This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: ret_pointer_1.f90 XPASSes everywhere; propose to remove xfailline.

Paul Brook wrote:

On Saturday 13 November 2004 14:27, Toon Moene wrote:

But wouldn't removing the xfail line at least concentrate our efforts on
those few cases where it indeed, still, errors ?

I thought gcc policy was that the testsuite should have zero unexpected FAILs. Ie. spurious XPASSes are preferable to FAILs for known-broken tests.

Well, yes, but that's unobtainable. The testsuite will always show more failures on (64-bit) big-endian systems, because 32-bit little-endian targets will hide 'm.

You were also proposing removing the comments mentioning the related PR, which has definitely not been fixed.

Ah, OK, I didn't know that - that might be a reason to keep them, although removing the xfail would still help in highlighting the problematic (target | options) couple.

It's not a regression (that particular code never worked), so IMHO it can wait in line with any other wrong-code bugs we have :(

I agree - I've no idea where this goes wrong ...

Toon Moene - e-mail: - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77:
A maintainer of GNU Fortran 95:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]