This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA] choose symbol from given block's objfile first.
> Joel> There are also some locations where the loop is inside a routine
> Joel> that does a lookup without "context". For instance,
> Joel> basic_lookup_transparent_type. In that case, the only real caller
> Joel> I could find was check_typedef, which I don't think we want to
> Joel> change.
>
> I'm not so sure. It seems like you could make a multi-objfile test case
> where an incomplete type is incorrectly resolved to a type in another
> objfile.
Yeah, I can see now that it should be relatively doable (using opaque
types). I'll keep that in mind for a rainy day. I still have to work
on enhancing the expresion parser for both C and Ada, as well as that PR
that I need to create... If I have a chance today, I'll try to create
a testcase that breaks check_typedef.
(someone is not looking forward to updating all calls to check_typedef :-))
So, just to be sure: Are we OK with this iteration for now?
--
Joel