Niclas, Yes, I disabled the docker based tests in our commit jobs, only the nightly job runs them, because of theses "too soon" problems I planned to tackle at some point. Thanks for looking into this!
As for Indexing-SQL ... yes it has been broken for months, see my previous emails to the list about it and https://issues.apache.org/jira/browse/POLYGENE-222 I did git bisect to find the offending commits, everything is in the issue. We either need to fix those or drop it for 3.0 ... Cheers /Paul Niclas Hedhman a écrit : > I have fixed most on my system now, but I have test errors in indexing-sql. > > Is that expected? Because it doesn't look like build related problems... > > For instance; > > <testcase name="script06" > classname="org.apache.polygene.index.sql.postgresql.PostgreSQLComplexQueryTest" > time="5.148"> > <failure message="java.lang.AssertionError: [Jack Doe] expected but > not returned" type="java.lang.AssertionError">java.lang.AssertionError: > [Jack Doe] expected but not returned > > > On Sun, Mar 19, 2017 at 2:42 PM, Niclas Hedhman <nic...@hedhman.org> wrote: > >> Also found a similar issue for both MySQL and JClouds... >> But MySQL sends info to stderr and not to stdout, so the "wait for log >> message" doesn't work, as that feature sits in tdomzal/junit-docker-rule >> repo and has not been released yet. >> >> Adding a DelayChecker and also an AndChecker. >> >> But still having problem in Gradle build, but not in the IDEA build.... >> WTF? To be continued... >> >> :extensions:entitystore-sql:compileTestJava - is not incremental (e.g. >> outputs have changed, no previous execution, etc.). >> /home/niclas/dev/polygene/java/extensions/entitystore- >> sql/src/test/java/org/apache/polygene/entitystore/sql/MySQLEntityStoreTest.java:46: >> error: cannot access StartCondition >> public static final DockerRule DOCKER = new DockerRule( >> ^ >> class file for pl.domzal.junit.docker.rule.wait.StartCondition not found >> 1 error >> >> >> On Sat, Mar 18, 2017 at 11:45 AM, Niclas Hedhman <nic...@hedhman.org> >> wrote: >> >>> Oh, Redis doesn't have this problem on my machine. So, perhaps there are >>> two independent issues. 1 being "wrong waitFor" for Cassandra and Riak on >>> my machine and 2 being "missing docker" on the Jenkins builds. Or? >>> >>> Niclas >>> >>> On Sat, Mar 18, 2017 at 11:36 AM, Niclas Hedhman <nic...@hedhman.org> >>> wrote: >>> >>>> Ok, I found the "WaitFor.logMessage()", which is probably what we need >>>> for these 3 containers. It works for Cassandra, and I'll see if I can do >>>> the same for the other two. >>>> >>>> On Sat, Mar 18, 2017 at 11:18 AM, Niclas Hedhman <nic...@hedhman.org> >>>> wrote: >>>> >>>>> Paul, >>>>> there seems to be a problem with Docker support in our ES tests. >>>>> >>>>> https://builds.apache.org/view/P/view/Polygene/job/Polygene( >>>>> JavaEdition)-develop-java8-check/286/testReport/ >>>>> >>>>> 1. Cassandra, Riak and Redis is missing from that report. >>>>> >>>>> 2. When I try to run them locally, they fail. >>>>> >>>>> Since I know Cassandra most, I have been looking at that. And it seems >>>>> to be a discrepancy between when the container is actually ready vs the >>>>> start up of the application in the container. >>>>> If I just wait more seconds during the assembly while debugging, then >>>>> the Cassandra tests succeeds. >>>>> >>>>> >>>>> Do you know anything about this? >>>>> >>>>> Cheers >>>>> -- >>>>> Niclas Hedhman, Software Developer >>>>> http://polygene.apache.org <http://zest.apache.org> - New Energy for >>>>> Java >>>>> >>>> -- >>>> Niclas Hedhman, Software Developer >>>> http://polygene.apache.org <http://zest.apache.org> - New Energy for >>>> Java >>>> >>> -- >>> Niclas Hedhman, Software Developer >>> http://polygene.apache.org <http://zest.apache.org> - New Energy for Java >>> >> -- >> Niclas Hedhman, Software Developer >> http://polygene.apache.org <http://zest.apache.org> - New Energy for Java >> > > >