X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=dict%2Flookgrep.c;h=bde37714457bb4346517134ce43492c00a81a2c1;hb=83533000f9456dcab2fc171abafd84d6104d4087;hp=2fd3533200d612af0f510c533923c035275aa1cf;hpb=b31a8654c5b7a597dd4a351f452ac91e5ef60798;p=idzebra-moved-to-github.git diff --git a/dict/lookgrep.c b/dict/lookgrep.c index 2fd3533..bde3771 100644 --- a/dict/lookgrep.c +++ b/dict/lookgrep.c @@ -1,10 +1,46 @@ /* - * Copyright (C) 1994, Index Data I/S + * Copyright (C) 1994-1999, Index Data * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: lookgrep.c,v $ - * Revision 1.14 1996-02-02 13:43:51 adam + * Revision 1.24 1999-05-26 07:49:12 adam + * C++ compilation. + * + * Revision 1.23 1999/05/15 14:36:37 adam + * Updated dictionary. Implemented "compression" of dictionary. + * + * Revision 1.22 1999/02/02 14:50:23 adam + * Updated WIN32 code specific sections. Changed header. + * + * 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. * @@ -91,13 +127,13 @@ static INLINE MatchWord get_bit (MatchContext *mc, MatchWord *m, int ch, static MatchContext *mk_MatchContext (struct DFA *dfa, int range) { - MatchContext *mc = xmalloc (sizeof(*mc)); + MatchContext *mc = (MatchContext *) xmalloc (sizeof(*mc)); int s; mc->n = (dfa->no_states+WORD_BITS) / WORD_BITS; mc->range = range; mc->fact = (range+1)*mc->n; - mc->match_mask = xcalloc (mc->n, sizeof(*mc->match_mask)); + mc->match_mask = (MatchWord *) xcalloc (mc->n, sizeof(*mc->match_mask)); for (s = 0; sno_states; s++) if (dfa->states[s]->rule_no) @@ -232,7 +268,8 @@ 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 +300,7 @@ 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; @@ -273,7 +310,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, dict_bf_readp (dict->dbf, ptr, &p); lo = 0; hi = DICT_nodir(p)-1; - indxp = (short*) ((char*) p+DICT_pagesize(dict)-sizeof(short)); + indxp = (short*) ((char*) p+DICT_bsize(p)-sizeof(short)); while (lo <= hi) { @@ -291,6 +328,7 @@ 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 +342,19 @@ 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 +366,7 @@ 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 +378,13 @@ 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 +392,7 @@ 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, @@ -359,11 +406,11 @@ 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) - -sizeof(short)); + indxp = (short*) ((char*) p+DICT_bsize(p)-sizeof(short)); } } } @@ -373,7 +420,7 @@ 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 +431,22 @@ 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) { @@ -395,7 +457,7 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, mc = mk_MatchContext (dfa, range); - Rj = xcalloc ((MAX_LENGTH+1) * mc->n, sizeof(*Rj)); + Rj = (MatchWord *) xcalloc ((MAX_LENGTH+1) * mc->n, sizeof(*Rj)); set_bit (mc, Rj, 0, 0); for (d = 1; d<=mc->range; d++) @@ -414,9 +476,10 @@ 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); + if (dict->head.root) + i = dict_grep (dict, dict->head.root, mc, Rj, 0, client, + userfunc, prefix, + dfa, max_pos, init_pos); else i = 0; logf (LOG_DEBUG, "max_pos = %d", *max_pos); @@ -425,3 +488,10 @@ 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; +}