X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=dict%2Flookgrep.c;h=39d2b3f4dcfe0e28a28fd43a4be71dd51955bc3d;hb=0216e1f261655bf8b32fd4878d25135751bdc6fa;hp=2fd3533200d612af0f510c533923c035275aa1cf;hpb=b31a8654c5b7a597dd4a351f452ac91e5ef60798;p=idzebra-moved-to-github.git diff --git a/dict/lookgrep.c b/dict/lookgrep.c index 2fd3533..39d2b3f 100644 --- a/dict/lookgrep.c +++ b/dict/lookgrep.c @@ -1,10 +1,37 @@ /* - * Copyright (C) 1994, Index Data I/S + * Copyright (C) 1994-1998, Index Data I/S * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: lookgrep.c,v $ - * Revision 1.14 1996-02-02 13:43:51 adam + * Revision 1.21 1998-06-24 12:16:12 adam + * Support for relations on text operands. Open range support in + * DFA module (i.e. [-j], [g-]). + * + * 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 + * Changed prototype for chr_map_input - added const. + * Added support for C++, headers uses extern "C" for public definitions. + * + * Revision 1.17 1996/06/04 10:20:06 adam + * Added support for character mapping. + * + * Revision 1.16 1996/05/24 14:46:04 adam + * Added dict_grep_cmap function to define user-mapping in grep lookups. + * + * Revision 1.15 1996/03/20 09:35:18 adam + * Function dict_lookup_grep got extra parameter, init_pos, which marks + * from which position in pattern approximate pattern matching should occur. + * + * Revision 1.14 1996/02/02 13:43:51 adam * The public functions simply use char instead of Dict_char to represent * search strings. Dict_char is used internally only. * @@ -232,7 +259,8 @@ static void or (MatchContext *mc, MatchWord *Rdst, } static INLINE int move (MatchContext *mc, MatchWord *Rj1, MatchWord *Rj, - Dict_char ch, struct DFA *dfa, MatchWord *Rtmp) + Dict_char ch, struct DFA *dfa, MatchWord *Rtmp, + int range) { int d; MatchWord *Rtmp_2 = Rtmp + mc->n; @@ -263,7 +291,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, MatchWord *Rj, int pos, void *client, int (*userfunc)(char *, const char *, void *), Dict_char *prefix, struct DFA *dfa, - int *max_pos) + int *max_pos, int init_pos) { int lo, hi, d; void *p; @@ -291,6 +319,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, MatchWord *Rj0 = Rj + j *mc->fact; MatchWord *Rj1 = Rj + (j+1)*mc->fact; MatchWord *Rj_tmp = Rj + (j+2)*mc->fact; + int range; memcpy (&ch, info+j*sizeof(Dict_char), sizeof(Dict_char)); prefix[pos+j] = ch; @@ -304,15 +333,19 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, return 1; break; } - move (mc, Rj1, Rj0, ch, dfa, Rj_tmp); + if (pos+j >= init_pos) + range = mc->range; + else + range = 0; + move (mc, Rj1, Rj0, ch, dfa, Rj_tmp, range); for (d = mc->n; --d >= 0; ) - if (Rj1[mc->range*mc->n + d]) + if (Rj1[range*mc->n + d]) break; if (d < 0) break; was_match = 0; for (d = mc->n; --d >= 0; ) - if (Rj1[mc->range*mc->n + d] & mc->match_mask[d]) + if (Rj1[range*mc->n + d] & mc->match_mask[d]) { was_match = 1; break; @@ -324,6 +357,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, MatchWord *Rj1 = Rj+ mc->fact; MatchWord *Rj_tmp = Rj+2*mc->fact; Dict_char ch; + int range; /* Dict_ptr subptr */ /* Dict_char sub char */ @@ -335,9 +369,13 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, if (pos > *max_pos) *max_pos = pos; - move (mc, Rj1, Rj, ch, dfa, Rj_tmp); + if (pos >= init_pos) + range = mc->range; + else + range = 0; + move (mc, Rj1, Rj, ch, dfa, Rj_tmp, range); for (d = mc->n; --d >= 0; ) - if (Rj1[mc->range*mc->n + d]) + if (Rj1[range*mc->n + d]) break; if (d >= 0) { @@ -345,7 +383,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, if (info[sizeof(Dict_ptr)+sizeof(Dict_char)]) { for (d = mc->n; --d >= 0; ) - if (Rj1[mc->range*mc->n + d] & mc->match_mask[d]) + if (Rj1[range*mc->n + d] & mc->match_mask[d]) { prefix[pos+1] = DICT_EOS; if ((*userfunc)((char*) prefix, @@ -359,7 +397,8 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, if (subptr) { if (dict_grep (dict, subptr, mc, Rj1, pos+1, - client, userfunc, prefix, dfa, max_pos)) + client, userfunc, prefix, dfa, max_pos, + init_pos)) return 1; dict_bf_readp (dict->dbf, ptr, &p); indxp = (short*) ((char*) p+DICT_pagesize(dict) @@ -373,7 +412,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, } int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, - int *max_pos, + int *max_pos, int init_pos, int (*userfunc)(char *name, const char *info, void *client)) { @@ -384,7 +423,22 @@ 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); +#if 0 + debug_dfa_trav = 1; + debug_dfa_tran = 1; + debug_dfa_followpos = 1; + dfa_verbose = 1; +#endif + + 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); + i = dfa_parse (dfa, &this_pattern); if (i || *this_pattern) { @@ -416,7 +470,7 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, *max_pos = 0; if (dict->head.last > 1) i = dict_grep (dict, 1, mc, Rj, 0, client, userfunc, prefix, - dfa, max_pos); + dfa, max_pos, init_pos); else i = 0; logf (LOG_DEBUG, "max_pos = %d", *max_pos); @@ -425,3 +479,10 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, rm_MatchContext (&mc); return i; } + +void dict_grep_cmap (Dict dict, void *vp, + const char **(*cmap)(void *vp, const char **from, int len)) +{ + dict->grep_cmap = cmap; + dict->grep_cmap_data = vp; +}