From: Adam Dickmeiss Date: Mon, 27 Oct 1997 14:33:03 +0000 (+0000) Subject: Moved towards generic character mapping depending on "structure" X-Git-Tag: ZEBRA.1.0~291 X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=commitdiff_plain;h=3d01abbe1238c82d62e4c01eb7cac760d3d53377 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. --- diff --git a/CHANGELOG b/CHANGELOG index 74a063c..ba53edc 100644 --- a/CHANGELOG +++ b/CHANGELOG @@ -1,3 +1,6 @@ +Moved towards generic character mapping. Type w, p uses +string.chr, u uses urx.chr and n uses numeric.chr. + Added support for C++, headers uses extern "C" for public definitions. New filter grs.marc. that reads MARC records in the ISO2709 diff --git a/dict/lookgrep.c b/dict/lookgrep.c index f3d8421..3e094e1 100644 --- a/dict/lookgrep.c +++ b/dict/lookgrep.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: lookgrep.c,v $ - * Revision 1.19 1997-09-18 08:59:18 adam + * Revision 1.20 1997-10-27 14:33:03 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 1.19 1997/09/18 08:59:18 adam * Extra generic handle for the character mapping routines. * * Revision 1.18 1997/09/05 15:29:58 adam @@ -413,7 +419,12 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, struct DFA *dfa = dfa_init(); int i, d; - logf (LOG_DEBUG, "dict_lookup_grep '%s' range=%d", pattern, range); + logf (LOG_DEBUG, "dict_lookup_grep range=%d", range); + for (i = 0; pattern[i]; i++) + { + logf (LOG_DEBUG, " %3d %c", pattern[i], + (pattern[i] > ' ' && pattern[i] < 127) ? pattern[i] : '?'); + } dfa_set_cmap (dfa, dict->grep_cmap_data, dict->grep_cmap); diff --git a/dict/scan.c b/dict/scan.c index 44ff661..d23e451 100644 --- a/dict/scan.c +++ b/dict/scan.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: scan.c,v $ - * Revision 1.8 1996-02-02 13:43:52 adam + * Revision 1.9 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 1.8 1996/02/02 13:43:52 adam * The public functions simply use char instead of Dict_char to represent * search strings. Dict_char is used internally only. * @@ -203,6 +209,14 @@ int dict_scan_r (Dict dict, Dict_ptr ptr, int pos, Dict_char *str, int dict_scan (Dict dict, char *str, int *before, int *after, void *client, int (*f)(char *name, const char *info, int pos, void *client)) { + int i; + + logf (LOG_DEBUG, "dict_scan"); + for (i = 0; str[i]; i++) + { + logf (LOG_DEBUG, " %3d %c", str[i], + (str[i] > ' ' && str[i] < 127) ? str[i] : '?'); + } if (dict->head.last <= 1) return 0; return dict_scan_r (dict, 1, 0, (Dict_char *) str, before, after, client, diff --git a/include/charmap.h b/include/charmap.h index b653e97..54a90d9 100644 --- a/include/charmap.h +++ b/include/charmap.h @@ -36,7 +36,13 @@ * OF THIS SOFTWARE. * * $Log: charmap.h,v $ - * Revision 1.3 1997-09-05 15:29:59 adam + * Revision 1.4 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 1.3 1997/09/05 15:29:59 adam * Changed prototype for chr_map_input - added const. * Added support for C++, headers uses extern "C" for public definitions. * @@ -56,18 +62,15 @@ extern const char *CHR_BASE; struct chr_t_entry; typedef struct chr_t_entry chr_t_entry; -typedef struct chrmaptab -{ - chr_t_entry *input; /* mapping table for input data */ - chr_t_entry *query_equiv; /* mapping table for queries */ - unsigned char *output[256]; /* return mapping - for display of registers */ - int base_uppercase; /* Start of upper-case ordinals */ -} chrmaptab, *CHRMAPTAB; +typedef struct chrmaptab_info *chrmaptab; + +chrmaptab chrmaptab_create(const char *tabpath, const char *name, + int map_only); +void chrmaptab_destroy (chrmaptab tab); + +const char **chr_map_input(chrmaptab t, const char **from, int len); -chrmaptab *chr_read_maptab(const char *tabpath, const char *name); -int chr_map_chrs(chr_t_entry *t, char **from, int len, int *read, char **to, - int max); -const char **chr_map_input(chr_t_entry *t, const char **from, int len); +const char *chr_map_output(chrmaptab t, const char **from, int len); #ifdef __cplusplus } diff --git a/include/recctrl.h b/include/recctrl.h index 93de77b..722bbf7 100644 --- a/include/recctrl.h +++ b/include/recctrl.h @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: recctrl.h,v $ - * Revision 1.21 1997-09-18 08:59:19 adam + * Revision 1.22 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 1.21 1997/09/18 08:59:19 adam * Extra generic handle for the character mapping routines. * * Revision 1.20 1997/09/17 12:19:10 adam @@ -85,33 +91,19 @@ #include #include #include +#include #ifdef __cplusplus extern "C" { #endif -typedef enum { - Word_String, - Word_Phrase, - Word_Numeric -} RecWordType; /* single word entity */ typedef struct { int attrSet; int attrUse; - RecWordType which; -#if 0 - enum { - Word_String, - Word_Phrase, - Word_Numeric - } which; -#endif - union { - char *string; - int numeric; - } u; + unsigned reg_type; + char *string; int seqno; } RecWord; @@ -126,7 +118,7 @@ struct recExtractCtrl { char *subType; void (*init)(RecWord *p); void (*add)(const RecWord *p); - const char **(*map_chrs_input)(void *vp, const char **from, int len); + ZebraMaps zebra_maps; int flagShowRecords; data1_handle dh; }; diff --git a/include/zebramap.h b/include/zebramap.h new file mode 100644 index 0000000..20acd94 --- /dev/null +++ b/include/zebramap.h @@ -0,0 +1,40 @@ +/* + * Copyright (C) 1994-1997, Index Data I/S + * All rights reserved. + * Sebastian Hammer, Adam Dickmeiss + * + * $Log: zebramap.h,v $ + * Revision 1.1 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. + * + */ + +#ifndef ZEBRAMAP_H +#define ZEBRAMAP_H + +#include + +#ifdef __cplusplus +extern "C" { +#endif + +typedef struct zebra_maps *ZebraMaps; +ZebraMaps zebra_maps_open (const char *tabpath); + +void zebra_maps_close (ZebraMaps zm); + +const char **zebra_maps_input (ZebraMaps zms, int reg_type, + const char **from, int len); +const char *zebra_maps_output(ZebraMaps, int reg_type, const char **from); + +int zebra_maps_attr (ZebraMaps zms, Z_AttributesPlusTerm *zapt, + int *reg_type, char **search_type, int *complete_flag); + +#ifdef __cplusplus +} +#endif + +#endif diff --git a/include/zebrautl.h b/include/zebrautl.h index 84e21f0..fd8aa20 100644 --- a/include/zebrautl.h +++ b/include/zebrautl.h @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zebrautl.h,v $ - * Revision 1.3 1997-09-17 12:19:11 adam + * Revision 1.4 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 1.3 1997/09/17 12:19:11 adam * Zebra version corresponds to YAZ version 1.4. * Changed Zebra server so that it doesn't depend on global common_resource. * @@ -21,10 +27,7 @@ #ifndef ZEBRA_UTIL_H #define ZEBRA_UTIL_H -#include -#include -#include +#include #include -#include #endif diff --git a/index/Makefile b/index/Makefile index 5107e04..b97c001 100644 --- a/index/Makefile +++ b/index/Makefile @@ -1,7 +1,7 @@ # Copyright (C) 1995-1996, Index Data I/S # All rights reserved. # Sebastian Hammer, Adam Dickmeiss -# $Id: Makefile,v 1.43 1997-09-04 13:59:28 adam Exp $ +# $Id: Makefile,v 1.44 1997-10-27 14:33:04 adam Exp $ SHELL=/bin/sh RANLIB=ranlib @@ -20,10 +20,10 @@ TPROG3=zebrasrv DEFS=$(INCLUDE) O1 = main.o dir.o dirs.o trav.o extract.o kinput.o kcompare.o \ symtab.o recindex.o recstat.o lockutil.o lockidx.o \ - zinfo.o invstat.o keychars.o + zinfo.o invstat.o O2 = kdump.o O3 = zserver.o kcompare.o zrpn.o zsets.o attribute.o recindex.o \ - zlogs.o lockutil.o locksrv.o zinfo.o keychars.o trunc.o + zlogs.o lockutil.o locksrv.o zinfo.o trunc.o CPP=$(CC) -E all: $(TPROG1) $(TPROG2) $(TPROG3) diff --git a/index/extract.c b/index/extract.c index a085d72..7374470 100644 --- a/index/extract.c +++ b/index/extract.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: extract.c,v $ - * Revision 1.75 1997-09-17 12:19:12 adam + * 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. * @@ -525,7 +531,7 @@ static void wordInit (RecWord *p) { p->attrSet = 1; p->attrUse = 1016; - p->which = Word_String; + p->reg_type = 'w'; } struct recKeys { @@ -589,19 +595,9 @@ 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) @@ -1002,7 +998,7 @@ static int recordExtract (SYSNO *sysno, const char *fname, extractCtrl.seekf = file_seek; extractCtrl.tellf = file_tell; extractCtrl.endf = file_end; - extractCtrl.map_chrs_input = map_chrs_input; + extractCtrl.zebra_maps = rGroup->zebra_maps; extractCtrl.flagShowRecords = rGroup->flagShowRecords; if (rGroup->flagShowRecords) printf ("File: %s %ld\n", fname, (long) recordOffset); diff --git a/index/index.h b/index/index.h index 2b5e1de..684b10e 100644 --- a/index/index.h +++ b/index/index.h @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: index.h,v $ - * Revision 1.54 1997-09-29 09:08:36 adam + * Revision 1.55 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 1.54 1997/09/29 09:08:36 adam * Revised locking system to be thread safe for the server. * * Revision 1.53 1997/09/25 14:54:43 adam @@ -195,6 +201,7 @@ #include #include #include +#include #include #include @@ -227,6 +234,7 @@ struct dirs_entry { time_t mtime; }; + struct recordGroup { char *groupName; char *databaseName; @@ -239,6 +247,7 @@ struct recordGroup { int fileVerboseLimit; data1_handle dh; BFiles bfs; + ZebraMaps zebra_maps; }; void getFnameTmp (char *fname, int no); @@ -315,8 +324,5 @@ int zebra_unlock (ZebraLockHandle h); int zebra_lock_fd (ZebraLockHandle h); void zebra_lock_prefix (Res res, char *dst); -void init_charmap(Res res); -const char **map_chrs_input(void *vp, const char **from, int len); -const char *map_chrs_output(const char **from); extern Res common_resource; diff --git a/index/kdump.c b/index/kdump.c index c4a7279..40ab26f 100644 --- a/index/kdump.c +++ b/index/kdump.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: kdump.c,v $ - * Revision 1.13 1997-09-09 13:38:07 adam + * Revision 1.14 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 1.13 1997/09/09 13:38:07 adam * Partial port to WIN95/NT. * * Revision 1.12 1997/09/05 09:52:32 adam @@ -128,7 +134,7 @@ int main (int argc, char **argv) char key_info[256]; FILE *inf; struct it_key prevk; - chrmaptab *map = 0; + chrmaptab map = 0; prevk.sysno = 0; prevk.seqno = 0; @@ -146,7 +152,7 @@ int main (int argc, char **argv) } else if (ret == 'm') { - if (!(map = chr_read_maptab (NULL, arg))) + if (!(map = chrmaptab_create (NULL, arg, 0))) { logf(LOG_FATAL, "Failed to open maptab"); exit(1); @@ -182,7 +188,7 @@ int main (int argc, char **argv) while (*from) { - char *res = (char*)map->output[(unsigned char) *(from++)]; + char *res = chr_map_output(map, from, 1); while (*res) *(to++) = *(res++); } diff --git a/index/main.c b/index/main.c index 4c5ae09..eb4533b 100644 --- a/index/main.c +++ b/index/main.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: main.c,v $ - * Revision 1.50 1997-09-25 14:55:52 adam + * Revision 1.51 1997-10-27 14:33:05 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 1.50 1997/09/25 14:55:52 adam * Minor changes. * * Revision 1.49 1997/09/17 12:19:15 adam @@ -227,6 +233,7 @@ int main (int argc, char **argv) rGroupDef.flagStoreKeys = -1; rGroupDef.flagShowRecords = 0; rGroupDef.fileVerboseLimit = 100000; + rGroupDef.zebra_maps = NULL; rGroupDef.dh = data1_create (); prog = *argv; @@ -280,7 +287,8 @@ int main (int argc, char **argv) bf_lockDir (rGroupDef.bfs, res_get (common_resource, "lockDir")); - init_charmap(common_resource); + rGroupDef.zebra_maps = zebra_maps_open (res_get( + common_resource, "profilePath")); } if (!strcmp (arg, "update")) cmd = 'u'; diff --git a/index/zinfo.c b/index/zinfo.c index 3aeeee2..17a20e4 100644 --- a/index/zinfo.c +++ b/index/zinfo.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zinfo.c,v $ - * Revision 1.4 1997-09-25 14:57:08 adam + * Revision 1.5 1997-10-27 14:33:05 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 1.4 1997/09/25 14:57:08 adam * Added string.h. * * Revision 1.3 1996/05/22 08:21:59 adam @@ -302,3 +308,4 @@ void zebTargetInfo_setDB (ZebTargetInfo *zti, ZebDatabaseInfo *zdi) zti->curDatabaseInfo->dirty = 1; memcpy (&zti->curDatabaseInfo->info, zdi, sizeof(*zdi)); } + diff --git a/index/zrpn.c b/index/zrpn.c index 7ec5e01..415200f 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zrpn.c,v $ - * Revision 1.67 1997-09-29 09:06:10 adam + * Revision 1.68 1997-10-27 14:33:06 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 1.67 1997/09/29 09:06:10 adam * Removed one static var in order to make this module thread safe. * * Revision 1.66 1997/09/25 14:58:03 adam @@ -247,6 +253,25 @@ #include #include +struct rpn_char_map_info { + ZebraMaps zm; + int reg_type; +}; + +static const char **rpn_char_map_handler (void *vp, const char **from, int len) +{ + struct rpn_char_map_info *p = vp; + return zebra_maps_input (p->zm, p->reg_type, from, len); +} + +static void rpn_char_map_prepare (ZServerInfo *zi, int reg_type, + struct rpn_char_map_info *map_info) +{ + map_info->zm = zi->zebra_maps; + map_info->reg_type = reg_type; + dict_grep_cmap (zi->dict, map_info, rpn_char_map_handler); +} + typedef struct { int type; int major; @@ -362,7 +387,8 @@ static int grep_handle (char *name, const char *info, void *p) return 0; } -static int term_pre (const char **src, const char *ct1, const char *ct2) +static int term_pre (ZebraMaps zebra_maps, int reg_type, const char **src, + const char *ct1, const char *ct2) { const char *s1, *s0 = *src; const char **map; @@ -375,7 +401,7 @@ static int term_pre (const char **src, const char *ct1, const char *ct2) if (ct2 && strchr (ct2, *s0)) break; s1 = s0; - map = map_chrs_input (0, &s1, strlen(s1)); + map = zebra_maps_input (zebra_maps, reg_type, &s1, strlen(s1)); if (**map != *CHR_SPACE) break; s0 = s1; @@ -384,24 +410,25 @@ static int term_pre (const char **src, const char *ct1, const char *ct2) return *s0; } -static int term_100 (const char **src, char *dst, int space_split) +static int term_100 (ZebraMaps zebra_maps, int reg_type, + const char **src, char *dst, int space_split) { const char *s0, *s1; const char **map; int i = 0; - if (!term_pre (src, NULL, NULL)) + if (!term_pre (zebra_maps, reg_type, src, NULL, NULL)) return 0; s0 = *src; while (*s0) { s1 = s0; - map = map_chrs_input (0, &s0, strlen(s0)); + map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); if (space_split && **map == *CHR_SPACE) break; while (s1 < s0) { - if (!isalnum (*s1)) + if (!isalnum (*s1) && *s1 != '-') dst[i++] = '\\'; dst[i++] = *s1++; } @@ -411,13 +438,14 @@ static int term_100 (const char **src, char *dst, int space_split) return i; } -static int term_101 (const char **src, char *dst, int space_split) +static int term_101 (ZebraMaps zebra_maps, int reg_type, + const char **src, char *dst, int space_split) { const char *s0, *s1; const char **map; int i = 0; - if (!term_pre (src, "#", "#")) + if (!term_pre (zebra_maps, reg_type, src, "#", "#")) return 0; s0 = *src; while (*s0) @@ -431,7 +459,7 @@ static int term_101 (const char **src, char *dst, int space_split) else { s1 = s0; - map = map_chrs_input (0, &s0, strlen(s0)); + map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); if (space_split && **map == *CHR_SPACE) break; while (s1 < s0) @@ -448,13 +476,14 @@ static int term_101 (const char **src, char *dst, int space_split) } -static int term_103 (const char **src, char *dst, int *errors, int space_split) +static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, + char *dst, int *errors, int space_split) { int i = 0; const char *s0, *s1; const char **map; - if (!term_pre (src, "^\\()[].*+?|", "(")) + if (!term_pre (zebra_maps, reg_type, src, "^\\()[].*+?|", "(")) return 0; s0 = *src; if (errors && *s0 == '+' && s0[1] && s0[2] == '+' && s0[3] && @@ -472,7 +501,7 @@ static int term_103 (const char **src, char *dst, int *errors, int space_split) else { s1 = s0; - map = map_chrs_input (0, &s0, strlen(s0)); + map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); if (**map == *CHR_SPACE) break; while (s1 < s0) @@ -488,9 +517,10 @@ static int term_103 (const char **src, char *dst, int *errors, int space_split) return i; } -static int term_102 (const char **src, char *dst, int space_split) +static int term_102 (ZebraMaps zebra_maps, int reg_type, const char **src, + char *dst, int space_split) { - return term_103 (src, dst, NULL, space_split); + return term_103 (zebra_maps, reg_type, src, dst, NULL, space_split); } /* gen_regular_rel - generate regular expression from relation @@ -508,21 +538,21 @@ static void gen_regular_rel (char *dst, int val, int islt) if (val >= 0) { if (islt) - strcpy (dst, "(-[0-9]+|"); + strcpy (dst, "(-[0-9]+|("); else - strcpy (dst, "("); + strcpy (dst, "(("); } else { if (!islt) { - strcpy (dst, "([0-9]+|-"); + strcpy (dst, "([0-9]+|-("); dst_p = strlen (dst); islt = 1; } else { - strcpy (dst, "(-"); + strcpy (dst, "((-"); islt = 0; } val = -val; @@ -599,7 +629,7 @@ static void gen_regular_rel (char *dst, int val, int islt) strcat (dst, "[0-9]"); strcat (dst, "[0-9]*"); } - strcat (dst, ")"); + strcat (dst, "))"); } static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, @@ -607,58 +637,53 @@ static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, char *term_dict, oid_value attributeSet, struct grep_info *grep_info, - int *max_pos) + int *max_pos, + int reg_type) { AttrType relation; int relation_value; int term_value; int r; + char *term_tmp = term_dict + strlen(term_dict); attr_init (&relation, zapt, 2); relation_value = attr_find (&relation, NULL); + logf (LOG_DEBUG, "relation value=%d", relation_value); switch (relation_value) { case 1: - if (!term_100 (term_sub, term_dict, 1)) + if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) return 0; - term_value = atoi (term_dict); - if (term_value <= 0) - return 1; + term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation <"); - gen_regular_rel (term_dict + strlen(term_dict), term_value-1, 1); + gen_regular_rel (term_tmp, term_value-1, 1); break; case 2: - if (!term_100 (term_sub, term_dict, 1)) + if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) return 0; - term_value = atoi (term_dict); - if (term_value < 0) - return 1; + term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation <="); - gen_regular_rel (term_dict + strlen(term_dict), term_value, 1); + gen_regular_rel (term_tmp, term_value, 1); break; case 4: - if (!term_100 (term_sub, term_dict, 1)) + if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) return 0; - term_value = atoi (term_dict); - if (term_value < 0) - term_value = 0; + term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation >="); - gen_regular_rel (term_dict + strlen(term_dict), term_value, 0); + gen_regular_rel (term_tmp, term_value, 0); break; case 5: - if (!term_100 (term_sub, term_dict, 1)) + if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) return 0; - term_value = atoi (term_dict); - if (term_value < 0) - term_value = 0; + term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation >"); - gen_regular_rel (term_dict + strlen(term_dict), term_value+1, 0); + gen_regular_rel (term_tmp, term_value+1, 0); break; default: return 0; } - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); + logf (LOG_DEBUG, "dict_lookup_grep: %s", term_tmp); r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, max_pos, 0, grep_handle); if (r) @@ -668,9 +693,10 @@ static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, } static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - const char **term_sub, int regType, + const char **term_sub, oid_value attributeSet, struct grep_info *grep_info, - int num_bases, char **basenames, int space_split) + int reg_type, int complete_flag, + int num_bases, char **basenames) { char term_dict[2*IT_MAX_WORD+2]; int j, r, base_no; @@ -680,7 +706,10 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, int use_value; oid_value curAttributeSet = attributeSet; const char *termp; + struct rpn_char_map_info rcmi; + int space_split = complete_flag ? 0 : 1; + rpn_char_map_prepare (zi, reg_type, &rcmi); attr_init (&use, zapt, 1); use_value = attr_find (&use, &curAttributeSet); logf (LOG_DEBUG, "field_term, use value %d", use_value); @@ -734,10 +763,11 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, } term_dict[prefix_len++] = ')'; term_dict[prefix_len++] = 1; - term_dict[prefix_len++] = regType; + term_dict[prefix_len++] = reg_type; + logf (LOG_LOG, "reg_type = %d", term_dict[prefix_len-1]); term_dict[prefix_len] = '\0'; if (!relational_term (zi, zapt, &termp, term_dict, - attributeSet, grep_info, &max_pos)) + attributeSet, grep_info, &max_pos, reg_type)) { j = prefix_len; switch (truncation_value) @@ -745,7 +775,8 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, case -1: /* not specified */ case 100: /* do not truncate */ term_dict[j++] = '('; - if (!term_100 (&termp, term_dict + j, space_split)) + if (!term_100 (zi->zebra_maps, reg_type, + &termp, term_dict + j, space_split)) return 0; strcat (term_dict, ")"); r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, @@ -755,7 +786,8 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, break; case 1: /* right truncation */ term_dict[j++] = '('; - if (!term_100 (&termp, term_dict + j, space_split)) + if (!term_100 (zi->zebra_maps, reg_type, + &termp, term_dict + j, space_split)) return 0; strcat (term_dict, ".*)"); dict_lookup_grep (zi->dict, term_dict, 0, grep_info, @@ -767,7 +799,8 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, return -1; case 101: /* process # in term */ term_dict[j++] = '('; - if (!term_101 (&termp, term_dict + j, space_split)) + if (!term_101 (zi->zebra_maps, reg_type, + &termp, term_dict + j, space_split)) return 0; strcat (term_dict, ")"); r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, @@ -777,7 +810,8 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, break; case 102: /* Regexp-1 */ term_dict[j++] = '('; - if (!term_102 (&termp, term_dict + j, space_split)) + if (!term_102 (zi->zebra_maps, reg_type, + &termp, term_dict + j, space_split)) return 0; strcat (term_dict, ")"); logf (LOG_DEBUG, "Regexp-1 tolerance=%d", r); @@ -787,10 +821,11 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", r); break; - case 103: /* Regexp-1 */ + case 103: /* Regexp-2 */ r = 1; term_dict[j++] = '('; - if (!term_103 (&termp, term_dict + j, &r, space_split)) + if (!term_103 (zi->zebra_maps, reg_type, + &termp, term_dict + j, &r, space_split)) return 0; strcat (term_dict, ")"); logf (LOG_DEBUG, "Regexp-2 tolerance=%d", r); @@ -822,7 +857,7 @@ static void trans_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, } static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - char *termz) + char *termz, int reg_type) { Z_Term *term = zapt->term; const char **map; @@ -835,7 +870,7 @@ static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, while ((len = (cp_end - cp)) > 0) { - map = map_chrs_input (0, &cp, len); + map = zebra_maps_input (zi->zebra_maps, reg_type, &cp, len); if (**map == *CHR_SPACE) space_map = *map; else @@ -854,6 +889,7 @@ static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, static RSET rpn_search_APT_relevance (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, oid_value attributeSet, + int reg_type, int complete_flag, int num_bases, char **basenames) { rset_relevance_parms parms; @@ -887,8 +923,8 @@ static RSET rpn_search_APT_relevance (ZServerInfo *zi, grep_info.isam_p_buf = NULL; while (1) { - r = field_term (zi, zapt, &termp, 'w', attributeSet, &grep_info, - num_bases, basenames, 1); + r = field_term (zi, zapt, &termp, attributeSet, &grep_info, + reg_type, complete_flag, num_bases, basenames); if (r <= 0) break; #ifdef TERM_COUNT @@ -911,41 +947,6 @@ static RSET rpn_search_APT_relevance (ZServerInfo *zi, return result; } -static RSET rpn_search_APT_cphrase (ZServerInfo *zi, - Z_AttributesPlusTerm *zapt, - oid_value attributeSet, - int num_bases, char **basenames) -{ - char termz[IT_MAX_WORD+1]; - struct grep_info grep_info; - RSET result; - const char *termp = termz; - int r; - - if (zapt->term->which != Z_Term_general) - { - zi->errCode = 124; - return NULL; - } - trans_term (zi, zapt, termz); - -#ifdef TERM_COUNT - grep_info.term_no = 0; -#endif - grep_info.isam_p_indx = 0; - grep_info.isam_p_size = 0; - grep_info.isam_p_buf = NULL; - - r = field_term (zi, zapt, &termp, 'p', attributeSet, &grep_info, - num_bases, basenames, 0); - result = rset_trunc (zi, grep_info.isam_p_buf, grep_info.isam_p_indx); -#ifdef TERM_COUNT - xfree(grep_info.term_no); -#endif - xfree (grep_info.isam_p_buf); - return result; -} - static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, int ordered, int exclusion, int relation, int distance) @@ -1127,6 +1128,7 @@ static RSET rpn_prox (ZServerInfo *zi, RSET *rset, int rset_no) static RSET rpn_search_APT_phrase (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, oid_value attributeSet, + int reg_type, int complete_flag, int num_bases, char **basenames) { char termz[IT_MAX_WORD+1]; @@ -1149,10 +1151,11 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, grep_info.isam_p_buf = NULL; while (1) - { - grep_info.isam_p_indx = 0; - r = field_term (zi, zapt, &termp, 'w', attributeSet, &grep_info, - num_bases, basenames, 1); + { + logf (LOG_LOG, "APT_phrase termp=%s", termp); + grep_info.isam_p_indx = 0; + r = field_term (zi, zapt, &termp, attributeSet, &grep_info, + reg_type, complete_flag, num_bases, basenames); if (r < 1) break; rset[rset_no] = rset_trunc (zi, grep_info.isam_p_buf, @@ -1208,80 +1211,31 @@ static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, oid_value attributeSet, int num_bases, char **basenames) { - AttrType relation; - AttrType structure; - AttrType completeness; - int relation_value, structure_value, completeness_value; + int reg_type; + char *search_type = NULL; + int complete_flag; - attr_init (&relation, zapt, 2); - attr_init (&structure, zapt, 4); - attr_init (&completeness, zapt, 6); + zebra_maps_attr (zi->zebra_maps, zapt, ®_type, &search_type, + &complete_flag); - relation_value = attr_find (&relation, NULL); - structure_value = attr_find (&structure, NULL); - completeness_value = attr_find (&completeness, NULL); - switch (structure_value) + logf (LOG_DEBUG, "reg_type=%c", reg_type); + logf (LOG_DEBUG, "complete_flag=%d", complete_flag); + logf (LOG_DEBUG, "search_type=%s", search_type); + if (!strcmp (search_type, "phrase")) + { + return rpn_search_APT_phrase (zi, zapt, attributeSet, + reg_type, complete_flag, + num_bases, basenames); + } + else if (!strcmp (search_type, "ranked")) { - case -1: - if (relation_value == 102) /* relevance relation */ - return rpn_search_APT_relevance (zi, zapt, attributeSet, - num_bases, basenames); - if (completeness_value == 2 || completeness_value == 3) - return rpn_search_APT_cphrase (zi, zapt, attributeSet, - num_bases, basenames); - return rpn_search_APT_phrase (zi, zapt, attributeSet, - num_bases, basenames); - case 1: /* phrase */ - if (relation_value == 102) /* relevance relation */ - return rpn_search_APT_relevance (zi, zapt, attributeSet, - num_bases, basenames); - if (completeness_value == 2 || completeness_value == 3) - return rpn_search_APT_cphrase (zi, zapt, attributeSet, - num_bases, basenames); - return rpn_search_APT_phrase (zi, zapt, attributeSet, - num_bases, basenames); - break; - case 2: /* word */ - if (relation_value == 102) /* relevance relation */ - return rpn_search_APT_relevance (zi, zapt, attributeSet, - num_bases, basenames); - if (completeness_value == 2 || completeness_value == 3) - return rpn_search_APT_cphrase (zi, zapt, attributeSet, - num_bases, basenames); - return rpn_search_APT_phrase (zi, zapt, attributeSet, - num_bases, basenames); - case 3: /* key */ - break; - case 4: /* year */ - break; - case 5: /* date - normalized */ - break; - case 6: /* word list */ - return rpn_search_APT_relevance (zi, zapt, attributeSet, - num_bases, basenames); - case 100: /* date - un-normalized */ - break; - case 101: /* name - normalized */ - break; - case 102: /* date - un-normalized */ - break; - case 103: /* structure */ - break; - case 104: /* urx */ - break; - case 105: /* free-form-text */ - return rpn_search_APT_relevance (zi, zapt, attributeSet, - num_bases, basenames); - case 106: /* document-text */ return rpn_search_APT_relevance (zi, zapt, attributeSet, + reg_type, complete_flag, num_bases, basenames); - case 107: /* local-number */ + } + else if (!strcmp (search_type, "local")) + { return rpn_search_APT_local (zi, zapt, attributeSet); - case 108: /* string */ - return rpn_search_APT_phrase (zi, zapt, attributeSet, - num_bases, basenames); - case 109: /* numeric string */ - break; } zi->errCode = 118; return NULL; @@ -1458,7 +1412,6 @@ int rpn_search (ZServerInfo *zi, oident *attrset; oid_value attributeSet; - dict_grep_cmap (zi->dict, 0, map_chrs_input); zlog_rpn (rpn); zi->errCode = 0; @@ -1513,14 +1466,15 @@ static int scan_handle (char *name, const char *info, int pos, void *client) } -static void scan_term_untrans (ODR odr, char **dstp, const char *src) +static void scan_term_untrans (ZServerInfo *zi, int reg_type, + char **dstp, const char *src) { - char *dst = odr_malloc (odr, strlen(src)*2+1); + char *dst = odr_malloc (zi->odr, strlen(src)*2+1); *dstp = dst; while (*src) { - const char *cp = map_chrs_output (&src); + const char *cp = zebra_maps_output (zi->zebra_maps, reg_type, &src); while (*cp) *dst++ = *cp++; } @@ -1542,13 +1496,15 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, char termz[IT_MAX_WORD+20]; AttrType use; int use_value; - AttrType completeness; - int completeness_value; struct scan_info *scan_info_array; struct scan_entry *glist; int ords[32], ord_no = 0; int ptr[32]; + int reg_type; + char *search_type = NULL; + int complete_flag; + logf (LOG_DEBUG, "scan, position = %d, num = %d", pos, num); if (attributeset == VAL_NONE) @@ -1558,9 +1514,12 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, use_value = attr_find (&use, &attributeset); logf (LOG_DEBUG, "use value %d", use_value); - attr_init (&completeness, zapt, 6); - completeness_value = attr_find (&completeness, NULL); - logf (LOG_DEBUG, "completeness value %d", completeness_value); + if (zebra_maps_attr (zi->zebra_maps, zapt, ®_type, &search_type, + &complete_flag)) + { + zi->errCode = 113; + return zi->errCode; + } if (use_value == -1) use_value = 1016; @@ -1601,6 +1560,9 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, int j, prefix_len = 0; int before_tmp = before, after_tmp = after; struct scan_info *scan_info = scan_info_array + i; + struct rpn_char_map_info rcmi; + + rpn_char_map_prepare (zi, reg_type, &rcmi); scan_info->before = before; scan_info->after = after; @@ -1611,12 +1573,11 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, for (j = 0; jlist[j].term = NULL; termz[prefix_len++] = ords[i]; - termz[prefix_len++] = - (completeness_value==2 || completeness_value==3) ? 'p': 'w'; + termz[prefix_len++] = reg_type; termz[prefix_len] = 0; strcpy (scan_info->prefix, termz); - trans_scan_term (zi, zapt, termz+prefix_len); + trans_scan_term (zi, zapt, termz+prefix_len, reg_type); dict_scan (zi->dict, termz, &before_tmp, &after_tmp, scan_info, scan_handle); @@ -1645,7 +1606,7 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, } if (j0 == -1) break; - scan_term_untrans (zi->odr, &glist[i+before].term, mterm); + scan_term_untrans (zi, reg_type, &glist[i+before].term, mterm); rset = rset_trunc (zi, &scan_info_array[j0].list[ptr[j0]].isam_p, 1); ptr[j0]++; @@ -1703,7 +1664,7 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, if (j0 == -1) break; - scan_term_untrans (zi->odr, &glist[before-1-i].term, mterm); + scan_term_untrans (zi, reg_type, &glist[before-1-i].term, mterm); rset = rset_trunc (zi, &scan_info_array[j0].list[before-1-ptr[j0]].isam_p, 1); diff --git a/index/zserver.c b/index/zserver.c index 12314b5..bd89d6e 100644 --- a/index/zserver.c +++ b/index/zserver.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zserver.c,v $ - * Revision 1.50 1997-09-29 09:08:36 adam + * Revision 1.51 1997-10-27 14:33:06 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 1.50 1997/09/29 09:08:36 adam * Revised locking system to be thread safe for the server. * * Revision 1.49 1997/09/25 14:57:23 adam @@ -259,7 +265,7 @@ static int register_lock (ZServerInfo *zi) return -1; } zi->zti = zebTargetInfo_open (zi->records, 0); - init_charmap (zi->res); + return 0; } @@ -321,6 +327,7 @@ bend_initresult *bend_init (bend_initrequest *q) zi->records = NULL; zi->odr = odr_createmem (ODR_ENCODE); zi->registered_sets = NULL; + zi->zebra_maps = zebra_maps_open (res_get(zi->res, "profilePath")); return r; } @@ -560,9 +567,11 @@ bend_scanresult *bend_scan (void *handle, bend_scanrequest *q, int *num) void bend_close (void *handle) { ZServerInfo *zi = handle; + if (zi->records) { resultSetDestroy (zi); + zebTargetInfo_close (zi->zti, 0); dict_close (zi->dict); if (zi->isam) is_close (zi->isam); @@ -571,18 +580,44 @@ void bend_close (void *handle) rec_close (&zi->records); register_unlock (zi); } + odr_destroy (zi->odr); + zebra_maps_close (zi->zebra_maps); bfs_destroy (zi->bfs); data1_destroy (zi->dh); zebra_server_lock_destroy (zi); - return; + + res_close (zi->res); + xfree (zi); } +#ifndef WINDOWS +static void pre_init (struct statserv_options_block *sob) +{ + char *pidfile = "zebrasrv.pid"; + int fd = creat (pidfile, 0666); + + if (fd == -1) + logf (LOG_WARN|LOG_ERRNO, "creat %s", pidfile); + else + { + char pidstr[30]; + + sprintf (pidstr, "%ld", (long) getpid ()); + write (fd, pidstr, strlen(pidstr)); + close (fd); + } +} +#endif + int main (int argc, char **argv) { struct statserv_options_block *sob; sob = statserv_getcontrol (); strcpy (sob->configname, FNAME_CONFIG); +#ifndef WINDOWS + sob->pre_init = pre_init; +#endif statserv_setcontrol (sob); return statserv_main (argc, argv); diff --git a/index/zserver.h b/index/zserver.h index ee609e2..2b25cdf 100644 --- a/index/zserver.h +++ b/index/zserver.h @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zserver.h,v $ - * Revision 1.26 1997-09-29 12:41:35 adam + * Revision 1.27 1997-10-27 14:33:06 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 1.26 1997/09/29 12:41:35 adam * Fixed bug regarding USE_TIMES var. * * Revision 1.25 1997/09/29 09:08:36 adam @@ -145,6 +151,7 @@ typedef struct { struct tms tms1; struct tms tms2; #endif + ZebraMaps zebra_maps; } ZServerInfo; int rpn_search (ZServerInfo *zi, diff --git a/recctrl/rectext.c b/recctrl/rectext.c index a317751..7c7c9c0 100644 --- a/recctrl/rectext.c +++ b/recctrl/rectext.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: rectext.c,v $ - * Revision 1.4 1996-11-04 14:09:16 adam + * Revision 1.5 1997-10-27 14:33:06 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 1.4 1996/11/04 14:09:16 adam * Minor changes. * * Revision 1.3 1996/11/01 09:00:33 adam @@ -99,7 +105,7 @@ static int text_extract (struct recExtractCtrl *p) struct buf_info *fi = buf_open (p); (*p->init)(&recWord); - recWord.which = Word_String; + recWord.reg_type = 'w'; do { int i = 0; @@ -117,7 +123,7 @@ static int text_extract (struct recExtractCtrl *p) w[j] = tolower(w[j]); w[i] = 0; recWord.seqno = seqno++; - recWord.u.string = w; + recWord.string = w; (*p->add)(&recWord); } } while (r > 0); diff --git a/util/Makefile b/util/Makefile index 3281cf5..fcdc9cd 100644 --- a/util/Makefile +++ b/util/Makefile @@ -1,7 +1,7 @@ # Copyright (C) 1994-1996, Index Data I/S # All rights reserved. # Sebastian Hammer, Adam Dickmeiss -# $Id: Makefile,v 1.26 1997-09-17 12:19:23 adam Exp $ +# $Id: Makefile,v 1.27 1997-10-27 14:33:06 adam Exp $ SHELL=/bin/sh RANLIB=ranlib @@ -14,7 +14,7 @@ TPROG=opt-test DEFS=$(INCLUDE) CPP=$(CC) -E LIB=../lib/zebrautl.a -PO = res.o charmap.o +PO = res.o charmap.o zebramap.o all: $(LIB) diff --git a/util/charmap.c b/util/charmap.c index 8e95b00..7104cd3 100644 --- a/util/charmap.c +++ b/util/charmap.c @@ -4,7 +4,13 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: charmap.c,v $ - * Revision 1.12 1997-09-05 15:30:11 adam + * Revision 1.13 1997-10-27 14:33:06 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 1.12 1997/09/05 15:30:11 adam * Changed prototype for chr_map_input - added const. * Added support for C++, headers uses extern "C" for public definitions. * @@ -53,18 +59,29 @@ #include #include -#include #include #include -#include #define CHR_MAXSTR 1024 #define CHR_MAXEQUIV 32 +int chr_map_chrs(chr_t_entry *t, char **from, int len, + int *read, char **to, int max); + const char *CHR_UNKNOWN = "\001"; const char *CHR_SPACE = "\002"; const char *CHR_BASE = "\003"; +struct chrmaptab_info +{ + chr_t_entry *input; /* mapping table for input data */ + chr_t_entry *query_equiv; /* mapping table for queries */ + unsigned char *output[256]; /* return mapping - for display of registers */ + int base_uppercase; /* Start of upper-case ordinals */ + char **tmp_buf; + NMEM nmem; +}; + /* * Character map trie node. */ @@ -80,19 +97,19 @@ struct chr_t_entry */ typedef struct chrwork { - chrmaptab *map; + chrmaptab map; char string[CHR_MAXSTR+1]; } chrwork; /* * Add an entry to the character map. */ -static chr_t_entry *set_map_string(chr_t_entry *root, const char *from, - int len, char *to) +static chr_t_entry *set_map_string(chr_t_entry *root, NMEM nmem, + const char *from, int len, char *to) { if (!root) { - root = xmalloc(sizeof(*root)); + root = nmem_malloc(nmem, sizeof(*root)); root->children = 0; root->target = 0; } @@ -100,7 +117,7 @@ static chr_t_entry *set_map_string(chr_t_entry *root, const char *from, { if (!root->target || (char*) root->target == CHR_SPACE || (char*) root->target == CHR_UNKNOWN) - root->target = (unsigned char *) xstrdup(to); + root->target = (unsigned char *) nmem_strdup(nmem, to); else if ((char*) to != CHR_SPACE) logf(LOG_DEBUG, "Character map overlap"); } @@ -110,13 +127,13 @@ static chr_t_entry *set_map_string(chr_t_entry *root, const char *from, { int i; - root->children = xmalloc(sizeof(chr_t_entry*) * 256); + root->children = nmem_malloc(nmem, sizeof(chr_t_entry*) * 256); for (i = 0; i < 256; i++) root->children[i] = 0; } if (!(root->children[(unsigned char) *from] = - set_map_string(root->children[(unsigned char) *from], from + 1, - len - 1, to))) + set_map_string(root->children[(unsigned char) *from], nmem, + from + 1, len - 1, to))) return 0; } return root; @@ -167,17 +184,24 @@ static chr_t_entry *find_entry(chr_t_entry *t, const char **from, int len) return t->target ? t : 0; } -const char **chr_map_input(chr_t_entry *t, const char **from, int len) +const char **chr_map_input(chrmaptab maptab, const char **from, int len) { - static const char *buf[2] = {0, 0}; + chr_t_entry *t = maptab->input; chr_t_entry *res; if (!(res = find_entry(t, from, len))) abort(); - buf[0] = (char *) res->target; - return buf; + maptab->tmp_buf[0] = (char*) res->target; + maptab->tmp_buf[1] = NULL; + return (const char **) maptab->tmp_buf; } +const char *chr_map_output(chrmaptab maptab, const char **from, int len) +{ + unsigned char c = ** (unsigned char **) from; + (*from)++; + return (const char*) maptab->output[c]; +} static unsigned char prim(char **s) { @@ -214,12 +238,13 @@ static unsigned char prim(char **s) */ static void fun_addentry(const char *s, void *data, int num) { - chrmaptab *tab = data; + chrmaptab tab = data; char tmp[2]; tmp[0] = num; tmp[1] = '\0'; - tab->input = set_map_string(tab->input, s, strlen(s), tmp); - tab->output[num + tab->base_uppercase] = (unsigned char *) xstrdup(s); + tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s), tmp); + tab->output[num + tab->base_uppercase] = + (unsigned char *) nmem_strdup(tab->nmem, s); } /* @@ -228,8 +253,9 @@ static void fun_addentry(const char *s, void *data, int num) */ static void fun_addspace(const char *s, void *data, int num) { - chrmaptab *tab = data; - tab->input = set_map_string(tab->input, s, strlen(s), (char*) CHR_SPACE); + chrmaptab tab = data; + tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s), + (char*) CHR_SPACE); } /* @@ -240,7 +266,7 @@ static void fun_mkstring(const char *s, void *data, int num) chrwork *arg = data; const char **res, *p = s; - res = chr_map_input(arg->map->input, &s, strlen(s)); + res = chr_map_input(arg->map, &s, strlen(s)); if (*res == (char*) CHR_UNKNOWN) logf(LOG_WARN, "Map: '%s' has no mapping", p); strncat(arg->string, *res, CHR_MAXSTR - strlen(arg->string)); @@ -255,7 +281,7 @@ static void fun_addmap(const char *s, void *data, int num) chrwork *arg = data; assert(arg->map->input); - set_map_string(arg->map->input, s, strlen(s), arg->string); + set_map_string(arg->map->input, arg->map->nmem, s, strlen(s), arg->string); } static int scan_string(char *s, void (*fun)(const char *c, void *data, int num), @@ -315,11 +341,11 @@ static int scan_string(char *s, void (*fun)(const char *c, void *data, int num), return 0; } -chrmaptab *chr_read_maptab(const char *tabpath, const char *name) +chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only) { FILE *f; char line[512], *argv[50]; - chrmaptab *res = xmalloc(sizeof(*res)); + chrmaptab res; int argc, num = (int) *CHR_BASE, i; if (!(f = yaz_path_fopen(tabpath, name, "r"))) @@ -328,21 +354,27 @@ chrmaptab *chr_read_maptab(const char *tabpath, const char *name) return 0; } res = xmalloc(sizeof(*res)); - res->input = xmalloc(sizeof(*res->input)); + res->nmem = nmem_create (); + res->tmp_buf = nmem_malloc (res->nmem, sizeof(*res->tmp_buf) * 100); + res->input = nmem_malloc(res->nmem, sizeof(*res->input)); res->input->target = (unsigned char*) CHR_UNKNOWN; res->input->equiv = 0; -#if 1 - res->input->children = xmalloc(sizeof(res->input) * 256); + res->input->children = nmem_malloc(res->nmem, sizeof(res->input) * 256); for (i = 0; i < 256; i++) { - res->input->children[i] = xmalloc(sizeof(*res->input)); + res->input->children[i] = nmem_malloc(res->nmem, sizeof(*res->input)); res->input->children[i]->children = 0; - res->input->children[i]->target = (unsigned char*) CHR_UNKNOWN; + if (map_only) + { + res->input->children[i]->target = nmem_malloc (res->nmem, + 2 * sizeof(char)); + res->input->children[i]->target[0] = i; + res->input->children[i]->target[1] = 0; + } + else + res->input->children[i]->target = (unsigned char*) CHR_UNKNOWN; res->input->children[i]->equiv = 0; } -#else - res->input->children = 0; -#endif res->query_equiv = 0; for (i = *CHR_BASE; i < 256; i++) res->output[i] = 0; @@ -351,7 +383,7 @@ chrmaptab *chr_read_maptab(const char *tabpath, const char *name) res->base_uppercase = 0; while ((argc = readconf_line(f, line, 512, argv, 50))) - if (!yaz_matchstr(argv[0], "lowercase")) + if (!map_only && !yaz_matchstr(argv[0], "lowercase")) { if (argc != 2) { @@ -370,7 +402,7 @@ chrmaptab *chr_read_maptab(const char *tabpath, const char *name) res->output[(int) *CHR_UNKNOWN + num] = (unsigned char*) "@"; num = (int) *CHR_BASE; } - else if (!yaz_matchstr(argv[0], "uppercase")) + else if (!map_only && !yaz_matchstr(argv[0], "uppercase")) { if (!res->base_uppercase) { @@ -391,7 +423,7 @@ chrmaptab *chr_read_maptab(const char *tabpath, const char *name) return 0; } } - else if (!yaz_matchstr(argv[0], "space")) + else if (!map_only && !yaz_matchstr(argv[0], "space")) { if (argc != 2) { @@ -433,12 +465,16 @@ chrmaptab *chr_read_maptab(const char *tabpath, const char *name) } else { -#if 0 - logf(LOG_WARN, "Syntax error at '%s' in %s", line, file); - fclose(f); - return 0; -#endif + logf(LOG_WARN, "Syntax error at '%s' in %s", line, name); } fclose(f); return res; } + +void chrmaptab_destroy(chrmaptab tab) +{ + nmem_destroy (tab->nmem); + xfree (tab); +} + + diff --git a/util/zebramap.c b/util/zebramap.c new file mode 100644 index 0000000..a925c79 --- /dev/null +++ b/util/zebramap.c @@ -0,0 +1,262 @@ +/* + * Copyright (C) 1994-1997, Index Data I/S + * All rights reserved. + * Sebastian Hammer, Adam Dickmeiss + * + * $Log: zebramap.c,v $ + * Revision 1.1 1997-10-27 14:33:06 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. + * + */ + +#include +#include + +#include +#include +#include + +struct zebra_map { + int reg_type; + chrmaptab maptab; + struct zebra_map *next; +}; + +struct zebra_maps { + char *tabpath; + NMEM nmem; + struct zebra_map *map_list; +}; + +void zebra_maps_close (ZebraMaps zms) +{ + struct zebra_map *zm = zms->map_list; + while (zm) + { + struct zebra_map *zm_next = zm->next; + + chrmaptab_destroy (zm->maptab); + xfree (zm); + zm = zm_next; + } + nmem_destroy (zms->nmem); + xfree (zms); +} + +ZebraMaps zebra_maps_open (const char *tabpath) +{ + ZebraMaps zms = xmalloc (sizeof(*zms)); + + zms->nmem = nmem_create (); + zms->tabpath = nmem_strdup (zms->nmem, tabpath); + zms->map_list = NULL; + return zms; +} + +chrmaptab zebra_map_get (ZebraMaps zms, int reg_type) +{ + char name[512]; + struct zebra_map *zm; + + for (zm = zms->map_list; zm; zm = zm->next) + { + if (reg_type == zm->reg_type) + return zm->maptab; + } + *name = '\0'; + switch (reg_type) + { + case 'w': + case 'p': + strcat (name, "string"); + break; + case 'n': + strcat (name, "numeric"); + break; + case 'u': + strcat (name, "urx"); + break; + default: + strcat (name, "null"); + } + strcat (name, ".chr"); + + zm = xmalloc (sizeof(*zm)); + zm->reg_type = reg_type; + zm->next = zms->map_list; + zms->map_list = zm; + if (!(zm->maptab = chrmaptab_create (zms->tabpath, name, 0))) + logf(LOG_WARN, "Failed to read character table %s", name); + else + logf(LOG_LOG, "Read table %s", name); + return zm->maptab; +} + +const char **zebra_maps_input (ZebraMaps zms, int reg_type, + const char **from, int len) +{ + static char str[2] = {0,0}; + static const char *buf[2] = {0,0}; + chrmaptab maptab; + + maptab = zebra_map_get (zms, reg_type); + if (maptab) + return chr_map_input(maptab, from, len); + + if (isalnum(**from)) + { + str[0] = isupper(**from) ? tolower(**from) : **from; + buf[0] = str; + } + else + buf[0] = (char*) CHR_SPACE; + (*from)++; + return buf; +} + +const char *zebra_maps_output(ZebraMaps zms, int reg_type, const char **from) +{ + chrmaptab maptab; + unsigned char i = (unsigned char) **from; + static char buf[2] = {0,0}; + + maptab = zebra_map_get (zms, reg_type); + if (maptab) + return chr_map_output (maptab, from, 1); + (*from)++; + buf[0] = i; + return buf; +} + + +/* ------------------------------------ */ + +typedef struct { + int type; + int major; + int minor; + Z_AttributesPlusTerm *zapt; +} AttrType; + +static int attr_find (AttrType *src, oid_value *attributeSetP) +{ + while (src->major < src->zapt->num_attributes) + { + Z_AttributeElement *element; + + element = src->zapt->attributeList[src->major]; + if (src->type == *element->attributeType) + { + switch (element->which) + { + case Z_AttributeValue_numeric: + ++(src->major); + if (element->attributeSet && attributeSetP) + { + oident *attrset; + + attrset = oid_getentbyoid (element->attributeSet); + *attributeSetP = attrset->value; + } + return *element->value.numeric; + break; + case Z_AttributeValue_complex: + if (src->minor >= element->value.complex->num_list || + element->value.complex->list[src->minor]->which != + Z_StringOrNumeric_numeric) + break; + ++(src->minor); + if (element->attributeSet && attributeSetP) + { + oident *attrset; + + attrset = oid_getentbyoid (element->attributeSet); + *attributeSetP = attrset->value; + } + return *element->value.complex->list[src->minor-1]->u.numeric; + default: + assert (0); + } + } + ++(src->major); + } + return -1; +} + +static void attr_init (AttrType *src, Z_AttributesPlusTerm *zapt, + int type) +{ + src->zapt = zapt; + src->type = type; + src->major = 0; + src->minor = 0; +} + +/* ------------------------------------ */ + +int zebra_maps_is_complete (ZebraMaps zms, int reg_type) +{ + if (reg_type == 'p') + return 1; + return 0; +} + +int zebra_maps_attr (ZebraMaps zms, Z_AttributesPlusTerm *zapt, + int *reg_type, char **search_type, int *complete_flag) +{ + AttrType completeness; + AttrType structure; + AttrType relation; + int completeness_value; + int structure_value; + int relation_value; + + attr_init (&structure, zapt, 4); + attr_init (&completeness, zapt, 6); + attr_init (&relation, zapt, 2); + + completeness_value = attr_find (&completeness, NULL); + structure_value = attr_find (&structure, NULL); + + if (completeness_value == 2 || completeness_value == 3) + *complete_flag = 1; + else + *complete_flag = 0; + *reg_type = 0; + + *search_type = "phrase"; + if (relation_value == 102) + *search_type = "ranked"; + + switch (structure_value) + { + case -1: + case 1: /* phrase */ + case 2: /* word */ + case 3: /* key */ + case 6: /* word list */ + case 105: /* free-form-text */ + case 106: /* document-text */ + case 108: /* string */ + if (*complete_flag) + *reg_type = 'p'; + else + *reg_type = 'w'; + break; + case 107: /* local-number */ + *search_type = "local"; + *reg_type = 0; + case 109: /* numeric string */ + *reg_type = 'n'; + break; + case 104: /* urx */ + *reg_type = 'u'; + break; + default: + return -1; + } + return 0; +}