This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: RFC: Components
- To: Fernando Nasser <fnasser at cygnus dot com>
- Subject: Re: RFC: Components
- From: jtc at redback dot com (J.T. Conklin)
- Date: 16 Nov 2000 15:27:51 -0800
- Cc: gdb at sources dot redhat dot com
- References: <3A0B3556.510B850A@cygnus.com>
- Reply-To: jtc at redback dot com
>>>>> "Fernando" == Fernando Nasser <fnasser@cygnus.com> writes:
Fernando> We got a submission from a contributor that would like to
Fernando> add the following facility to GDB, which can be useful when
Fernando> debugging Operating Systems.
"Useful for debugging operating systems" may be stretching the truth a
bit, as it's only useful for those that bind objects into multiple
'components' linked at different addressed but are loaded into memory
at the same time. Seems a bit odd to me, but I can't discount it
completely. I once used a system that had assembly language 'stubs'
for calling real functions/system calls in a different component. But
since they were assembly and only a few lines long, there was no need
for debugger support.
Yet something bugs me about the implementation. The word "component"
doesn't seem right, but I can't think of a good alternative. "module"
seems too general as well. The other thing I'm uneasy about is that
component names are inherited from the names of the symbol files.
Again, I don't have any alternative to propose.
I'd suggest that the contributor feel free to come and participate in
this discussion. I think it will be more efficent and yeild better
results than design by proxy. Besides, we don't bite.
--jtc
--
J.T. Conklin
RedBack Networks