Re: RFC: implement "catch signal"

[Dodji Seketeli <dodji at redhat dot com>]

Tom Tromey <tromey@redhat.com> writes:

> Jan> I do not know how serious is the backward compatibility of
> Jan> type="catchpoint".
>
> Tom> Me neither.
>
> On irc, Andrà said that at least his front end isn't handling catchpoint
> at all.  So I would guess that changing this, at least for him, would be
> fine.
>
> CCing Marc Khouzam, to get the Eclipse perspective, and Dodji for
> nemiver.

Actually, Nemiver doesn't handle catchpoint either.  It's on my todo
list.

> The background is that right now all catchpoints report
> type="catchpoint" in MI; but it seems better to me to make this report
> the real catchpoint type, e.g., type="catch-load" or something like
> that.
>
> The alternative is something like type="catchpoint",catch-type="load".

I tend to prefer this alternative, but really, I don't have any strong
opinion about this.  I'd say that if the information about the different
types of catchpoints is present, it's fine.

Just curious, what would be the drawback of this alternative, compared
to, say, type='catch-load'?

> I sometimes think we should have a special "MI discuss" list just so we
> can work these things out without requiring all the MI consumer authors
> to filter through the main lists.  Just CCing a couple people whose
> names I remember isn't very good... :)

Hehe, yeah, i'd welcome such a list.

Thanks.

-- 
		Dodji