Hi Vitalii,
I did run all the tests on my machine and it completed without error after 42
min. Is the PR ok to proceed, or what do I need to do in order for the PR to
be reviewed here and approved?
—C
> On Nov 12, 2018, at 15:18, Vitalii Diravka wrote:
>
> Hi Charles,
>
> Actually 45-50
Hi Charles,
Actually 45-50 mins is the Travis time limit for the job, but anyway you
can see the output and find the result of the run.
Also you can try to enable CircleCI for your own Drill repo. It can run
tests for you similar to Travis, but without time limits.
But neither TravisCI nor
Hi Paul,
I fixed the issue with the unit test, updated the PR but now Tavis is timing
out when building Kudu or something not related to my PR. Is there anything I
can do about that?
The job exceeded the maximum time limit for jobs, and has been terminated.
> On Nov 11, 2018, at 14:32, Paul
Hi Charles,
This error is Drill's long-winded way of saying you have a direct memory (value
vector) memory leak: you allocated a vector (or buffer) which was never passed
downstream or freed. Looks like quite a few leaked so should be fairly easy to
track down.
On your last batch, did you
Hi Paul,
Thanks for this. I updated the test and I think I’m close but how I’m getting
different errors. It looks like it isn’t closing the cluster properly, but I’m
not quite sure what to do about this.
— C
Tests run: 2, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 1.817 sec <<<
Hi Charles,
Thanks for the PR. Two suggestions for your test. First, use TupleSchema:
TupleSchema schema = new SchemaBuilder() ... .buildSchema().
BatchSchema has some limitations that TupleSchema overcomes.
Second, when I did a PR that added unions, I normalized the "buildFoo()"
methods.
