On Wed, Feb 3, 2016 at 6:20 PM, John Sirois <jsir...@apache.org> wrote:
> I did some more digging on the `make check` hang I encountered which led > to this proposal. The originating github PR actually failed TravisCI in > the go shard with exactly this same hang: > https://travis-ci.org/apache/thrift/jobs/101493938 > > I'm an Apache Thrift dev newcomer (though long-time user back to ~2010), > so I have almost no context, but is it typical to merge PRs that fail > either TravisCI or Jenkins checks? > I'd be willing to invest a bit of time into process docs, CI setups, etc > to make sure the checks already in-place are actually used to keep master > green. > > I may have dipped my toes into the dev waters with simple bad timing, but > pretending that is not the case and I'm observing status quo, things like a > red master are a big detractor to contribution enthusiasm. They give me > less faith in the product as a long-time user, and more work to do to get > to the the work I want to do to develop my own features. > > On Wed, Feb 3, 2016 at 5:37 PM, John Sirois <jsir...@apache.org> wrote: > >> I just noticed there is a Thrift-precommit job that _does_ have `make >> check`: >> >> $ curl --netrc -sS >> https://builds.apache.org/job/Thrift-precommit/config.xml | xmllint >> --xpath "//command/text()" - >> sh bootstrap.sh >> sloccount --duplicates --wide --details lib tutorial test > >> sloccount.sc >> cppcheck --error-exitcode=0 --force --xml lib/cpp/src/ lib/c_glib/src/ >> tutorial/cpp/ test/cpp/ 2> cppcheck-result.xml >> sh bootstrap.sh >> sh configure --without-erlang --without-haskell >> make >> make check >> make dist >> cd lib/java >> ant javadocsh test/test.sh | tee test.log >> >> Perhaps `Thrift-all` is obsolete? >> >> Any background on the jenkins jobs for Thrift is appreaciated. There are >> more than a few jobs and maybe some could be deleted? >> >> On Wed, Feb 3, 2016 at 5:34 PM, John Sirois <jsir...@apache.org> wrote: >> >>> >>> >>> On Wed, Feb 3, 2016 at 5:33 PM, John Sirois <jsir...@apache.org> wrote: >>> >>>> In the course of working on a new feature for thrift java, I rebased >>>> against master and hit a test hang runing `make check`. I tracked this >>>> down and have a fix out [1], but I was curious how this hang slipped by. I >>>> happen to have jenkins rights through Apache Aurora to view the job >>>> configuration for Thrift-all and I found it runs: >>>> >>> >>> where [1] is: >>> >>> [1] https://issues.apache.org/jira/browse/THRIFT-3597 >>> >>> >>>> >>>> $ curl --netrc -sS https://builds.apache.org/job/Thrift-all/config.xml >>>> | xmllint --xpath "//command/text()" - >>>> #sh bootstrap.sh >>>> #sloccount --duplicates --wide --details lib tutorial test > >>>> sloccount.sc >>>> #cppcheck --error-exitcode=0 --force --xml lib/cpp/src/ lib/c_glib/src/ >>>> tutorial/cpp/ test/cpp/ 2> cppcheck-result.xml >>>> sh bootstrap.sh >>>> sh configure >>>> make >>>> make dist >>>> cd lib/java >>>> ant javadoc >>>> >>>> So `make` and `make dist`, but no `make check`. >>>> >>>> I'd like to propose adding make check to this set of commands to ensure >>>> make check is green. >>>> >>>> If folks agree this is a good idea, I can go ahead and make the edit. >>>> >>>> Thanks for your feedback. >>>> >>> There has been no feedback on this proposal and so I'd like to proceed on lazy consensus (Although I'd rather proceed on active consensus!) and modify the jenkins Thrift-all job to include a `make check` like the `Thrift-precommit` job. I'll do this this afternoon unless I hear objections.
