On 27/05/2015 14:11, Mark Thomas wrote: > On 26/05/2015 13:23, Konstantin Kolinko wrote: >> 2015-05-26 13:39 GMT+03:00 Rainer Jung <rainer.j...@kippdata.de>: >>> Apart from the one failed test (BIO TestWebSocketFrameClient) CI logs all >>> tests as having skipped tests, because the test run output does not contain >>> "Skipped: 0" (checked in our build.xml). >>> >>> Locally executed tests print: >>> >>> [junit] Tests run: 23, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: >>> 0.615 sec >>> >>> on the CI system: >>> >>> [junit] Tests run: 23, Failures: 0, Errors: 0, Time elapsed: 0.095 sec >>> >>> Is that a version inconsistency? Both should run with our configured junit >>> 4.11, but maybe the CI system has another junit in the CLASSPATH? >> >> It is odd. >> >> Looking at logs there. The oldest logs have "Skipped" in them. >> http://ci.apache.org/projects/tomcat/tomcat8/logs/ >> >> By dichotomy: >> http://ci.apache.org/projects/tomcat/tomcat8/logs/1667637/ >> - Good (19-Mar-2015 01:45) >> http://ci.apache.org/projects/tomcat/tomcat8/logs/1667747/ >> - Bad (19-Mar-2015 14:27) >> >> http://svn.apache.org/viewvc?view=revision&revision=r1667747 >> is "Trivial Javadoc fix (mainly to test CI system)" >> >> The oldest build available at the server is >> http://ci.apache.org/builders/tomcat-8-trunk/builds/172 >> (Mon Mar 23 08:33:14 2015) >> >> From mail archive: >> [http://mail-archives.apache.org/mod_mbox/tomcat-dev/201503.mbox/%3C550AA0CB.80009%40apache.org%3E] >> "Buildbot slave changes" >> >> [quote, Mark wrote] >> I have just moved trunk, tomcat-8 and tomcat-6 (tomcat-7 was there >> already) buildbot builds to a new slave. There main aim is to get faster >> builds. A combination of a physical machine rather than an overstretched >> VM and not being on the same slave as the resource intensive sling >> builds should result is a significant speed up. >> >> I imagine I broke a bunch of stuff with the move. Apologies in advance >> and I'll get it fixed as quickly as I can. >> >> At the same time I fixed the issues that meant the tomcat-7 builds >> weren't building the Windows installer and weren't running RAT. >> [/quote] >> >> >> Tomcat trunk, Tomcat 7 builders have the same problem. >> >> (Trunk uses a different version of Cobertura, I have not backported r1663472. >> I mean that the local build likely has not enabled Cobertura. >> So it is not it.) >> >> Maybe the version of Apache Ant? >> E.g. Ant having junit jars in its lib directory? >> >> Or just an old version of Ant? > > I think this is: > https://bz.apache.org/bugzilla/show_bug.cgi?id=43969 > > silvanus.a.o is running Ant 1.8.2. I'll get it updated to 1.9.3 and see > where we are.
Confirmed. That fixed it. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
