Community
Participate
Working Groups
3.2 M3 OpenCloseWindowTest#testOpenCloseWindows:org.eclipse.ui.resourcePerspective() This test fails on three platforms and has had the most significant jump on 20050808
Might be related to the increased cost of opening the problems view.
When I open the resource perspective I see Navigator, Outline and Tasks views - no Problems view.
Well, Task and Problems view do share some code, that might be what Kim means.
That is indeed what Kim means.
After running the performance tests against the 0802 code base, then adding in changes that were made to a couple of classes in the 0808 build and running the tests again, I found an increase in the elapsed time to run the test and came to the following conclusion. There were changes to WorkbenchWindow that were rolled back by Doug in revision 1.294 that were made by Stefan in 1.293 in order to enhance performance, as far as I can tell. This seems to be a cause of the first jump in elapsed time to run these tests. Note there were also changes made to WorkbenchPage at the same time too that could also be contributing to the jump but not as much as the WorkbenchWindow change - compare revisions 1.294 and 1.293. Bug 90355 seems to be the root of the issue. Doug, can you shed some light as to why these changes were rolled out?
The changes were initially committed early in 3.2, and then rolled back. Those changes should not be responsible for a regression over 3.1. They were rolled back because they introduced bugs, which is talked about through Bug 90355. Also, when I look at the performance results, I see a small spike in October and a much larger spike in November. These are both well past when the changes (and unchanges) for Bug 90355 were made.
The rollback gave us performance the same as 3.1 (as expected). The spike at 1122 is the one we should keep an eye on.
Oh - nevermind Doug. I just got the story (short version) from Tod.
The jump between 20050802 and 20050808 coincides with the release of the some TrimLayout code. I ran the performance tests against the 0802 and 0808 builds and saw an increase in elapsed time to run the tests. When I took the 0802 code and loaded just the TrimLayout code changes from 0808 (and other supporting classes so it would compile) and ran the tests, there was a jump in the elapsed time, though not as much as the increase between the 0802 and 0808 builds. So I think the TrimLayout code changes are contributing to this jump, but may not be responsible for all of it.
This bug hasn't had any activity in quite some time. Maybe the problem got resolved, was a duplicate of something else, or became less pressing for some reason - or maybe it's still relevant but just hasn't been looked at yet. If you have further information on the current state of the bug, please add it and remove the stalebug whiteboard tag. The information can be, for example, that the problem still occurs, that you still want the feature, that more information is needed, or that the bug is (for whatever reason) no longer relevant. --
