> On March 18, 2016, 11:32 p.m., Jie Yu wrote: > > src/docker/executor.cpp, lines 588-590 > > <https://reviews.apache.org/r/44709/diff/4/?file=1306254#file1306254line588> > > > > This is more like a question: are you talking about the case where the > > agent binary is updated, while the executor binary is not?
Exactly! Is it not clear from the comment or you think it is rather unprobable? - Alexander ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/44709/#review124330 ----------------------------------------------------------- On March 18, 2016, 5:24 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/44709/ > ----------------------------------------------------------- > > (Updated March 18, 2016, 5:24 p.m.) > > > Review request for mesos, Ben Mahler and Jie Yu. > > > Repository: mesos > > > Description > ------- > > We allow unknown flags to ensure that when a new flag is added > in Mesos version N, agent of version N works with the executor > of version N-1, i.e. the latter does not segfault when observing > a newly introduced flag. > > > Diffs > ----- > > src/docker/executor.cpp afc769d0887e3842106e4c350e94c95c8ffc085e > src/launcher/executor.cpp 4149f084b0b234fb995df04e9d1ca704feec082e > > Diff: https://reviews.apache.org/r/44709/diff/ > > > Testing > ------- > > `make check` on Mac OS 10.10.4 > > > Thanks, > > Alexander Rukletsov > >
