On Tue, 13 May 2014, Joseph S. Myers wrote:
> On Mon, 12 May 2014, Hans-Peter Nilsson wrote:
> > On Thu, 24 Apr 2014, Jeff Law wrote:
> > > On 04/16/14 18:20, seg...@kernel.crashing.org wrote:
> > > > PR target/60822
> > > > 2014-04-16 Segher Boessenkool <seg...@kernel.crashing.org>
> > > >
> > > > * config/m68k/m68k.md (extendplussidi): Don't allow memory for
> > > > operand 1.
> > > Thanks. I tweaked the comment and added the testcase to the regression
> > > suite
> > > and installed the fix on the trunk.
> > In the commit message:
> > Added:
> > trunk/gcc/testsuite/gcc.c-torture/execute/pr60822.c
> > trunk/gcc/testsuite/gcc.c-torture/execute/pr60822.x
> >
> > I hope it's not too officious to remind people that dg- markup
> > can be used *even* in the c-torture test-suite these days;
> > gating in .x files is not necessary (since a few years, IIRC).
>
> I believe bug 20567 is still current: gcc.c-torture/execute ignores dg-
> directives and still needs .x files, until someone converts it to the dg
> harness.
I didn't believe that to be correct, so I checked; it is.
To wit, removing the .x file and adding a top line
"/* { dg-require-effective-target int16 } */"
didn't stop the test from being executed for an "int32plus"
target. Also, no sign of dg-anything in c-torture.exp.
Sorry for being wrong.
Let me instead suggest sticking to gcc.dg/torture for new tests.
There I'd be right. :)
brgds, H-P
