On Tue, 21 Mar 2006 at 16:13:18 -0500, Alan Ruttenberg wrote:
> My patch clobbers decent printing of error messages in the debugger.
> Don't have time to track it down now. Immediate symptoms can be
> prevented by changing
>
> ((typep object 'standard-object)
>
> to
>
> (and (typep object 'standard-object) (not (typep object 'condition)))
Yeah, I noticed that too, when I was first looking at your patch, so I
tracked it down for the specific case I noticed and fixed it. The
problem, loosely speaking, is that the Java code to initialize the
condition object doesn't always follow the MAKE-INSTANCE initialization
protocol closely enough. Unfortunately, there's no single place to fix
this.
Most likely you've run into another specific case that needs fixing. If
you let me know the details (i.e. the specific types of the conditions
that don't print right), I'll try to clean things up.
I know this can be avoided by special-casing condition objects in
OUTPUT-UGLY-OBJECT, but I'd much rather fix the underlying brokenness.
Thanks.
-Peter
|