X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fretrieve.c;h=650eb82b38158b59854a67ac5ac1f612dc58d699;hb=a20e59c5087fb92c419f2330c786367f9ce8ccd2;hp=dcc17b9018ed624e56f10aa683b6ce4b44043dc5;hpb=e9b33fe049e35d06dfeedb830feea32553d26cee;p=idzebra-moved-to-github.git diff --git a/index/retrieve.c b/index/retrieve.c index dcc17b9..650eb82 100644 --- a/index/retrieve.c +++ b/index/retrieve.c @@ -4,7 +4,38 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: retrieve.c,v $ - * Revision 1.7 1999-03-02 16:15:43 quinn + * Revision 1.17 2002-05-03 13:49:04 adam + * play with shellsort + * + * Revision 1.16 2002/04/04 20:50:37 adam + * Multi register works with record paths and data1 profile path + * + * Revision 1.15 2002/04/04 14:14:13 adam + * Multiple registers (alpha early) + * + * Revision 1.14 2001/01/22 11:41:41 adam + * Added support for raw retrieval (element set name "R"). + * + * Revision 1.13 2000/03/20 19:08:36 adam + * Added remote record import using Z39.50 extended services and Segment + * Requests. + * + * Revision 1.12 2000/03/15 15:00:30 adam + * First work on threaded version. + * + * Revision 1.11 1999/10/29 10:00:00 adam + * Fixed minor bug where database name wasn't set in zebra_record_fetch. + * + * Revision 1.10 1999/05/26 07:49:13 adam + * C++ compilation. + * + * Revision 1.9 1999/05/20 12:57:18 adam + * Implemented TCL filter. Updated recctrl system. + * + * Revision 1.8 1999/03/09 16:27:49 adam + * More work on SDRKit integration. + * + * Revision 1.7 1999/03/02 16:15:43 quinn * Added "tagsysno" and "tagrank" directives to zebra.cfg. * * Revision 1.6 1999/02/18 15:01:25 adam @@ -39,58 +70,42 @@ #include #endif -#include -#include "zserver.h" - -#ifndef ZEBRASDR -#define ZEBRASDR 0 -#endif - -#if ZEBRASDR -#include "zebrasdr.h" -#endif - -struct fetch_control { - int record_offset; - int record_int_pos; - char *record_int_buf; - int record_int_len; - int fd; -}; +#include "index.h" +#include -static int record_ext_read (void *fh, char *buf, size_t count) +int zebra_record_ext_read (void *fh, char *buf, size_t count) { - struct fetch_control *fc = fh; + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; return read (fc->fd, buf, count); } -static off_t record_ext_seek (void *fh, off_t offset) +off_t zebra_record_ext_seek (void *fh, off_t offset) { - struct fetch_control *fc = fh; + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; return lseek (fc->fd, offset + fc->record_offset, SEEK_SET); } -static off_t record_ext_tell (void *fh) +off_t zebra_record_ext_tell (void *fh) { - struct fetch_control *fc = fh; + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; return lseek (fc->fd, 0, SEEK_CUR) - fc->record_offset; } -static off_t record_int_seek (void *fh, off_t offset) +off_t zebra_record_int_seek (void *fh, off_t offset) { - struct fetch_control *fc = fh; + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; return (off_t) (fc->record_int_pos = offset); } -static off_t record_int_tell (void *fh) +off_t zebra_record_int_tell (void *fh) { - struct fetch_control *fc = fh; + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; return (off_t) fc->record_int_pos; } -static int record_int_read (void *fh, char *buf, size_t count) +int zebra_record_int_read (void *fh, char *buf, size_t count) { - struct fetch_control *fc = fh; + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; int l = fc->record_int_len - fc->record_int_pos; if (l <= 0) return 0; @@ -100,6 +115,12 @@ static int record_int_read (void *fh, char *buf, size_t count) return l; } +void zebra_record_int_end (void *fh, off_t off) +{ + struct zebra_fetch_control *fc = (struct zebra_fetch_control *) fh; + fc->offset_end = off; +} + int zebra_record_fetch (ZebraHandle zh, int sysno, int score, ODR stream, oid_value input_format, Z_RecordComposition *comp, oid_value *output_format, char **rec_bufp, @@ -110,24 +131,33 @@ int zebra_record_fetch (ZebraHandle zh, int sysno, int score, ODR stream, RecType rt; struct recRetrieveCtrl retrieveCtrl; char subType[128]; - struct fetch_control fc; + struct zebra_fetch_control fc; RecordAttr *recordAttr; + void *clientData; - rec = rec_get (zh->records, sysno); + rec = rec_get (zh->reg->records, sysno); if (!rec) { logf (LOG_DEBUG, "rec_get fail on sysno=%d", sysno); + *basenamep = 0; return 14; } - recordAttr = rec_init_attr (zh->zei, rec); + recordAttr = rec_init_attr (zh->reg->zei, rec); file_type = rec->info[recInfo_fileType]; fname = rec->info[recInfo_filename]; basename = rec->info[recInfo_databaseName]; - *basenamep = odr_malloc (stream, strlen(basename)+1); + *basenamep = (char *) odr_malloc (stream, strlen(basename)+1); strcpy (*basenamep, basename); - if (!(rt = recType_byName (zh->recTypes, file_type, subType))) + if (comp && comp->which == Z_RecordComp_simple && + comp->u.simple->which == Z_ElementSetNames_generic) + { + if (!strcmp (comp->u.simple->u.generic, "R")) + file_type = "text"; + } + if (!(rt = recType_byName (zh->reg->recTypes, + file_type, subType, &clientData))) { logf (LOG_WARN, "Retrieve: Cannot handle type %s", file_type); return 14; @@ -137,95 +167,42 @@ int zebra_record_fetch (ZebraHandle zh, int sysno, int score, ODR stream, fc.fd = -1; if (rec->size[recInfo_storeData] > 0) { - retrieveCtrl.readf = record_int_read; - retrieveCtrl.seekf = record_int_seek; - retrieveCtrl.tellf = record_int_tell; + retrieveCtrl.readf = zebra_record_int_read; + retrieveCtrl.seekf = zebra_record_int_seek; + retrieveCtrl.tellf = zebra_record_int_tell; fc.record_int_len = rec->size[recInfo_storeData]; fc.record_int_buf = rec->info[recInfo_storeData]; fc.record_int_pos = 0; logf (LOG_DEBUG, "Internal retrieve. %d bytes", fc.record_int_len); } -#if ZEBRASDR - else if (*fname == '%') - { - ZebraSdrHandle h; - int segment = 0, r; - char *cp, xname[128]; - unsigned char *buf; - - logf (LOG_DEBUG, "SDR"); - strcpy (xname, fname+1); - if ((cp = strrchr (xname, '.'))) - { - *cp++ = 0; - segment = atoi(cp); - } - h = zebraSdr_open (xname); - if (!h) - { - logf (LOG_WARN, "sdr open %s", xname); - return 0; - } - if (zebraSdr_segment (h, &segment) < 0) - { - logf (LOG_WARN, "zebraSdr_segment fail segment=%d", - segment); - rec_rm (&rec); - return 14; - } - r = zebraSdr_read (h, &buf); - if (r < 1) - { - logf (LOG_WARN, "zebraSdr_read fail segment=%d", - segment); - rec_rm (&rec); - return 14; - } - zebraSdr_close (h); - - fc.record_int_len = recordAttr->recordSize; - fc.record_int_buf = buf + recordAttr->recordOffset; - fc.record_int_pos = 0; - - logf (LOG_LOG, "segment = %d len=%d off=%d", - segment, - recordAttr->recordSize, - recordAttr->recordOffset); - if (fc.record_int_len > 180) - { - logf (LOG_LOG, "%0.70s", fc.record_int_buf); - logf (LOG_LOG, "%0.70s", fc.record_int_buf + - (fc.record_int_len - 70)); - } - else - logf (LOG_LOG, "%0.*s", - fc.record_int_len, fc.record_int_buf); - - /* the following two lines makes rec_rm delete buf */ - rec->size[recInfo_storeData] = r; - rec->info[recInfo_storeData] = buf; - - retrieveCtrl.readf = record_int_read; - retrieveCtrl.seekf = record_int_seek; - retrieveCtrl.tellf = record_int_tell; - } -#endif else { - if ((fc.fd = open (fname, O_BINARY|O_RDONLY)) == -1) + char full_rep[1024]; + + if (zh->path_reg && !yaz_is_abspath (fname)) + { + strcpy (full_rep, zh->path_reg); + strcat (full_rep, "/"); + strcat (full_rep, fname); + } + else + strcpy (full_rep, fname); + + + if ((fc.fd = open (full_rep, O_BINARY|O_RDONLY)) == -1) { logf (LOG_WARN|LOG_ERRNO, "Retrieve fail; missing file: %s", - fname); + full_rep); rec_rm (&rec); return 14; } fc.record_offset = recordAttr->recordOffset; - retrieveCtrl.readf = record_ext_read; - retrieveCtrl.seekf = record_ext_seek; - retrieveCtrl.tellf = record_ext_tell; + retrieveCtrl.readf = zebra_record_ext_read; + retrieveCtrl.seekf = zebra_record_ext_seek; + retrieveCtrl.tellf = zebra_record_ext_tell; - record_ext_seek (retrieveCtrl.fh, 0); + zebra_record_ext_seek (retrieveCtrl.fh, 0); } retrieveCtrl.subType = subType; retrieveCtrl.localno = sysno; @@ -235,11 +212,14 @@ int zebra_record_fetch (ZebraHandle zh, int sysno, int score, ODR stream, retrieveCtrl.input_format = retrieveCtrl.output_format = input_format; retrieveCtrl.comp = comp; retrieveCtrl.diagnostic = 0; - retrieveCtrl.dh = zh->dh; + retrieveCtrl.dh = zh->reg->dh; retrieveCtrl.res = zh->res; - (*rt->retrieve)(&retrieveCtrl); + retrieveCtrl.rec_buf = 0; + retrieveCtrl.rec_len = -1; + + (*rt->retrieve)(clientData, &retrieveCtrl); *output_format = retrieveCtrl.output_format; - *rec_bufp = retrieveCtrl.rec_buf; + *rec_bufp = (char *) retrieveCtrl.rec_buf; *rec_lenp = retrieveCtrl.rec_len; if (fc.fd != -1) close (fc.fd);