-/*
- * Copyright (C) 1994-1995, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: trav.c,v $
- * 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 1.41 2002-09-03 11:44:54 adam Exp $
+ Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002
+ 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)
+#else
+#include <unistd.h>
+#endif
+#include <direntz.h>
#include <fcntl.h>
-#include <ctype.h>
+#include <time.h>
-#include <alexutil.h>
#include "index.h"
-#if 0
-static void repository_extract_r (int cmd, char *rep, char *databaseName)
+static int repComp (const char *a, const char *b, size_t len)
+{
+ if (!len)
+ return 0;
+ return memcmp (a, b, len);
+}
+
+static void repositoryExtractR (ZebraHandle zh, int deleteFlag, char *rep,
+ struct recordGroup *rGroup,
+ int level)
{
struct dir_entry *e;
int i;
- struct stat fs;
size_t rep_len = strlen (rep);
- e = dir_open (rep);
+ e = dir_open (rep, zh->path_reg, rGroup->followLinks);
if (!e)
return;
+ logf (LOG_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);
- stat (rep, &fs);
- switch (fs.st_mode & S_IFMT)
+ if ((ecp = strrchr (e[i].name, '/')))
+ *ecp = '\0';
+ if (level == 0 && rGroup->databaseNamePath)
+ rGroup->databaseName = e[i].name;
+
+ switch (e[i].kind)
{
- case S_IFREG:
- file_extract (cmd, rep, rep, databaseName);
+ case dirs_file:
+ fileExtract (zh, NULL, rep, rGroup, deleteFlag);
break;
- case S_IFDIR:
- repository_extract_r (cmd, rep, databaseName);
+ case dirs_dir:
+ repositoryExtractR (zh, deleteFlag, rep, rGroup, level+1);
break;
}
}
dir_free (&e);
-}
-
-void copy_file (const char *dst, const char *src)
-{
- int d_fd = open (dst, O_WRONLY|O_CREAT, 0666);
- int s_fd = open (src, O_RDONLY);
- char *buf;
- size_t i, r, w;
-
- if (d_fd == -1)
- {
- logf (LOG_FATAL|LOG_ERRNO, "Cannot create %s", dst);
- exit (1);
- }
- if (s_fd == -1)
- {
- logf (LOG_FATAL|LOG_ERRNO, "Cannot open %s", src);
- exit (1);
- }
- buf = xmalloc (4096);
- while ((r=read (s_fd, buf, 4096))>0)
- for (w = 0; w < r; w += i)
- {
- i = write (d_fd, buf + w, r - w);
- if (i == -1)
- {
- logf (LOG_FATAL|LOG_ERRNO, "write");
- exit (1);
- }
- }
- if (r)
- {
- logf (LOG_FATAL|LOG_ERRNO, "read");
- exit (1);
- }
- xfree (buf);
- close (d_fd);
- close (s_fd);
-}
-
-void del_file (const char *dst)
-{
- unlink (dst);
-}
-void del_dir (const char *dst)
-{
- logf (LOG_DEBUG, "rmdir of %s", dst);
- if (rmdir (dst) == -1)
- logf (LOG_ERRNO|LOG_WARN, "rmdir");
-}
-
-void repository_update_r (int cmd, char *dst, char *src, char *databaseName);
-
-void repository_add_tree (int cmd, char *dst, char *src, char *databaseName)
-{
- mkdir (dst, 0755);
- repository_update_r (cmd, dst, src, databaseName);
}
-void repository_del_tree (int cmd, char *dst, char *src, char *databaseName)
+static void fileDeleteR (ZebraHandle zh,
+ struct dirs_info *di, struct dirs_entry *dst,
+ const char *base, char *src,
+ struct recordGroup *rGroup)
{
- size_t dst_len = strlen (dst);
+ char tmppath[1024];
size_t src_len = strlen (src);
- struct dir_entry *e_dst;
- int i_dst = 0;
- struct stat fs_dst;
- e_dst = dir_open (dst);
-
- dir_sort (e_dst);
-
- if (src[src_len-1] != '/')
- src[src_len] = '/';
- else
- --src_len;
- if (dst[dst_len-1] != '/')
- dst[dst_len] = '/';
- else
- --dst_len;
- while (e_dst[i_dst].name)
+ while (dst && !repComp (dst->path, src, src_len+1))
{
- strcpy (dst +dst_len+1, e_dst[i_dst].name);
- strcpy (src +src_len+1, e_dst[i_dst].name);
-
- stat (dst, &fs_dst);
- switch (fs_dst.st_mode & S_IFMT)
+ switch (dst->kind)
{
- case S_IFREG:
- file_extract ('d', dst, dst, databaseName);
- del_file (dst);
+ case dirs_file:
+ sprintf (tmppath, "%s%s", base, dst->path);
+ fileExtract (zh, &dst->sysno, tmppath, rGroup, 1);
+
+ strcpy (tmppath, dst->path);
+ dst = dirs_read (di);
+ dirs_del (di, tmppath);
break;
- case S_IFDIR:
- repository_del_tree (cmd, dst, src, databaseName);
+ case dirs_dir:
+ strcpy (tmppath, dst->path);
+ dst = dirs_read (di);
+ dirs_rmdir (di, tmppath);
break;
+ default:
+ dst = dirs_read (di);
}
- i_dst++;
- }
- dir_free (&e_dst);
- if (dst_len > 0)
- {
- dst[dst_len] = '\0';
- del_dir (dst);
}
}
-void repository_update_r (int cmd, char *dst, char *src, char *databaseName)
-{
- struct dir_entry *e_dst, *e_src;
- int i_dst = 0, i_src = 0;
- struct stat fs_dst, fs_src;
- size_t dst_len = strlen (dst);
- size_t src_len = strlen (src);
-
- e_dst = dir_open (dst);
- e_src = dir_open (src);
-
- if (!e_dst && !e_src)
- return;
- if (!e_dst)
- {
- dir_free (&e_src);
- repository_add_tree (cmd, dst, src, databaseName);
- return;
- }
- else if (!e_src)
- {
- dir_free (&e_dst);
- repository_del_tree (cmd, dst, src, databaseName);
- return;
- }
-
- dir_sort (e_src);
- dir_sort (e_dst);
-
- if (src[src_len-1] != '/')
- src[src_len] = '/';
- else
- --src_len;
- if (dst[dst_len-1] != '/')
- dst[dst_len] = '/';
- else
- --dst_len;
- while (e_dst[i_dst].name || e_src[i_src].name)
- {
- int sd;
-
- if (e_dst[i_dst].name && e_src[i_src].name)
- sd = strcmp (e_dst[i_dst].name, e_src[i_src].name);
- else if (e_src[i_src].name)
- sd = 1;
- else
- sd = -1;
-
- if (sd == 0)
- {
- strcpy (dst +dst_len+1, e_dst[i_dst].name);
- strcpy (src +src_len+1, e_src[i_src].name);
-
- /* check type, date, length */
-
- stat (dst, &fs_dst);
- stat (src, &fs_src);
-
- switch (fs_dst.st_mode & S_IFMT)
- {
- case S_IFREG:
- if (fs_src.st_ctime > fs_dst.st_ctime)
- {
- file_extract ('d', dst, dst, databaseName);
- file_extract ('a', src, dst, databaseName);
- copy_file (dst, src);
- }
- break;
- case S_IFDIR:
- repository_update_r (cmd, dst, src, databaseName);
- break;
- }
- i_src++;
- i_dst++;
- }
- else if (sd > 0)
- {
- strcpy (dst +dst_len+1, e_src[i_src].name);
- strcpy (src +src_len+1, e_src[i_src].name);
-
- stat (src, &fs_src);
- switch (fs_src.st_mode & S_IFMT)
- {
- case S_IFREG:
- file_extract ('a', src, dst, databaseName);
- copy_file (dst, src);
- break;
- case S_IFDIR:
- repository_add_tree (cmd, dst, src, databaseName);
- break;
- }
- i_src++;
- }
- else
- {
- strcpy (dst +dst_len+1, e_dst[i_dst].name);
- strcpy (src +src_len+1, e_dst[i_dst].name);
-
- stat (dst, &fs_dst);
- switch (fs_dst.st_mode & S_IFMT)
- {
- case S_IFREG:
- file_extract ('d', dst, dst, databaseName);
- del_file (dst);
- break;
- case S_IFDIR:
- repository_del_tree (cmd, dst, src, databaseName);
- break;
- }
- i_dst++;
- }
- }
- dir_free (&e_dst);
- dir_free (&e_src);
-}
-
-void repository (int cmd, const char *rep, const char *base_path,
- char *databaseName)
-{
- char rep_tmp1[2048];
- char rep_tmp2[2048];
-
- strcpy (rep_tmp1, rep);
- if (base_path)
- {
- strcpy (rep_tmp2, base_path);
- repository_update_r (cmd, rep_tmp2, rep_tmp1, databaseName);
- }
- else
- repository_extract_r (cmd, rep_tmp1, databaseName);
-}
-#endif
-
-static int repComp (const char *a, const char *b, size_t len)
-{
- if (!len)
- return 0;
- return memcmp (a, b, len);
-}
-
-static void repositoryUpdateR (struct dirs_info *di, struct dirs_entry *dst,
- const char *base, char *src, char *databaseName)
+static void fileUpdateR (ZebraHandle zh,
+ struct dirs_info *di, struct dirs_entry *dst,
+ const char *base, char *src,
+ struct recordGroup *rGroup,
+ int level)
{
struct dir_entry *e_src;
int i_src = 0;
- static char tmppath[256];
+ static char tmppath[1024];
size_t src_len = strlen (src);
sprintf (tmppath, "%s%s", base, src);
- e_src = dir_open (tmppath);
+ e_src = dir_open (tmppath, zh->path_reg, rGroup->followLinks);
+ logf (LOG_LOG, "dir %s", tmppath);
-#if 1
+#if 0
if (!dst || repComp (dst->path, src, src_len))
#else
if (!dst || strcmp (dst->path, src))
{
if (!e_src)
return;
- if (src_len && src[src_len-1] == '/')
- --src_len;
- else
+
+ if (src_len && src[src_len-1] != '/')
+ {
src[src_len] = '/';
- src[src_len+1] = '\0';
+ src[++src_len] = '\0';
+ }
dirs_mkdir (di, src, 0);
- dst = NULL;
+ if (dst && repComp (dst->path, src, src_len))
+ dst = NULL;
}
else if (!e_src)
{
- /* delete tree dst */
+ strcpy (src, dst->path);
+ fileDeleteR (zh, di, dst, base, src, rGroup);
return;
}
else
{
- if (src_len && src[src_len-1] == '/')
- --src_len;
- else
+ if (src_len && src[src_len-1] != '/')
+ {
src[src_len] = '/';
- src[src_len+1] = '\0';
+ src[++src_len] = '\0';
+ }
dst = dirs_read (di);
}
dir_sort (e_src);
{
int sd;
- if (dst && !repComp (dst->path, src, src_len+1))
+ if (dst && !repComp (dst->path, src, src_len))
{
if (e_src[i_src].name)
{
- logf (LOG_DEBUG, "dst=%s src=%s", dst->path + src_len+1,
+ logf (LOG_DEBUG, "dst=%s src=%s", dst->path + src_len,
e_src[i_src].name);
- sd = strcmp (dst->path + src_len+1, e_src[i_src].name);
+ sd = strcmp (dst->path + src_len, e_src[i_src].name);
}
else
sd = -1;
else
break;
logf (LOG_DEBUG, "trav sd=%d", sd);
+
+ if (level == 0 && rGroup->databaseNamePath)
+ rGroup->databaseName = e_src[i_src].name;
if (sd == 0)
{
- strcpy (src + src_len+1, e_src[i_src].name);
+ strcpy (src + src_len, e_src[i_src].name);
sprintf (tmppath, "%s%s", base, src);
switch (e_src[i_src].kind)
{
case dirs_file:
- if (e_src[i_src].ctime > dst->ctime)
+ if (e_src[i_src].mtime > dst->mtime)
{
- if (fileExtract (&dst->sysno, tmppath, databaseName, 0))
- dirs_add (di, src, dst->sysno, e_src[i_src].ctime);
+ if (fileExtract (zh, &dst->sysno, tmppath, rGroup, 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));
}
dst = dirs_read (di);
break;
case dirs_dir:
- repositoryUpdateR (di, dst, base, src, databaseName);
+ fileUpdateR (zh, di, dst, base, src, rGroup, level+1);
dst = dirs_last (di);
logf (LOG_DEBUG, "last is %s", dst ? dst->path : "null");
break;
else if (sd > 0)
{
SYSNO sysno = 0;
- strcpy (src + src_len+1, e_src[i_src].name);
+ strcpy (src + src_len, e_src[i_src].name);
sprintf (tmppath, "%s%s", base, src);
switch (e_src[i_src].kind)
{
case dirs_file:
- if (fileExtract (&sysno, tmppath, databaseName, 0))
- dirs_add (di, src, sysno, e_src[i_src].ctime);
+ if (fileExtract (zh, &sysno, tmppath, rGroup, 0))
+ dirs_add (di, src, sysno, e_src[i_src].mtime);
break;
case dirs_dir:
- repositoryUpdateR (di, dst, base, src, databaseName);
+ fileUpdateR (zh, di, dst, base, src, rGroup, level+1);
if (dst)
dst = dirs_last (di);
break;
}
else /* sd < 0 */
{
- assert (0);
+ strcpy (src, dst->path);
+ sprintf (tmppath, "%s%s", base, dst->path);
+
+ switch (dst->kind)
+ {
+ case dirs_file:
+ fileExtract (zh, &dst->sysno, tmppath, rGroup, 1);
+ dirs_del (di, dst->path);
+ dst = dirs_read (di);
+ break;
+ case dirs_dir:
+ fileDeleteR (zh, di, dst, base, src, rGroup);
+ dst = dirs_last (di);
+ }
}
}
dir_free (&e_src);
}
-void repositoryUpdate (const char *path, char *databaseName)
+static void groupRes (ZebraHandle zh, struct recordGroup *rGroup)
{
- struct dirs_info *di;
- char src[256];
- Dict dict;
+ char resStr[256];
+ char gPrefix[256];
- dict = dict_open ("repdict", 40, 1);
+ if (!rGroup->groupName || !*rGroup->groupName)
+ *gPrefix = '\0';
+ else
+ sprintf (gPrefix, "%s.", rGroup->groupName);
+
+ sprintf (resStr, "%srecordId", gPrefix);
+ rGroup->recordId = res_get (zh->res, resStr);
+ sprintf (resStr, "%sdatabasePath", gPrefix);
+ rGroup->databaseNamePath =
+ atoi (res_get_def (zh->res, resStr, "0"));
+
+ rGroup->databaseNamePath =
+ atoi (res_get_def (zh->res, resStr, "0"));
+
+ if (rGroup->followLinks == -1)
+ {
+ sprintf (resStr, "%sfollowLinks", gPrefix);
+ rGroup->followLinks =
+ atoi (res_get_def (zh->res, resStr, "1"));
+ }
+}
+
+void repositoryShow (ZebraHandle zh)
+
+{
+ struct recordGroup *rGroup = &zh->rGroup;
+ char src[1024];
+ int src_len;
+ struct dirs_entry *dst;
+ Dict dict;
+ struct dirs_info *di;
+
+ if (!(dict = dict_open (zh->reg->bfs, FMATCH_DICT, 50, 0, 0)))
+ {
+ logf (LOG_FATAL, "dict_open fail of %s", FMATCH_DICT);
+ return;
+ }
- di = dirs_open (dict, path);
- strcpy (src, "");
- repositoryUpdateR (di, dirs_read (di), path, src, databaseName);
+ assert (rGroup->path);
+ strcpy (src, rGroup->path);
+ src_len = strlen (src);
+
+ if (src_len && src[src_len-1] != '/')
+ {
+ src[src_len] = '/';
+ src[++src_len] = '\0';
+ }
+
+ di = dirs_open (dict, src, rGroup->flagRw);
+
+ while ( (dst = dirs_read (di)) )
+ logf (LOG_LOG, "%s", dst->path);
dirs_free (&di);
-
dict_close (dict);
}
+
+static void fileUpdate (ZebraHandle zh,
+ Dict dict, struct recordGroup *rGroup,
+ const char *path)
+{
+ struct dirs_info *di;
+ struct stat sbuf;
+ char src[1024];
+ char dst[1024];
+ int src_len;
+
+ assert (path);
+
+ if (zh->path_reg && !yaz_is_abspath(path))
+ {
+ strcpy (src, zh->path_reg);
+ strcat (src, "/");
+ }
+ else
+ *src = '\0';
+ strcat (src, path);
+ zebra_file_stat (src, &sbuf, rGroup->followLinks);
+
+ strcpy (src, path);
+ src_len = strlen (src);
+
+ if (S_ISREG(sbuf.st_mode))
+ {
+ struct dirs_entry *e_dst;
+ di = dirs_fopen (dict, src);
+
+ e_dst = dirs_read (di);
+ if (e_dst)
+ {
+ if (sbuf.st_mtime > e_dst->mtime)
+ if (fileExtract (zh, &e_dst->sysno, src, rGroup, 0))
+ dirs_add (di, src, e_dst->sysno, sbuf.st_mtime);
+ }
+ else
+ {
+ SYSNO sysno = 0;
+ if (fileExtract (zh, &sysno, src, rGroup, 0))
+ dirs_add (di, src, sysno, sbuf.st_mtime);
+ }
+ dirs_free (&di);
+ }
+ else if (S_ISDIR(sbuf.st_mode))
+ {
+ if (src_len && src[src_len-1] != '/')
+ {
+ src[src_len] = '/';
+ src[++src_len] = '\0';
+ }
+ di = dirs_open (dict, src, rGroup->flagRw);
+ *dst = '\0';
+ fileUpdateR (zh, di, dirs_read (di), src, dst, rGroup, 0);
+ dirs_free (&di);
+ }
+ else
+ {
+ logf (LOG_WARN, "Ignoring path %s", src);
+ }
+}
+
+
+static void repositoryExtract (ZebraHandle zh,
+ int deleteFlag, struct recordGroup *rGroup,
+ const char *path)
+{
+ struct stat sbuf;
+ char src[1024];
+
+ assert (path);
+
+ if (zh->path_reg && !yaz_is_abspath(path))
+ {
+ strcpy (src, zh->path_reg);
+ strcat (src, "/");
+ }
+ else
+ *src = '\0';
+ strcat (src, path);
+ zebra_file_stat (src, &sbuf, rGroup->followLinks);
+
+ strcpy (src, path);
+
+ if (S_ISREG(sbuf.st_mode))
+ fileExtract (zh, NULL, src, rGroup, deleteFlag);
+ else if (S_ISDIR(sbuf.st_mode))
+ repositoryExtractR (zh, deleteFlag, src, rGroup, 0);
+ else
+ logf (LOG_WARN, "Ignoring path %s", src);
+}
+
+static void repositoryExtractG (ZebraHandle zh,
+ int deleteFlag, struct recordGroup *rGroup)
+{
+ if (*rGroup->path == '\0' || !strcmp(rGroup->path, "-"))
+ {
+ char src[1024];
+
+ while (scanf ("%1020s", src) == 1)
+ repositoryExtract (zh, deleteFlag, rGroup, src);
+ }
+ else
+ repositoryExtract (zh, deleteFlag, rGroup, rGroup->path);
+}
+
+void repositoryUpdate (ZebraHandle zh)
+{
+ struct recordGroup *rGroup = &zh->rGroup;
+ groupRes (zh, rGroup);
+ assert (rGroup->path);
+ if (rGroup->recordId && !strcmp (rGroup->recordId, "file"))
+ {
+ Dict dict;
+ if (!(dict = dict_open (zh->reg->bfs, FMATCH_DICT, 50,
+ rGroup->flagRw, 0)))
+ {
+ logf (LOG_FATAL, "dict_open fail of %s", FMATCH_DICT);
+ return ;
+ }
+ if (*rGroup->path == '\0' || !strcmp(rGroup->path, "-"))
+ {
+ char src[1024];
+ while (scanf ("%s", src) == 1)
+ fileUpdate (zh, dict, rGroup, src);
+ }
+ else
+ fileUpdate (zh, dict, rGroup, rGroup->path);
+ dict_close (dict);
+ }
+ else
+ repositoryExtractG (zh, 0, rGroup);
+}
+
+void repositoryDelete (ZebraHandle zh)
+{
+ repositoryExtractG (zh, 1, &zh->rGroup);
+}
+