X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=dict%2Flookgrep.c;h=bde37714457bb4346517134ce43492c00a81a2c1;hb=df117f4652b6999cf689cf5e2e65f8d8541ab4b0;hp=13eb8cb1804356f038bc49c1256d6478e8b68021;hpb=c0c85097014e5b007075ab71d9b5bee412e28dd5;p=idzebra-moved-to-github.git

diff --git a/dict/lookgrep.c b/dict/lookgrep.c
index 13eb8cb..bde3771 100644
--- a/dict/lookgrep.c
+++ b/dict/lookgrep.c
@@ -1,10 +1,32 @@
 /*
- * Copyright (C) 1994-1996, Index Data I/S 
+ * Copyright (C) 1994-1999, Index Data
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: lookgrep.c,v $
- * Revision 1.18  1997-09-05 15:29:58  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.
  *
@@ -105,13 +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; s<dfa->no_states; s++)
         if (dfa->states[s]->rule_no)
@@ -288,7 +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)
     {
@@ -388,8 +410,7 @@ static int dict_grep (Dict dict, Dict_ptr ptr, MatchContext *mc,
                                    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));
                 }
             }
         }
@@ -410,9 +431,21 @@ 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);
+    dfa_set_cmap (dfa, dict->grep_cmap_data, dict->grep_cmap);
 
     i = dfa_parse (dfa, &this_pattern);
     if (i || *this_pattern)
@@ -424,7 +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++)
@@ -443,8 +476,9 @@ 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,
+    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;
@@ -455,8 +489,9 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client,
     return i;
 }
 
-void dict_grep_cmap (Dict dict,
-		     const char **(*cmap)(const char **from, int len))
+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;
 }