----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/71908/#review219121 -----------------------------------------------------------
src/common/http.cpp Line 210 (original), 210 (patched) <https://reviews.apache.org/r/71908/#comment307221> Do I understand correctly that this behavior detail is not covered by any Mesos tests? Maybe it is worth adding something similar to HTTP.AsV1Protobuf but with empty map/repeated field? - Andrei Sekretenko On Dec. 12, 2019, 11:12 p.m., Benjamin Mahler wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/71908/ > ----------------------------------------------------------- > > (Updated Dec. 12, 2019, 11:12 p.m.) > > > Review request for mesos, Andrei Sekretenko and Greg Mann. > > > Bugs: MESOS-6568 > https://issues.apache.org/jira/browse/MESOS-6568 > > > Repository: mesos > > > Description > ------- > > Per MESOS-6568, the omission of empty repeated fields / maps is > inconsistent with the standard proto3 -> JSON mapping. An empty > array is also more convenient to program against since you don't > need to special case the empty case with an absence check. > > This may break any tests that rely on exact string outputs, but > it generally should not break any logic. > > > Diffs > ----- > > src/common/http.cpp c5b2a91958c870e272895520ba04fc5287891c3c > > > Diff: https://reviews.apache.org/r/71908/diff/1/ > > > Testing > ------- > > make check > > > Thanks, > > Benjamin Mahler > >
