Re: GNOME::Debugger::Commander proposal

[Martin Baulig <martin home-of-linux org>]

Elliot Lee <sopwith@redhat.com> writes:

> On 3 Oct 1999, Dave Camp wrote:
> 
> > Elliot Lee <sopwith@redhat.com> writes:
> > 
> > > > We did this because this interface aims to be asynchronous.  You will send
> > > > a set_breakpoint command, and at some point in the future, you will get
> > > > a "breakpoint set" event on the event channel. 
> > > 
> > > This is going to add a lot of complexity for no apparent reason.
> > 
> > There is a very good reason.  There are a number of things that need to 
> > be informed of an added breakpoint.  The editor needs to know, so that it
> > can put up a breakpoint indicator.  The breakpoint manager needs to know.
> > 
> > The one object that always knows when a breakpoint is added is the program
> > object.  So it becomes the program object's responsibility to notify 
> > interested components (via the event channel) that a breakpoint has been 
> > added.
> 
> This isn't a good reason to make it asynchronous, though. You can still do
> notification without making it async.
> 
> By returning the breakpoint number from the operation invocation, you give
> the caller the option of using the return value OR the async notification.
> I like having options. :)

Oh, so it probably wasn't such a "silly" question from me ...

-- 
Martin Baulig - martin@home-of-linux.org - http://www.home-of-linux.org