[lto][patch] Add comdat support on the symbol table

[Janis Johnson]

On Thu, 2008-08-21 at 09:37 -0400, Diego Novillo wrote:
> On Thu, Aug 21, 2008 at 09:29, Rafael Espindola <espindola@google.com> wrote:
> >> Patch is OK.  Thanks.
> >>
> >> I'm wondering, though, if we shouldn't have some kind of validation
> >> built in parse_table_entry().
> >
> > I can add some tests. The test system has support for assembling a .s
> > file and running a test program on the resulting .o?
> 
> I meant asserts in the parser, making sure what is read has some
> semblance of validity.  But, yes, dejagnu tests should be possible.  I
> think we can access the .o files, but am not sure.  Janis?

A test can compile and assemble a source file to a .o with
"dg-do assemble".  If you really want to start with a .s file then
you'd need a .exp file to handle that, but it should be possible.
A quick scan through tests that use "dg-do assemble" didn't find
anything that looks at the .o but it would certainly be possible
with a procedure invoked from dg-final.

Janis