Community
Participate
Working Groups
Since build N20090424-1216, the scenario OpenClosePerspectiveTest#testOpenClosePerspectives:org.eclipse.ui.tests.performancePerspective1() has got a big performance regression. Looking at the Status Table, it seems to be the same for *all* perspective scenarios...
It seems that there's also an impact on other scenarios, typically OpenCloseViewTest#showView:org.eclipse.ui.navigator.ProjectExplorer() which also got a regression since N20090424-1216...
Maybe a regression opening the Project Explorer?
On possibility, Bug 273066 PW
Perspective activation now performs a full update on both the main menu and the CoolBar. Also we may have (temporarily) have introduced a lazy loading regression (see bug 262636). Paul, is there any way to get around the lazy loading issue without losing the update completely?
I can confirm that change in the bug 262636 added significant regression to this test. CPU time: with change for the bug 273066 reversed: 75ms; with the change: 290ms. That's about 300% increase on my Win XP machine.
(In reply to comment #5) > I can confirm that change in the bug 262636 added significant regression to > this test. That should be bug 273066.
The bugs Oleg mentions also cause flashing during perspective switching. Should we consider backing them out for M7? PW
>Should we consider backing them out for M7? Nothing is seriously broken, so I think it's good enough if we make sure to take a close look during RC1.
Will do, thanks Dani
ping
We'll be looking at this as part of bug 262636 PW
Paul, we seem to be able to avoid the 'flashing' during the perspective switch simply by not updating the TB's in the Perspective's 'onActivate'. They still appear to behave normally (i.e. I can have two different sets of hidden tool items in two perspectives and switching between them updates the TB's in a manner consistent with the overrides.
Marking as a DUP since the results are on the other defect... *** This bug has been marked as a duplicate of bug 262636 ***