-/*
- * Copyright (C) 1994-1996, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: trav.c,v $
- * Revision 1.28 1996-11-01 08:58:44 adam
- * Interface to isamc system now includes update and delete.
- *
- * Revision 1.27 1996/10/29 14:06:56 adam
- * Include zebrautl.h instead of alexutil.h.
- *
- * Revision 1.26 1996/06/04 10:19:01 adam
- * Minor changes - removed include of ctype.h.
- *
- * Revision 1.25 1996/05/01 13:46:37 adam
- * First work on multiple records in one file.
- * New option, -offset, to the "unread" command in the filter module.
- *
- * Revision 1.24 1996/04/26 10:00:23 adam
- * Added option -V to zebraidx to display version information.
- * Removed stupid warnings from file update.
- *
- * Revision 1.23 1996/04/12 07:02:25 adam
- * File update of single files.
- *
- * Revision 1.22 1996/04/09 06:50:50 adam
- * Bug fix: bad reference in function fileUpdateR.
- *
- * Revision 1.21 1996/03/22 15:34:18 quinn
- * Fixed bad reference
- *
- * Revision 1.20 1996/03/21 14:50:10 adam
- * File update uses modify-time instead of change-time.
- *
- * Revision 1.19 1996/03/20 16:16:55 quinn
- * Added diagnostic output
- *
- * Revision 1.18 1996/03/19 12:43:27 adam
- * Bug fix: File update traversal didn't handle trailing slashes correctly.
- * Bug fix: Update of sub directory groups wasn't handled correctly.
- *
- * Revision 1.17 1996/02/12 18:45:17 adam
- * Changed naming of some functions.
- *
- * Revision 1.16 1996/02/05 12:30:02 adam
- * Logging reduced a bit.
- * The remaining running time is estimated during register merge.
- *
- * Revision 1.15 1995/12/07 17:38:48 adam
- * Work locking mechanisms for concurrent updates/commit.
- *
- * Revision 1.14 1995/12/06 12:41:26 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 1.13 1995/11/28 09:09:46 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 1.12 1995/11/24 11:31:37 adam
- * Commands add & del read filenames from stdin if source directory is
- * empty.
- * Match criteria supports 'constant' strings.
- *
- * Revision 1.11 1995/11/22 17:19:19 adam
- * Record management uses the bfile system.
- *
- * Revision 1.10 1995/11/21 15:01:16 adam
- * New general match criteria implemented.
- * New feature: document groups.
- *
- * Revision 1.9 1995/11/21 09:20:32 adam
- * Yet more work on record match.
- *
- * Revision 1.8 1995/11/20 16:59:46 adam
- * New update method: the 'old' keys are saved for each records.
- *
- * Revision 1.7 1995/11/20 11:56:28 adam
- * Work on new traversal.
- *
- * Revision 1.6 1995/11/17 15:54:42 adam
- * Started work on virtual directory structure.
- *
- * Revision 1.5 1995/10/17 18:02:09 adam
- * New feature: databases. Implemented as prefix to words in dictionary.
- *
- * Revision 1.4 1995/09/28 09:19:46 adam
- * xfree/xmalloc used everywhere.
- * Extract/retrieve method seems to work for text records.
- *
- * Revision 1.3 1995/09/06 16:11:18 adam
- * Option: only one word key per file.
- *
- * Revision 1.2 1995/09/04 12:33:43 adam
- * Various cleanup. YAZ util used instead.
- *
- * Revision 1.1 1995/09/01 14:06:36 adam
- * Split of work into more files.
- *
- */
+/* $Id: trav.c,v 1.49 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 <stdio.h>
#include <assert.h>
-#include <unistd.h>
-#include <dirent.h>
-#include <sys/stat.h>
#include <sys/types.h>
+#ifdef WIN32
+#include <io.h>
+#define S_ISREG(x) (x & _S_IFREG)
+#define S_ISDIR(x) (x & _S_IFDIR)
+#endif
+#if HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+#include <direntz.h>
#include <fcntl.h>
#include <time.h>
return memcmp (a, b, len);
}
-static void repositoryExtractR (int deleteFlag, char *rep,
- struct recordGroup *rGroup)
+static void repositoryExtractR (ZebraHandle zh, int deleteFlag, char *rep,
+ int level)
{
struct dir_entry *e;
int i;
size_t rep_len = strlen (rep);
- e = dir_open (rep);
+ e = dir_open (rep, zh->path_reg, zh->m_follow_links);
if (!e)
return;
- logf (LOG_LOG, "Dir: %s", rep);
+ yaz_log (YLOG_LOG, "dir %s", rep);
if (rep[rep_len-1] != '/')
rep[rep_len] = '/';
else
--rep_len;
+
for (i=0; e[i].name; i++)
{
+ char *ecp;
strcpy (rep +rep_len+1, e[i].name);
+ if ((ecp = strrchr (e[i].name, '/')))
+ *ecp = '\0';
+
switch (e[i].kind)
{
case dirs_file:
- fileExtract (NULL, rep, rGroup, deleteFlag);
+ fileExtract (zh, NULL, rep, deleteFlag);
break;
case dirs_dir:
- repositoryExtractR (deleteFlag, rep, rGroup);
+ repositoryExtractR (zh, deleteFlag, rep, level+1);
break;
}
}
}
-static void fileDeleteR (struct dirs_info *di, struct dirs_entry *dst,
- const char *base, char *src,
- struct recordGroup *rGroup)
+static void fileDeleteR (ZebraHandle zh,
+ struct dirs_info *di, struct dirs_entry *dst,
+ const char *base, char *src)
{
char tmppath[1024];
size_t src_len = strlen (src);
{
case dirs_file:
sprintf (tmppath, "%s%s", base, dst->path);
- fileExtract (&dst->sysno, tmppath, rGroup, 1);
+ fileExtract (zh, &dst->sysno, tmppath, 1);
strcpy (tmppath, dst->path);
dst = dirs_read (di);
}
}
-static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst,
- const char *base, char *src,
- struct recordGroup *rGroup)
+static void fileUpdateR (ZebraHandle zh,
+ struct dirs_info *di, struct dirs_entry *dst,
+ const char *base, char *src,
+ int level)
{
struct dir_entry *e_src;
int i_src = 0;
size_t src_len = strlen (src);
sprintf (tmppath, "%s%s", base, src);
- e_src = dir_open (tmppath);
- logf (LOG_LOG, "Dir: %s", tmppath);
+ e_src = dir_open (tmppath, zh->path_reg, zh->m_follow_links);
+ yaz_log (YLOG_LOG, "dir %s", tmppath);
#if 0
if (!dst || repComp (dst->path, src, src_len))
else if (!e_src)
{
strcpy (src, dst->path);
- fileDeleteR (di, dst, base, src, rGroup);
+ fileDeleteR (zh, di, dst, base, src);
return;
}
else
{
if (e_src[i_src].name)
{
- logf (LOG_DEBUG, "dst=%s src=%s", dst->path + src_len,
+ yaz_log (YLOG_DEBUG, "dst=%s src=%s", dst->path + src_len,
e_src[i_src].name);
sd = strcmp (dst->path + src_len, e_src[i_src].name);
}
sd = 1;
else
break;
- logf (LOG_DEBUG, "trav sd=%d", sd);
+ yaz_log (YLOG_DEBUG, "trav sd=%d", sd);
+
if (sd == 0)
{
strcpy (src + src_len, e_src[i_src].name);
case dirs_file:
if (e_src[i_src].mtime > dst->mtime)
{
- if (fileExtract (&dst->sysno, tmppath, rGroup, 0))
+ if (fileExtract (zh, &dst->sysno, tmppath, 0))
{
dirs_add (di, src, dst->sysno, e_src[i_src].mtime);
}
- logf (LOG_DEBUG, "old: %s", ctime (&dst->mtime));
- logf (LOG_DEBUG, "new: %s", ctime (&e_src[i_src].mtime));
+ yaz_log (YLOG_DEBUG, "old: %s", ctime (&dst->mtime));
+ yaz_log (YLOG_DEBUG, "new: %s", ctime (&e_src[i_src].mtime));
}
dst = dirs_read (di);
break;
case dirs_dir:
- fileUpdateR (di, dst, base, src, rGroup);
+ fileUpdateR (zh, di, dst, base, src, level+1);
dst = dirs_last (di);
- logf (LOG_DEBUG, "last is %s", dst ? dst->path : "null");
+ yaz_log (YLOG_DEBUG, "last is %s", dst ? dst->path : "null");
break;
default:
dst = dirs_read (di);
switch (e_src[i_src].kind)
{
case dirs_file:
- if (fileExtract (&sysno, tmppath, rGroup, 0))
+ if (fileExtract (zh, &sysno, tmppath, 0))
dirs_add (di, src, sysno, e_src[i_src].mtime);
break;
case dirs_dir:
- fileUpdateR (di, dst, base, src, rGroup);
+ fileUpdateR (zh, di, dst, base, src, level+1);
if (dst)
dst = dirs_last (di);
break;
switch (dst->kind)
{
case dirs_file:
- fileExtract (&dst->sysno, tmppath, rGroup, 1);
+ fileExtract (zh, &dst->sysno, tmppath, 1);
dirs_del (di, dst->path);
dst = dirs_read (di);
break;
case dirs_dir:
- fileDeleteR (di, dst, base, src, rGroup);
+ fileDeleteR (zh, di, dst, base, src);
dst = dirs_last (di);
}
}
dir_free (&e_src);
}
-static void groupRes (struct recordGroup *rGroup)
-{
- char resStr[256];
- char gPrefix[256];
-
- if (!rGroup->groupName || !*rGroup->groupName)
- *gPrefix = '\0';
- else
- sprintf (gPrefix, "%s.", rGroup->groupName);
-
- sprintf (resStr, "%srecordId", gPrefix);
- rGroup->recordId = res_get (common_resource, resStr);
-}
-
-void repositoryShow (struct recordGroup *rGroup)
+void repositoryShow (ZebraHandle zh, const char *path)
{
char src[1024];
int src_len;
struct dirs_entry *dst;
Dict dict;
struct dirs_info *di;
-
- if (!(dict = dict_open (FMATCH_DICT, 50, 1)))
+
+ if (!(dict = dict_open_res (zh->reg->bfs, FMATCH_DICT, 50, 0, 0, zh->res)))
{
- logf (LOG_FATAL, "dict_open fail of %s", FMATCH_DICT);
- exit (1);
+ yaz_log (YLOG_FATAL, "dict_open fail of %s", FMATCH_DICT);
+ return;
}
- assert (rGroup->path);
- strcpy (src, rGroup->path);
+ strncpy(src, path, sizeof(src)-1);
+ src[sizeof(src)-1]='\0';
src_len = strlen (src);
if (src_len && src[src_len-1] != '/')
src[++src_len] = '\0';
}
- di = dirs_open (dict, src);
+ di = dirs_open (dict, src, zh->m_flag_rw);
while ( (dst = dirs_read (di)) )
- logf (LOG_LOG, "%s", dst->path);
+ yaz_log (YLOG_LOG, "%s", dst->path);
dirs_free (&di);
dict_close (dict);
}
-static void fileUpdate (Dict dict, struct recordGroup *rGroup,
- const char *path)
+static void fileUpdate (ZebraHandle zh, Dict dict, const char *path)
{
struct dirs_info *di;
struct stat sbuf;
char src[1024];
char dst[1024];
- int src_len;
+ int src_len, ret;
assert (path);
+
+ if (zh->path_reg && !yaz_is_abspath(path))
+ {
+ strcpy (src, zh->path_reg);
+ strcat (src, "/");
+ }
+ else
+ *src = '\0';
+ strcat (src, path);
+ ret = zebra_file_stat (src, &sbuf, zh->m_follow_links);
+
strcpy (src, path);
src_len = strlen (src);
- stat (src, &sbuf);
- if (S_ISREG(sbuf.st_mode))
+ if (ret == -1)
+ {
+ yaz_log (YLOG_WARN|YLOG_ERRNO, "Cannot access path %s", src);
+ }
+ else if (S_ISREG(sbuf.st_mode))
{
struct dirs_entry *e_dst;
- di = dirs_fopen (dict, src);
+ di = dirs_fopen (dict, src, zh->m_flag_rw);
e_dst = dirs_read (di);
if (e_dst)
{
if (sbuf.st_mtime > e_dst->mtime)
- if (fileExtract (&e_dst->sysno, src, rGroup, 0))
+ if (fileExtract (zh, &e_dst->sysno, src, 0))
dirs_add (di, src, e_dst->sysno, sbuf.st_mtime);
}
else
{
SYSNO sysno = 0;
- if (fileExtract (&sysno, src, rGroup, 0))
+ if (fileExtract (zh, &sysno, src, 0))
dirs_add (di, src, sysno, sbuf.st_mtime);
}
dirs_free (&di);
src[src_len] = '/';
src[++src_len] = '\0';
}
- di = dirs_open (dict, src);
+ di = dirs_open (dict, src, zh->m_flag_rw);
*dst = '\0';
- fileUpdateR (di, dirs_read (di), src, dst, rGroup);
+ fileUpdateR (zh, di, dirs_read (di), src, dst, 0);
dirs_free (&di);
}
else
{
- logf (LOG_WARN, "Cannot handle file %s", src);
+ yaz_log (YLOG_WARN, "Skipping path %s", src);
}
}
-
-static void repositoryExtract (int deleteFlag, struct recordGroup *rGroup,
- const char *path)
+static void repositoryExtract (ZebraHandle zh,
+ int deleteFlag, const char *path)
{
struct stat sbuf;
char src[1024];
+ int ret;
assert (path);
+
+ if (zh->path_reg && !yaz_is_abspath(path))
+ {
+ strcpy (src, zh->path_reg);
+ strcat (src, "/");
+ }
+ else
+ *src = '\0';
+ strcat (src, path);
+ ret = zebra_file_stat (src, &sbuf, zh->m_follow_links);
+
strcpy (src, path);
- stat (src, &sbuf);
- if (S_ISREG(sbuf.st_mode))
- fileExtract (NULL, src, rGroup, deleteFlag);
+ if (ret == -1)
+ yaz_log (YLOG_WARN|YLOG_ERRNO, "Cannot access path %s", src);
+ else if (S_ISREG(sbuf.st_mode))
+ fileExtract (zh, NULL, src, deleteFlag);
else if (S_ISDIR(sbuf.st_mode))
- repositoryExtractR (deleteFlag, src, rGroup);
+ repositoryExtractR (zh, deleteFlag, src, 0);
else
- logf (LOG_WARN, "Cannot handle file %s", src);
+ yaz_log (YLOG_WARN, "Skipping path %s", src);
}
-static void repositoryExtractG (int deleteFlag, struct recordGroup *rGroup)
+static void repositoryExtractG (ZebraHandle zh, const char *path,
+ int deleteFlag)
{
- if (*rGroup->path == '\0' || !strcmp(rGroup->path, "-"))
+ if (!strcmp(path, "") || !strcmp(path, "-"))
{
char src[1024];
-
- while (scanf ("%s", src) == 1)
- repositoryExtract (deleteFlag, rGroup, src);
+
+ while (scanf ("%1020s", src) == 1)
+ repositoryExtract (zh, deleteFlag, src);
}
else
- repositoryExtract (deleteFlag, rGroup, rGroup->path);
+ repositoryExtract (zh, deleteFlag, path);
}
-void repositoryUpdate (struct recordGroup *rGroup)
+#if 0
+static int dump_file_dict_func(char *name, const char *info, int pos,
+ void *client)
{
- groupRes (rGroup);
- assert (rGroup->path);
- if (rGroup->recordId && !strcmp (rGroup->recordId, "file"))
+ yaz_log(YLOG_LOG, "%s", name);
+ return 0;
+}
+static void dump_file_dict(Dict dict)
+{
+ int before = 10;
+ int after = 1000;
+ char term[1000];
+
+ strcpy(term, "0");
+ dict_scan (dict, term, &before, &after, 0, dump_file_dict_func);
+}
+#endif
+
+void repositoryUpdate (ZebraHandle zh, const char *path)
+{
+ assert (path);
+ if (zh->m_record_id && !strcmp (zh->m_record_id, "file"))
{
Dict dict;
- if (!(dict = dict_open (FMATCH_DICT, 50, 1)))
+ if (!(dict = dict_open_res (zh->reg->bfs, FMATCH_DICT, 50,
+ zh->m_flag_rw, 0, zh->res)))
{
- logf (LOG_FATAL, "dict_open fail of %s", FMATCH_DICT);
- exit (1);
+ yaz_log (YLOG_FATAL, "dict_open fail of %s", FMATCH_DICT);
+ return ;
}
- if (*rGroup->path == '\0' || !strcmp(rGroup->path, "-"))
+ if (!strcmp(path, "") || !strcmp(path, "-"))
{
char src[1024];
while (scanf ("%s", src) == 1)
- fileUpdate (dict, rGroup, src);
+ fileUpdate (zh, dict, src);
}
else
- fileUpdate (dict, rGroup, rGroup->path);
+ fileUpdate (zh, dict, path);
+#if 0
+ dump_file_dict(dict);
+#endif
dict_close (dict);
+
}
else
- repositoryExtractG (0, rGroup);
+ repositoryExtractG (zh, path, 0);
}
-void repositoryDelete (struct recordGroup *rGroup)
+void repositoryDelete (ZebraHandle zh, const char *path)
{
- repositoryExtractG (1, rGroup);
+ assert (path);
+ repositoryExtractG (zh, path, 1);
}