X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Ftrav.c;h=c4768fe9ec0f0fd56ec98b89d61b8263fd05ff56;hb=5e7516d00451ada59a311fad05e8e3959242cdc7;hp=c929e51de2145b9433ec5fcd903e879cc33724ee;hpb=9f9b9eeb79d92c7793ffaba8805537bd401bdfbe;p=idzebra-moved-to-github.git diff --git a/index/trav.c b/index/trav.c index c929e51..c4768fe 100644 --- a/index/trav.c +++ b/index/trav.c @@ -1,52 +1,26 @@ /* - * Copyright (C) 1994-1995, Index Data I/S + * Copyright (C) 1994-2002, Index Data * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: trav.c,v $ - * 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 1.39 2002-04-04 20:50:37 adam Exp $ */ + + #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) @@ -56,29 +30,39 @@ static int repComp (const char *a, const char *b, size_t len) return memcmp (a, b, len); } -static void repositoryExtractR (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); 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, 0); + fileExtract (zh, NULL, rep, rGroup, deleteFlag); break; case dirs_dir: - repositoryExtractR (rep, rGroup); + repositoryExtractR (zh, deleteFlag, rep, rGroup, level+1); break; } } @@ -86,19 +70,53 @@ static void repositoryExtractR (char *rep, struct recordGroup *rGroup) } -static void repositoryUpdateR (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); + + while (dst && !repComp (dst->path, src, src_len+1)) + { + switch (dst->kind) + { + 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 dirs_dir: + strcpy (tmppath, dst->path); + dst = dirs_read (di); + dirs_rmdir (di, tmppath); + break; + default: + dst = dirs_read (di); + } + } +} + +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); + 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)) @@ -106,26 +124,29 @@ static void repositoryUpdateR (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) { - /* 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); @@ -134,13 +155,13 @@ static void repositoryUpdateR (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; @@ -150,23 +171,30 @@ static void repositoryUpdateR (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)) - 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, 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; @@ -178,17 +206,17 @@ static void repositoryUpdateR (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: - repositoryUpdateR (di, dst, base, src, rGroup); + fileUpdateR (zh, di, dst, base, src, rGroup, level+1); if (dst) dst = dirs_last (di); break; @@ -197,34 +225,213 @@ static void repositoryUpdateR (struct dirs_info *di, struct dirs_entry *dst, } 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 (struct recordGroup *rGroup) +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); - assert (rGroup->path); - di = dirs_open (dict, rGroup->path); - strcpy (src, ""); - repositoryUpdateR (di, dirs_read (di), rGroup->path, src, rGroup); - dirs_free (&di); + 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")); +} +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; + } + + 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); } -void repositoryExtract (struct recordGroup *rGroup) +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); + stat (src, &sbuf); + + 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) { - char src[256]; + 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); + 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); - strcpy (src, rGroup->path); - repositoryExtractR (src, rGroup); + 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); +} +