Community
Participate
Working Groups
We need to add a test case to the performance test suite that covers the scenario fixed by bug# 315978. A program that generates the test case similar to the reported supplied test case is available in bug# 315978 comment# 24. Satyam, please follow up.
Created attachment 172571 [details] Patch for the Performance test Frederic, Incorporated the copyright as per your instructions. Please look at the copyright to be sure that it is good and correct.
(In reply to comment #1) > Created an attachment (id=172571) [details] > Patch for the Performance test > > Frederic, > Incorporated the copyright as per your instructions. Please look at the > copyright to be sure that it is good and correct. Patch looks good to me. However, there's a typo in EclipseVisitorBug.java. The copyright comment ends with two slash, hence you get a compiler error while running the test. Please fix it. Also, I would like to have a preview of times in R3_6_maintenance and on top of 3.6.0 (v_A58) to have an idea of what is the fix perf improvement. Could you put some numbers before we release the patch? TIA
(In reply to comment #2) Frederic, Thanks for your comments. I will fix the copyright. Without the patch it takes 1.6 minutes and with the patch it takes around 1.5 seconds.
Created attachment 172740 [details] Patch for the Performance test Fixed the copyright typo!
Great :-) Hence, +1 for the patch
Released in HEAD and perf_36x
Verified for 3.6.1 RC2 by code inspection.
(In reply to comment #7) > Verified for 3.6.1 RC2 by code inspection. I disagree, this test was not released in R3_6_maintenance branch. It was released in perf_36x branch which is the branch for the 3.6 baseline and which is used for 3.7 performance tests.
To be backported for 3.6.1, the patch needs to be released both in R3_6_maintenance branch *and* perf_35x (which is the branch for the baseline of 3.6 performance tests...).
Note also that this test has an invalid duration in HEAD stream, hence bug 322325 needs to be fixed before releasing this test into R3_6_maintenance stream...
(In reply to comment #9) > To be backported for 3.6.1, the patch needs to be released both in > R3_6_maintenance branch *and* perf_35x (which is the branch for the baseline of > 3.6 performance tests...). Frederic, good catch. Please proceed with releasing the patch appropriately for the 3.6 branch. This is already targetted for and has a +1 for 3.6.1 backport -- Thanks!
(In reply to comment #11) > (In reply to comment #9) > > To be backported for 3.6.1, the patch needs to be released both in > > R3_6_maintenance branch *and* perf_35x (which is the branch for the baseline of > > 3.6 performance tests...). > > Frederic, good catch. Please proceed with releasing the patch > appropriately for the 3.6 branch. This is already targetted > for and has a +1 for 3.6.1 backport -- Thanks! I will take care of this as well as for bug 322325... I also change the target to 3.6.2 as bug 315978 was manually verified (by me), hence there's no real urgency for 3.6.1...
Satyam, Unfortunately, I won't have time to look at this test, hence I put it back to you...
ping! This is targeted for 3.6.2 and we have RC1 this week.
Satyam, could you please add the test for tomorrow's M-build. Thanks.
Created attachment 186515 [details] Patch for 3.6 Maintenance Modified the build.properties also accordingly.
+1.
Released on 3.6 maintenance branch and perf_35x branch
Verified for 3.6.2 using code inspection