> On June 27, 2016, 8:29 p.m., Vinod Kone wrote: > > src/master/http.cpp, line 2706 > > <https://reviews.apache.org/r/48400/diff/4/?file=1423251#file1423251line2706> > > > > I was looking at this code today and realized this might not be safe? > > What's the guarantee that a given `Task*` is still valid and not being > > accessed by other actor (master) when we are executing the lambda? Is the > > lambda guaranteed to be executed in master's context? cc @anand @mcypark > > Anand Mazumdar wrote: > Good catch! This needs to go inside `defer`.
`defer`ing solves the concurrent access problem, but not the `Task*` becoming invalid problem? - Vinod ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/48400/#review139640 ----------------------------------------------------------- On June 18, 2016, 2:33 a.m., Jay Guo wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/48400/ > ----------------------------------------------------------- > > (Updated June 18, 2016, 2:33 a.m.) > > > Review request for mesos, Anand Mazumdar and Vinod Kone. > > > Bugs: MESOS-5493 > https://issues.apache.org/jira/browse/MESOS-5493 > > > Repository: mesos > > > Description > ------- > > Add a helper function _tasks() to collect tasks. This is > to be used by both tasks() and getTasks() in operator API. > > > Diffs > ----- > > src/master/http.cpp dca814b070e513420175a2df6db7844fd425c3fa > src/master/master.hpp 2bddfe1c6638281e3ed34146490bc33f66d87d5c > > Diff: https://reviews.apache.org/r/48400/diff/ > > > Testing > ------- > > make check on ubuntu 14.04 > > > Thanks, > > Jay Guo > >