Extra generic handle for the character mapping routines.

[idzebra-moved-to-github.git] / dfa / dfa.c

diff --git a/dfa/dfa.c b/dfa/dfa.c
index 276ca06..f55b69a 100644 (file)
--- a/dfa/dfa.c
+++ b/dfa/dfa.c
@@ -1,10 +1,20 @@
 /*
- * Copyright (C) 1994-1996, Index Data I/S 
+ * Copyright (C) 1994-1997, Index Data I/S 
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: dfa.c,v $
- * Revision 1.14  1996-10-29 13:57:22  adam
+ * Revision 1.17  1997-09-18 08:59:17  adam
+ * Extra generic handle for the character mapping routines.
+ *
+ * Revision 1.16  1997/09/05 15:29:57  adam
+ * Changed prototype for chr_map_input - added const.
+ * Added support for C++, headers uses extern "C" for public definitions.
+ *
+ * Revision 1.15  1997/02/10 10:19:20  adam
+ * Added facility for open character sets, eg [a-].
+ *
+ * Revision 1.14  1996/10/29 13:57:22  adam
  * Include of zebrautl.h instead of alexutil.h.
  *
  * Revision 1.13  1996/06/17 14:24:08  adam
@@ -61,6 +71,8 @@
 #include "dfap.h"
 #include "imalloc.h"
 
+#define DFA_OPEN_RANGE 1
+
 #define CAT     16000
 #define OR      16001
 #define STAR    16002
@@ -413,10 +425,11 @@ static int read_charset (void)
             break;
         if (parse_info->cmap)
         {
-            char **mapto, mapfrom[2];
+            const char **mapto;
+           char mapfrom[2];
             const char *mcp = mapfrom;
             mapfrom[0] = ch0;
-            mapto = (*parse_info->cmap)(&mcp, 1);
+            mapto = (*parse_info->cmap)(parse_info->cmap_data, &mcp, 1);
             assert (mapto);
             ch0 = mapto[0][0];
         }
@@ -424,25 +437,38 @@ static int read_charset (void)
         ch1 = nextchar_set (&esc1);
         if (!esc1 && ch1 == '-')
         {
+            int open_range = 0;
             if ((ch1 = nextchar_set (&esc1)) == 0)
                 break;
+#if DFA_OPEN_RANGE
+            if (!esc1 && ch1 == ']')
+            {
+                ch1 = 255;
+                open_range = 1;
+            }
+#else
             if (!esc1 && ch1 == ']')
             {
                 add_BSet (parse_info->charset, look_chars, '-');
                 break;
             }
-            if (parse_info->cmap)
+#endif
+            if (!open_range && parse_info->cmap)
             {
-                char **mapto, mapfrom[2];
+                const char **mapto;
+               char mapfrom[2];
                 const char *mcp = mapfrom;
                 mapfrom[0] = ch1;
-                mapto = (*parse_info->cmap) (&mcp, 1);
+                mapto = (*parse_info->cmap) (parse_info->cmap_data, &mcp, 1);
                 assert (mapto);
                 ch1 = mapto[0][0];
             }
             for (i=ch0; ++i<=ch1;)
                 add_BSet (parse_info->charset, look_chars, i);
-            ch0 = nextchar_set (&esc0);
+            if (!open_range)
+                ch0 = nextchar_set (&esc0);
+            else
+                break;
         }
         else
         {
@@ -457,7 +483,7 @@ static int read_charset (void)
 
 static int map_l_char (void)
 {
-    char **mapto;
+    const char **mapto;
     const char *cp0 = (const char *) (expr_ptr-1);
     int i = 0, len = strlen(cp0);
 
@@ -470,7 +496,7 @@ static int map_l_char (void)
     if (!parse_info->cmap)
         return L_CHAR;
 
-    mapto = (*parse_info->cmap) (&cp0, len);
+    mapto = (*parse_info->cmap) (parse_info->cmap_data, &cp0, len);
     assert (mapto);
     
     expr_ptr = (const unsigned char *) cp0;
@@ -1084,9 +1110,11 @@ struct DFA *dfa_init (void)
     return dfa;
 }
 
-void dfa_set_cmap (struct DFA *dfa, char **(*cmap)(const char **from, int len))
+void dfa_set_cmap (struct DFA *dfa, void *vp,
+                  const char **(*cmap)(void *vp, const char **from, int len))
 {
     dfa->parse_info->cmap = cmap;
+    dfa->parse_info->cmap_data = vp;
 }
 
 int dfa_parse (struct DFA *dfa, const char **pattern)