> On June 7, 2016, 3:18 p.m., Vinod Kone wrote: > > src/tests/partition_tests.cpp, line 574 > > <https://reviews.apache.org/r/48343/diff/1/?file=1408804#file1408804line574> > > > > Hmm. Instead of moving them to master_tests.cpp, keep them in > > partition_tests.cpp and have them inherit from a different testfixture than > > `PartitionTest`. These are partiton tests after all.
Arguably they aren't necessarily partition tests, if we say that an agent is "partitioned" if it has failed health checks (which is what the rest of the `PartitionTest` cases cover). The TCP socket breaking between master and agent isn't treated as indicating the agent is partitioned / due for removal, it just means the agent is currently disconnected. Happy to keep them in `partition_tests.cpp` if you'd prefer though. - Neil ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/48343/#review136478 ----------------------------------------------------------- On June 7, 2016, 1:45 p.m., Neil Conway wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/48343/ > ----------------------------------------------------------- > > (Updated June 7, 2016, 1:45 p.m.) > > > Review request for mesos, Benjamin Mahler, Jan Schlicht, and Vinod Kone. > > > Bugs: MESOS-5547 > https://issues.apache.org/jira/browse/MESOS-5547 > > > Repository: mesos > > > Description > ------- > > PartitionTest will soon be parameterized with whether to run > the registry in "strict" mode; since these tests don't > depend on the registry, there's no point running them twice. > > > Diffs > ----- > > src/tests/master_tests.cpp 34be015aa314a7574e9065efb7b1bb8e1570c5b7 > src/tests/partition_tests.cpp 956493d5b2fb531dcedc477b2fdbbc0609e01412 > > Diff: https://reviews.apache.org/r/48343/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Neil Conway > >
