10743 2016-11-30 14:01:09 +0000 build-perf-test: Add test to verify if parsing time is improved with a hot codeparser cache 2017-12-06 17:18:58 +0000 1 1 10 QA/Testing Build Testing general 2.3 x86 Multiple RESOLVED FIXED Medium enhancement 2.3 1 jose.perez.carranza markus.lehtonen markus.lehtonen --- --- No (bug/feature does not impact docs) oldest_to_newest 68718 0 jose.perez.carranza 2016-11-30 14:01:09 +0000 A proposal to implement this is to take measurements of `bitbake -e` first time and then run again `bitbake -e` and check the improvement on the parsing time. 68732 1 benjamin.esquivel 2016-11-30 16:30:14 +0000 Measuring the improvement of the parsing time has to happen between the previous revision and the current one. It should be part of the performance numbers where you can check historically how this parsing time is behaving. Hot parsing time has to be measured against previous measurements, not against cold parsing times. Same thing for cold parsing times, it has to be measured against previous of-its-kind measurements as well. 68766 2 markus.lehtonen 2016-12-01 09:02:47 +0000 We basically already do this in oeqa.buildperf.test_basic.Test3. We do "bitbake -p" with cold, "warm" and hot cache. If I'm not mistaken we should close this one as already implemented. 68775 3 benjamin.esquivel 2016-12-01 15:51:25 +0000 marking this as resolved, waiting for QA verification.