This is the mail archive of the mailing list for the glibc 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: [PATCH roland/test-unsupported] Let tests result in UNSUPPORTED; use that for unbuildable C++ cases

On Sun, 8 Mar 2015, Carlos O'Donell wrote:

> I'm less fine with XPASS failing the entire make check run. There are cases
> that I want to XFAIL for Fedora that include intermittent test failures related
> to kernel or glibc issues. If those tests pass and become XPASS, then `make check`

I'd like to see such XFAILs go in the main glibc repository, not just in 
distribution versions, if the cause of the failure isn't distribution 
glibc changes.  (It would be desirable to comment them e.g. "this fails 
intermittently with kernels before commit abcdef, which was in kernel g.h 
and later", so we can tell when such XFAILs are no longer needed - or if 
the failure is well-enough understood to know there's a glibc bug 
involved, the comment should point to that bug in Bugzilla.  We could do 
with a policy on how conditional to try to make such XFAILs.)

Generically, I think listing known failures on the per-release wiki pages 
is worse than XFAILing them, at least after some analysis has been done to 
understand the cause of the failure and how to tell when it's been fixed.

Joseph S. Myers

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