-/*
- * Copyright (C) 1994-1995, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: main.c,v $
- * Revision 1.38 1996-03-26 16:01:14 adam
- * New setting lockPath: directory of various lock files.
- *
- * Revision 1.37 1996/03/19 12:43:26 adam
- * Bug fix: File update traversal didn't handle trailing slashes correctly.
- * Bug fix: Update of sub directory groups wasn't handled correctly.
- *
- * Revision 1.36 1996/02/12 18:45:37 adam
- * New fileVerboseFlag in record group control.
- *
- * Revision 1.35 1996/02/12 15:56:11 adam
- * New code command: unread.
- *
- * Revision 1.34 1996/02/07 14:06:39 adam
- * Better progress report during register merge.
- * New command: clean - removes temporary shadow files.
- *
- * Revision 1.33 1996/02/06 17:11:18 adam
- * Minor changes.
- *
- * Revision 1.32 1996/02/01 20:50:04 adam
- * Bug fix: zebraIndexUnlock was always called even though zebraIndexLock
- * was never called - happens when no commands are specified.
- *
- * Revision 1.31 1996/01/08 19:15:46 adam
- * New input filter that works!
- *
- * Revision 1.30 1995/12/12 16:00:59 adam
- * System call sync(2) used after update/commit.
- * Locking (based on fcntl) uses F_EXLCK and F_SHLCK instead of F_WRLCK
- * and F_RDLCK.
- *
- * Revision 1.29 1995/12/11 11:43:30 adam
- * Locking based on fcntl instead of flock.
- * Setting commitEnable removed. Command line option -n can be used to
- * prevent commit if commit setting is defined in the configuration file.
- *
- * Revision 1.28 1995/12/08 16:22:56 adam
- * Work on update while servers are running. Three lock files introduced.
- * The servers reload their registers when necessary, but they don't
- * reestablish result sets yet.
- *
- * Revision 1.27 1995/12/07 17:38:47 adam
- * Work locking mechanisms for concurrent updates/commit.
- *
- * Revision 1.26 1995/12/06 12:41:23 adam
- * New command 'stat' for the index program.
- * Filenames can be read from stdin by specifying '-'.
- * Bug fix/enhancement of the transformation from terms to regular
- * expressons in the search engine.
- *
- * Revision 1.25 1995/12/01 16:24:39 adam
- * Commit files use separate meta file area.
- *
- * Revision 1.24 1995/11/30 17:01:38 adam
- * New setting commitCache: points to commit directories/files.
- * New command commit: commits at the end of a zebraidx run.
- *
- * Revision 1.23 1995/11/30 08:34:31 adam
- * Started work on commit facility.
- * Changed a few malloc/free to xmalloc/xfree.
- *
- * Revision 1.22 1995/11/28 09:09:42 adam
- * Zebra config renamed.
- * Use setting 'recordId' to identify record now.
- * Bug fix in recindex.c: rec_release_blocks was invokeded even
- * though the blocks were already released.
- * File traversal properly deletes records when needed.
- *
- * Revision 1.21 1995/11/27 14:27:39 adam
- * Renamed 'update' command to 'dir'.
- *
- * Revision 1.20 1995/11/27 13:58:53 adam
- * New option -t. storeStore data implemented in server.
- *
- * Revision 1.19 1995/11/25 10:24:06 adam
- * More record fields - they are enumerated now.
- * New options: flagStoreData flagStoreKey.
- *
- * Revision 1.18 1995/11/22 17:19:17 adam
- * Record management uses the bfile system.
- *
- * Revision 1.17 1995/11/21 15:01:16 adam
- * New general match criteria implemented.
- * New feature: document groups.
- *
- * Revision 1.16 1995/11/20 11:56:27 adam
- * Work on new traversal.
- *
- * Revision 1.15 1995/11/01 16:25:51 quinn
- * *** empty log message ***
- *
- * Revision 1.14 1995/10/17 18:02:09 adam
- * New feature: databases. Implemented as prefix to words in dictionary.
- *
- * Revision 1.13 1995/10/10 12:24:39 adam
- * Temporary sort files are compressed.
- *
- * Revision 1.12 1995/10/04 16:57:20 adam
- * Key input and merge sort in one pass.
- *
- * Revision 1.11 1995/09/29 14:01:45 adam
- * Bug fixes.
- *
- * Revision 1.10 1995/09/28 14:22:57 adam
- * Sort uses smaller temporary files.
- *
- * Revision 1.9 1995/09/14 07:48:24 adam
- * Record control management.
- *
- * Revision 1.8 1995/09/06 16:11:18 adam
- * Option: only one word key per file.
- *
- * Revision 1.7 1995/09/05 15:28:39 adam
- * More work on search engine.
- *
- * Revision 1.6 1995/09/04 12:33:43 adam
- * Various cleanup. YAZ util used instead.
- *
- * Revision 1.5 1995/09/04 09:10:39 adam
- * More work on index add/del/update.
- * Merge sort implemented.
- * Initial work on z39 server.
- *
- * Revision 1.4 1995/09/01 14:06:36 adam
- * Split of work into more files.
- *
- * Revision 1.3 1995/09/01 10:57:07 adam
- * Minor changes.
- *
- * Revision 1.2 1995/09/01 10:30:24 adam
- * More work on indexing. Not working yet.
- *
- * Revision 1.1 1995/08/31 14:50:24 adam
- * New simple file index tool.
- *
- */
+/* $Id: main.c,v 1.126 2005-08-26 10:11:49 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 <stdio.h>
+#include <stdlib.h>
+#include <string.h>
#include <assert.h>
+#ifdef WIN32
+#include <io.h>
+#endif
+#if HAVE_UNISTD_H
#include <unistd.h>
+#endif
+#if HAVE_SYS_TIME_H
+#include <sys/time.h>
+#endif
+#include <time.h>
+#if HAVE_SYS_TIMES_H
+#include <sys/times.h>
+#endif
-#include <alexutil.h>
-#include <data1.h>
-#include "index.h"
+#include <yaz/log.h>
+#include <yaz/options.h>
+#include <idzebra/api.h>
char *prog;
-size_t mem_max = 4*1024*1024;
-extern char *data1_tabpath;
-static void abort_func (int level, const char *msg, void *info)
+static void filter_cb(void *cd, const char *name)
+{
+ puts (name);
+}
+
+static void show_filters(ZebraService zs)
{
- if (level & LOG_FATAL)
- abort ();
+ zebra_filter_info(zs, 0, filter_cb);
}
int main (int argc, char **argv)
int ret;
int cmd = 0;
char *arg;
- char *configName = NULL;
- int nsections;
- int disableCommit = 0;
-
- struct recordGroup rGroupDef;
+ char *configName = 0;
+ int nsections = 0;
+ int enable_commit = 1;
+ char *database = 0;
+ Res res = res_open(0, 0, 0);
- rGroupDef.groupName = NULL;
- rGroupDef.databaseName = NULL;
- rGroupDef.path = NULL;
- rGroupDef.recordId = NULL;
- rGroupDef.recordType = NULL;
- rGroupDef.flagStoreData = -1;
- rGroupDef.flagStoreKeys = -1;
- rGroupDef.fileVerboseFlag = 1;
+ int trans_started=0;
+#if HAVE_SYS_TIMES_H
+ struct tms tms1, tms2;
+ double usec;
+#endif
+#if HAVE_SYS_TIME_H
+ struct timeval start_time, end_time;
+#endif
+#ifndef WIN32
+ char nbuf[100];
+#endif
+ ZebraService zs = 0;
+ ZebraHandle zh = 0;
+ nmem_init ();
+
+#ifdef WIN32
+#else
+ sprintf(nbuf, "%.40s(%ld)", *argv, (long) getpid());
+ yaz_log_init_prefix (nbuf);
+#endif
+#if HAVE_SYS_TIMES_H
+ times(&tms1);
+#endif
+#if HAVE_SYS_TIME_H
+ gettimeofday(&start_time, 0);
+#endif
prog = *argv;
if (argc < 2)
{
- fprintf (stderr, "zebraidx [options] command <dir> ...\n"
+ fprintf (stderr, "%s [options] command <dir> ...\n"
"Commands:\n"
" update <dir> Update index with files below <dir>.\n"
" If <dir> is empty filenames are read from stdin.\n"
" delete <dir> Delete index with files below <dir>.\n"
+ " create <db> Create database <db>\n"
+ " drop <db> Drop database <db>\n"
" commit Commit changes\n"
" clean Clean shadow files\n"
"Options:\n"
" -g <group> Index files according to group settings.\n"
" -d <database> Records belong to Z39.50 database <database>.\n"
" -m <mbytes> Use <mbytes> before flushing keys to disk.\n"
- " -n Don't use shadow system\n"
- " -v <level> Set logging to <level>.\n");
+ " -n Don't use shadow system.\n"
+ " -s Show analysis on stdout, but do no work.\n"
+ " -v <level> Set logging to <level>.\n"
+ " -l <file> Write log to <file>.\n"
+ " -L Don't follow symbolic links.\n"
+ " -f <n> Display information for the first <n> records.\n"
+ " -V Show version.\n", *argv
+ );
exit (1);
}
- log_event_end (abort_func, NULL);
- while ((ret = options ("t:c:g:d:m:v:n", argv, argc, &arg)) != -2)
+ while ((ret = options("sVt:c:g:d:m:v:nf:l:L", argv, argc, &arg)) != -2)
{
if (ret == 0)
{
- const char *rval;
if(cmd == 0) /* command */
{
- if (!common_resource)
+ if (!zs)
{
- common_resource = res_open (configName ?
- configName : FNAME_CONFIG);
- if (!common_resource)
+ const char *config = configName ? configName : "zebra.cfg";
+ zs = zebra_start_res (config, 0, res);
+ if (!zs)
{
- logf (LOG_FATAL, "Cannot open resource `%s'",
- configName);
+ yaz_log (YLOG_FATAL, "Cannot read config %s", config);
exit (1);
- }
- data1_tabpath = res_get (common_resource, "profilePath");
- bf_lockDir (res_get (common_resource, "lockPath"));
+ }
+ zh = zebra_open (zs);
+ zebra_shadow_enable (zh, enable_commit);
}
+
+ if (database &&
+ zebra_select_database (zh, database) == ZEBRA_FAIL)
+ {
+ yaz_log(YLOG_FATAL, "Could not select database %s "
+ "errCode=%d",
+ database, zebra_errCode(zh) );
+ exit (1);
+ }
if (!strcmp (arg, "update"))
cmd = 'u';
+ else if (!strcmp (arg, "update1"))
+ cmd = 'U';
+ else if (!strcmp (arg, "update2"))
+ cmd = 'm';
else if (!strcmp (arg, "dump"))
cmd = 's';
else if (!strcmp (arg, "del") || !strcmp(arg, "delete"))
cmd = 'd';
+ else if (!strcmp (arg, "init"))
+ {
+ zebra_init (zh);
+ }
+ else if (!strcmp(arg, "drop"))
+ {
+ cmd = 'D';
+ }
+ else if (!strcmp(arg, "create"))
+ {
+ cmd = 'C';
+ }
else if (!strcmp (arg, "commit"))
{
- zebraIndexLock (1);
- rval = res_get (common_resource, "shadow");
- if (rval && *rval)
- bf_cache (1);
- else
- {
- logf (LOG_FATAL, "Cannot perform commit");
- logf (LOG_FATAL, "No shadow area defined");
- exit (1);
- }
- if (bf_commitExists ())
- {
- logf (LOG_LOG, "Commit start");
- zebraIndexLockMsg ("c");
- zebraIndexWait (1);
- logf (LOG_LOG, "Commit execute");
- bf_commitExec ();
- sync ();
- zebraIndexLockMsg ("d");
- zebraIndexWait (0);
- logf (LOG_LOG, "Commit clean");
- bf_commitClean ();
- }
- else
- logf (LOG_LOG, "Nothing to commit");
+ zebra_commit (zh);
}
else if (!strcmp (arg, "clean"))
{
- zebraIndexLock (1);
- if (bf_commitExists ())
- {
- zebraIndexLockMsg ("d");
- zebraIndexWait (0);
- logf (LOG_LOG, "Commit clean");
- bf_commitClean ();
- }
- else
- logf (LOG_LOG, "Nothing to clean");
+ zebra_clean (zh);
}
else if (!strcmp (arg, "stat") || !strcmp (arg, "status"))
{
- zebraIndexLock (0);
- rval = res_get (common_resource, "shadow");
- if (rval && *rval)
- {
- bf_cache (1);
- zebraIndexLockMsg ("r");
- }
- rec_prstat ();
+ zebra_register_statistics (zh,0);
}
- else
+ else if (!strcmp (arg, "dumpdict"))
{
- logf (LOG_FATAL, "Unknown command: %s", arg);
- exit (1);
+ zebra_register_statistics (zh,1);
}
- }
- else
- {
- struct recordGroup rGroup;
-
- zebraIndexLock (0);
- rval = res_get (common_resource, "shadow");
- if (rval && *rval && !disableCommit)
+ else if (!strcmp (arg, "compact"))
{
- bf_cache (1);
- zebraIndexLockMsg ("r");
+ zebra_compact (zh);
}
- else
+ else if (!strcmp (arg, "filters"))
{
- bf_cache (0);
- zebraIndexLockMsg ("w");
+ show_filters(zs);
}
- zebraIndexWait (0);
-
- memcpy (&rGroup, &rGroupDef, sizeof(rGroup));
- key_open (mem_max);
- rGroup.path = arg;
- if (cmd == 'u')
- {
- logf (LOG_LOG, "Updating %s", rGroup.path);
- repositoryUpdate (&rGroup);
- }
- else if (cmd == 'd')
- {
- logf (LOG_LOG, "Deleting %s", rGroup.path);
- repositoryDelete (&rGroup);
- }
- else if (cmd == 's')
+ else
{
- logf (LOG_LOG, "Dumping %s", rGroup.path);
- repositoryShow (&rGroup);
+ yaz_log (YLOG_FATAL, "unknown command: %s", arg);
+ exit (1);
}
-
- cmd = 0;
- nsections = key_close ();
- if (nsections)
+ }
+ else
+ {
+ ZEBRA_RES res = ZEBRA_OK;
+ if (!trans_started)
+ {
+ trans_started=1;
+ if (zebra_begin_trans (zh, 1) != ZEBRA_OK)
+ exit(1);
+ }
+ switch (cmd)
{
- logf (LOG_LOG, "Merging with index");
- key_input (FNAME_WORD_DICT, FNAME_WORD_ISAM, nsections,
- 60);
- sync ();
+ case 'u':
+ res = zebra_repository_update (zh, arg);
+ break;
+ case 'd':
+ res = zebra_repository_delete (zh, arg);
+ break;
+ case 's':
+ res = zebra_repository_show (zh, arg);
+ nsections = 0;
+ break;
+ case 'C':
+ res = zebra_create_database(zh, arg);
+ break;
+ case 'D':
+ res = zebra_drop_database(zh, arg);
+ break;
+ default:
+ nsections = 0;
}
+ if (res != ZEBRA_OK)
+ {
+ yaz_log(YLOG_WARN, "Operation failed");
+ exit(1);
+ }
+ log_event_end (NULL, NULL);
}
}
- else if (ret == 'v')
+ else if (ret == 'V')
{
- log_init (log_mask_str(arg), prog, NULL);
+ printf("Zebra %s %s\n", ZEBRAVER, ZEBRADATE);
+ printf(" (C) 1994-2005, Index Data ApS\n");
+#ifdef WIN32
+#ifdef _DEBUG
+ printf(" WIN32 Debug\n");
+#else
+ printf(" WIN32 Release\n");
+#endif
+#endif
+#if HAVE_BZLIB_H
+ printf("Using: libbzip2, (C) 1996-1999 Julian R Seward. All rights reserved.\n");
+#endif
}
+ else if (ret == 'v')
+ yaz_log_init_level (yaz_log_mask_str(arg));
+ else if (ret == 'l')
+ yaz_log_init_file (arg);
else if (ret == 'm')
- {
- mem_max = 1024*1024*atoi(arg);
- }
+ res_set(res, "memMax", arg);
else if (ret == 'd')
- {
- rGroupDef.databaseName = arg;
- }
+ database = arg;
+ else if (ret == 's')
+ res_set(res, "openRW", "0");
else if (ret == 'g')
- {
- rGroupDef.groupName = arg;
- }
+ res_set(res, "group", arg);
+ else if (ret == 'f')
+ res_set(res, "fileVerboseLimit", arg);
else if (ret == 'c')
configName = arg;
else if (ret == 't')
- rGroupDef.recordType = arg;
+ res_set(res, "recordType", arg);
else if (ret == 'n')
- disableCommit = 1;
+ enable_commit = 0;
+ else if (ret == 'L')
+ res_set(res, "followLinks", "0");
else
- {
- logf (LOG_FATAL, "Unknown option '-%s'", arg);
- exit (1);
- }
+ yaz_log (YLOG_WARN, "unknown option '-%s'", arg);
+ } /* while arg */
+
+ if (trans_started)
+ zebra_end_trans (zh);
+
+ zebra_close (zh);
+ zebra_stop (zs);
+#if HAVE_SYS_TIMES_H
+#if HAVE_SYS_TIME_H
+ if (trans_started)
+ {
+ gettimeofday(&end_time, 0);
+ usec = (end_time.tv_sec - start_time.tv_sec) * 1000000.0 +
+ end_time.tv_usec - start_time.tv_usec;
+ times(&tms2);
+ yaz_log (YLOG_LOG, "zebraidx times: %5.2f %5.2f %5.2f",
+ usec / 1000000,
+ (double) (tms2.tms_utime - tms1.tms_utime)/100,
+ (double) (tms2.tms_stime - tms1.tms_stime)/100);
}
- if (common_resource)
- zebraIndexUnlock ();
+#endif
+#endif
+ nmem_exit();
exit (0);
+ return 0;
}