This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: Two feature suggestions
- From: Daniel Jacobowitz <drow at mvista dot com>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: Gdb List <gdb at sources dot redhat dot com>
- Date: Sat, 9 Nov 2002 17:39:58 -0500
- Subject: Re: Two feature suggestions
- References: <87lm42bct4.fsf@fleche.redhat.com>
On Sat, Nov 09, 2002 at 03:25:43PM -0700, Tom Tromey wrote:
> While debugging libgcj today I ran across a couple features I would
> find useful. These are applicable to both C++ and Java.
>
> First, if you have a large inheritance tree, sometimes you don't know
> the class which actually defines a given method. For instance, in my
> case I want to set a breakpoint on the `validate' method. I'm looking
> at an object of type `Window'. Window doesn't implement validate, it
> inherits it from its superclass.
>
> I'd like to be able to type `b java.awt.Window.validate' and have gdb
> find the actual implementation for me. Traversing the inheritance
> tree is more easily done by gdb than by me.
It should already do this. If it doesn't then that's just one of many
ways our C++ support isn't coping with Java.
> Also, in my case I only want a breakpoint for that method for my
> particular window. I'd like to be able to type something like:
>
> b <expression>.validate
>
> and have it do something like:
>
> b java.awt.Window.validate
> cond this == <value of expression>
>
> That is, set an object-specific breakpoint on the actual `validate'
> method that will be called on the object I'm interested in.
>
> These are both convenience operations -- I can do them by hand, but
> I'd prefer that gdb do the lookups and such for me.
I'm not so sure about this one; since you can already say this and it
will just breakpoint on the method...
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer