----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/46491/#review129902 -----------------------------------------------------------
Fix it, then Ship it! src/launcher/executor.cpp (lines 90 - 93) <https://reviews.apache.org/r/46491/#comment193451> Now that we add a fourth state documenting e.g., which states are initial or terminal might help maintenance. I personally feel that if we continue to implement an ad hoc state machine here raising the level of abstraction could both help readibility and avoid implementation errors. src/launcher/executor.cpp (line 557) <https://reviews.apache.org/r/46491/#comment193452> Did you mean `killTask()` here? src/launcher/executor.cpp (line 677) <https://reviews.apache.org/r/46491/#comment193465> Not yours, but could please confirm that this really needs to perform a copy? - Benjamin Bannier On April 21, 2016, 4:28 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/46491/ > ----------------------------------------------------------- > > (Updated April 21, 2016, 4:28 p.m.) > > > Review request for mesos, Benjamin Bannier and Ben Mahler. > > > Bugs: MESOS-5240 > https://issues.apache.org/jira/browse/MESOS-5240 > > > Repository: mesos > > > Description > ------- > > In command executor, capture the state when a task is killed (i.e., > reaped) in a flag and use this flag to prevent calls to `killTask()` > and `escalated()` when they are executed after the task is killed. > > > Diffs > ----- > > src/launcher/executor.cpp bec9cba091df2d1b340e68e67966ec322558c315 > > Diff: https://reviews.apache.org/r/46491/diff/ > > > Testing > ------- > > `make check` on Mac OS 10.10.4 > > > Thanks, > > Alexander Rukletsov > >
