Re: [testsuite]: Skip tests for targets with int < 32 bits

[Mike Stump <mikestump@comcast.net>]

On Jun 13, 2011, at 11:18 AM, Georg-Johann Lay wrote:
> Who is a "specific maintainer" here?

I'd be happy to have the author of the testcase weigh in, or someone that cares about int32plus, or the the person that fixed the bug in the compiler for which the testcase was written...

> I found you (and Rainer Orth) as "testsuite" maintainers.
> Or does that just refer to the test harness and not to specific test cases?

I consider the situation much like the role of a global write privs person.  They, in theory, can approve a C++ change, but such changes are at least at times, better reviewed and approved by a C++ person.  I'd rather haver a arm person ok the gcc.target/arm/* testcases, I'd rather have an x86_64 person review and approve gcc.target/i386, I'd rather have a fortran person review gcc.fotran changes.

> Then I observed that
> /* { dg-require-effective-target int32plus } */
> does not work as intended for all test cases.

Ah, yes, right.  Longer term, I think that's a bug we should fix.  For now, as Jakub points out, you need to create a .x file for them.  I'd like to see the .x files go away.

> For exammple, I added this line to, e.g.
> * gcc.c-torture/execute/cmpsi-2.c
> * gcc.c-torture/execute/pr45262.c
> in trunk r172757
> http://gcc.gnu.org/viewcvs?view=revision&revision=172757
> 
> However, these tests are still executed as you can see in gcc-testresults for trunk r174959:
> http://gcc.gnu.org/ml/gcc-testresults/2011-06/msg01429.html

I don't understand why you'd propose a change that doesn't work?  In general, maintainers rely upon contributors to test and ensure that a change does something worthwhile.  So, yes, I agree with Jakub, this part of the patch should be reverted, and an .x file created/modified, or possibly the testcase modified to be more portable.  If you want to enhance the driver to process the dg stuff, I like that direction.

> Is this a bug resp. worth a bug report?

If you want, though, I think it might make more sense in a forward looking design document, or a open projects file.  It isn't a bug, because it was never a feature.