Hi Tobias! On 2019-10-25T18:17:26+0200, Tobias Burnus <tob...@codesourcery.com> wrote: > This patch is about: libgomp/testsuite/libgomp.fortran/, only > > The two test cases I added recently use 'call abort()', which is > nowadays frowned on as that's a ventor extension. Hence, I change it to > 'stop'. > > Additionally, the 'fortran.exp' in the directory states: "For Fortran > we're doing torture testing, as Fortran has far more tests with arrays > etc. that testing just -O0 or -O2 is insufficient, that is typically not > the case for C/C++." > > The torture testing is only done if there is a "dg-do run"; without > dg-do, it also does run, but only with a single compile flag setting.
(It was me who suggested that task.)
> I have only selectively added it; I think 'dg-do run' does not make
> sense for:
> * condinc*.f – only do uses '!$ include'
> * omp_cond*.f* – only tests '!$' code, including comments.
> Hence, I excluded those and only changed the others. (However, one can
> still argue about the remaining ones – such as 'omp_hello.f' or tabs*.f*.)
But why shouldn't these still be torture tested?
Anyway -- as usual ;-) -- I'd prefer if any such rationale ("not doing
torture testing because [...]") was put into the respective testsuite
files, so that it's obvious to the next person reading that file.
> --- a/libgomp/testsuite/libgomp.fortran/strassen.f90
> +++ b/libgomp/testsuite/libgomp.fortran/strassen.f90
> @@ -0,0 +1 @@
> +! { dg-do run }
| ! { dg-options "-O2" }
That's not a useful combination, is it? (Just noticed this one here;
didn't verify all files suggested to be changed.)
Not sure into which direction this should be fixed: continue to just test
'-O2', or remove the '-O2' override.
> --- a/libgomp/testsuite/libgomp.fortran/use_device_addr-1.f90
> +++ b/libgomp/testsuite/libgomp.fortran/use_device_addr-1.f90
> @@ -0,0 +1 @@
> +! { dg-do run }
> --- a/libgomp/testsuite/libgomp.fortran/use_device_addr-2.f90
> +++ b/libgomp/testsuite/libgomp.fortran/use_device_addr-2.f90
> @@ -0,0 +1 @@
> +! { dg-do run }
On powerpc64le-unknown-linux-gnu without offloading, I'm seeing (only)
the '-O0' execution tests FAIL for both these, with 'STOP 1' message.
Grüße
Thomas
signature.asc
Description: PGP signature
