[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [cdt-debug-dev] About working with memory in debugger
|
Thomas Fletcher writes:
> On Tue, 29 Oct 2002, Elena Zannoni wrote:
>
> > Thomas Fletcher writes:
> > > On Tue, 29 Oct 2002, Igor S Zamjatin wrote:
> > >
> > > > Hi all!
> > > >
> > > > At first I want to say that all CDT Developers do a great job and want to
> > > > thank them a lot.
> > > >
> > > > I want to ask when approximately we can expect some version of Memory
> > > > View. May be prior variant is in CVS Repository already?
> > >
> > > It is already in CVS and if you look at the history on this list it
> > > is undergoing development right now!
> > >
> > > > And another question is will it possible to change memory from Eclipse-CDT
> > > > and how it will be connected with GDB/MI protocol? As I can see there is
> > > > no any command that change memory. Some -var* commands do similar actions
> > > > but I'm not shure that it is possible to write something directly by
> > > > memory address using that commands.
> > >
> > > I haven't looked at the code for the Memory View specifically
> > > but you can always do a manual set of a random memory address
> > > to a value if I remember correctly. Something along the lines
> > > of:
> > > set *(int*)(0x8034565) = 12
> > > should work (I'm certain I've used this in the past though the syntax
> > > looks a little gunky).
> > >
> >
> > You should use the memory commands:
> > mi_cmd_data_write_memory
> > mi_cmd_data_read_memory
> >
> > they are in mi-main.c (which admittedly is not the right place). They
> > haven't been extensively tested. As you find bugs/problems you can
> > report them on gdb@xxxxxxxxxxxxxxxxxx and enter bugs in the gdb PR
> > database at:
> > http://sources.redhat.com/gdb/bugs/
> > there is an MI category for the bugs.
>
> Elena,
>
> Thanks for the update =;-) I actually am not conversant in
> MI, I was just indicating how as a user you could conceivably
> do such a thing from the gdb command line and naturally assumed
> that if there wasn't a better way to do it (which you've pointed
> out) that this could be one way to try it.
Yes, sorry if my tone was a bit upset, I realized just today that my
mail didn't get out (I was travelling and had a bad connection). No
wonder nobody aknowledged my first post. :-(
>
> Of course this is why having the source is so valuable ... since
> you can just dive into the CDT MI layer and check out the command
> which is being transmitted.
>
Yes. I think there were thoughts of implementing memory read and
writes using varobj objects, but I don't remember the details, it was
a long time back. I would be interested in any bugs you guys find in
the current implementation, though.
Elena
> Thanks,
> Thomas
> -------------------------------------------------------------
> Thomas (toe-mah) Fletcher QNX Software Systems
> thomasf@xxxxxxx OS Technology Developer
> (613)-591-0931 http://www.qnx.com/
>
> _______________________________________________
> cdt-debug-dev mailing list
> cdt-debug-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/cdt-debug-dev