We should not block the release for the Windows failure since it only involves test input and the main changes were trivial, propagating a different exception type. However, I am unclear on why the test fails since the filesystem blobstore should convert user input / to backend fs separator \ on Windows. That is to say, code which runs on Linux and Windows *should* behave identically, although in previous releases they did not.
Unfortunately these Windows filesystem issues are common since most development happens on Linux. JCLOUDS-754 might give us a path forward. On Mon, Nov 27, 2017 at 02:20:43PM +0100, Ignasi Barrera wrote: > Andrew, regarding the test, I'd say an input string for each OS would make > sense. @gaul, WDYT? > Regarding the release, if this is just about the test configuration but not > the behavior of the filesystem API itself, I'd say we should not block the > release for this reason? > > @andrea, thanks for running it on Java 7. May you post your vote by > replying to the [VOTE] thread? It is a PITA that Gmail collapses both > threads as one. One way to see the threads properly is to temporarily go to > the Gmail settings, and turn the conversation view off. > Regarding the failure in Java 8, I think it is some configuration in the > vagrant-java-bindings library we use. Perhaps @neykov can help here? > > *Caused by: org.osgi.framework.BundleException: Unresolved constraint in > bundle name.neykov.vagrant-java-bindings [87]: Unable to resolve 87.0: > missing requirement [87.0] osgi.ee <http://osgi.ee>; (&(osgi.ee > <http://osgi.ee>=JavaSE)(version=1.6))* > > > On 26 November 2017 at 21:35, Andrea Turli <andrea.tu...@gmail.com> wrote: > > > hi, > > > > I've just rebuilt with java 7 > > > > Apache Maven 3.5.2 (138edd61fd100ec658bfa2d307c43b76940a5d7d; > > 2017-10-18T07:58:13Z) > > Maven home: /usr/share/maven > > Java version: 1.7.0_151, vendor: Oracle Corporation > > Java home: /usr/lib/jvm/java-7-openjdk-amd64/jre > > Default locale: en, platform encoding: UTF-8 > > OS name: "linux", version: "4.9.44-linuxkit-aufs", arch: "amd64", family: > > "unix" > > > > and jclouds-karaf has no issues > > > > [INFO] Reactor Summary: > > [INFO] > > [INFO] Apache jclouds :: Karaf ............................ SUCCESS [ > > 46.836 s] > > [INFO] jclouds :: Karaf :: Core ........................... SUCCESS [ > > 5.985 s] > > [INFO] jclouds :: Karaf :: Bundles ........................ SUCCESS [ > > 0.172 s] > > [INFO] jclouds :: Karaf :: JSch Agentproxy for JSch (shaded bundle) SUCCESS > > [ 5.848 s] > > [INFO] jclouds :: Karaf :: Utils .......................... SUCCESS [ > > 2.720 s] > > [INFO] jclouds :: Karaf :: Recipe ......................... SUCCESS [ > > 2.926 s] > > [INFO] jclouds :: Karaf :: Cache .......................... SUCCESS [ > > 3.302 s] > > [INFO] jclouds :: Karaf :: Services ....................... SUCCESS [ > > 7.548 s] > > [INFO] jclouds :: Karaf :: Commands ....................... SUCCESS [ > > 13.752 s] > > [INFO] jclouds :: Karaf :: URL Handler .................... SUCCESS [ > > 2.393 s] > > [INFO] jclouds :: Karaf :: Chef ........................... SUCCESS [ > > 0.370 s] > > [INFO] jclouds :: Karaf :: Chef :: Core ................... SUCCESS [ > > 1.955 s] > > [INFO] jclouds :: Karaf :: Chef :: Services ............... SUCCESS [ > > 4.195 s] > > [INFO] jclouds :: Karaf :: Chef :: Commands ............... SUCCESS [ > > 8.073 s] > > [INFO] jclouds :: Karaf :: Chef :: Cache .................. SUCCESS [ > > 2.842 s] > > [INFO] jclouds :: Karaf :: Feature ........................ SUCCESS [ > > 7.439 s] > > [INFO] jclouds :: Karaf :: Integration Tests .............. SUCCESS [17:05 > > min] > > [INFO] > > ------------------------------------------------------------------------ > > [INFO] BUILD SUCCESS > > [INFO] > > ------------------------------------------------------------------------ > > [INFO] Total time: 19:04 min > > [INFO] Finished at: 2017-11-26T20:28:31Z > > [INFO] Final Memory: 42M/280M > > [INFO] > > ------------------------------------------------------------------------ > > > > > > I change my vote to +1 > > > > On Sun, Nov 26, 2017 at 9:20 PM, Andrew Phillips <aphill...@qrmedia.com> > > wrote: > > > > > Please use this thread for discussion of issues uncovered in the RC, > > >> questions you may have about the RC, etc. > > >> > > > > > > The following test fails repeatedly for me on Windows: > > > > > > Failed tests: > > > FilesystemStorageStrategyImplTest.testDeletingInvalidPathFil > > > eEndsNormally:695 > > > Deleting an invalid path ended with an InvalidPathException. > > > > > > Specifically, it's the "/" character in the test input "A<!:!@#$%^&*?]8 > > > /\0"; removing that works. > > > > > > Would having an OS-specific test input seem like the correct solution? Or > > > should the test be modified to expect a test failure on Windows? > > > > > > Regards > > > > > > ap > > > > > -- Andrew Gaul http://gaul.org/
