> On March 15, 2018, 3:36 p.m., Jan Schlicht wrote: > > src/common/http.cpp > > Lines 885-886 (original), 885-886 (patched) > > <https://reviews.apache.org/r/66088/diff/1/?file=1976840#file1976840line885> > > > > If it's the ordering of `ObjectApprovers` that's important, then we > > shouldn't use a `hashmap` for `approvers`.
The order is only relevant for the `map` and `zip` states, we want that the element order match in both operations. Afterwards, it doesn't matter. - Alexander ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/66088/#review199259 ----------------------------------------------------------- On March 15, 2018, 3:27 p.m., Alexander Rojas wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/66088/ > ----------------------------------------------------------- > > (Updated March 15, 2018, 3:27 p.m.) > > > Review request for mesos, Benno Evers and Jan Schlicht. > > > Bugs: MESOS-8677 > https://issues.apache.org/jira/browse/MESOS-8677 > > > Repository: mesos > > > Description > ------- > > The method in charge of creating `ObjectApprovers` was relying that > iterating over equivalent hashset's would yield the same order of > results. > > While that seemed to be true in different architectures, the macOS > implementation did not follow the same pattern showing a bug in the > code. > > > Diffs > ----- > > src/common/http.cpp 03f539f8942d50de34f42fd711053e94d0a1c95d > > > Diff: https://reviews.apache.org/r/66088/diff/2/ > > > Testing > ------- > > ```sh > # In multiple platforms. > make check > ``` > > > Thanks, > > Alexander Rojas > >
