I have the performance tests excluded for test, but included as the only tests for verify (compile->test->package->verify->install). I don't think it's happening, but it would have to be properly verified.
On 21 May 2014 20:33, Remko Popma <[email protected]> wrote: > I see, so instead of using "mvn clean install" for building the project, I > should use "mvn clean package". > > I'm still wondering if the performance tests are not run twice: once in > the test phase (by the surefire plugin) and once more in the verify phase > (by the failsafe plugin). That would explain the increased build time. > > Matt, could this be what's happening? > > Sent from my iPhone > > On 2014/05/22, at 9:14, Ralph Goers <[email protected]> wrote: > > I thought Matt said he changed some things to only run in the verify > phase. I think the performance tests were put there. > > Sent from my iPad > > On May 21, 2014, at 4:21 PM, Remko Popma <[email protected]> wrote: > > Interestingly, there seems to be a big difference between > mvn clean install -> 8.5 minutes > and > mvn test -> 5.5 - 6 minutes > > So when just running "mvn test" we get the "fast" build time (which works > out about right if the new deadlock test takes 10-20 secs). > > What is the difference between mvn clean install and mvn test? > I doubt that compiling takes 2-3 minutes, especially because previously > mvn clean install took 5-6 minutes. > > Could it be that the maven-failsafe-plugin redundantly runs some tests > (performance tests) that are also executed in the surefire test run? > So basically the performance tests are run twice? > > > > > On Thu, May 22, 2014 at 1:18 AM, Ralph Goers > <[email protected]>wrote: > >> I added a test that checks for a deadlock. I recall it taking about 10-20 >> seconds on my machine. I tried to make it be as fast as it could be and >> still catch the deadlock when it occurred. >> >> Ralph >> >> On May 21, 2014, at 8:42 AM, Remko Popma <[email protected]> wrote: >> >> > What happened with the build? >> > I brought it down from 8 minutes to 5.5 minutes on my PC >> > and now it is 8.5 minutes again! One step forward, two steps back... >> > >> > Is this because the maven-failsafe-plugin tests were added? If so, can >> we take that out? >> > >> > Or set it up so that the performance tests are only run during >> continuous integration. I thought that's why the performance tests were >> separated out, or did I misunderstand? >> > >> > Remko >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >> > -- Matt Sicker <[email protected]>
