[DGD] error recovery with atomic functions

[bart at wotf.org]

On Thu, 25 Jun 2015 10:47:56 +0200, Felix A. Croes wrote
> bart at wotf.org wrote:
> 
> >[...]
> > On another note, I noticed you made a change to no longer call operator
> > functions in builtin object types. Seems I did have a use for that, I like
> > ob<-"/builtin/function" over object_name(ob) == "/builtin/function#-1", but
> > this no longer works. Oh well, guess I'll deal :-)
> 
> That is a bug, actually.  When I added the new INSTANCEOF 
> instruction, I failed to copy this functionality from the old kfun.  
> Pull requests which address this issue will be carefully considered. 
> :)

Not a pull request, but a stab at this anyway. Quick test suggests this works,
but it has been a while since I last meddled with such things :-)

Bart.

diff --git a/src/interpret.c b/src/interpret.c
index 18f596a..762b81c 100644
--- a/src/interpret.c
+++ b/src/interpret.c
@@ -2807,7 +2807,12 @@ static void i_interpret1(frame *f, char *pc)
                break;

            case T_LWOBJECT:
-               instance = i_instanceof(f, f->sp->u.array->elts->oindex, l);
+                if (f->sp->u.array->elts->type != T_OBJECT) {
+                    instance =
(strcmp(o_builtin_name(f->sp->u.array->elts->u.number),
+                               i_classname(f, l)) == 0);
+                } else {
+                   instance = i_instanceof(f, f->sp->u.array->elts->oindex, l);
+                }
                arr_del(f->sp->u.array);
                break;



> 
> Regards,
> Felix Croes
> ____________________________________________
> https://mail.dworkin.nl/mailman/listinfo/dgd


--
http://www.flickr.com/photos/mrobjective/
http://www.om-d.org/