Re: Debugging plans

[Martin Grabmueller]

> From: Neil Jerram <address@hidden>
> Date: 29 Jun 2001 17:26:37 +0100
> 
> >>>>> "Martin" == Martin Grabmueller <address@hidden> writes:
> 
>     Martin> I like this very much.  It's schemey and looks like it's
>     Martin> quite flexible.
> 
> Thanks!  I think Marius is also correct, that it's inefficient for
> some simple breakpoints, but we can mix and match.
> 
>     >> (Perhaps it would be better if these procedures simply returned
>     >> non-#f to indicate that their caller should then do
>     >> (debug-stack stack); this way the procedures could be logically
>     >> composed.)
> 
>     Martin> I'm not sure about it.  Maybe we should try out the
>     Martin> #f/non-#f returning variants first, because it offers more
>     Martin> flexibility.
> 
> Sorry, which way are you arguing for ??.  (I think you mean the
> non-(#f/non-#f returning) variants.)

I'm arguing for first trying #f/non-#f, and if that does not work out,
switch to the other variant.  That is because the latter can be
emulated with the former (invoking the debugger and then return #f).

>     Martin> First I thought about implementing the hook procedures
>     Martin> like this [...]
> 
>     Martin> but then I noticed that `proc' above is not known, because
>     Martin> it is any of the elements of the hook.  Hmmm.  It seems
>     Martin> that the hook implementation is not the Right Thing for
>     Martin> breakpoints.
> 
> Yes, perhaps we need something other than hooks.  But it seems like an
> uncharacteristic (for Scheme) language limitation that it isn't
> possible to construct a closure that includes a reference to itself.
> Are we missing something?

No, it is possible to for a closure to reference itself (as was
already pointed out by Thomas Bushnell).  I think I was on the wrong
track in my previous post.  It was only a few hours ago, and I can't
remember my own problem anymore...

Confused,
  'martin