----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/6505/#review10106 -----------------------------------------------------------
src/master/allocator.hpp <https://reviews.apache.org/r/6505/#comment21402> A fundamental constructor-like thing taking flags always worries me - can you have an overload - perhaps in the tested impl only - take the delay directly? I suppose you can just create a mesos master Flags and pass in - but even then only one value is needed here do mixing in a smaller Flags seems more appropriate. Either way - the new batch behavior is untested - can you leverage a Clock here? I think this concept is already covered for libprocess processes - so maybe the only thing blocking a test is someone pushing you to add one for this ;) - John Sirois On Aug. 10, 2012, 3 a.m., Benjamin Hindman wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/6505/ > ----------------------------------------------------------- > > (Updated Aug. 10, 2012, 3 a.m.) > > > Review request for mesos, Thomas Marshall, John Sirois, and Vinod Kone. > > > Description > ------- > > See summary. > > > Diffs > ----- > > src/master/allocator.hpp c6273b5 > src/master/dominant_share_allocator.hpp ba5dc29 > src/master/dominant_share_allocator.cpp 00a47f7 > src/master/flags.hpp 1227ccc > src/master/master.cpp 4f62687 > src/tests/allocator_tests.cpp b3db13d > src/tests/resource_offers_tests.cpp c004772 > src/tests/utils.hpp a768360 > > Diff: https://reviews.apache.org/r/6505/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Benjamin Hindman > >
