On Fri, Apr 4, 2014, at 1:11, Aurélien Gâteau wrote: > On Sun, Mar 30, 2014, at 16:42, Albert Astals Cid wrote: > > Hello people, at the moment we have various 4.13 projects failing. > > > > Since we promised a while ago to not release without failing tests we > > have > > 10 days until tagging to fix the tests or we'll have to delay the > > release. > > > > If you need some environment to be running, it'd be good if your test > > could > > bring it up automagically, if not please talk with Ben I'm sure he'll > > be more than happy to try to help you in providing you that environment. > > > > I am also open to be convinced that the test is right and that it's > > unfixable > > to run correctly on jenkins, but make sure you are really convincing if > > you > > resort to that ;-) > > > > Gwenview (2 failing tests) > > > > http://build.kde.org/view/KDE%20SC%20stable/job/gwenview_stable/196/testReport/ > > I disabled the failing test in semanticinfobackendtest (that test did > not > make sense anymore with Baloo), but I can't reproduce the failure in > placetreemodeltest. > > The failure happens at setup time: the tests tries to rm kde > configuration > dir (ie $KDEHOME, after checking it ends with ".kde-unit-test"). Failing > snippet is the following: > > QString confDir = qgetenv("KDEHOME"); > QVERIFY(confDir.endsWith(".kde-unit-test")); // Better safe than > sorry > if (QFileInfo(confDir).isDir()) { > bool ok = KIO::NetAccess::del(KUrl::fromPath(confDir), 0); > QVERIFY(ok); > } > > the QVERIFY(ok) fails on CI, but not on my box. It sounds like another > situation where QFileInfo and KIO disagree. Ideas are welcome.
FYI, that test just passed. Looks like a race condition :/. On the other hand, Gwenview is green now \o/. Aurélien