X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=dict%2Flookgrep.c;h=d4e9873b3259fb28692d42da0c2f7c5403199a35;hp=f3a55047885472b3a1ead0c30b6040c20a7a5e2b;hb=3c5f6226f97612c0d6ac40591f600587c5ffa858;hpb=29357d2be4bc20d412d38bcb29876daedda28637 diff --git a/dict/lookgrep.c b/dict/lookgrep.c index f3a5504..d4e9873 100644 --- a/dict/lookgrep.c +++ b/dict/lookgrep.c @@ -1,10 +1,66 @@ /* - * 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 1.6 1995-10-09 16:18:32 adam + * 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. + * + * Revision 1.13 1996/01/08 09:09:30 adam + * Function dfa_parse got 'const' string argument. + * + * Revision 1.12 1995/12/11 09:04:48 adam + * Bug fix: the lookup/scan/lookgrep didn't handle empty dictionary. + * + * Revision 1.11 1995/12/06 14:43:02 adam + * New function: dict_delete. + * + * Revision 1.10 1995/11/16 17:00:44 adam + * Changed stupid log. + * + * Revision 1.9 1995/10/27 13:58:09 adam + * Makes 'Database unavailable' diagnostic. + * + * Revision 1.8 1995/10/19 14:57:21 adam + * New feature: grep lookup saves length of longest prefix match. + * + * Revision 1.7 1995/10/17 18:01:22 adam + * Userfunc may return non-zero in which case the the grepping stops + * immediately. + * + * Revision 1.6 1995/10/09 16:18:32 adam * Function dict_lookup_grep got extra client data parameter. * * Revision 1.5 1995/09/14 11:52:59 adam @@ -206,7 +262,8 @@ 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; @@ -235,8 +292,9 @@ static INLINE int move (MatchContext *mc, MatchWord *Rj1, MatchWord *Rj, static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, MatchWord *Rj, int pos, void *client, - int (*userfunc)(Dict_char *, const char *, void *), - Dict_char *prefix, struct DFA *dfa) + int (*userfunc)(char *, const char *, void *), + Dict_char *prefix, struct DFA *dfa, + int *max_pos, int init_pos) { int lo, hi, d; void *p; @@ -264,25 +322,33 @@ 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; + if (pos+j > *max_pos) + *max_pos = pos+j; if (ch == DICT_EOS) { if (was_match) - (*userfunc)(prefix, info+(j+1)*sizeof(Dict_char), - client); + if ((*userfunc)((char*) prefix, + info+(j+1)*sizeof(Dict_char), client)) + 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; @@ -294,6 +360,7 @@ 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 */ @@ -303,9 +370,15 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, memcpy (&ch, info+sizeof(Dict_ptr), sizeof(Dict_char)); prefix[pos] = ch; - move (mc, Rj1, Rj, ch, dfa, Rj_tmp); + if (pos > *max_pos) + *max_pos = pos; + 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) { @@ -313,19 +386,23 @@ 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; - (*userfunc)(prefix, info+sizeof(Dict_ptr)+ - sizeof(Dict_char), client); + if ((*userfunc)((char*) prefix, + info+sizeof(Dict_ptr)+ + sizeof(Dict_char), client)) + return 1; break; } } memcpy (&subptr, info, sizeof(Dict_ptr)); if (subptr) { - dict_grep (dict, subptr, mc, Rj1, pos+1, - client, userfunc, prefix, dfa); + if (dict_grep (dict, subptr, mc, Rj1, pos+1, + 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)); @@ -337,17 +414,34 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc, return 0; } -int dict_lookup_grep (Dict dict, Dict_char *pattern, int range, void *client, - int (*userfunc)(Dict_char *name, const char *info, - void *client)) +int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client, + int *max_pos, int init_pos, + int (*userfunc)(char *name, const char *info, + void *client)) { MatchWord *Rj; Dict_char prefix[MAX_LENGTH+1]; - char *this_pattern = pattern; + const char *this_pattern = pattern; MatchContext *mc; struct DFA *dfa = dfa_init(); int i, d; +#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) { @@ -376,10 +470,22 @@ int dict_lookup_grep (Dict dict, Dict_char *pattern, int range, void *client, } } } - i = dict_grep (dict, 1, mc, Rj, 0, client, userfunc, prefix, dfa); - + *max_pos = 0; + if (dict->head.last > 1) + i = dict_grep (dict, 1, mc, Rj, 0, client, userfunc, prefix, + dfa, max_pos, init_pos); + else + i = 0; + logf (LOG_DEBUG, "max_pos = %d", *max_pos); dfa_delete (&dfa); xfree (Rj); 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; +}