This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA] [patch] 'info symbol' to print more info
On Sat, Nov 15, 2008 at 11:01 AM, Joel Brobecker <brobecker@adacore.com> wrote:
>
> > But here, the message is already built up from 3 separate conditions,
> > and I am adding one more. What you are proposing then leads to a
> > chain of 16 if/then/else clauses. I don't think that's really
> > appropriate ...
>
> I haven't looked very closely at the details of the patch since Michael
> and Eli already did, so I can't comment on the exact number of if
> branches.
Note that the patch merely adds one more clause to existing three.
What Andreas is asking me to do here is rewrite existing code ...
> But, generally speaking, we just don't have much choice if
> we want to support i18n well.
Maybe 'info symbol' is sufficiently obscure that we don't need to
support i18n well for it?
> That being said, I agree that 16 branches is a large number, and perhaps
> we should let go of some of them.
If we always print symbol offset (even when 0), that eliminates
one branch.
> For instance, there was this discussion
> about not printing the name of the objfile if MULTI_OBJFILE_P. If we
> get rid of that, does it reduce the number of cases?
I think the answer is no, because we still have to check for:
&& osect->objfile && osect->objfile->name
> The idea itself
> is nice, but perhaps code simplicity is more important in this case.
> Are there any other cosmectic features that we can get rid of to
> reduce the number of cases further?
Except for the offset mentioned above, I don't see what else could
be cut.
Thanks,
--
Paul Pluzhnikov