, int mk)
+static void r_flush(RSFD rfd, int mk)
{
- struct rset_temp_info *info = ((struct rset_temp_rfd*) rfd)->info;
+ struct rset_private *info = rfd->rset->priv;
if (!info->fname && mk)
{
- char *s = (char*) tempnam (NULL, "zrs");
+#if HAVE_MKSTEMP
+ char template[1024];
+ if (info->temp_path)
+ sprintf(template, "%s/zrsXXXXXX", info->temp_path);
+ else
+ sprintf(template, "zrsXXXXXX");
+
+ info->fd = mkstemp(template);
- info->fname = xmalloc (strlen(s)+1);
- strcpy (info->fname, s);
+ if (info->fd == -1)
+ {
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: mkstemp %s", template);
+ exit(1);
+ }
+ info->fname = nmem_strdup(rfd->rset->nmem, template);
+#else
+ char *s = (char*) tempnam(info->temp_path, "zrs");
+ info->fname= nmem_strdup(rfd->rset->nmem, s);
- logf (LOG_DEBUG, "creating tempfile %s", info->fname);
- info->fd = open (info->fname, O_RDWR|O_CREAT, 0666);
+ yaz_log(log_level, "creating tempfile %s", info->fname);
+ info->fd = open(info->fname, O_BINARY|O_RDWR|O_CREAT, 0666);
if (info->fd == -1)
{
- logf (LOG_FATAL|LOG_ERRNO, "open %s", info->fname);
- exit (1);
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: open %s", info->fname);
+ exit(1);
}
+#endif
}
if (info->fname && info->fd != -1 && info->dirty)
{
- size_t r, count;
+ size_t count;
+ int r;
- if (lseek (info->fd, info->pos_buf, SEEK_SET) == -1)
+ if (lseek(info->fd, info->pos_buf, SEEK_SET) == -1)
{
- logf (LOG_FATAL|LOG_ERRNO, "lseek %s", info->fname);
- exit (1);
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: lseek (1) %s", info->fname);
+ exit(1);
}
count = info->buf_size;
if (count > info->pos_end - info->pos_buf)
count = info->pos_end - info->pos_buf;
- if ((r = write (info->fd, info->buf_mem, count)) < count)
+ if ((r = write(info->fd, info->buf_mem, count)) < (int) count)
{
if (r == -1)
- logf (LOG_FATAL|LOG_ERRNO, "read %s", info->fname);
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: write %s", info->fname);
else
- logf (LOG_FATAL, "write of %ld but got %ld",
+ yaz_log(YLOG_FATAL, "rstemp: write of %ld but got %ld",
(long) count, (long) r);
- exit (1);
+ exit(1);
}
info->dirty = 0;
}
}
-static void r_close (RSFD rfd)
+static void r_close(RSFD rfd)
{
- struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info;
-
- r_flush (rfd, 0);
- if (info->fname && info->fd != -1)
+ struct rset_private *info = (struct rset_private *)rfd->rset->priv;
+ if (rfd_is_last(rfd))
{
- close (info->fd);
- info->fd = -1;
+ r_flush(rfd, 0);
+ if (info->fname && info->fd != -1)
+ {
+ close(info->fd);
+ info->fd = -1;
+ }
}
}
-static void r_delete (struct rset_control *ct)
-{
- struct rset_temp_info *info = ct->buf;
-
- if (info->fname)
- unlink (info->fname);
- free (info->buf_mem);
- logf (LOG_DEBUG, "r_delete: set size %ld", (long) info->pos_end);
- if (info->fname)
- {
- logf (LOG_DEBUG, "r_delete: unlink %s", info->fname);
- unlink (info->fname);
- free (info->fname);
- }
- free (info);
-}
/* r_reread:
read from file to window if file is assocated with set -
indicated by fname
*/
-static void r_reread (RSFD rfd)
+static void r_reread(RSFD rfd)
{
- struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info;
+ struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
+ struct rset_private *info = (struct rset_private *)rfd->rset->priv;
if (info->fname)
{
- size_t r, count;
+ size_t count;
+ int r;
- info->pos_border = info->pos_cur + info->buf_size;
+ info->pos_border = mrfd->pos_cur +
+ info->buf_size;
if (info->pos_border > info->pos_end)
info->pos_border = info->pos_end;
count = info->pos_border - info->pos_buf;
if (count > 0)
{
- if (lseek (info->fd, info->pos_buf, SEEK_SET) == -1)
+ if (lseek(info->fd, info->pos_buf, SEEK_SET) == -1)
{
- logf (LOG_FATAL|LOG_ERRNO, "lseek %s", info->fname);
- exit (1);
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: lseek (2) %s fd=%d", info->fname, info->fd);
+ exit(1);
}
- if ((r = read (info->fd, info->buf_mem, count)) < count)
+ if ((r = read(info->fd, info->buf_mem, count)) < (int) count)
{
if (r == -1)
- logf (LOG_FATAL|LOG_ERRNO, "read %s", info->fname);
+ yaz_log(YLOG_FATAL|YLOG_ERRNO, "rstemp: read %s", info->fname);
else
- logf (LOG_FATAL, "read of %ld but got %ld",
+ yaz_log(YLOG_FATAL, "read of %ld but got %ld",
(long) count, (long) r);
- exit (1);
+ exit(1);
}
}
}
info->pos_border = info->pos_end;
}
-static void r_rewind (RSFD rfd)
+static int r_read(RSFD rfd, void *buf, TERMID *term)
{
- struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info;
-
- r_flush (rfd, 0);
- info->pos_cur = 0;
- info->pos_buf = 0;
- r_reread (rfd);
-}
+ struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
+ struct rset_private *info = (struct rset_private *)rfd->rset->priv;
-static int r_count (struct rset_control *ct)
-{
- struct rset_temp_info *info = ct->buf;
-
- return info->pos_end / info->key_size;
-}
+ size_t nc = mrfd->pos_cur + rfd->rset->keycontrol->key_size;
-static int r_read (RSFD rfd, void *buf)
-{
- struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info;
-
- size_t nc = info->pos_cur + info->key_size;
-
- if (nc > info->pos_border)
+ if (mrfd->pos_cur < info->pos_buf || nc > info->pos_border)
{
if (nc > info->pos_end)
return 0;
- r_flush (rfd, 0);
- info->pos_buf = info->pos_cur;
- r_reread (rfd);
+ r_flush(rfd, 0);
+ info->pos_buf = mrfd->pos_cur;
+ r_reread(rfd);
}
- memcpy (buf, info->buf_mem + (info->pos_cur - info->pos_buf),
- info->key_size);
- info->pos_cur = nc;
+ memcpy(buf, info->buf_mem + (mrfd->pos_cur - info->pos_buf),
+ rfd->rset->keycontrol->key_size);
+ if (term)
+ *term = rfd->rset->term;
+ /* FIXME - should we store and return terms ?? */
+ mrfd->pos_cur = nc;
+ mrfd->cur++;
return 1;
}
-static int r_write (RSFD rfd, const void *buf)
+static int r_write(RSFD rfd, const void *buf)
{
- struct rset_temp_info *info = ((struct rset_temp_rfd*)rfd)->info;
+ struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
+ struct rset_private *info = (struct rset_private *)rfd->rset->priv;
- size_t nc = info->pos_cur + info->key_size;
+ size_t nc = mrfd->pos_cur + rfd->rset->keycontrol->key_size;
if (nc > info->pos_buf + info->buf_size)
{
- r_flush (rfd, 1);
- info->pos_buf = info->pos_cur;
+ r_flush(rfd, 1);
+ info->pos_buf = mrfd->pos_cur;
if (info->pos_buf < info->pos_end)
- r_reread (rfd);
+ r_reread(rfd);
}
info->dirty = 1;
- memcpy (info->buf_mem + (info->pos_cur - info->pos_buf), buf,
- info->key_size);
- info->pos_cur = nc;
+ memcpy(info->buf_mem + (mrfd->pos_cur - info->pos_buf), buf,
+ rfd->rset->keycontrol->key_size);
+ mrfd->pos_cur = nc;
if (nc > info->pos_end)
info->pos_border = info->pos_end = nc;
+ info->hits++;
return 1;
}
+static void r_pos(RSFD rfd, double *current, double *total)
+{
+ struct rfd_private *mrfd = (struct rfd_private*) rfd->priv;
+ struct rset_private *info = (struct rset_private *)rfd->rset->priv;
+
+ *current = (double) mrfd->cur;
+ *total = (double) info->hits;
+}