This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [testsuite] Require effective target weak_undefined for visibility-22.c
- From: Rainer Orth <ro at CeBiTec dot Uni-Bielefeld dot DE>
- To: Jakub Jelinek <jakub at redhat dot com>
- Cc: Tom de Vries <Tom_deVries at mentor dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, Mike Stump <mikestump at comcast dot net>
- Date: Thu, 14 Dec 2017 15:18:21 +0100
- Subject: Re: [testsuite] Require effective target weak_undefined for visibility-22.c
- Authentication-results: sourceware.org; auth=none
- References: <dbc6cd45-9318-4048-45f7-b469ebbef8b7@mentor.com> <20171214134735.GR2353@tucnak> <cf6efd8a-9362-78fb-7150-d939b4c35256@mentor.com> <20171214141425.GS2353@tucnak>
Hi Jakub,
> On Thu, Dec 14, 2017 at 03:09:02PM +0100, Tom de Vries wrote:
>> On 12/14/2017 02:47 PM, Jakub Jelinek wrote:
>> > On Thu, Dec 14, 2017 at 02:40:12PM +0100, Tom de Vries wrote:
>> > > --- a/gcc/testsuite/gcc.dg/visibility-22.c
>> > > +++ b/gcc/testsuite/gcc.dg/visibility-22.c
>> > > @@ -1,6 +1,7 @@
>> > > /* PR target/32219 */
>> > > /* { dg-do run } */
>> > > /* { dg-require-visibility "" } */
>> > > +/* { dg-require-effective-target weak_undefined } */
>> > > /* { dg-options "-O2 -fPIC" { target fpic } } */
>> > > /* This test requires support for undefined weak symbols. This support
>> > > is not available on hppa*-*-hpux*. The test is skipped rather than
>> >
>> > Shouldn't then the:
>> > /* This test requires support for undefined weak symbols. This support
>> > is not available on hppa*-*-hpux*. The test is skipped rather than
>> > xfailed to suppress the warning that would otherwise arise. */
>> > /* { dg-skip-if "" { "hppa*-*-hpux*" "*-*-aix*" "*-*-darwin*" } } */
>> > stuff be dropped too?
>>
>> I don't know whether the new effective target test will fail for each of
>> these 3 targets. But the warning mentioned for hppa*-*-hpux* will make the
>> effective target test fail, so I think that one can be removed.
>
> Or you can remove all 3, and if unsure, just add those to the weak_undefined
> effective target (return 0 for them). And ask target maintainers to verify
> and perhaps remove.
I'd do it the other way round: remove dg-skip-if completely and ping the
target maintainers to check (and eventually improve) the proc.
Otherwise those (probably unnecessary) special cases tend to stay
around forever ;-)
Rainer
--
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University