On Wed, Jul 23, 2014 at 4:47 PM, Bernd Schmidt <ber...@codesourcery.com> wrote:
> This is an idea I discussed with a few folks at the Cauldron, and since they
> made supportive noises, I decided to work on it. The problem I'm trying to
> solve is that for ptx, I'll have to mark a lot of testcases as unsupported
> (uses of things such as indirect jumps, alloca, and sometimes K&R-style
> function definitions). In c-torture/execute, we can't use
> dg-require-effective-target directives, and using .x files feels a little
> antiquated. So I've started to work on a little shell script which migrates
> files out of c-torture/{compile,execute} and into gcc.dg/torture, or
> c-c++-common/torture if the test seems to work for C++ as well.
>
> Below is the current script, which produced no new failures on my latest run
> of testing on x86_64-linux. Some tests remain in c-torture (mostly because
> of warnings), but their number can be somewhat reduced by applying a
> preliminary patch first. This fixes up warnings in some testcases when doing
> so does not appear to affect what they are testing for. We can decide later
> what to do with the remaining tests (such as adding dg-warning as
> appropriate).
>
> One thing to note is that this script will migrate most of the limits tests
> to c-c++-common, increasing runtime of the testsuite a bit.
>
> Comments, objections? Ok to apply the preliminary patch?Yes, what if you don't move the tests but just change how the .exp to use the same infrastructure as gcc.dg/torture instead? I think changing the testcase in some cases is changing what is being tested so a script doing it automatically is not a good thing. Each testcase should be audited instead. Thanks, Andrew Pinski > > Any other targets I should add to the list of compilers that are run before > deciding where to move the test? Does anyone want to see the full patch in > the final submission (git doesn't seem to produce something nice for the > renames unfortunately) or just the script and generated ChangeLog? > > > Bernd
