X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fextract.c;h=db1c9c7d457bd2ac5c9094d2f5027e460adf3723;hb=5b886a7358c16434d5c9f5f0d4af2f6e2e5d5c73;hp=df17b2d601ba7c6c27ddc656a50eced6da25f8ef;hpb=80eac939036e614a3c09c84b85569b9318da5288;p=idzebra-moved-to-github.git diff --git a/index/extract.c b/index/extract.c index df17b2d..db1c9c7 100644 --- a/index/extract.c +++ b/index/extract.c @@ -1,10 +1,38 @@ /* - * Copyright (C) 1994-1996, Index Data I/S + * Copyright (C) 1994-1998, Index Data I/S * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: extract.c,v $ - * Revision 1.70 1997-07-01 13:00:42 adam + * Revision 1.77 1998-01-12 15:04:08 adam + * The test option (-s) only uses read-lock (and not write lock). + * + * Revision 1.76 1997/10/27 14:33:04 adam + * Moved towards generic character mapping depending on "structure" + * field in abstract syntax file. Fixed a few memory leaks. Fixed + * bug with negative integers when doing searches with relational + * operators. + * + * Revision 1.75 1997/09/17 12:19:12 adam + * Zebra version corresponds to YAZ version 1.4. + * Changed Zebra server so that it doesn't depend on global common_resource. + * + * Revision 1.74 1997/09/09 13:38:06 adam + * Partial port to WIN95/NT. + * + * Revision 1.73 1997/09/04 13:57:20 adam + * New file extract/retrieve method tellf (added). + * Added O_BINARY for open calls. + * + * Revision 1.72 1997/07/15 16:32:29 adam + * Bug fix: Match handler didn't terminate the resulting string! + * + * Revision 1.71 1997/07/15 16:28:41 adam + * Bug fix: storeData didn't work with files with multiple records. + * Bug fix: fixed memory management with records; not really well + * thought through. + * + * Revision 1.70 1997/07/01 13:00:42 adam * Bug fix in routine searchRecordKey: uninitialized variables. * * Revision 1.69 1997/04/29 09:26:03 adam @@ -254,7 +282,11 @@ */ #include #include +#ifdef WINDOWS +#include +#else #include +#endif #include #include @@ -291,7 +323,7 @@ static void logRecord (int showFlag) } } -void key_open (int mem) +int key_open (BFiles bfs, int mem, int rw) { if (!mem) mem = atoi(res_get_def (common_resource, "memMax", "4"))*1024*1024; @@ -304,16 +336,26 @@ void key_open (int mem) key_buf_used = 0; key_file_no = 0; - if (!(matchDict = dict_open (GMATCH_DICT, 50, 1))) + if (!(matchDict = dict_open (bfs, GMATCH_DICT, 50, rw))) { logf (LOG_FATAL, "dict_open fail of %s", GMATCH_DICT); - exit (1); + return -1; } assert (!records); - records = rec_open (1); -#if 1 - zti = zebTargetInfo_open (records, 1); -#endif + records = rec_open (bfs, rw); + if (!records) + { + dict_close (matchDict); + return -1; + } + zti = zebTargetInfo_open (records, rw); + if (!zti) + { + rec_close (&records); + dict_close (matchDict); + return -1; + } + return 0; } struct encode_info { @@ -418,7 +460,7 @@ void key_flush (void) if (!(outf = fopen (out_fname, "w"))) { - logf (LOG_FATAL|LOG_ERRNO, "fopen (4) %s", out_fname); + logf (LOG_FATAL|LOG_ERRNO, "fopen %s", out_fname); exit (1); } logf (LOG_LOG, "writing section %d", key_file_no); @@ -444,7 +486,7 @@ void key_flush (void) if (!(outf = fopen (out_fname, "w"))) { - logf (LOG_FATAL|LOG_ERRNO, "fopen (4) %s", out_fname); + logf (LOG_FATAL|LOG_ERRNO, "fopen %s", out_fname); exit (1); } logf (LOG_LOG, "writing section %d", key_file_no); @@ -484,7 +526,7 @@ void key_flush (void) key_buf_used = 0; } -int key_close (void) +int key_close () { key_flush (); xfree (key_buf); @@ -502,7 +544,7 @@ static void wordInit (RecWord *p) { p->attrSet = 1; p->attrUse = 1016; - p->which = Word_String; + p->reg_type = 'w'; } struct recKeys { @@ -566,19 +608,9 @@ static void addRecordKey (const RecWord *p) memcpy (dst, &attrUse, sizeof(attrUse)); dst += sizeof(attrUse); } - switch (p->which) - { - case Word_String: - *dst++ = 'w'; - break; - case Word_Phrase: - *dst++ = 'p'; - break; - case Word_Numeric: - *dst++ = 'n'; - } - for (i = 0; p->u.string[i] && i < IT_MAX_WORD-3; i++) - *dst++ = p->u.string[i]; + *dst++ = p->reg_type; + for (i = 0; p->string[i] && i < IT_MAX_WORD-3; i++) + *dst++ = p->string[i]; *dst++ = '\0'; if (!diff) @@ -744,6 +776,12 @@ static off_t file_seek (void *handle, off_t offset) return lseek (p->fd, offset, SEEK_SET); } +static off_t file_tell (void *handle) +{ + struct file_read_info *p = handle; + return p->file_offset; +} + static int file_read (void *handle, char *buf, size_t count) { struct file_read_info *p = handle; @@ -918,6 +956,7 @@ static char *fileMatchStr (struct recKeys *reckeys, struct recordGroup *rGroup, fname, rGroup->groupName ? rGroup->groupName : "none"); return NULL; } + *dst = '\0'; return dstBuf; } @@ -959,6 +998,7 @@ static int recordExtract (SYSNO *sysno, const char *fname, extractCtrl.subType = subType; extractCtrl.init = wordInit; extractCtrl.add = addRecordKey; + extractCtrl.dh = rGroup->dh; reckeys.buf_used = 0; reckeys.prevAttrUse = -1; @@ -969,10 +1009,11 @@ static int recordExtract (SYSNO *sysno, const char *fname, extractCtrl.offset = recordOffset; extractCtrl.readf = file_read; extractCtrl.seekf = file_seek; + extractCtrl.tellf = file_tell; extractCtrl.endf = file_end; - extractCtrl.map_chrs_input = map_chrs_input; - extractCtrl.flagShowRecords = rGroup->flagShowRecords; - if (rGroup->flagShowRecords) + extractCtrl.zebra_maps = rGroup->zebra_maps; + extractCtrl.flagShowRecords = !rGroup->flagRw; + if (!rGroup->flagRw) printf ("File: %s %ld\n", fname, (long) recordOffset); logInfo.fname = fname; @@ -987,8 +1028,8 @@ static int recordExtract (SYSNO *sysno, const char *fname, if (r) { /* error occured during extraction ... */ - if (!rGroup->flagShowRecords && - records_processed < rGroup->fileVerboseLimit) + if (rGroup->flagRw && + records_processed < rGroup->fileVerboseLimit) { logf (LOG_WARN, "fail %s %s %ld code = %d", rGroup->recordType, fname, (long) recordOffset, r); @@ -999,7 +1040,7 @@ static int recordExtract (SYSNO *sysno, const char *fname, { /* the extraction process returned no information - the record is probably empty - unless flagShowRecords is in use */ - if (rGroup->flagShowRecords) + if (!rGroup->flagRw) return 1; logf (LOG_WARN, "No keys generated for file %s", fname); logf (LOG_WARN, " The file is probably empty"); @@ -1039,7 +1080,9 @@ static int recordExtract (SYSNO *sysno, const char *fname, /* new record */ if (deleteFlag) { - logf (LOG_LOG, "Cannot delete new record"); + logf (LOG_LOG, "delete %s %s %ld", rGroup->recordType, + fname, (long) recordOffset); + logf (LOG_WARN, "cannot delete record above (seems new)"); return 1; } if (records_processed < rGroup->fileVerboseLimit) @@ -1143,16 +1186,18 @@ static int recordExtract (SYSNO *sysno, const char *fname, xfree (rec->info[recInfo_storeData]); if (rGroup->flagStoreData == 1) { - rec->size[recInfo_storeData] = fi->file_max; - rec->info[recInfo_storeData] = xmalloc (fi->file_max); + int size = fi->file_moffset - recordOffset; + if (!size) + size = fi->file_max - recordOffset; + rec->size[recInfo_storeData] = size; + rec->info[recInfo_storeData] = xmalloc (size); if (lseek (fi->fd, recordOffset, SEEK_SET) < 0) { logf (LOG_ERRNO|LOG_FATAL, "seek to %ld in %s", fname, (long) recordOffset); exit (1); } - if (read (fi->fd, rec->info[recInfo_storeData], fi->file_max) - < fi->file_max) + if (read (fi->fd, rec->info[recInfo_storeData], size) < size) { logf (LOG_ERRNO|LOG_FATAL, "read %d bytes of %s", fi->file_max, fname); @@ -1298,7 +1343,7 @@ int fileExtract (SYSNO *sysno, const char *fname, fd = -1; else { - if ((fd = open (fname, O_RDONLY)) == -1) + if ((fd = open (fname, O_BINARY|O_RDONLY)) == -1) { logf (LOG_WARN|LOG_ERRNO, "open %s", fname); return 0;