Re: [Freetype] Another Indic GSUB problem

[Werner LEMBERG]

> The result is that when ContextSubst2() calls Get_Class() on the
> initial glyph of the context, it gets a Not_Covered error and gives
> up.  It seems from how this table is formed that it's supposed to
> assume class 0.

Yes.

> This sounds like the first glyph's class is assumed.  Modifying
> ftxgsub.c to make this assumption causes it to work.

I'll apply the patch below -- please check.


    Werner


======================================================================

--- ftxgsub.c.orig      Sat Mar 17 12:21:02 2001
+++ ftxgsub.c   Thu Aug  2 02:30:16 2001
@@ -2167,7 +2167,7 @@
 
     error = Get_Class( &csf2->ClassDef, in->string[in->pos],
                        &classes[0], NULL );
-    if ( error )
+    if ( error && error != TTO_Err_Not_Covered )
       goto End;
     known_classes = 0;
 
@@ -3506,7 +3506,7 @@
 
     error = Get_Class( &ccsf2->InputClassDef, in->string[in->pos],
                        &input_classes[0], NULL );
-    if ( error )
+    if ( error && error != TTO_Err_Not_Covered )
       goto End1;
 
     cscs = &ccsf2->ChainSubClassSet[input_classes[0]];