----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/69205/#review210177 -----------------------------------------------------------
Fix it, then Ship it! src/common/type_utils.cpp Lines 86 (patched) <https://reviews.apache.org/r/69205/#comment294805> Not yours, but it would be great to document that we don't want to include `cache` explicitly. src/v1/mesos.cpp Lines 85 (patched) <https://reviews.apache.org/r/69205/#comment294806> Ditto. - Benjamin Bannier On Oct. 30, 2018, 12:02 a.m., Joseph Wu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/69205/ > ----------------------------------------------------------- > > (Updated Oct. 30, 2018, 12:02 a.m.) > > > Review request for mesos and Benjamin Bannier. > > > Bugs: MESOS-9357 > https://issues.apache.org/jira/browse/MESOS-9357 > > > Repository: mesos > > > Description > ------- > > This test seems to fail because of a mismatch between the implementation > of std::hash and operator== of the CommandInfo::URI object. The hash > accounts for all fields except for 'cache'. The equality operator > accounts for all fields except 'cache' **and** 'output_file'. > > This mismatch results in replacing the URI value in the fetcher's > mapping, rather than adding a value. Why the std::unordered_map > checks the equality operator rather than the hash value is unknown, > but may be an implementation detail of clang's stdlib. > > > Diffs > ----- > > src/common/type_utils.cpp 33d63809b61a18e03ff745c88f024c71dd221ca2 > src/v1/mesos.cpp 9b2df2dd798dff24a91a2604ab53c7fbb5ecfbcf > > > Diff: https://reviews.apache.org/r/69205/diff/1/ > > > Testing > ------- > > make check > > > Thanks, > > Joseph Wu > >
