Kai, that test will always fail on some systems. How about this: just test for whether that function completes at all, rather than testing for success. - Dan
On Wednesday 14 May 2008 18:37:56 Dan Kegel wrote:
Kai, that test will always fail on some systems. How about this: just test for whether that function completes at all, rather than testing for success.
Seriously, if an ISP gets you to a spam page for nonexistant.winehq.org, can't we sue them for abusing the winehq.org trade mark? This really seems like breaking a valid test just to work around broken ISPs.
I admit that I'm a bit miffed that I didn't catch this in the 1-hour window between you sending that patch and Alexandre committing it, as it's now my time that's wasted by having to try again and again for a patch that gets past the discussion the initial patch should have gotten.
But yeah, I agree removing the ok() is acceptable as I'm mostly checking for a crash bug. I'll send another patch.
Cheers, Kai
-
Dan Kegel -
Kai Blin