> On July 15, 2016, 5:41 p.m., Greg Mann wrote: > > Regarding the regression test: several of our tests use this "local mode" > > to test example frameworks (see 'src/tests/examples_tests.cpp'), but they > > all set the MESOS_WORK_DIR env var before running. I also noticed that we > > don't have any tests which run the 'mesos-local' binary; I created a > > separate ticket for that: https://issues.apache.org/jira/browse/MESOS-5850. > > > > One easy thing you could do to cover the issue which led to these patches > > is to modify one of the existing example framework test scripts to _not_ > > set MESOS_WORK_DIR before running. This would just involve removing lines > > like > > [these](https://github.com/apache/mesos/blob/4764768fc781bb99294039cebf072748f190bfba/src/tests/test_framework_test.sh#L25-L26).
Done, as a side note, I don't think the `atexit` stuff is working because if I do `ls /tmp/mesos*` after running `make check`, I still see a bunch of folders. - Ammar ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/50003/#review142393 ----------------------------------------------------------- On July 15, 2016, 9:01 p.m., Ammar Askar wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/50003/ > ----------------------------------------------------------- > > (Updated July 15, 2016, 9:01 p.m.) > > > Review request for mesos, Greg Mann and Vinod Kone. > > > Bugs: MESOS-5613 > https://issues.apache.org/jira/browse/MESOS-5613 > > > Repository: mesos > > > Description > ------- > > Propagate work_dir flag from local runs to agents/masters. > > > Diffs > ----- > > src/local/flags.hpp f0af0d2 > src/local/local.cpp a543aef > src/tests/test_framework_test.sh 617ca52 > > Diff: https://reviews.apache.org/r/50003/diff/ > > > Testing > ------- > > Manually tested that `mesos local` and `mesos-local` run now. Not sure if a > regression test for this would be good, please advise. > > > Thanks, > > Ammar Askar > >
