drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Muhammad Gelbana <m.gelb...@gmail.com>
Subject Re: Drill test cases failures after a fresh clone
Date Sat, 24 Jun 2017 12:58:03 GMT
Thank you all for your replies. Here is the jira issue as requested.
https://issues.apache.org/jira/browse/DRILL-5606

Please tell me right away if you need more information.

*---------------------*
*Muhammad Gelbana*
http://www.linkedin.com/in/mgelbana

On Fri, Jun 23, 2017 at 7:52 PM, Paul Rogers <progers@mapr.com> wrote:

> Hi Muhammad,
>
> The tests have worked for me. That said, we do want them to work “out of
> the box” for everyone, so let’s see if we can track down the issues you are
> having.
>
> TestClassTransformation.testCompilationNoDebug is a timeout issue.
> Sometimes this just means that tests run slowly, perhaps due to other load
> on the system. When that happens, JUnit fails them due to a hard-coded,
> fixed timeout. Was there other load on our machine while running the tests?
>
> The others look like functional issues. We need more details since the
> summary messages are truncated. Did you happen to capture the entire test
> output? If so, you can find the full details for each test so we can see
> what’s happening.
>
> Otherwise, you can run the test individually:
>
> > cd exec/java-exec
> > mvn surefire:test -Dtest=TestName
>
> Try this with the four tests that failed (in one case, there were two
> failures in the same test class.)
>
> You can also run the four tests in our IDE. Now, sometimes I’ve found that
> tests work fine in one of the three cases, but fail in others. (Maven set
> up a bunch of system properties, for example. Maven runs tests in a single
> JVM, which can occasionally cause problems, but your IDE will launch a new
> JVM for each run…)
>
> Once we see what you find, we can start to figure out what is going on.
>
> To make this easier to track, I’d suggest opening a JIRA ticket to report
> the issue, then posting your test results there. We’ll need that JIRA if we
> have to make any fixes to tests.
>
> Thanks,
>
> - Paul
>
> > On Jun 23, 2017, at 4:35 AM, Muhammad Gelbana <m.gelbana@gmail.com>
> wrote:
> >
> > I intend to create a pull request, so I forked Drill on github and cloned
> > the repository locally on my machine.
> >
> > I built the project by running
> > *mvn clean install -DskipTests*
> >
> > Then I tried running the tests by running
> > *mvn clean install*
> >
> > Then I started facing test failures. The first time, the following tests
> > failed:
> >
> > Tests in error:
> >>
> >> TestClassTransformation.testJDKClassCompiler:73->
> compilationInnerClass:118
> >> »  ...
> >>  TestClassTransformation.testCompilationNoDebug:88 »  test timed out
> >> after 5000...
> >> *  TestCastFunctions.testToDateForTimeStamp:79 »  at position 0 column
> >> '`col`' mi...*
> >>  TestNewMathFunctions.testTrigoMathFunc:111->runTest:85 »
> ExecutionSetup
> >> Failur...
> >> *  TestNewDateFunctions.testIsDate:61 »  After matching 0 records, did
> not
> >> find e...*
> >>
> >
> > I thought to run the tests again because I found a test case failing due
> to
> > a timeout, so after the second run, the following tests failed *again*:
> >
> > Tests in error:
> >> *  TestCastFunctions.testToDateForTimeStamp:79 »  at position 0 column
> >> '`col`' mi...*
> >> *  TestNewDateFunctions.testIsDate:61 »  After matching 0 records, did
> not
> >> find e...*
> >
> >
> > Running tests take around half an hour and I was wondering if there is
> > something I need to setup or configure before I can run the tests and get
> > 100% success rate since I haven't modified anything yet.
> >
> > *---------------------*
> > *Muhammad Gelbana*
> > http://www.linkedin.com/in/mgelbana
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message