X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=recctrl%2Frecgrs.c;h=982573be6539b9687e0d17e967cd8721003aefd1;hp=d23a3d70bbd8283566f5cb2a7e673d3227b1d6e6;hb=6c9fcd3b5d3108702fa1ffc92dab4ab6060f9a19;hpb=7a49c3db444b475f63722c3da03e15c6db93f1a9 diff --git a/recctrl/recgrs.c b/recctrl/recgrs.c index d23a3d7..982573b 100644 --- a/recctrl/recgrs.c +++ b/recctrl/recgrs.c @@ -1,222 +1,24 @@ -/* - * Copyright (C) 1994-2001, Index Data - * All rights reserved. - * - * $Log: recgrs.c,v $ - * Revision 1.41 2001-05-22 21:01:47 adam - * Removed print of data1 tree on stdout so that inetd works again. - * - * Revision 1.40 2001/03/29 21:31:31 adam - * Fixed "record begin" for Tcl filter. - * - * Revision 1.39 2000/12/05 19:09:15 adam - * Fixed problem where indexer could crash if abstract syntax was undefined. - * - * Revision 1.38 2000/12/05 14:44:58 adam - * Fixed minor bug that could cause zmbol to break it data were emitted - * with not parent tags. - * - * Revision 1.37 2000/12/05 12:22:53 adam - * Termlist source implemented (so that we can index values of XML/SGML - * attributes). - * - * Revision 1.36 2000/12/05 10:01:44 adam - * Fixed bug regarding user-defined attribute sets. - * - * Revision 1.35 2000/11/29 15:21:31 adam - * Fixed problem with passwd db. - * - * Revision 1.34 2000/02/25 13:24:49 adam - * Fixed bug regarding pointer conversion that showed up on OSF V5. - * - * Revision 1.33 1999/11/30 13:48:04 adam - * Improved installation. Updated for inclusion of YAZ header files. - * - * Revision 1.32 1999/09/07 07:19:21 adam - * Work on character mapping. Implemented replace rules. - * - * Revision 1.31 1999/07/14 10:56:43 adam - * Fixed potential memory leak. - * - * Revision 1.30 1999/07/06 12:26:41 adam - * Retrieval handler obeys schema and handles XML transfer syntax. - * - * Revision 1.29 1999/05/26 07:49:14 adam - * C++ compilation. - * - * Revision 1.28 1999/05/21 12:00:17 adam - * Better diagnostics for extraction process. - * - * Revision 1.27 1999/05/20 12:57:18 adam - * Implemented TCL filter. Updated recctrl system. - * - * Revision 1.26 1999/03/02 16:15:44 quinn - * Added "tagsysno" and "tagrank" directives to zebra.cfg. - * - * Revision 1.25 1999/02/18 15:01:26 adam - * Minor changes. - * - * Revision 1.24 1999/02/02 14:51:28 adam - * Updated WIN32 code specific sections. Changed header. - * - * Revision 1.23 1998/10/18 07:51:10 adam - * Changed one logf call. - * - * Revision 1.22 1998/10/16 08:14:37 adam - * Updated record control system. - * - * Revision 1.21 1998/07/01 09:16:10 adam - * Element localno only added when it's greater than 0. - * - * Revision 1.20 1998/05/20 10:12:26 adam - * Implemented automatic EXPLAIN database maintenance. - * Modified Zebra to work with ASN.1 compiled version of YAZ. - * - * Revision 1.19 1998/03/11 11:19:05 adam - * Changed the way sequence numbers are generated. - * - * Revision 1.18 1998/03/05 08:41:31 adam - * Minor changes. - * - * Revision 1.17 1998/02/10 12:03:06 adam - * Implemented Sort. - * - * Revision 1.16 1998/01/29 13:38:17 adam - * Fixed problem with mapping to record with unknown schema. - * - * Revision 1.15 1998/01/26 10:37:57 adam - * Better diagnostics. - * - * Revision 1.14 1997/11/06 11:41:01 adam - * Implemented "begin variant" for the sgml.regx filter. - * - * Revision 1.13 1997/10/31 12:35:44 adam - * Added a few log statements. - * - * Revision 1.12 1997/10/29 12:02:22 adam - * Using oid_ent_to_oid used instead of the non thread-safe oid_getoidbyent. - * - * Revision 1.11 1997/10/27 14:34:00 adam - * Work on generic character mapping depending on "structure" field - * in abstract syntax file. - * - * Revision 1.10 1997/09/18 08:59:21 adam - * Extra generic handle for the character mapping routines. - * - * Revision 1.9 1997/09/17 12:19:21 adam - * Zebra version corresponds to YAZ version 1.4. - * Changed Zebra server so that it doesn't depend on global common_resource. - * - * Revision 1.8 1997/09/09 13:38:14 adam - * Partial port to WIN95/NT. - * - * Revision 1.7 1997/09/05 15:30:10 adam - * Changed prototype for chr_map_input - added const. - * Added support for C++, headers uses extern "C" for public definitions. - * - * Revision 1.6 1997/09/04 13:54:40 adam - * Added MARC filter - type grs.marc. where syntax refers - * to abstract syntax. New method tellf in retrieve/extract method. - * - * Revision 1.5 1997/07/15 16:29:03 adam - * Initialized dummy variable to keep checker gcc happy. - * - * Revision 1.4 1997/04/30 08:56:08 quinn - * null - * - * Revision 1.2 1996/10/11 16:06:43 quinn - * Revision 1.3 1997/02/24 10:41:50 adam - * Cleanup of code and commented out the "end element-end-record" code. - * - * Revision 1.2 1996/10/11 16:06:43 quinn - * Fixed arguments to nodetogr - * - * Revision 1.1 1996/10/11 10:57:25 adam - * New module recctrl. Used to manage records (extract/retrieval). - * - * Revision 1.29 1996/10/08 10:30:21 quinn - * Fixed type mismatch - * - * Revision 1.28 1996/10/07 16:06:40 quinn - * Added SOIF support - * - * Revision 1.27 1996/06/11 10:54:12 quinn - * Relevance work - * - * Revision 1.26 1996/06/06 12:08:45 quinn - * Added showRecord function - * - * Revision 1.25 1996/06/04 14:18:53 quinn - * Charmap work - * - * Revision 1.24 1996/06/04 13:27:54 quinn - * More work on charmapping - * - * Revision 1.23 1996/06/04 10:19:01 adam - * Minor changes - removed include of ctype.h. - * - * Revision 1.22 1996/06/03 10:15:27 quinn - * Various character-mapping. - * - * Revision 1.21 1996/05/31 13:27:24 quinn - * Character-conversion in phrases, too. - * - * Revision 1.19 1996/05/16 15:31:14 quinn - * a7 - * - * Revision 1.18 1996/05/09 07:28:56 quinn - * Work towards phrases and multiple registers - * - * Revision 1.17 1996/05/01 13:46:37 adam - * First work on multiple records in one file. - * New option, -offset, to the "unread" command in the filter module. - * - * Revision 1.16 1996/01/17 14:57:54 adam - * Prototype changed for reader functions in extract/retrieve. File - * is identified by 'void *' instead of 'int. - * - * Revision 1.15 1996/01/08 19:15:47 adam - * New input filter that works! - * - * Revision 1.14 1995/12/15 12:36:11 adam - * Retrieval calls data1_read_regx when subType is specified. - * - * Revision 1.13 1995/12/15 12:24:43 quinn - * *** empty log message *** - * - * Revision 1.12 1995/12/15 12:20:28 quinn - * *** empty log message *** - * - * Revision 1.11 1995/12/15 12:07:57 quinn - * Changed extraction strategy. - * - * Revision 1.10 1995/12/14 11:10:48 quinn - * Explain work - * - * Revision 1.9 1995/12/13 17:14:05 quinn - * *** empty log message *** - * - * Revision 1.8 1995/12/13 15:33:18 quinn - * *** empty log message *** - * - * Revision 1.7 1995/12/13 13:45:39 quinn - * Changed data1 to use nmem. - * - * Revision 1.6 1995/12/04 14:22:30 adam - * Extra arg to recType_byName. - * Started work on new regular expression parsed input to - * structured records. - * - * Revision 1.5 1995/11/28 14:18:37 quinn - * Set output_format. - * - * Revision 1.4 1995/11/21 13:14:49 quinn - * Fixed end-of-data-field problem (maybe). - * - * Revision 1.3 1995/11/15 19:13:09 adam - * Work on record management. - * - */ +/* $Id: recgrs.c,v 1.98 2005-01-15 19:38:32 adam Exp $ + Copyright (C) 1995-2005 + 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 @@ -228,98 +30,645 @@ #include #include -#include -#include "grsread.h" +#include +#include #define GRS_MAX_WORD 512 -struct grs_handler { - RecTypeGrs type; - void *clientData; - int initFlag; - struct grs_handler *next; +struct source_parser { + int len; + const char *tok; + const char *src; + int lookahead; }; -struct grs_handlers { - struct grs_handler *handlers; -}; - -static int read_grs_type (struct grs_handlers *h, - struct grs_read_info *p, const char *type, - data1_node **root) +static int sp_lex(struct source_parser *sp) { - struct grs_handler *gh = h->handlers; - const char *cp = strchr (type, '.'); - - if (cp == NULL || cp == type) + while (*sp->src == ' ') + (sp->src)++; + sp->tok = sp->src; + sp->len = 0; + while (*sp->src && !strchr("<>();,-: ", *sp->src)) { - cp = strlen(type) + type; - *p->type = 0; + sp->src++; + sp->len++; } + if (sp->len) + sp->lookahead = 't'; else - strcpy (p->type, cp+1); - for (gh = h->handlers; gh; gh = gh->next) { - if (!memcmp (type, gh->type->type, cp-type)) + sp->lookahead = *sp->src; + if (*sp->src) + sp->src++; + } + return sp->lookahead; +} + + +static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd) +{ + if (sp->lookahead != 't') + return 0; + if (sp->len == 4 && !memcmp(sp->tok, "data", sp->len)) + { + if (n->which == DATA1N_data) { - if (!gh->initFlag) + wrd->string = n->u.data.data; + wrd->length = n->u.data.len; + } + sp_lex(sp); + } + else if (sp->len == 3 && !memcmp(sp->tok, "tag", sp->len)) + { + if (n->which == DATA1N_tag) + { + wrd->string = n->u.tag.tag; + wrd->length = strlen(n->u.tag.tag); + } + sp_lex(sp); + } + else if (sp->len == 4 && !memcmp(sp->tok, "attr", sp->len)) + { + sp_lex(sp); + if (sp->lookahead != '(') + return 0; + sp_lex(sp); + if (sp->lookahead != 't') + return 0; + + if (n->which == DATA1N_tag) + { + data1_xattr *p = n->u.tag.attributes; + while (p && strlen(p->name) != sp->len && + memcmp (p->name, sp->tok, sp->len)) + p = p->next; + if (p) { - gh->initFlag = 1; - gh->clientData = (*gh->type->init)(); + wrd->string = p->value; + wrd->length = strlen(p->value); } - p->clientData = gh->clientData; - *root = (gh->type->read)(p); - gh->clientData = p->clientData; + } + sp_lex(sp); + if (sp->lookahead != ')') return 0; + sp_lex(sp); + } + else if (sp->len == 5 && !memcmp(sp->tok, "range", sp->len)) + { + int start, len; + sp_lex(sp); + if (sp->lookahead != '(') + return 0; + + sp_lex(sp); + sp_expr(sp, n, wrd); + if (sp->lookahead != ',') + return 0; + + sp_lex(sp); + if (sp->lookahead != 't') + return 0; + start = atoi_n(sp->tok, sp->len); + + sp_lex(sp); + if (sp->lookahead != ',') + return 0; + + sp_lex(sp); + if (sp->lookahead != 't') + return 0; + len = atoi_n(sp->tok, sp->len); + + sp_lex(sp); + if (sp->lookahead != ')') + return 0; + + sp_lex(sp); + if (wrd->string && wrd->length) + { + wrd->string += start; + wrd->length -= start; + if (wrd->length > len) + wrd->length = len; } } return 1; } -static void grs_add_handler (struct grs_handlers *h, RecTypeGrs t) +static int sp_parse(data1_node *n, RecWord *wrd, const char *src) { - struct grs_handler *gh = (struct grs_handler *) malloc (sizeof(*gh)); - gh->next = h->handlers; - h->handlers = gh; - gh->initFlag = 0; - gh->clientData = 0; - gh->type = t; + struct source_parser sp; + sp.len = 0; + sp.tok = 0; + sp.src = src; + sp.lookahead = 0; + sp_lex(&sp); + + return sp_expr(&sp, n, wrd); } -static void *grs_init(RecType recType) +int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) +{ + int res = 1; + char *attname; + data1_xattr *attr; + + if (!p) { + return 1; + } else { + if (p->which == XPATH_PREDICATE_RELATION) { + if (p->u.relation.name[0]) { + if (*p->u.relation.name != '@') { + yaz_log(YLOG_WARN, + " Only attributes (@) are supported in xelm xpath predicates"); + yaz_log(YLOG_WARN, "predicate %s ignored", p->u.relation.name); + return (1); + } + attname = p->u.relation.name + 1; + res = 0; + /* looking for the attribute with a specified name */ + for (attr = n->u.tag.attributes; attr; attr = attr->next) { + yaz_log(YLOG_DEBUG," - attribute %s <-> %s", attname, attr->name ); + + if (!strcmp(attr->name, attname)) { + if (p->u.relation.op[0]) { + if (*p->u.relation.op != '=') { + yaz_log(YLOG_WARN, + "Only '=' relation is supported (%s)",p->u.relation.op); + yaz_log(YLOG_WARN, "predicate %s ignored", p->u.relation.name); + res = 1; break; + } else { + yaz_log(YLOG_DEBUG," - value %s <-> %s", + p->u.relation.value, attr->value ); + if (!strcmp(attr->value, p->u.relation.value)) { + res = 1; break; + } + } + } else { + /* attribute exists, no value specified */ + res = 1; break; + } + } + } + yaz_log(YLOG_DEBUG, "return %d", res); + return res; + } else { + return 1; + } + } + else if (p->which == XPATH_PREDICATE_BOOLEAN) { + if (!strcmp(p->u.boolean.op,"and")) { + return d1_check_xpath_predicate(n, p->u.boolean.left) + && d1_check_xpath_predicate(n, p->u.boolean.right); + } + else if (!strcmp(p->u.boolean.op,"or")) { + return (d1_check_xpath_predicate(n, p->u.boolean.left) + || d1_check_xpath_predicate(n, p->u.boolean.right)); + } else { + yaz_log(YLOG_WARN, "Unknown boolean relation %s, ignored",p->u.boolean.op); + return 1; + } + } + } + return 0; +} + + +/* *ostrich* + +New function, looking for xpath "element" definitions in abs, by +tagpath, using a kind of ugly regxp search.The DFA was built while +parsing abs, so here we just go trough them and try to match +against the given tagpath. The first matching entry is returned. + +pop, 2002-12-13 + +Added support for enhanced xelm. Now [] predicates are considered +as well, when selecting indexing rules... (why the hell it's called +termlist???) + +pop, 2003-01-17 + +*/ + +data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n) { - struct grs_handlers *h = (struct grs_handlers *) malloc (sizeof(*h)); - h->handlers = 0; + data1_absyn *abs = n->root->u.root.absyn; + data1_xpelement *xpe = abs->xp_elements; + data1_node *nn; +#ifdef ENHANCED_XELM + struct xpath_location_step *xp; - grs_add_handler (h, recTypeGrs_sgml); - grs_add_handler (h, recTypeGrs_regx); -#if HAVE_TCL_H - grs_add_handler (h, recTypeGrs_tcl); #endif - grs_add_handler (h, recTypeGrs_marc); - return h; + char *pexpr = xmalloc(strlen(tagpath)+2); + int ok = 0; + + sprintf (pexpr, "%s\n", tagpath); + yaz_log(YLOG_DEBUG,"Checking tagpath %s",tagpath); + while (xpe) + { + struct DFA_state **dfaar = xpe->dfa->states; + struct DFA_state *s=dfaar[0]; + struct DFA_tran *t; + const char *p; + int i; + unsigned char c; + int start_line = 1; + + c = *pexpr++; t = s->trans; i = s->tran_no; + if ((c >= t->ch[0] && c <= t->ch[1]) || (!t->ch[0])) { + p = pexpr; + do { + if ((s = dfaar[t->to])->rule_no && + (start_line || s->rule_nno)) { + ok = 1; + break; + } + for (t=s->trans, i=s->tran_no; --i >= 0; t++) { + if ((unsigned) *p >= t->ch[0] && (unsigned) *p <= t->ch[1]) + break; + } + p++; + } while (i >= 0); + } + if (ok) + yaz_log(YLOG_DEBUG," xpath match %s",xpe->xpath_expr); + else + yaz_log(YLOG_DEBUG," xpath no match %s",xpe->xpath_expr); + + pexpr--; + if (ok) { +#ifdef ENHANCED_XELM + /* we have to check the perdicates up to the root node */ + xp = xpe->xpath; + + /* find the first tag up in the node structure */ + nn = n; while (nn && nn->which != DATA1N_tag) { + nn = nn->parent; + } + + /* go from inside out in the node structure, while going + backwards trough xpath location steps ... */ + for (i=xpe->xpath_len - 1; i>0; i--) { + + yaz_log(YLOG_DEBUG,"Checking step %d: %s on tag %s", + i,xp[i].part,nn->u.tag.tag); + + if (!d1_check_xpath_predicate(nn, xp[i].predicate)) { + yaz_log(YLOG_DEBUG," Predicates didn't match"); + ok = 0; + break; + } + + if (nn->which == DATA1N_tag) { + nn = nn->parent; + } + } +#endif + if (ok) { + break; + } + } + xpe = xpe->next; + } + + xfree(pexpr); + + if (ok) { + yaz_log(YLOG_DEBUG,"Got it"); + return xpe->termlists; + } else { + return NULL; + } } -static void grs_destroy(void *clientData) +/* use + 1 start element (tag) + 2 end element + 3 start attr (and attr-exact) + 4 end attr + + 1016 cdata + 1015 attr data + + *ostrich* + + Now, if there is a matching xelm described in abs, for the + indexed element or the attribute, then the data is handled according + to those definitions... + + modified by pop, 2002-12-13 +*/ + +/* add xpath index for an attribute */ +static void index_xpath_attr (char *tag_path, char *name, char *value, + char *structure, struct recExtractCtrl *p, + RecWord *wrd) { - struct grs_handlers *h = (struct grs_handlers *) clientData; - struct grs_handler *gh = h->handlers, *gh_next; - while (gh) + wrd->attrSet = VAL_IDXPATH; + wrd->attrUse = 1; + wrd->reg_type = '0'; + wrd->string = tag_path; + wrd->length = strlen(tag_path); + (*p->tokenAdd)(wrd); + + if (value) { + wrd->attrUse = 1015; + wrd->reg_type = 'w'; + wrd->string = value; + wrd->length = strlen(value); + (*p->tokenAdd)(wrd); + } + + wrd->attrUse = 2; + wrd->reg_type = '0'; + wrd->string = tag_path; + wrd->length = strlen(tag_path); + (*p->tokenAdd)(wrd); +} + + +static void index_xpath (data1_node *n, struct recExtractCtrl *p, + int level, RecWord *wrd, int use) +{ + int i; + char tag_path_full[1024]; + size_t flen = 0; + data1_node *nn; + int termlist_only = 1; + data1_termlist *tl; + int xpdone = 0; + + yaz_log(YLOG_DEBUG, "index_xpath level=%d use=%d", level, use); + if ((!n->root->u.root.absyn) || + (n->root->u.root.absyn->enable_xpath_indexing)) { + termlist_only = 0; + } + + switch (n->which) { - gh_next = gh->next; - if (gh->initFlag) - (*gh->type->destroy)(gh->clientData); - free (gh); - gh = gh_next; + case DATA1N_data: + wrd->string = n->u.data.data; + wrd->length = n->u.data.len; + xpdone = 0; + flen = 0; + + /* we have to fetch the whole path to the data tag */ + for (nn = n; nn; nn = nn->parent) { + if (nn->which == DATA1N_tag) { + size_t tlen = strlen(nn->u.tag.tag); + if (tlen + flen > (sizeof(tag_path_full)-2)) return; + memcpy (tag_path_full + flen, nn->u.tag.tag, tlen); + flen += tlen; + tag_path_full[flen++] = '/'; + } + else if (nn->which == DATA1N_root) break; + } + + tag_path_full[flen] = 0; + + /* If we have a matching termlist... */ + if (n->root->u.root.absyn && + (tl = xpath_termlist_by_tagpath(tag_path_full, n))) + { + for (; tl; tl = tl->next) + { + /* need to copy recword because it may be changed */ + RecWord wrd_tl; + wrd->reg_type = *tl->structure; + /* this is the ! case, so structure is for the xpath index */ + memcpy (&wrd_tl, wrd, sizeof(*wrd)); + if (tl->source) + sp_parse(n, &wrd_tl, tl->source); + if (!tl->att) { + wrd_tl.attrSet = VAL_IDXPATH; + wrd_tl.attrUse = use; + if (p->flagShowRecords) + { + int i; + printf("%*sXPath index", (level + 1) * 4, ""); + printf (" XData:\""); + for (i = 0; i 40) + printf (" ..."); + fputc ('\n', stdout); + } + else + (*p->tokenAdd)(&wrd_tl); + xpdone = 1; + } else { + /* this is just the old fashioned attribute based index */ + wrd_tl.attrSet = (int) (tl->att->parent->reference); + wrd_tl.attrUse = tl->att->locals->local; + if (p->flagShowRecords) + { + int i; + printf("%*sIdx: [%s]", (level + 1) * 4, "", + tl->structure); + printf("%s:%s [%d] %s", + tl->att->parent->name, + tl->att->name, tl->att->value, + tl->source); + printf (" XData:\""); + for (i = 0; i 40) + printf (" ..."); + fputc ('\n', stdout); + } + else + (*p->tokenAdd)(&wrd_tl); + } + } + } + /* xpath indexing is done, if there was no termlist given, + or no ! in the termlist, and default indexing is enabled... */ + if (!p->flagShowRecords && !xpdone && !termlist_only) + { + wrd->attrSet = VAL_IDXPATH; + wrd->attrUse = use; + wrd->reg_type = 'w'; + (*p->tokenAdd)(wrd); + } + break; + case DATA1N_tag: + flen = 0; + for (nn = n; nn; nn = nn->parent) + { + if (nn->which == DATA1N_tag) + { + size_t tlen = strlen(nn->u.tag.tag); + if (tlen + flen > (sizeof(tag_path_full)-2)) + return; + memcpy (tag_path_full + flen, nn->u.tag.tag, tlen); + flen += tlen; + tag_path_full[flen++] = '/'; + } + else if (nn->which == DATA1N_root) + break; + } + + + wrd->reg_type = '0'; + wrd->string = tag_path_full; + wrd->length = flen; + wrd->attrSet = VAL_IDXPATH; + wrd->attrUse = use; + if (p->flagShowRecords) + { + printf("%*s tag=", (level + 1) * 4, ""); + for (i = 0; ilength && i < 40; i++) + fputc (wrd->string[i], stdout); + if (i == 40) + printf (" .."); + printf("\n"); + } + else + { + data1_xattr *xp; + data1_termlist *tl; + int do_xpindex; + + tag_path_full[flen] = 0; + + /* Add tag start/end xpath index, only when there is a ! in the apropriate xelm + directive, or default xpath indexing is enabled */ + if (!(do_xpindex = 1 - termlist_only)) { + if ((tl = xpath_termlist_by_tagpath(tag_path_full, n))) { + for (; tl; tl = tl->next) { if (!tl->att) {do_xpindex = 1;} } + } + } + if (do_xpindex) { + (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */ + } + + if (use == 1) /* only for the starting tag... */ + { +#define MAX_ATTR_COUNT 50 + data1_termlist *tll[MAX_ATTR_COUNT]; + + int i = 0; + + /* get termlists for attributes, and find out, if we have to do xpath indexing */ + for (xp = n->u.tag.attributes; xp; xp = xp->next) { + i++; + } + + i = 0; + for (xp = n->u.tag.attributes; xp; xp = xp->next) { + char comb[512]; + int do_xpindex = 1 - termlist_only; + data1_termlist *tl; + char attr_tag_path_full[1024]; + int int_len = flen; + + /* this could be cached as well */ + sprintf (attr_tag_path_full, "@%s/%.*s", + xp->name, int_len, tag_path_full); + + tll[i] = xpath_termlist_by_tagpath(attr_tag_path_full,n); + + /* if there is a ! in the xelm termlist, or default indexing is on, + proceed with xpath idx */ + if ((tl = tll[i])) + { + for (; tl; tl = tl->next) + { + if (!tl->att) + do_xpindex = 1; + } + } + + if (do_xpindex) { + + /* attribute (no value) */ + wrd->reg_type = '0'; + wrd->attrUse = 3; + wrd->string = xp->name; + wrd->length = strlen(xp->name); + + wrd->seqno--; + (*p->tokenAdd)(wrd); + + if (xp->value && + strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2) { + + /* attribute value exact */ + strcpy (comb, xp->name); + strcat (comb, "="); + strcat (comb, xp->value); + + wrd->attrUse = 3; + wrd->reg_type = '0'; + wrd->string = comb; + wrd->length = strlen(comb); + wrd->seqno--; + + (*p->tokenAdd)(wrd); + } + } + i++; + } + + i = 0; + for (xp = n->u.tag.attributes; xp; xp = xp->next) { + data1_termlist *tl; + char attr_tag_path_full[1024]; + int int_len = flen; + int xpdone = 0; + + sprintf (attr_tag_path_full, "@%s/%.*s", + xp->name, int_len, tag_path_full); + + if ((tl = tll[i])) + { + /* If there is a termlist given (=xelm directive) */ + for (; tl; tl = tl->next) + { + if (!tl->att) { + /* add xpath index for the attribute */ + index_xpath_attr (attr_tag_path_full, xp->name, + xp->value, tl->structure, + p, wrd); + xpdone = 1; + } else { + /* add attribute based index for the attribute */ + if (xp->value) { + wrd->attrSet = (int) + (tl->att->parent->reference); + wrd->attrUse = tl->att->locals->local; + wrd->reg_type = *tl->structure; + wrd->string = xp->value; + wrd->length = strlen(xp->value); + (*p->tokenAdd)(wrd); + } + } + } + } + /* if there was no termlist for the given path, + or the termlist didn't have a ! element, index + the attribute as "w" */ + if ((!xpdone) && (!termlist_only)) + { + index_xpath_attr (attr_tag_path_full, xp->name, + xp->value, "w", p, wrd); + } + i++; + } + } + } } - free (h); } -static void index_tag (data1_node *par, data1_node *n, - struct recExtractCtrl *p, int level, RecWord *wrd) +static void index_termlist (data1_node *par, data1_node *n, + struct recExtractCtrl *p, int level, RecWord *wrd) { data1_termlist *tlist = 0; data1_datatype dtype = DATA1K_string; + /* * cycle up towards the root until we find a tag with an att.. * this has the effect of indexing locally defined tags with @@ -327,36 +676,20 @@ static void index_tag (data1_node *par, data1_node *n, */ while (!par->u.tag.element) - if (!par->parent || !(par=get_parent_tag(p->dh, par->parent))) - break; + if (!par->parent || !(par=get_parent_tag(p->dh, par->parent))) + break; if (!par || !(tlist = par->u.tag.element->termlists)) - return; + return; if (par->u.tag.element->tag) - dtype = par->u.tag.element->tag->kind; + dtype = par->u.tag.element->tag->kind; for (; tlist; tlist = tlist->next) { - char xattr[512]; /* consider source */ wrd->string = 0; - - if (!strcmp (tlist->source, "data") && n->which == DATA1N_data) - { - wrd->string = n->u.data.data; - wrd->length = n->u.data.len; - } - else if (sscanf (tlist->source, "attr(%511[^)])", xattr) == 1 && - n->which == DATA1N_tag) - { - data1_xattr *p = n->u.tag.attributes; - while (p && strcmp (p->name, xattr)) - p = p->next; - if (p) - { - wrd->string = p->value; - wrd->length = strlen(p->value); - } - } + assert(tlist->source); + sp_parse(n, wrd, tlist->source); + if (wrd->string) { if (p->flagShowRecords) @@ -368,11 +701,11 @@ static void index_tag (data1_node *par, data1_node *n, tlist->att->parent->name, tlist->att->name, tlist->att->value, tlist->source); - printf (" data=\""); - for (i = 0; ilength && i < 8; i++) + printf (" XData:\""); + for (i = 0; ilength && i < 40; i++) fputc (wrd->string[i], stdout); fputc ('"', stdout); - if (wrd->length > 8) + if (wrd->length > 40) printf (" ..."); fputc ('\n', stdout); } @@ -387,10 +720,9 @@ static void index_tag (data1_node *par, data1_node *n, } } -static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level) +static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level, + RecWord *wrd) { - RecWord wrd; - (*p->init)(p, &wrd); /* set defaults */ for (; n; n = n->next) { if (p->flagShowRecords) /* display element description to user */ @@ -398,7 +730,7 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level) if (n->which == DATA1N_root) { printf("%*s", level * 4, ""); - printf("Record type: '%s'\n", n->u.root.absyn->name); + printf("Record type: '%s'\n", n->u.root.type); } else if (n->which == DATA1N_tag) { @@ -430,14 +762,18 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level) } } + if (n->which == DATA1N_tag) + { + index_termlist (n, n, p, level, wrd); + /* index start tag */ + if (n->root->u.root.absyn) + index_xpath (n, p, level, wrd, 1); + } + if (n->child) - if (dumpkeys(n->child, p, level + 1) < 0) + if (dumpkeys(n->child, p, level + 1, wrd) < 0) return -1; - if (n->which == DATA1N_tag) - { - index_tag (n, n, p, level, &wrd); - } if (n->which == DATA1N_data) { @@ -447,9 +783,9 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level) { printf("%*s", level * 4, ""); printf("Data: "); - if (n->u.data.len > 32) - printf("'%.24s ... %.6s'\n", n->u.data.data, - n->u.data.data + n->u.data.len-6); + if (n->u.data.len > 256) + printf("'%.170s ... %.70s'\n", n->u.data.data, + n->u.data.data + n->u.data.len-70); else if (n->u.data.len > 0) printf("'%.*s'\n", n->u.data.len, n->u.data.data); else @@ -457,8 +793,17 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level) } if (par) - index_tag (par, n, p, level, &wrd); + index_termlist (par, n, p, level, wrd); + + index_xpath (n, p, level, wrd, 1016); } + + if (n->which == DATA1N_tag) + { + /* index end tag */ + index_xpath (n, p, level, wrd, 2); + } + if (p->flagShowRecords && n->which == DATA1N_root) { printf("%*s-------------\n\n", level * 4, ""); @@ -471,24 +816,31 @@ int grs_extract_tree(struct recExtractCtrl *p, data1_node *n) { oident oe; int oidtmp[OID_SIZE]; + RecWord wrd; oe.proto = PROTO_Z3950; oe.oclass = CLASS_SCHEMA; - oe.value = n->u.root.absyn->reference; - - if ((oid_ent_to_oid (&oe, oidtmp))) - (*p->schemaAdd)(p, oidtmp); + if (n->u.root.absyn) + { + oe.value = n->u.root.absyn->reference; + + if ((oid_ent_to_oid (&oe, oidtmp))) + (*p->schemaAdd)(p, oidtmp); + } + (*p->init)(p, &wrd); - return dumpkeys(n, p, 0); + return dumpkeys(n, p, 0, &wrd); } -static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p, - NMEM mem) +static int grs_extract_sub(void *clientData, struct recExtractCtrl *p, + NMEM mem, + data1_node *(*grs_read)(struct grs_read_info *)) { data1_node *n; struct grs_read_info gri; oident oe; int oidtmp[OID_SIZE]; + RecWord wrd; gri.readf = p->readf; gri.seekf = p->seekf; @@ -498,35 +850,48 @@ static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p, gri.offset = p->offset; gri.mem = mem; gri.dh = p->dh; + gri.clientData = clientData; - if (read_grs_type (h, &gri, p->subType, &n)) - return RECCTRL_EXTRACT_ERROR; + n = (*grs_read)(&gri); if (!n) return RECCTRL_EXTRACT_EOF; oe.proto = PROTO_Z3950; oe.oclass = CLASS_SCHEMA; +#if 0 if (!n->u.root.absyn) return RECCTRL_EXTRACT_ERROR; - oe.value = n->u.root.absyn->reference; - if ((oid_ent_to_oid (&oe, oidtmp))) - (*p->schemaAdd)(p, oidtmp); +#endif + if (n->u.root.absyn) + { + oe.value = n->u.root.absyn->reference; + if ((oid_ent_to_oid (&oe, oidtmp))) + (*p->schemaAdd)(p, oidtmp); + } + data1_concat_text(p->dh, mem, n); - if (dumpkeys(n, p, 0) < 0) + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, n, "UTF-8", data1_get_encoding(p->dh, n)); + +#if 0 + data1_pr_tree (p->dh, n, stdout); +#endif + + (*p->init)(p, &wrd); + if (dumpkeys(n, p, 0, &wrd) < 0) { data1_free_tree(p->dh, n); - return RECCTRL_EXTRACT_ERROR; + return RECCTRL_EXTRACT_ERROR_GENERIC; } data1_free_tree(p->dh, n); return RECCTRL_EXTRACT_OK; } -static int grs_extract(void *clientData, struct recExtractCtrl *p) +int zebra_grs_extract(void *clientData, struct recExtractCtrl *p, + data1_node *(*grs_read)(struct grs_read_info *)) { int ret; NMEM mem = nmem_create (); - struct grs_handlers *h = (struct grs_handlers *) clientData; - - ret = grs_extract_sub(h, p, mem); + ret = grs_extract_sub(clientData, p, mem, grs_read); nmem_destroy(mem); return ret; } @@ -534,7 +899,8 @@ static int grs_extract(void *clientData, struct recExtractCtrl *p) /* * Return: -1: Nothing done. 0: Ok. >0: Bib-1 diagnostic. */ -static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c) +static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c, + char **addinfo, ODR o) { data1_esetname *eset; Z_Espec1 *espec = 0; @@ -548,10 +914,11 @@ static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c) if (!(eset = data1_getesetbyname(dh, n->u.root.absyn, c->u.simple->u.generic))) { - logf(LOG_LOG, "Unknown esetname '%s'", c->u.simple->u.generic); + yaz_log(YLOG_LOG, "Unknown esetname '%s'", c->u.simple->u.generic); + *addinfo = odr_strdup(o, c->u.simple->u.generic); return 25; /* invalid esetname */ } - logf(LOG_DEBUG, "Esetname '%s' in simple compspec", + yaz_log(YLOG_DEBUG, "Esetname '%s' in simple compspec", c->u.simple->u.generic); espec = eset->spec; break; @@ -568,23 +935,24 @@ static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c) data1_getesetbyname(dh, n->u.root.absyn, p->u.elementSetName))) { - logf(LOG_LOG, "Unknown esetname '%s'", + yaz_log(YLOG_DEBUG, "Unknown esetname '%s'", p->u.elementSetName); + *addinfo = odr_strdup(o, p->u.elementSetName); return 25; /* invalid esetname */ } - logf(LOG_DEBUG, "Esetname '%s' in complex compspec", + yaz_log(YLOG_DEBUG, "Esetname '%s' in complex compspec", p->u.elementSetName); espec = eset->spec; break; case Z_ElementSpec_externalSpec: if (p->u.externalSpec->which == Z_External_espec1) { - logf(LOG_DEBUG, "Got Espec-1"); + yaz_log(YLOG_DEBUG, "Got Espec-1"); espec = p->u.externalSpec-> u.espec1; } else { - logf(LOG_LOG, "Unknown external espec."); + yaz_log(YLOG_LOG, "Unknown external espec."); return 25; /* bad. what is proper diagnostic? */ } break; @@ -596,27 +964,78 @@ static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c) } if (espec) { - logf (LOG_DEBUG, "Element: Espec-1 match"); + yaz_log(YLOG_DEBUG, "Element: Espec-1 match"); return data1_doespec1(dh, n, espec); } else { - logf (LOG_DEBUG, "Element: all match"); + yaz_log(YLOG_DEBUG, "Element: all match"); return -1; } } -static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) +/* Add Zebra info in separate namespace ... + + 359 + 447 + records/genera.xml + + +*/ + +static void zebra_xml_metadata (struct recRetrieveCtrl *p, data1_node *top, + NMEM mem) +{ + const char *idzebra_ns[3]; + const char *i2 = "\n "; + const char *i4 = "\n "; + data1_node *n; + + idzebra_ns[0] = "xmlns"; + idzebra_ns[1] = "http://www.indexdata.dk/zebra/"; + idzebra_ns[2] = 0; + + data1_mk_text (p->dh, mem, i2, top); + + n = data1_mk_tag (p->dh, mem, "idzebra", idzebra_ns, top); + + data1_mk_text (p->dh, mem, "\n", top); + + data1_mk_text (p->dh, mem, i4, n); + + data1_mk_tag_data_int (p->dh, n, "size", p->recordSize, mem); + + if (p->score != -1) + { + data1_mk_text (p->dh, mem, i4, n); + data1_mk_tag_data_int (p->dh, n, "score", p->score, mem); + } + data1_mk_text (p->dh, mem, i4, n); + data1_mk_tag_data_zint (p->dh, n, "localnumber", p->localno, mem); + if (p->fname) + { + data1_mk_text (p->dh, mem, i4, n); + data1_mk_tag_data_text(p->dh, n, "filename", p->fname, mem); + } + data1_mk_text (p->dh, mem, i2, n); +} + +int zebra_grs_retrieve(void *clientData, struct recRetrieveCtrl *p, + data1_node *(*grs_read)(struct grs_read_info *)) { - data1_node *node = 0, *onode = 0; + data1_node *node = 0, *onode = 0, *top; data1_node *dnew; data1_maptab *map; int res, selected = 0; NMEM mem; struct grs_read_info gri; - char *tagname; - struct grs_handlers *h = (struct grs_handlers *) clientData; + const char *tagname; + int requested_schema = VAL_NONE; + data1_marctab *marctab; + int dummy; mem = nmem_create(); gri.readf = p->readf; @@ -627,70 +1046,79 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) gri.offset = 0; gri.mem = mem; gri.dh = p->dh; + gri.clientData = clientData; - logf (LOG_DEBUG, "grs_retrieve"); - if (read_grs_type (h, &gri, p->subType, &node)) - { - p->diagnostic = 14; - nmem_destroy (mem); - return 0; - } + yaz_log(YLOG_DEBUG, "grs_retrieve"); + node = (*grs_read)(&gri); if (!node) { p->diagnostic = 14; nmem_destroy (mem); return 0; } + data1_concat_text(p->dh, mem, node); + #if 0 data1_pr_tree (p->dh, node, stdout); #endif - logf (LOG_DEBUG, "grs_retrieve: size"); - if ((dnew = data1_insert_taggeddata(p->dh, node, node, - "size", mem))) + top = data1_get_root_tag (p->dh, node); + + yaz_log(YLOG_DEBUG, "grs_retrieve: size"); + tagname = data1_systag_lookup(node->u.root.absyn, "size", "size"); + if (tagname && + (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem))) { dnew->u.data.what = DATA1I_text; dnew->u.data.data = dnew->lbuf; sprintf(dnew->u.data.data, "%d", p->recordSize); dnew->u.data.len = strlen(dnew->u.data.data); } - - tagname = res_get_def(p->res, "tagrank", "rank"); - if (strcmp(tagname, "0") && p->score >= 0 && - (dnew = data1_insert_taggeddata(p->dh, node, node, tagname, mem))) + + tagname = data1_systag_lookup(node->u.root.absyn, "rank", "rank"); + if (tagname && p->score >= 0 && + (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem))) { - logf (LOG_DEBUG, "grs_retrieve: %s", tagname); + yaz_log(YLOG_DEBUG, "grs_retrieve: %s", tagname); dnew->u.data.what = DATA1I_num; dnew->u.data.data = dnew->lbuf; sprintf(dnew->u.data.data, "%d", p->score); dnew->u.data.len = strlen(dnew->u.data.data); } - tagname = res_get_def(p->res, "tagsysno", "localControlNumber"); - if (strcmp(tagname, "0") && p->localno > 0 && - (dnew = data1_insert_taggeddata(p->dh, node, node, tagname, mem))) + tagname = data1_systag_lookup(node->u.root.absyn, "sysno", + "localControlNumber"); + if (tagname && p->localno > 0 && + (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem))) { - logf (LOG_DEBUG, "grs_retrieve: %s", tagname); + yaz_log(YLOG_DEBUG, "grs_retrieve: %s", tagname); dnew->u.data.what = DATA1I_text; dnew->u.data.data = dnew->lbuf; - sprintf(dnew->u.data.data, "%d", p->localno); + + sprintf(dnew->u.data.data, ZINT_FORMAT, p->localno); dnew->u.data.len = strlen(dnew->u.data.data); } + if (p->input_format == VAL_TEXT_XML) + zebra_xml_metadata (p, top, mem); + +#if 0 + data1_pr_tree (p->dh, node, stdout); +#endif if (p->comp && p->comp->which == Z_RecordComp_complex && p->comp->u.complex->generic && - p->comp->u.complex->generic->schema) + p->comp->u.complex->generic->which == Z_Schema_oid && + p->comp->u.complex->generic->schema.oid) { - oident *oe = oid_getentbyoid (p->comp->u.complex->generic->schema); + oident *oe = oid_getentbyoid (p->comp->u.complex->generic->schema.oid); if (oe) requested_schema = oe->value; } - /* If schema has been specified, map if possible, then check that * we got the right one */ if (requested_schema != VAL_NONE) { - logf (LOG_DEBUG, "grs_retrieve: schema mapping"); + yaz_log(YLOG_DEBUG, "grs_retrieve: schema mapping"); for (map = node->u.root.absyn->maptabs; map; map = map->next) { if (map->target_absyn_ref == requested_schema) @@ -718,22 +1146,23 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) * the overlap of schema and formatting which is inherent in the MARC * family) */ - logf (LOG_DEBUG, "grs_retrieve: syntax mapping"); - for (map = node->u.root.absyn->maptabs; map; map = map->next) - { - if (map->target_absyn_ref == p->input_format) - { - onode = node; - if (!(node = data1_map_record(p->dh, onode, map, mem))) - { - p->diagnostic = 14; - nmem_destroy (mem); - return 0; - } - break; - } - } - logf (LOG_DEBUG, "grs_retrieve: schemaIdentifier"); + yaz_log(YLOG_DEBUG, "grs_retrieve: syntax mapping"); + if (node->u.root.absyn) + for (map = node->u.root.absyn->maptabs; map; map = map->next) + { + if (map->target_absyn_ref == p->input_format) + { + onode = node; + if (!(node = data1_map_record(p->dh, onode, map, mem))) + { + p->diagnostic = 14; + nmem_destroy (mem); + return 0; + } + break; + } + } + yaz_log(YLOG_DEBUG, "grs_retrieve: schemaIdentifier"); if (node->u.root.absyn && node->u.root.absyn->reference != VAL_NONE && p->input_format == VAL_GRS1) @@ -760,10 +1189,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) sprintf(p, "%d", *ii); p += strlen(p); } - *(p++) = '\0'; - - if ((dnew = data1_insert_taggeddata(dh, node, node, - "schemaIdentifier", mem))) + if ((dnew = data1_mk_tag_data_wd(dh, top, + "schemaIdentifier", mem))) { dnew->u.data.what = DATA1I_oid; dnew->u.data.data = (char *) nmem_malloc(mem, p - tmp); @@ -773,8 +1200,9 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) } } - logf (LOG_DEBUG, "grs_retrieve: element spec"); - if (p->comp && (res = process_comp(p->dh, node, p->comp)) > 0) + yaz_log(YLOG_DEBUG, "grs_retrieve: element spec"); + if (p->comp && (res = process_comp(p->dh, node, p->comp, &p->addinfo, + p->odr)) > 0) { p->diagnostic = res; if (onode) @@ -789,14 +1217,19 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) #if 0 data1_pr_tree (p->dh, node, stdout); #endif - logf (LOG_DEBUG, "grs_retrieve: transfer syntax mapping"); + yaz_log(YLOG_DEBUG, "grs_retrieve: transfer syntax mapping"); switch (p->output_format = (p->input_format != VAL_NONE ? p->input_format : VAL_SUTRS)) { - data1_marctab *marctab; - int dummy; - case VAL_TEXT_XML: +#if 0 + data1_pr_tree (p->dh, node, stdout); +#endif + /* default output encoding for XML is UTF-8 */ + data1_iconv (p->dh, mem, node, + p->encoding ? p->encoding : "UTF-8", + data1_get_encoding(p->dh, node)); + if (!(p->rec_buf = data1_nodetoidsgml(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -808,6 +1241,7 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) } break; case VAL_GRS1: + data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node)); dummy = 0; if (!(p->rec_buf = data1_nodetogr(p->dh, node, selected, p->odr, &dummy))) @@ -816,6 +1250,9 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->rec_len = (size_t) (-1); break; case VAL_EXPLAIN: + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node)); + if (!(p->rec_buf = data1_nodetoexplain(p->dh, node, selected, p->odr))) p->diagnostic = 238; @@ -823,6 +1260,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->rec_len = (size_t) (-1); break; case VAL_SUMMARY: + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node)); if (!(p->rec_buf = data1_nodetosummary(p->dh, node, selected, p->odr))) p->diagnostic = 238; @@ -830,6 +1269,9 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->rec_len = (size_t) (-1); break; case VAL_SUTRS: + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, + data1_get_encoding(p->dh, node)); if (!(p->rec_buf = data1_nodetobuf(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -841,6 +1283,9 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) } break; case VAL_SOIF: + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, + data1_get_encoding(p->dh, node)); if (!(p->rec_buf = data1_nodetosoif(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -866,6 +1311,9 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->diagnostic = 238; break; } + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, + data1_get_encoding(p->dh, node)); if (!(p->rec_buf = data1_nodetomarc(p->dh, marctab, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -884,13 +1332,3 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) return 0; } -static struct recType grs_type = -{ - "grs", - grs_init, - grs_destroy, - grs_extract, - grs_retrieve -}; - -RecType recTypeGrs = &grs_type;