Re: Regression for gdb.fortran/library-module.exp [Re: [RFA] choose symbol from given block's objfile first.]

[Joel Brobecker <brobecker@adacore.com>]

> > There is an intermediate solution, which is to always search the
> > main objfile first, and then the current DSO, and then the rest.
> 
> Some such order may be what ld.so does for the symbol search which
> should be the real right solution.

Is it possible to link an executable that defines a global variable
against a shared library that also defines a global variable with
the same name? It's something I tried on GNU/Linux, but the linker
rejected the link.

If that's something that's not possible, then I am OK with that
intermediate solution, because it means that real duplicates are
going to be only within shared libraries. If it's defined inside
the main objfile, then it cannot be defined as a global in any other
shared library, and we should be OK.

> Except for that there is probably wrong debug info in some cases from
> GCC and (a) I believe GDB should try to be compatible with current
> widespread buggy GCC debug info, (b) We may try to propose fixing the
> GCC produced debug info but I have no idea how yet, some such
> possibilities were described in
> 	http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40040

I agree with both.

-- 
Joel