X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Ftrav.c;h=1143d1eb6b40dac9e4450785509567254abba0d4;hp=4d087bd33677a6e98135e3e10313154580d75d69;hb=896c0427df9d8eff5de6a1735dcd992e067df844;hpb=b1a4c757e1728f5b036cc39e5b990faa83b73bfc diff --git a/index/trav.c b/index/trav.c index 4d087bd..1143d1e 100644 --- a/index/trav.c +++ b/index/trav.c @@ -1,83 +1,43 @@ -/* - * Copyright (C) 1994-1995, Index Data I/S - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: trav.c,v $ - * 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 1.40 2002-08-02 19:26:55 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 #include -#include -#include -#include #include +#include +#ifdef WIN32 +#include +#define S_ISREG(x) (x & _S_IFREG) +#define S_ISDIR(x) (x & _S_IFDIR) +#else +#include +#endif +#include #include -#include +#include -#include #include "index.h" static int repComp (const char *a, const char *b, size_t len) @@ -87,31 +47,39 @@ static int repComp (const char *a, const char *b, size_t len) return memcmp (a, b, len); } -static void repositoryExtractR (int deleteFlag, char *rep, - struct recordGroup *rGroup) +static void repositoryExtractR (ZebraHandle zh, int deleteFlag, char *rep, + struct recordGroup *rGroup, + 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); if (!e) return; - logf (LOG_LOG, "Dir: %s", rep); + 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); + if ((ecp = strrchr (e[i].name, '/'))) + *ecp = '\0'; + if (level == 0 && rGroup->databaseNamePath) + rGroup->databaseName = e[i].name; + switch (e[i].kind) { case dirs_file: - fileExtract (NULL, rep, rGroup, deleteFlag); + fileExtract (zh, NULL, rep, rGroup, deleteFlag); break; case dirs_dir: - repositoryExtractR (deleteFlag, rep, rGroup); + repositoryExtractR (zh, deleteFlag, rep, rGroup, level+1); break; } } @@ -119,18 +87,10 @@ static void repositoryExtractR (int deleteFlag, char *rep, } -static void stdinExtractR (int deleteFlag, struct recordGroup *rGroup) -{ - char tmppath[1024]; - - logf (LOG_LOG, "stdinExtractR"); - while (scanf ("%s", tmppath) == 1) - fileExtract (NULL, tmppath, rGroup, deleteFlag); -} - -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, + struct recordGroup *rGroup) { char tmppath[1024]; size_t src_len = strlen (src); @@ -141,7 +101,7 @@ static void fileDeleteR (struct dirs_info *di, struct dirs_entry *dst, { case dirs_file: sprintf (tmppath, "%s%s", base, dst->path); - fileExtract (&dst->sysno, tmppath, rGroup, 1); + fileExtract (zh, &dst->sysno, tmppath, rGroup, 1); strcpy (tmppath, dst->path); dst = dirs_read (di); @@ -158,9 +118,11 @@ static void fileDeleteR (struct dirs_info *di, struct dirs_entry *dst, } } -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, + struct recordGroup *rGroup, + int level) { struct dir_entry *e_src; int i_src = 0; @@ -168,9 +130,10 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, size_t src_len = strlen (src); sprintf (tmppath, "%s%s", base, src); - e_src = dir_open (tmppath); - logf (LOG_LOG, "Dir: %s", tmppath); -#if 1 + e_src = dir_open (tmppath, zh->path_reg); + logf (LOG_LOG, "dir %s", tmppath); + +#if 0 if (!dst || repComp (dst->path, src, src_len)) #else if (!dst || strcmp (dst->path, src)) @@ -178,27 +141,29 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, { 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) { strcpy (src, dst->path); - fileDeleteR (di, dst, base, src, rGroup); + 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); @@ -207,13 +172,13 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, { 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; @@ -223,25 +188,30 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, 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, rGroup, 0)) + if (fileExtract (zh, &dst->sysno, tmppath, rGroup, 0)) { - dirs_add (di, src, dst->sysno, e_src[i_src].ctime); + 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: - fileUpdateR (di, dst, base, src, rGroup); + fileUpdateR (zh, di, dst, base, src, rGroup, level+1); dst = dirs_last (di); logf (LOG_DEBUG, "last is %s", dst ? dst->path : "null"); break; @@ -253,17 +223,17 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, 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, rGroup, 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: - fileUpdateR (di, dst, base, src, rGroup); + fileUpdateR (zh, di, dst, base, src, rGroup, level+1); if (dst) dst = dirs_last (di); break; @@ -278,12 +248,12 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, switch (dst->kind) { case dirs_file: - fileExtract (&dst->sysno, tmppath, rGroup, 1); + fileExtract (zh, &dst->sysno, tmppath, rGroup, 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, rGroup); dst = dirs_last (di); } } @@ -291,7 +261,7 @@ static void fileUpdateR (struct dirs_info *di, struct dirs_entry *dst, dir_free (&e_src); } -static void groupRes (struct recordGroup *rGroup) +static void groupRes (ZebraHandle zh, struct recordGroup *rGroup) { char resStr[256]; char gPrefix[256]; @@ -302,51 +272,183 @@ static void groupRes (struct recordGroup *rGroup) sprintf (gPrefix, "%s.", rGroup->groupName); sprintf (resStr, "%srecordId", gPrefix); - rGroup->recordId = res_get (common_resource, resStr); + rGroup->recordId = res_get (zh->res, resStr); + sprintf (resStr, "%sdatabasePath", gPrefix); + rGroup->databaseNamePath = + atoi (res_get_def (zh->res, resStr, "0")); } -void repositoryUpdate (struct recordGroup *rGroup) +void repositoryShow (ZebraHandle zh) + { - char src[256]; + 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; + } + + 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); +} - groupRes (rGroup); - if (rGroup->recordId && !strcmp (rGroup->recordId, "file")) +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)) { - Dict dict; - struct dirs_info *di; + strcpy (src, zh->path_reg); + strcat (src, "/"); + } + else + *src = '\0'; + strcat (src, path); + stat (src, &sbuf); - if (!(dict = dict_open (FMATCH_DICT, 50, 1))) + 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) { - logf (LOG_FATAL, "dict_open fail of %s", FMATCH_DICT); - exit (1); + 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); } - assert (rGroup->path); - di = dirs_open (dict, rGroup->path); - strcpy (src, ""); - fileUpdateR (di, dirs_read (di), rGroup->path, src, rGroup); dirs_free (&di); - dict_close (dict); } - else + else if (S_ISDIR(sbuf.st_mode)) { - strcpy (src, rGroup->path); - if (*src == '\0' || !strcmp (src, "-")) - stdinExtractR (0, rGroup); - else - repositoryExtractR (0, src, rGroup); + 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); } } -void repositoryDelete (struct recordGroup *rGroup) + +static void repositoryExtract (ZebraHandle zh, + int deleteFlag, struct recordGroup *rGroup, + const char *path) { - char src[256]; + struct stat sbuf; + char src[1024]; - assert (rGroup->path); - groupRes (rGroup); - strcpy (src, rGroup->path); - if (*src == '\0' || !strcmp(src, "-")) - stdinExtractR (1, rGroup); + assert (path); + + if (zh->path_reg && !yaz_is_abspath(path)) + { + strcpy (src, zh->path_reg); + strcat (src, "/"); + } else - repositoryExtractR (1, src, rGroup); + *src = '\0'; + strcat (src, path); + stat (src, &sbuf); + + 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); }