X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fmain.c;h=c29e4c0a9bc8787dd4ff74d44ec48487052df7eb;hb=8d3211326d448452d82a869143ed6d08d463954b;hp=1a3fc7665504e39238aa8d435f9962bf46165a9c;hpb=5a47fb16ab0fc9574660a0ed04151f359c826cbf;p=idzebra-moved-to-github.git diff --git a/index/main.c b/index/main.c index 1a3fc76..c29e4c0 100644 --- a/index/main.c +++ b/index/main.c @@ -1,172 +1,59 @@ /* - * Copyright (C) 1994-1995, Index Data I/S + * Copyright (C) 1994-2001, Index Data * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: main.c,v $ - * 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 1.80 2001-11-19 23:05:22 adam Exp $ */ #include +#include #include +#ifdef WIN32 +#include +#else #include +#endif -#include -#include +#include #include "index.h" +#include "recindex.h" + +#ifndef ZEBRASDR +#define ZEBRASDR 0 +#endif + +#if ZEBRASDR +#include "zebrasdr.h" +#endif char *prog; -size_t mem_max = 4*1024*1024; -extern char *data1_tabpath; -static void abort_func (int level, const char *msg, void *info) -{ - if (level & LOG_FATAL) - abort (); -} +Res common_resource = 0; + int main (int argc, char **argv) { int ret; int cmd = 0; char *arg; - char *configName = NULL; - int nsections; + char *configName = FNAME_CONFIG; + int nsections = 0; int disableCommit = 0; - + size_t mem_max = 0; + char nbuf[100]; struct recordGroup rGroupDef; - + + nmem_init (); + +#ifdef WIN32 +#else + sprintf(nbuf, "%.40s(%d)", *argv, getpid()); + yaz_log_init_prefix (nbuf); +#endif + +#if ZEBRASDR + zebraSdr_std (); + rGroupDef.useSDR = 0; +#endif rGroupDef.groupName = NULL; rGroupDef.databaseName = NULL; rGroupDef.path = NULL; @@ -174,12 +61,19 @@ int main (int argc, char **argv) rGroupDef.recordType = NULL; rGroupDef.flagStoreData = -1; rGroupDef.flagStoreKeys = -1; - rGroupDef.fileVerboseFlag = 1; + rGroupDef.flagRw = 1; + rGroupDef.databaseNamePath = 0; + rGroupDef.explainDatabase = 0; + rGroupDef.fileVerboseLimit = 100000; + rGroupDef.zebra_maps = NULL; + rGroupDef.dh = data1_create (); + rGroupDef.recTypes = recTypes_init (rGroupDef.dh); + recTypes_default_handlers (rGroupDef.recTypes); prog = *argv; if (argc < 2) { - fprintf (stderr, "zebraidx [options] command ...\n" + fprintf (stderr, "%s [options] command ...\n" "Commands:\n" " update Update index with files below .\n" " If is empty filenames are read from stdin.\n" @@ -192,12 +86,23 @@ int main (int argc, char **argv) " -g Index files according to group settings.\n" " -d Records belong to Z39.50 database .\n" " -m Use before flushing keys to disk.\n" - " -n Don't use shadow system\n" - " -v Set logging to .\n"); + " -n Don't use shadow system.\n" + " -s Show analysis on stdout, but do no work.\n" + " -v Set logging to .\n" + " -l Write log to .\n" + " -f Display information for the first records.\n" +#if ZEBRASDR + " -S Use SDRKit\n" +#endif + " -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:" +#if ZEBRASDR + "S" +#endif + , argv, argc, &arg)) != -2) { if (ret == 0) { @@ -206,158 +111,266 @@ int main (int argc, char **argv) { if (!common_resource) { +#if ZMBOL + logf (LOG_LOG, "zmbol version %s %s", + ZEBRAVER, ZEBRADATE); +#else + logf (LOG_LOG, "zebra version %s %s", + ZEBRAVER, ZEBRADATE); +#endif common_resource = res_open (configName ? configName : FNAME_CONFIG); if (!common_resource) { - logf (LOG_FATAL, "Cannot open resource `%s'", - configName); + logf (LOG_FATAL, "cannot read file `%s'", configName); exit (1); } - data1_tabpath = res_get (common_resource, "profilePath"); + data1_set_tabpath (rGroupDef.dh, res_get (common_resource, + "profilePath")); + + rGroupDef.bfs = + bfs_create (res_get (common_resource, "register")); + if (!rGroupDef.bfs) + { + logf (LOG_FATAL, "Cannot access register"); + exit(1); + } + + bf_lockDir (rGroupDef.bfs, + res_get (common_resource, "lockDir")); + rGroupDef.zebra_maps = zebra_maps_open (common_resource); } 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")) + { + zebraIndexUnlock(); + rval = res_get (common_resource, "shadow"); + zebraIndexLock (rGroupDef.bfs, 0, rval); + if (rval && *rval) + bf_cache (rGroupDef.bfs, rval); + zebraIndexLockMsg ("w"); + bf_reset (rGroupDef.bfs); + } else if (!strcmp (arg, "commit")) { - zebraIndexLock (1); rval = res_get (common_resource, "shadow"); + zebraIndexLock (rGroupDef.bfs, 1, rval); if (rval && *rval) - bf_cache (1); + bf_cache (rGroupDef.bfs, rval); else { logf (LOG_FATAL, "Cannot perform commit"); logf (LOG_FATAL, "No shadow area defined"); exit (1); } - if (bf_commitExists ()) + if (bf_commitExists (rGroupDef.bfs)) { - logf (LOG_LOG, "Commit start"); + logf (LOG_LOG, "commit start"); zebraIndexLockMsg ("c"); zebraIndexWait (1); - logf (LOG_LOG, "Commit execute"); - bf_commitExec (); + logf (LOG_LOG, "commit execute"); + bf_commitExec (rGroupDef.bfs); +#ifndef WIN32 sync (); +#endif zebraIndexLockMsg ("d"); zebraIndexWait (0); - logf (LOG_LOG, "Commit clean"); - bf_commitClean (); + logf (LOG_LOG, "commit clean"); + bf_commitClean (rGroupDef.bfs, rval); } else - logf (LOG_LOG, "Nothing to commit"); + logf (LOG_LOG, "nothing to commit"); } else if (!strcmp (arg, "clean")) { - zebraIndexLock (1); - if (bf_commitExists ()) + rval = res_get (common_resource, "shadow"); + zebraIndexLock (rGroupDef.bfs, 1, rval); + if (bf_commitExists (rGroupDef.bfs)) { zebraIndexLockMsg ("d"); zebraIndexWait (0); - logf (LOG_LOG, "Commit clean"); - bf_commitClean (); + logf (LOG_LOG, "commit clean"); + bf_commitClean (rGroupDef.bfs, rval); } else - logf (LOG_LOG, "Nothing to clean"); + logf (LOG_LOG, "nothing to clean"); } else if (!strcmp (arg, "stat") || !strcmp (arg, "status")) { - zebraIndexLock (0); + Records records; rval = res_get (common_resource, "shadow"); + zebraIndexLock (rGroupDef.bfs, 0, rval); if (rval && *rval) { - bf_cache (1); + bf_cache (rGroupDef.bfs, rval); zebraIndexLockMsg ("r"); } - rec_prstat (); + records = rec_open (rGroupDef.bfs, 0, 0); + rec_prstat (records); + rec_close (&records); + inv_prstat (rGroupDef.bfs); + } + else if (!strcmp (arg, "compact")) + { + rval = res_get (common_resource, "shadow"); + zebraIndexLock (rGroupDef.bfs, 0, rval); + if (rval && *rval) + { + bf_cache (rGroupDef.bfs, rval); + zebraIndexLockMsg ("r"); + } + inv_compact(rGroupDef.bfs); } else { - logf (LOG_FATAL, "Unknown command: %s", arg); + logf (LOG_FATAL, "unknown command: %s", arg); exit (1); } } - else + else { struct recordGroup rGroup; - - zebraIndexLock (0); +#if ZMBOL +#else + /* For zebra, delete lock file and reset register */ + if (rGroupDef.flagRw) + { + zebraIndexUnlock(); + bf_reset (rGroupDef.bfs); + } +#endif rval = res_get (common_resource, "shadow"); - if (rval && *rval && !disableCommit) - { - bf_cache (1); - zebraIndexLockMsg ("r"); - } - else - { - bf_cache (0); - zebraIndexLockMsg ("w"); - } - zebraIndexWait (0); - + zebraIndexLock (rGroupDef.bfs, 0, rval); + if (rGroupDef.flagRw) + { + if (rval && *rval && !disableCommit) + { + bf_cache (rGroupDef.bfs, rval); + zebraIndexLockMsg ("r"); + } + else + { + bf_cache (rGroupDef.bfs, 0); + zebraIndexLockMsg ("w"); + } + 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') + switch (cmd) { - logf (LOG_LOG, "Deleting %s", rGroup.path); - repositoryDelete (&rGroup); - } - else if (cmd == 's') - { - logf (LOG_LOG, "Dumping %s", rGroup.path); + case 'u': + if (!key_open (&rGroup, mem_max)) + { + logf (LOG_LOG, "updating %s", rGroup.path); + repositoryUpdate (&rGroup); + nsections = key_close (&rGroup); + } + break; + case 'U': + if (!key_open (&rGroup, mem_max)) + { + logf (LOG_LOG, "updating (pass 1) %s", rGroup.path); + repositoryUpdate (&rGroup); + key_close (&rGroup); + } + nsections = 0; + break; + case 'd': + if (!key_open (&rGroup,mem_max)) + { + logf (LOG_LOG, "deleting %s", rGroup.path); + repositoryDelete (&rGroup); + nsections = key_close (&rGroup); + } + break; + case 's': + logf (LOG_LOG, "dumping %s", rGroup.path); repositoryShow (&rGroup); + nsections = 0; + break; + case 'm': + nsections = -1; + break; + default: + nsections = 0; } - cmd = 0; - nsections = key_close (); if (nsections) { - logf (LOG_LOG, "Merging with index"); - key_input (FNAME_WORD_DICT, FNAME_WORD_ISAM, nsections, - 60); + logf (LOG_LOG, "merging with index"); + key_input (rGroup.bfs, nsections, 60, common_resource); +#ifndef WIN32 sync (); +#endif } + log_event_end (NULL, NULL); } } - else if (ret == 'v') + else if (ret == 'V') { - log_init (log_mask_str(arg), prog, NULL); +#if ZMBOL + fprintf (stderr, "Z'mbol %s %s\n", ZEBRAVER, ZEBRADATE); +#else + fprintf (stderr, "Zebra %s %s\n", ZEBRAVER, ZEBRADATE); +#endif + fprintf (stderr, " (C) 1994-2001, Index Data ApS\n"); +#ifdef WIN32 +#ifdef _DEBUG + fprintf (stderr, " WIN32 Debug\n"); +#else + fprintf (stderr, " WIN32 Release\n"); +#endif +#endif +#if HAVE_BZLIB_H + fprintf (stderr, "libbzip2\n" + " (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); - } else if (ret == 'd') - { rGroupDef.databaseName = arg; - } + else if (ret == 's') + rGroupDef.flagRw = 0; else if (ret == 'g') - { rGroupDef.groupName = arg; - } + else if (ret == 'f') + rGroupDef.fileVerboseLimit = atoi(arg); else if (ret == 'c') configName = arg; else if (ret == 't') rGroupDef.recordType = arg; else if (ret == 'n') disableCommit = 1; +#if ZEBRASDR + else if (ret == 'S') + rGroupDef.useSDR = 1; +#endif else - { - logf (LOG_FATAL, "Unknown option '-%s'", arg); - exit (1); - } + logf (LOG_WARN, "unknown option '-%s'", arg); } + recTypes_destroy (rGroupDef.recTypes); if (common_resource) + { zebraIndexUnlock (); + bfs_destroy (rGroupDef.bfs); + } + data1_destroy (rGroupDef.dh); exit (0); + return 0; }