Rainer Orth <[email protected]> writes:
> Mike Stump <[email protected]> writes:
>
>> On Apr 5, 2011, at 1:56 AM, Rainer Orth wrote:
>>> * lto.c (promote_var): Only set VISIBILITY_HIDDEN if
>>> HAVE_GAS_HIDDEN.
>>
>> Oh, at a minimum, if TARGET_ASM_ASSEMBLE_VISIBILITY is set, doing this stuff
>> I think is useful?
>
> No, this won't work. E.g. on Solaris with an assembler without
> visibility support, TARGET_ASM_ASSEMBLE_VISIBILITY is set, but just
> emits a warning. This is similiar to default_assemble_visibility with
> HAVE_GAS_HIDDEN undefined.
>
> Right now, there are four definitions of TARGET_ASM_ASSEMBLE_VISIBILITY:
>
> * i386/cygming.h: i386_pe_assemble_visibility only warns about
> visibility attributes, so no problem here.
>
> * darwin.h: darwin_assemble_visibility is the only implementation which
> can handle VISIBILITY_HIDDEN (only), but doesn't define
> HAVE_GAS_HIDDEN. Maybe it should, but one would have to check every
> instance of the macro to make sure there are no ill effects.
>
> * rs6000/rs6000.c: protected by HAVE_GAS_HIDDEN.
>
> * sol2.h: warns unless HAVE_GAS_HIDDEN.
I've had a closer look now and think it's possible (and desirable) to
define HAVE_GAS_HIDDEN for Darwin, too. I've now (after lots of
trouble, and without success getting Ada to bootstrap on PowerPC Darwin)
set up a development environment on Mac OS X 10.5, both i386 and powerpc.
My current plan (though this may be slow) is to define HAVE_GAS_HIDDEN
for Darwin in gcc/configure.ac and check what else is necessary to make
this work. Once that is done, my patch can probably go in.
Additionally, one might want to rename HAVE_GAS_HIDDEN to
HAVE_AS_VISIBILITY since that's what the macro really means. (Actually,
that's a lie: it means HAVE_AS_LD_VISIBILITY, but I don't think we need
to become that verbose.)
Does this sound reasonable?
Thanks.
Rainer
--
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University
