Hi Martin,
> On 10/9/18 11:03 AM, Rainer Orth wrote:
>> Hi Martin,
>>
>>> rename from gcc/testsuite/g++.dg/ext/pr82625.C
>>> rename to gcc/testsuite/g++.target/i386/pr82625.C
>>> index 59b174f8c51..0eb70baed5e 100644
>>> --- a/gcc/testsuite/g++.dg/ext/pr82625.C
>>> +++ b/gcc/testsuite/g++.target/i386/pr82625.C
>>> @@ -1,7 +1,7 @@
>>> /* { dg-do compile } */
>>> /* { dg-require-ifunc "" } */
>>> /* { dg-options "-O2 -fdump-tree-optimized" } */
>>> -/* { dg-do compile { target i?86-*-* x86_64-*-* } } */
>>> +/* { dg-do compile } */
>>
>> this is wrong: two instances of dg-do and dg-do after target selections
>> don't play well with one another. When you added it yesterday, you
>> badly broke the test on all x86 targets without ifunc support (probably
>> everything non-Linux):
>>
>> +UNRESOLVED: g++.dg/ext/pr82625.C -std=gnu++11 scan-tree-dump-times
>> optimized "return 4096;" 1
>> +UNRESOLVED: g++.dg/ext/pr82625.C -std=gnu++11 scan-tree-dump-times
>> optimized "return 8192;" 1
>> +FAIL: g++.dg/ext/pr82625.C -std=gnu++11 (test for excess errors)
>>
>> Excess errors:
>> /vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/ext/pr82625.C:7:17:
>> error: multiversioning needs ifunc which is not supported on this target
>> /vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/ext/pr82625.C:24:1:
>> error: use of multiversioned function without a default
>>
>> Just remove the second dg-do line.
>
> Thanks for testing!
>
> I'm sending updated version of the patch.
I noticed that before the tests were run with all of
-std=(c|gnu)++(98|11|14), now with no explict -std option. I wonder if
this is an issue.
Rainer
--
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University
