> On Jan. 12, 2016, 10:56 p.m., Adam B wrote: > > src/tests/executor_http_api_tests.cpp, line 394 > > <https://reviews.apache.org/r/41806/diff/2/?file=1194782#file1194782line394> > > > > Do you think there's any test overhead in doing another AWAIT for a > > `response` that has already been awaited? I'd think it would be a near > > instantaneous noop, but if not, the EXPECT_SOME_EQ could be faster. > > Maybe you could time a few test runs before/after this patch to make > > sure we're not noticeably slowing things down.
Seems like it doesn't make a noticeable performance difference. I ran `time ./src/mesos-tests --gtest_filter="MasterTest.SlavesEndpointWithoutSlaves" --gtest_repeat=100`. With the patch applied, I got the following elapsed times: ``` 5.269 4.663 4.803 4.672 ``` (The first time probably suffered from cold cache/warmup effects.) Without the patch applied, I get: ``` 4.867 4.989 4.728 4.756 ``` - Neil ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/41806/#review114086 ----------------------------------------------------------- On Jan. 12, 2016, 8:47 p.m., Neil Conway wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/41806/ > ----------------------------------------------------------- > > (Updated Jan. 12, 2016, 8:47 p.m.) > > > Review request for mesos and Adam B. > > > Repository: mesos > > > Description > ------- > > Used `AWAIT_EXPECT_RESPONSE_HEADER_EQ()` to check the "Content-Type" of > the response, rather than accessing the "headers" field directly, and > used the symbol `APPLICATION_JSON` rather than a string literal. Also > added "Content-Type" checks to a few places that had neglected to make > them, and cleaned up some whitespace style. > > > Diffs > ----- > > src/tests/executor_http_api_tests.cpp > 952aacb720715ad26f91fa08bc386b242b7e007b > src/tests/fault_tolerance_tests.cpp > fd1c3c90101eed9ef9352511e0c72a463cca965c > src/tests/master_tests.cpp 223b9d20a3a8a8194a3a6a605ec2394c37ab5957 > src/tests/metrics_tests.cpp f081fb9b68f25c6d6005f195c34253fba8abc146 > src/tests/monitor_tests.cpp a848d14ebb6cab79c06bcf55bd39f044b41a006e > src/tests/repair_tests.cpp 63ec889c4954c2c60d3466952551aa25b3284ddf > src/tests/scheduler_driver_tests.cpp > 1365d21ccad87923b862fb4942f1fd51630a62b7 > src/tests/scheduler_http_api_tests.cpp > 4d23a5a8368e0ed126469fa4a90a889b339ad004 > src/tests/slave_tests.cpp e4fb490a1d877547fe883c22dbc47bb4969ecef6 > src/tests/status_update_manager_tests.cpp > bd34b97a3559a5fea9a7a253a89e0ac3029f4a33 > src/tests/utils.cpp 877139e97249761658dce3b1058cdc2e2a52367b > > Diff: https://reviews.apache.org/r/41806/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Neil Conway > >