X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fzrpn.c;h=a5da2449f25c068a629f4391a992e2d89337fe01;hb=bf1f271315a6acfd52d8d7abf1aee2e0502bdbf2;hp=409cea097d68610004cc0a28c62bef115df3459c;hpb=4e45c8e2e01c52e8d03dbd9a0873a9e2ff7e5f91;p=idzebra-moved-to-github.git diff --git a/index/zrpn.c b/index/zrpn.c index 409cea0..a5da244 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -1,10 +1,37 @@ /* - * Copyright (C) 1994-1995, Index Data I/S + * Copyright (C) 1994-1996, Index Data I/S * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: zrpn.c,v $ - * Revision 1.44 1996-05-14 06:16:44 adam + * Revision 1.53 1996-06-26 09:21:43 adam + * Bug fix: local attribute set wasn't obeyed in scan. + * + * Revision 1.52 1996/06/17 14:26:20 adam + * Function gen_regular_rel changed to handle negative numbers. + * + * Revision 1.51 1996/06/11 10:54:15 quinn + * Relevance work + * + * Revision 1.50 1996/06/07 08:51:53 adam + * Bug fix: Character mapping was broken (introducued by last revision). + * + * Revision 1.49 1996/06/04 10:18:11 adam + * Search/scan uses character mapping module. + * + * Revision 1.48 1996/05/28 15:15:01 adam + * Bug fix: Didn't handle unknown database correctly. + * + * Revision 1.47 1996/05/15 18:36:28 adam + * Function trans_term transforms unsearchable characters to blanks. + * + * Revision 1.46 1996/05/15 11:57:56 adam + * Fixed bug introduced by set/field mapping in search operations. + * + * Revision 1.45 1996/05/14 11:34:00 adam + * Scan support in multiple registers/databases. + * + * Revision 1.44 1996/05/14 06:16:44 adam * Compact use/set bytes used in search service. * * Revision 1.43 1996/05/09 09:54:43 adam @@ -163,26 +190,13 @@ #include "zserver.h" #include "attribute.h" +#include #include #include #include #include #include -int index_word_prefix_map (char *string, oid_value attrSet, int attrUse, - char *basename) -{ - attent *attp; - - logf (LOG_DEBUG, "oid_value attrSet = %d, attrUse = %d", attrSet, attrUse); - attp = att_getentbyatt (attrSet, attrUse); - if (!attp) - return -1; - logf (LOG_DEBUG, "ord=%d", attp->attset_ordinal); - return index_word_prefix (string, attp->attset_ordinal, - attp->local_attributes->local, basename); -} - typedef struct { int type; int major; @@ -486,17 +500,25 @@ static RSET rset_trunc (ISAM isam, ISAM_P *isam_p, int no) return rset_trunc_r (isam, isam_p, 0, no, 100); } -struct grep_info { - ISAM_P *isam_p_buf; - int isam_p_size; - int isam_p_indx; -}; +#define TERM_COUNT + +struct grep_info { +#ifdef TERM_COUNT + int *term_no; +#endif + ISAM_P *isam_p_buf; + int isam_p_size; + int isam_p_indx; +}; static void add_isam_p (const char *info, struct grep_info *p) { if (p->isam_p_indx == p->isam_p_size) { ISAM_P *new_isam_p_buf; +#ifdef TERM_COUNT + int *new_term_no; +#endif p->isam_p_size = 2*p->isam_p_size + 100; new_isam_p_buf = xmalloc (sizeof(*new_isam_p_buf) * @@ -508,6 +530,18 @@ static void add_isam_p (const char *info, struct grep_info *p) xfree (p->isam_p_buf); } p->isam_p_buf = new_isam_p_buf; + +#ifdef TERM_COUNT + new_term_no = xmalloc (sizeof(*new_term_no) * + p->isam_p_size); + if (p->term_no) + { + memcpy (new_term_no, p->isam_p_buf, + p->isam_p_indx * sizeof(*p->term_no)); + xfree (p->term_no); + } + p->term_no = new_term_no; +#endif } assert (*info == sizeof(*p->isam_p_buf)); memcpy (p->isam_p_buf + p->isam_p_indx, info+1, sizeof(*p->isam_p_buf)); @@ -516,19 +550,45 @@ static void add_isam_p (const char *info, struct grep_info *p) static int grep_handle (char *name, const char *info, void *p) { - logf (LOG_DEBUG, "dict name: %s", name); add_isam_p (info, p); return 0; } +/* gen_regular_rel - generate regular expression from relation + * val: border value (inclusive) + * islt: 1 if <=; 0 if >=. + */ static void gen_regular_rel (char *dst, int val, int islt) { - int dst_p = 1; + int dst_p; int w, d, i; int pos = 0; char numstr[20]; - *dst = '('; + logf (LOG_DEBUG, "gen_regular_rel. val=%d, islt=%d", val, islt); + if (val >= 0) + { + if (islt) + strcpy (dst, "(-[0-9]+|"); + else + strcpy (dst, "("); + } + else + { + if (!islt) + { + strcpy (dst, "([0-9]+|-"); + dst_p = strlen (dst); + islt = 1; + } + else + { + strcpy (dst, "(-"); + islt = 0; + } + val = -val; + } + dst_p = strlen (dst); sprintf (numstr, "%d", val); for (w = strlen(numstr); --w >= 0; pos++) { @@ -665,7 +725,7 @@ static void verbatim_char (int ch, int *indx, char *dst) } static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - const char *term_sub, + const char *term_sub, int regType, oid_value attributeSet, struct grep_info *grep_info, int num_bases, char **basenames) { @@ -689,7 +749,6 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, for (base_no = 0; base_no < num_bases; base_no++) { -#if 1 attent *attp; data1_local_attribute *local_attr; int max_pos, prefix_len = 0; @@ -697,6 +756,8 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, attp = att_getentbyatt (curAttributeSet, use_value); if (!attp) { + logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d", + curAttributeSet, use_value); zi->errCode = 114; return -1; } @@ -704,6 +765,7 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, { zi->errCode = 109; /* Database unavailable */ zi->errString = basenames[base_no]; + return -1; } for (local_attr = attp->local_attributes; local_attr; local_attr = local_attr->next) @@ -718,20 +780,17 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, term_dict[prefix_len++] = '|'; else term_dict[prefix_len++] = '('; - if ((ord >= 'A' && ord <= 'Z') || (ord >= 'a' && ord <= 'z')) - term_dict[prefix_len++] = ord; - else - { - term_dict[prefix_len++] = '\\'; - term_dict[prefix_len++] = ord; - } + term_dict[prefix_len++] = 1; + term_dict[prefix_len++] = ord; } if (!prefix_len) { zi->errCode = 114; return -1; } - term_dict[prefix_len++] = ')'; + term_dict[prefix_len++] = ')'; + term_dict[prefix_len++] = 1; + term_dict[prefix_len++] = regType; term_dict[prefix_len] = '\0'; if (!relational_term (zi, zapt, term_sub, term_dict, attributeSet, grep_info, &max_pos)) @@ -806,214 +865,51 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, break; } } -#else - int max_pos; -#if 1 - attent *attp; - data1_local_attribute *local_attr; - int prefix_len; - - attp = att_getentbyatt (curAttributeSet, use_value); - if (!attp) - { - zi->errCode = 114; - return -1; - } - for (local_attr = attp->local_attributes; local_attr; - local_attr = local_attr->next) - { - prefix_len = index_word_prefix (term_dict, attp->attset_ordinal, - local_attr->local, - basenames[base_no]); - - if (!relational_term (zi, zapt, term_sub, term_dict, - attributeSet, grep_info, &max_pos)) - { - const char *cp; - - j = prefix_len; - switch (truncation_value) - { - case -1: /* not specified */ - case 100: /* do not truncate */ - term_dict[j++] = '('; - for (i = 0; term_sub[i]; i++) - verbatim_char (term_sub[i], &j, term_dict); - strcpy (term_dict+j, ")"); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r); - break; - case 1: /* right truncation */ - term_dict[j++] = '('; - for (i = 0; term_sub[i]; i++) - verbatim_char (term_sub[i], &j, term_dict); - strcpy (term_dict+j, ".*)"); - dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - break; - case 2: /* left truncation */ - case 3: /* left&right truncation */ - zi->errCode = 120; - return -1; - case 101: /* process # in term */ - term_dict[j++] = '('; - for (i=0; term_sub[i]; i++) - if (term_sub[i] == '#' && i > 2) - { - term_dict[j++] = '.'; - term_dict[j++] = '*'; - } - else - verbatim_char (term_sub[i], &j, term_dict); - strcpy (term_dict+j, ")"); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", - r); - break; - case 102: /* regular expression */ - sprintf (term_dict + j, "(%s)", term_sub); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", - r); - break; - case 103: /* regular expression with error correction */ - cp = term_sub; - r = 0; - if (*cp == '*' && cp[1] && cp[2]) - { - r = atoi (cp+1); - cp += 2; - } - sprintf (term_dict + j, "(%s)", cp); - r = dict_lookup_grep (zi->wordDict, term_dict, r, grep_info, - &max_pos, j, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", - r); - break; - } - } - if (max_pos <= strlen(basenames[base_no])) - { - zi->errCode = 109; /* Database unavailable */ - zi->errString = basenames[base_no]; - return -1; - } - } -#else - int prefix_len = index_word_prefix_map (term_dict, curAttributeSet, - use_value, - basenames[base_no]); - if (prefix_len < 0) - { - zi->errCode = 114; - return -1; - } - if (!relational_term (zi, zapt, term_sub, term_dict, - attributeSet, grep_info, &max_pos)) - { - const char *cp; - - j = prefix_len; - switch (truncation_value) - { - case -1: /* not specified */ - case 100: /* do not truncate */ - term_dict[j++] = '('; - for (i = 0; term_sub[i]; i++) - verbatim_char (term_sub[i], &j, term_dict); - strcpy (term_dict+j, ")"); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r); - break; - case 1: /* right truncation */ - term_dict[j++] = '('; - for (i = 0; term_sub[i]; i++) - verbatim_char (term_sub[i], &j, term_dict); - strcpy (term_dict+j, ".*)"); - dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - break; - case 2: /* left truncation */ - case 3: /* left&right truncation */ - zi->errCode = 120; - return -1; - case 101: /* process # in term */ - term_dict[j++] = '('; - for (i=0; term_sub[i]; i++) - if (term_sub[i] == '#' && i > 2) - { - term_dict[j++] = '.'; - term_dict[j++] = '*'; - } - else - verbatim_char (term_sub[i], &j, term_dict); - strcpy (term_dict+j, ")"); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", - r); - break; - case 102: /* regular expression */ - sprintf (term_dict + j, "(%s)", term_sub); - r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", - r); - break; - case 103: /* regular expression with error correction */ - cp = term_sub; - r = 0; - if (*cp == '*' && cp[1] && cp[2]) - { - r = atoi (cp+1); - cp += 2; - } - sprintf (term_dict + j, "(%s)", cp); - r = dict_lookup_grep (zi->wordDict, term_dict, r, grep_info, - &max_pos, j, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", - r); - break; - } - } - if (max_pos <= strlen(basenames[base_no])) - { - zi->errCode = 109; /* Database unavailable */ - zi->errString = basenames[base_no]; - return -1; - } -#endif -#endif } logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); return 0; } static void trans_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - int regType, char *termz) + char *termz) { - size_t i, sizez; + size_t sizez; Z_Term *term = zapt->term; sizez = term->u.general->len; if (sizez > IT_MAX_WORD-1) sizez = IT_MAX_WORD-1; - termz[0] = regType; - for (i = 0; i < sizez; i++) - termz[i+1] = index_char_cvt (term->u.general->buf[i]); - termz[i+1] = '\0'; + memcpy (termz, term->u.general->buf, sizez); + termz[sizez] = '\0'; +} + +static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, + char *termz) +{ + Z_Term *term = zapt->term; + char **map; + const char *cp = (const char *) term->u.general->buf; + const char *cp_end = cp + term->u.general->len; + const char *src; + int i = 0; + int prev_space = 0; + int len; + + while ((len = (cp_end - cp)) > 0) + { + map = map_chrs_input (&cp, len); + if (**map == *CHR_SPACE) + { + if (prev_space) + continue; + prev_space = 1; + } + else + prev_space = 0; + for (src = *map; *src; src++) + termz[i++] = *src; + } + termz[i] = '\0'; } static RSET rpn_search_APT_relevance (ZServerInfo *zi, @@ -1025,48 +921,75 @@ static RSET rpn_search_APT_relevance (ZServerInfo *zi, char termz[IT_MAX_WORD+1]; char term_sub[IT_MAX_WORD+1]; struct grep_info grep_info; - char *p0 = termz, *p1 = NULL; + char *p0 = termz; RSET result; + int term_index = 0; parms.key_size = sizeof(struct it_key); parms.max_rec = 100; parms.cmp = key_compare; parms.is = zi->wordIsam; + parms.no_terms = 0; if (zapt->term->which != Z_Term_general) { zi->errCode = 124; return NULL; } - trans_term (zi, zapt, 'w', termz); + 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; while (1) { - if ((p1 = strchr (p0, ' '))) + char **map; + char *p2, *p1; + + p1 = p0; + while (*(p0 = p1)) { - memcpy (term_sub, p0, p1-p0); - term_sub[p1-p0] = '\0'; + map = map_chrs_input (&p1, strlen(p1)); + if (**map != *CHR_SPACE) + break; } - else - strcpy (term_sub, p0); - if (field_term (zi, zapt, term_sub, attributeSet, &grep_info, + if (!*p0) + break; + + p1 = p0; + while (*(p2 = p1)) + { + map = map_chrs_input (&p1, strlen(p1)); + if (**map == *CHR_SPACE) + break; + } + if (p2 == p0) + break; + memcpy (term_sub, p0, p2-p0); + term_sub[p2-p0] = '\0'; + p0 = p2; + if (field_term (zi, zapt, term_sub, 'w', attributeSet, &grep_info, num_bases, basenames)) return NULL; - if (!p1) - break; - p0 = p1; - while (*++p0 == ' ') - ; +#ifdef TERM_COUNT + for (; term_index < grep_info.isam_p_indx; term_index++) + grep_info.term_no[term_index] = parms.no_terms; + parms.no_terms++; +#endif } + parms.term_no = grep_info.term_no; parms.isam_positions = grep_info.isam_p_buf; parms.no_isam_positions = grep_info.isam_p_indx; if (grep_info.isam_p_indx > 0) result = rset_create (rset_kind_relevance, &parms); else result = rset_create (rset_kind_null, NULL); +#ifdef TERM_COUNT + xfree(grep_info.term_no); +#endif xfree (grep_info.isam_p_buf); return result; } @@ -1086,51 +1009,16 @@ static RSET rpn_search_APT_cphrase (ZServerInfo *zi, zi->errCode = 124; return NULL; } - trans_term (zi, zapt, 'p', termz); - - grep_info.isam_p_indx = 0; - grep_info.isam_p_size = 0; - grep_info.isam_p_buf = NULL; - - if (field_term (zi, zapt, termz, attributeSet, &grep_info, - num_bases, basenames)) - return NULL; - if (grep_info.isam_p_indx < 1) - result = rset_create (rset_kind_null, NULL); - else if (grep_info.isam_p_indx == 1) - { - parms.is = zi->wordIsam; - parms.pos = *grep_info.isam_p_buf; - result = rset_create (rset_kind_isam, &parms); - } - else - result = rset_trunc (zi->wordIsam, grep_info.isam_p_buf, - grep_info.isam_p_indx); - xfree (grep_info.isam_p_buf); - return result; -} -static RSET rpn_search_APT_word (ZServerInfo *zi, - Z_AttributesPlusTerm *zapt, - oid_value attributeSet, - int num_bases, char **basenames) -{ - rset_isam_parms parms; - char termz[IT_MAX_WORD+1]; - struct grep_info grep_info; - RSET result; - - if (zapt->term->which != Z_Term_general) - { - zi->errCode = 124; - return NULL; - } - trans_term (zi, zapt, 'w', termz); + 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; - if (field_term (zi, zapt, termz, attributeSet, &grep_info, + if (field_term (zi, zapt, termz, 'p', attributeSet, &grep_info, num_bases, basenames)) return NULL; if (grep_info.isam_p_indx < 1) @@ -1144,6 +1032,9 @@ static RSET rpn_search_APT_word (ZServerInfo *zi, else result = rset_trunc (zi->wordIsam, 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; } @@ -1241,7 +1132,7 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, { char termz[IT_MAX_WORD+1]; char term_sub[IT_MAX_WORD+1]; - char *p0 = termz, *p1 = NULL; + char *p0 = termz; RSET rset[60], result; int i, rset_no = 0; struct grep_info grep_info; @@ -1251,23 +1142,45 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, zi->errCode = 124; return NULL; } - trans_term (zi, zapt, 'w', termz); + trans_term (zi, zapt, termz); +#ifdef TERM_COUNT + grep_info.term_no = 0; +#endif grep_info.isam_p_size = 0; grep_info.isam_p_buf = NULL; while (1) { - if ((p1 = strchr (p0, ' '))) + char **map; + char *p2, *p1; + + p1 = p0; + while (*(p0 = p1)) { - memcpy (term_sub, p0, p1-p0); - term_sub[p1-p0] = '\0'; + map = map_chrs_input (&p1, strlen(p1)); + if (**map != *CHR_SPACE) + break; } - else - strcpy (term_sub, p0); + if (!*p0) + break; + + p1 = p0; + while (*(p2 = p1)) + { + map = map_chrs_input (&p1, strlen(p1)); + if (**map == *CHR_SPACE) + break; + } + if (p2 == p0) + break; + + memcpy (term_sub, p0, p2-p0); + term_sub[p2-p0] = '\0'; + p0 = p2; grep_info.isam_p_indx = 0; - if (field_term (zi, zapt, term_sub, attributeSet, &grep_info, + if (field_term (zi, zapt, term_sub, 'w', attributeSet, &grep_info, num_bases, basenames)) return NULL; if (grep_info.isam_p_indx == 0) @@ -1287,12 +1200,10 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, assert (rset[rset_no]); if (++rset_no >= sizeof(rset)/sizeof(*rset)) break; - if (!p1) - break; - p0 = p1; - while (*++p0 == ' ') - ; } +#ifdef TERM_COUNT + xfree(grep_info.term_no); +#endif xfree (grep_info.isam_p_buf); if (rset_no == 0) return rset_create (rset_kind_null, NULL); @@ -1323,9 +1234,9 @@ static RSET rpn_search_APT_local (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, result = rset_create (rset_kind_temp, &parms); rsfd = rset_open (result, RSETF_WRITE|RSETF_SORT_SYSNO); - trans_term (zi, zapt, 'w', termz); + trans_term (zi, zapt, termz); - key.sysno = atoi (termz+1); + key.sysno = atoi (termz); if (key.sysno <= 0) key.sysno = 1; rset_write (result, rsfd, &key); @@ -1374,8 +1285,11 @@ static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, if (relation_value == 102) /* relevance relation */ return rpn_search_APT_relevance (zi, zapt, attributeSet, num_bases, basenames); - return rpn_search_APT_word (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 */ @@ -1404,8 +1318,8 @@ static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, case 107: /* local-number */ return rpn_search_APT_local (zi, zapt, attributeSet); case 108: /* string */ - return rpn_search_APT_word (zi, zapt, attributeSet, - num_bases, basenames); + return rpn_search_APT_phrase (zi, zapt, attributeSet, + num_bases, basenames); case 109: /* numeric string */ break; } @@ -1555,6 +1469,7 @@ int rpn_search (ZServerInfo *zi, oident *attrset; oid_value attributeSet; + dict_grep_cmap (zi->wordDict, map_chrs_input); zlog_rpn (rpn); zi->errCode = 0; @@ -1576,23 +1491,23 @@ int rpn_search (ZServerInfo *zi, return zi->errCode; } +struct scan_info_entry { + char *term; + ISAM_P isam_p; +}; + struct scan_info { - struct scan_entry *list; + struct scan_info_entry *list; ODR odr; int before, after; - ISAM isam; char prefix[20]; }; static int scan_handle (char *name, const char *info, int pos, void *client) { int len_prefix, idx; - ISAM_P isam_p; - RSET rset; struct scan_info *scan_info = client; - rset_isam_parms parms; - len_prefix = strlen(scan_info->prefix); if (memcmp (name, scan_info->prefix, len_prefix)) return 1; @@ -1600,113 +1515,261 @@ static int scan_handle (char *name, const char *info, int pos, void *client) idx = scan_info->after - pos + scan_info->before; else idx = - pos - 1; + logf (LOG_DEBUG, "%-3d %s", idx, name+len_prefix); scan_info->list[idx].term = odr_malloc (scan_info->odr, strlen(name + len_prefix)+1); strcpy (scan_info->list[idx].term, name + len_prefix); - assert (*info == sizeof(isam_p)); - memcpy (&isam_p, info+1, sizeof(isam_p)); - parms.is = scan_info->isam; - parms.pos = isam_p; -#if 1 - rset = rset_create (rset_kind_isam, &parms); - count_set (rset, &scan_info->list[idx].occurrences); - rset_delete (rset); -#else - scan_info->list[idx].occurrences = 1; -#endif - logf (LOG_DEBUG, "pos=%3d idx=%3d name=%s", pos, idx, name); + assert (*info == sizeof(ISAM_P)); + memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAM_P)); return 0; } -static int dummy_handle (char *name, const char *info, void *p) -{ - return 0; +static void scan_term_untrans (ODR odr, char **dstp, const char *src) +{ + char *dst = odr_malloc (odr, strlen(src)*2+1); + *dstp = dst; + + while (*src) + { + const char *cp = map_chrs_output (&src); + while (*cp) + *dst++ = *cp++; + } + *dst = '\0'; } int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, + oid_value attributeset, int num_bases, char **basenames, int *position, int *num_entries, struct scan_entry **list, int *status) { - int i, j, sizez, max_pos; + int i; int pos = *position; int num = *num_entries; int before; int after; + int base_no; char termz[IT_MAX_WORD+20]; AttrType use; int use_value; - Z_Term *term = zapt->term; - struct scan_info scan_info; + AttrType completeness; + int completeness_value; + struct scan_info *scan_info_array; + struct scan_entry *glist; + int ords[32], ord_no = 0; + int ptr[32]; logf (LOG_DEBUG, "scan, position = %d, num = %d", pos, num); - if (num_bases != 1) - return 111; - scan_info.before = before = pos-1; - scan_info.after = after = 1+num-pos; - scan_info.odr = zi->odr; - - logf (LOG_DEBUG, "scan, before = %d, after = %d", before, after); - - scan_info.isam = zi->wordIsam; - scan_info.list = odr_malloc (zi->odr, (before+after)* - sizeof(*scan_info.list)); - for (j = 0; jerrCode = 114; + } + if (zebTargetInfo_curDatabase (zi->zti, basenames[base_no])) + { + zi->errString = basenames[base_no]; + return zi->errCode = 109; /* Database unavailable */ + } + for (local_attr = attp->local_attributes; local_attr && ord_no < 32; + local_attr = local_attr->next) + { + int ord; - dict_lookup_grep (zi->wordDict, termz, 0, NULL, &max_pos, 0, - dummy_handle); - if (max_pos <= strlen(*basenames)) + ord = zebTargetInfo_lookupSU (zi->zti, attp->attset_ordinal, + local_attr->local); + if (ord > 0) + ords[ord_no++] = ord; + } + } + if (ord_no == 0) + return zi->errCode = 113; + before = pos-1; + after = 1+num-pos; + scan_info_array = odr_malloc (zi->odr, ord_no * sizeof(*scan_info_array)); + for (i = 0; i < ord_no; i++) { - zi->errString = *basenames; - return zi->errCode = 109; /* Database unavailable */ + int j, prefix_len = 0; + int before_tmp = before, after_tmp = after; + struct scan_info *scan_info = scan_info_array + i; + + scan_info->before = before; + scan_info->after = after; + scan_info->odr = zi->odr; + + scan_info->list = odr_malloc (zi->odr, (before+after)* + sizeof(*scan_info->list)); + 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] = 0; + strcpy (scan_info->prefix, termz); + + trans_scan_term (zi, zapt, termz+prefix_len); + + dict_scan (zi->wordDict, termz, &before_tmp, &after_tmp, scan_info, + scan_handle); } - strcpy (scan_info.prefix, termz); - sizez = term->u.general->len; - if (sizez > IT_MAX_WORD) - sizez = IT_MAX_WORD; - for (j = 0; ju.general->buf[j]); - termz[j+i] = '\0'; + glist = odr_malloc (zi->odr, (before+after)*sizeof(*glist)); + for (i = 0; i < ord_no; i++) + ptr[i] = before; - dict_scan (zi->wordDict, termz, &before, &after, &scan_info, scan_handle); - *status = BEND_SCAN_SUCCESS; - - for (i = 0; iodr, &glist[i+before].term, mterm); + parms.is = zi->wordIsam; + parms.pos = scan_info_array[j0].list[ptr[j0]].isam_p; + rset = rset_create (rset_kind_isam, &parms); + + ptr[j0]++; + for (j = j0+1; jwordIsam; + parms.pos = scan_info_array[j].list[ptr[j]].isam_p; + rset2 = rset_create (rset_kind_isam, &parms); + + bool_parms.key_size = sizeof(struct it_key); + bool_parms.cmp = key_compare; + bool_parms.rset_l = rset; + bool_parms.rset_r = rset2; + + rset = rset_create (rset_kind_or, &bool_parms); + + ptr[j]++; + } + } + count_set (rset, &glist[i+before].occurrences); + rset_delete (rset); + } + if (i < after) + { + *num_entries -= (after-i); *status = BEND_SCAN_PARTIAL; + } + + for (i = 0; i 0)) + { + j0 = j; + mterm = tst; + } + } + if (j0 == -1) break; - if (i) - *status = BEND_SCAN_PARTIAL; - *position -= i; - *num_entries -= i; - *list = scan_info.list+i; /* list is set to first 'real' entry */ + scan_term_untrans (zi->odr, &glist[before-1-i].term, mterm); - if (*num_entries == 0) /* signal 'unsupported use-attribute' */ - zi->errCode = 114; /* if no entries was found */ + parms.is = zi->wordIsam; + parms.pos = scan_info_array[j0].list[before-1-ptr[j0]].isam_p; + rset = rset_create (rset_kind_isam, &parms); + + ptr[j0]++; + + for (j = j0+1; jwordIsam; + parms.pos = scan_info_array[j].list[before-1-ptr[j]].isam_p; + rset2 = rset_create (rset_kind_isam, &parms); + + bool_parms.key_size = sizeof(struct it_key); + bool_parms.cmp = key_compare; + bool_parms.rset_l = rset; + bool_parms.rset_r = rset2; + + rset = rset_create (rset_kind_or, &bool_parms); + + ptr[j]++; + } + } + count_set (rset, &glist[before-1-i].occurrences); + rset_delete (rset); + } + i = before-i; + if (i) + { + *status = BEND_SCAN_PARTIAL; + *position -= i; + *num_entries -= i; + } + *list = glist + i; /* list is set to first 'real' entry */ + logf (LOG_DEBUG, "position = %d, num_entries = %d", *position, *num_entries); if (zi->errCode) logf (LOG_DEBUG, "scan error: %d", zi->errCode); - return 0; + return zi->errCode; } - -