Same here. I don't have any Windows machines for development available, so I brought this up here to see if any Windows devs had any ideas. I'd set up a macOS CI build on Travis as well, but half of us are developing on macOS anyways, so it seems less important. Testing on GNU/Linux, however, has brought up some interesting issues regarding file name case sensitivity and timezones in the past which we fixed.
On 7 April 2018 at 19:14, Remko Popma <[email protected]> wrote: > Carter, > > To me it’s sufficient due diligence when you’ve run the tests on your > machine. > > The continuous integration builds are a smoke test to flush out issues on > other environments. They sometimes show issues none of us see when building > locally. > > I would only consider a failing test to break the build if other > committers cannot build locally anymore. > > Cheers, > Remko > > > > > On Apr 8, 2018, at 4:28, Matt Sicker <[email protected]> wrote: > > > > Everything builds and tests fine for me on macOS using master. > Interesting. > > > >> On 7 April 2018 at 14:06, Carter Kozak <[email protected]> wrote: > >> > >> I've been having issues with EquinoxLoadApiBundleTest, > >> FelixLoadApiBundleTest, and FileAppenderPermissionsTest (has some > >> special casing on osx which may not be compatible with my environment, > >> works on linux though) on osx, haven't had a chance to dig into them > >> quite yet. > >> > >> I've executed the tests on my changes prior to committing, but may > >> have missed something. If my change caused a regression or test flake, > >> please revert and I'll debug without blocking master. > >> > >> -ck > >> > >> On Sat, Apr 7, 2018 at 2:31 PM, Gary Gregory <[email protected]> > >> wrote: > >>> On Sat, Apr 7, 2018 at 12:15 PM, Matt Sicker <[email protected]> wrote: > >>> > >>>> In this build [1], it appears to be failing in some Windows MongoDB > >>>> integration tests. Might be a file permission assumption? > >>>> > >>> > >>> Could be a corrupt download: > >>> > >>> java.io.IOException: File > >>> C:\Users\jenkins\.embedmongo\win32\mongodb-win32-x86_64- > >> 2008plus-ssl-3.6.2.zip > >>> at de.flapdoodle.embed.process.extract.AbstractExtractor. > >> archiveStreamWithExceptionHint(AbstractExtractor.java:53) > >>> ~[de.flapdoodle.embed.process-2.0.2.jar:?] > >>> ...<snip>... > >>> > >>> at org.apache.maven.surefire.booter.ForkedBooter.main( > >> ForkedBooter.java:407) > >>> [surefire-booter-2.20.1.jar:2.20.1] > >>> Caused by: java.util.zip.ZipException: archive is not a ZIP archive > >>> > >>> > >>> Clear C:\Users\jenkins\.embedmongo\ and try again? > >>> > >>> Gary > >>> > >>>> > >>>> [1]: > >>>> https://builds.apache.org/job/Log4jCore/job/master/17/ > >>>> execution/node/27/log/ > >>>> > >>>>> On 7 April 2018 at 11:48, Gary Gregory <[email protected]> > wrote: > >>>>> > >>>>> There was a commit today or last night, I think about messages and > >>>>> parameters, maybe it is related. > >>>>> > >>>>> Gary > >>>>> > >>>>>> On Sat, Apr 7, 2018, 10:22 Matt Sicker <[email protected]> wrote: > >>>>>> > >>>>>> I'm actually getting the same test failures on macOS now. Might just > >>>> be a > >>>>>> regression. > >>>>>> > >>>>>>> On 7 April 2018 at 11:15, Matt Sicker <[email protected]> wrote: > >>>>>>> > >>>>>>> https://builds.apache.org/job/Log4jCore/job/master/16/ > >>>>>>> execution/node/38/log/?consoleFull > >>>>>>> > >>>>>>> I'm not exactly sure what this error is. Did I misconfigure this > >> job > >>>>>>> entirely? This test doesn't seem to fail on the Ubuntu build. > >>>>>>> > >>>>>>> -- > >>>>>>> Matt Sicker <[email protected]> > >>>>>>> > >>>>>> > >>>>>> > >>>>>> > >>>>>> -- > >>>>>> Matt Sicker <[email protected]> > >>>>>> > >>>>> > >>>> > >>>> > >>>> > >>>> -- > >>>> Matt Sicker <[email protected]> > >>>> > >> > > > > > > > > -- > > Matt Sicker <[email protected]> > -- Matt Sicker <[email protected]>
