On Wed, Jan 31, 2018 at 11:28:26AM +0100, Jan Hubicka wrote:
> > On Wed, Jan 31, 2018 at 10:16:10AM +0100, Jan Hubicka wrote:
> > > > On Wed, Jan 31, 2018 at 09:51:31AM +0100, Christophe Lyon wrote:
> > > > > Even with Rainer's fix, I've noticed that the new test fails on
> > > > > arm/aarch64:
> > > > > FAIL: g++.dg/torture/pr81360.C -O2 -flto -fuse-linker-plugin
> > > > > -fno-fat-lto-objects scan-ipa-dump icf "Equal symbols: 0"
> > > >
> > > > With -flto -fno-fat-lto-objects I believe we just write the bytecode
> > > > before
> > > > IPA passes and stop, so dg-do compile test won't really do icf at
> > > > compile
> > > > time, but at link time.
> > > > The more reason to move the test to g++.dg/ipa/ ...
> > >
> > > The testcase was about ICE on another testcase and I have copied the
> > > Equal symbols
> > > check by accident. We can drop all the dump scanning from the test, just
> > > we want
> > > to see if it builds.
> >
> > If you want to torture test it, please drop the -O2 from dg-options though
> > too.
> Like this?
>
> Index: testsuite/g++.dg/torture/pr81360.C
> ===================================================================
> --- testsuite/g++.dg/torture/pr81360.C (revision 257226)
> +++ testsuite/g++.dg/torture/pr81360.C (working copy)
> @@ -1,5 +1,5 @@
> /* { dg-do compile } */
> -/* { dg-options "-O2 -fno-early-inlining -fdump-ipa-icf" } */
> +/* { dg-options "-fno-early-inlining" } */
>
> template <int dim> class B;
> template <int, int dim> class TriaObjectAccessor;
> @@ -76,4 +76,3 @@ int main()
> return 0;
> }
>
> -/* { dg-final { scan-ipa-dump "Equal symbols: 0" "icf" } } */
Yes, thanks (with ChangeLog of course).
Jakub
