I'd say whoever broke the build should take a look at it, but that's
hard to tell now I guess.
I'm working on setting up build automation on Travis CI which will help
with seeing what or who broke the build, but that will only help in the
future. For now I'd say, depending on whether the MongoDB adapter is
considered production ready, ignore the problems or cancel the release.
Mit freundlichen Grüßen,
------------------------------------------------------------------------
*Christian Beikov*
Am 20.09.2017 um 04:18 schrieb Jesus Camacho Rodriguez:
Michael,
It seems nobody is willing to tackle CALCITE-1985 shortly, thus
we need to unblock the situation.
I would suggest we move it to 1.15.0 instead of qualifying it
as blocker for 1.14.0, since it seems it may cause a performance
regression for the MongoDB adapter, but there is no loss of
functionality and rest of tests are passing. We could add a
release note if necessary.
Otherwise, we need to cancel the release and wait for it to
be fixed, since we have been in the release cycle for a while
now holding all commits to master branch.
I would like to hear what others think about this too?
Other ideas?
Thanks,
-Jesús
On 9/20/17, 12:08 AM, "Michael Mior" <[email protected]> wrote:
Anyone willing to look at https://issues.apache.org/jira/browse/CALCITE-1985?
As far as I'm concerned, this is the only blocker for the release since it
seems I'm the only one who has issues with the Cassandra adapter.
--
Michael Mior
[email protected]
2017-09-14 14:50 GMT-04:00 Michael Mior <[email protected]>:
LGTM. I'm just waiting on resolving CALCITE-1981 and CALCITE-1985. Jesús,
any changes related to MV rewriting that may have caused CALCITE-1981?
Also, would appreciate anyone having a look at the Mongo failures in
CALCITE-1985. Other than that, I think we're good to release.
--
Michael Mior
[email protected]
2017-09-14 14:46 GMT-04:00 Julian Hyde <[email protected]>:
Michael,
May I commit https://github.com/apache/calcite/pull/537? It's just a
javadoc change.
Julian
On Wed, Sep 13, 2017 at 9:25 PM, Christian Beikov
<[email protected]> wrote:
I could setup test automation on TravisCI for most of the different
databases if you want with some description amd scripts for how to run
the
docker containers for DBs locally. Would that help?
It's almost as flexible as Jenkins.
Regards,
Christian
Am 13.09.2017 22:46 schrieb "Michael Mior" <[email protected]>:
Created https://issues.apache.org/jira/browse/CALCITE-1985 to track
the
issues with the MongoDB adapter. If someone else could have a quick
look
that would be great. Failed test output is in there.
--
Michael Mior
[email protected]
2017-09-13 14:29 GMT-04:00 Michael Mior <[email protected]>:
That's the major blocker here. I don't want to make a release when
the
Cassandra adapter is completely broken for me. Although if others can
confirm that there are no issues, then maybe we should proceed and
I'll
figure out the issues with my environment later.
There are also a couple integration test failures on MongoDB that I'd
like
to resolve, but at least one of those seems spurious as the generated
plan
also seems ok. I'll open up JIRA for those and any failures I
encounter
with details in case someone else is able to have a quick look.
Other than test failures, I don't believe there are any outstanding
bug
fixes or PRs that need to be merged. As mentioned earlier, if someone
could
run integration tests for Oracle, that would be great. I had a brief
look
over the Coverity scan results and didn't see anything that looks
worth
blocking over, but I think it would be a good idea to have a more
thorough
review in the future and also to set up some models to prune spurious
warnings so it's easier to take action in the future.
--
Michael Mior
[email protected]
2017-09-13 14:10 GMT-04:00 Julian Hyde <[email protected]>:
How close are we to a release candidate? Commits to master are
paused,
so let's either make an RC soon or re-open the branch to commits.
Michael and I seem to be deadlocked on
https://issues.apache.org/jira/browse/CALCITE-1981. Michael gets a
build error every time, I never get a build error or a runtime
error.
(I got one once - due to stale jars on my class path, I think.)
Can someone else please try to reproduce the build error? Then we'll
know whether it's me or Michael who has a messed-up environment.
Julian
On Fri, Sep 8, 2017 at 1:59 PM, Julian Hyde <[email protected]>
wrote:
I don’t think there is anyone who has knowledge of MongoDB and
time to
make the fixes.
On Sep 7, 2017, at 10:06 AM, Michael Mior <[email protected]>
wrote:
Thanks. I'm also seeing some integration test failures for the
MongoDB
adapter. If someone more familiar with Mongo could check that
out,
that
would be great.
--
Michael Mior
[email protected]
2017-09-07 12:51 GMT-04:00 Julian Hyde <[email protected]>:
OK, doing that now. The coverity config is a little rusty, so
it may
take a while before I have it working. I'll let you know.
I discovered a couple of tests that were failing on windows, so
I'll
be committing fixes for those also.
Other PRs should wait until after the release.
On Wed, Sep 6, 2017 at 6:21 PM, Michael Mior <
[email protected]>
wrote:
Could you also trigger a new Coverity scan? I don't have write
access to
your repo.
--
Michael Mior
[email protected]
2017-09-06 15:50 GMT-04:00 Julian Hyde <[email protected]>:
I’m good to go.
Today, I will check that Calcite build / test on still works
Windows.
Julian
On Sep 6, 2017, at 11:48 AM, Michael Mior <
[email protected]>
wrote:
As far as I know, all changes people are hoping to have in
the
1.14.0
release have landed. Please speak up if that is not the case,
otherwise
I'm
hoping to prepare RC0 for tomorrow.
--
Michael Mior
[email protected]
