Community
Participate
Working Groups
The virtual viewer tests have been failing on linux due to an update service dialog being open on the machine when the tests run. If possible, we should force the viewer shell to be on top/in the foreground.
I converted the shells used in the viewer to use SWT.ON_TOP. The down-side is that on Linux/GTK, this removes the shell's trim controls, but the tests still work. Mike, please have a look.
It seems that changing the shell style to ON_TOP had a bad effect on the Linux tests. I'm going to make a slight adjustment: instead of calling shell.setMaximized(), i'll call shell.setSize(800,600). We'll see if this has a positive effect on the over night tests. The Mac tests also ended in DNF. I'll run those locally.
I'll leave the bug open until the nightly tests pass.
The tests on Linux passed last night so that's some progress. However, I'm seeing something very odd on Mac. I managed to get my home Mac (10.6 x86_64) set up to run the tests and I get a behavior where the tests repeatedly get stuck with no updates in the viewers. If I manually change the focus away from the shell with the viewer, the viewer refreshes and the test continues. It's as if the lazy loading viewer thinks that it's covered by another window and simply doesn't refresh. I found bug 307322 but it doesn't explain why the nightly tests end in DNF when they were running fine in M5 on the same hardware.
My theory about finding a duplicate of bug 307322 didn't pan out because I cannot reproduce that failure on my Mac. I'll try reverting the new shell behavior (SWT.ON_TOP) for macs and see if that cures the DNF in nightly build.
The mac tests started passing again as so did others. So I'm marking as fixed. Although the win32 tests ended with a dnf on the 21st so if a popup window was responsible then the ON_TOP style didn't help there. I'm closing this bug for now. If we get persistent failures we may need to consider some other alternatives for these tests.
I think we need to revisit the tests, the Win test DNF'd again in the N20110222-2000 build
I still can't reproduce the DNF locally, so I'm kind of shooting in the dark: I tried one theory last night - by disabling a test that was recently changed, but that didn't help. Now I backed out the change made originally for this fix for the Windows platform (I already backed it out for Mac). Let's see if that cures it.
(In reply to comment #8) The tests still ended with DNF over the weekend. I'm out of ideas on this since I can't reproduce the failure locally. Mike do you know if there is any way we could "see" the tests failing on the test machine.
(In reply to comment #9) > (In reply to comment #8) > The tests still ended with DNF over the weekend. I'm out of ideas on this > since I can't reproduce the failure locally. Mike do you know if there is any > way we could "see" the tests failing on the test machine. There is no way that I am aware of, the person to ask would be Kim. CC'ing Kim for comment.
We are having problems with the Windows machine being logged off as the build user intermittently which can cause these types of tests. I'm working right now with the sysadmin to try to solve this. It's impacting other teams. Very frustrating.
Thank you Kim for the clarification. I wish I asked sooner :-)
The tests are passing again, and the changes from this bug are mostly disabled. All in all, I think I'm starting to think that the ON_TOP style is not really buying us any advantage. For example it didn't help any on the Windows machine that was logging out the user. Any opinions?
(In reply to comment #13) > The tests are passing again, and the changes from this bug are mostly disabled. > All in all, I think I'm starting to think that the ON_TOP style is not really > buying us any advantage. I agree, I don't think there is much we can do, if even forcing the viewer to be ON_TOP does not prevent the tests failing.
I backed out the ON_TOP change.
We still get DNF failures. Had another failure on I20130730-0800. Could we not run the tests inside a 'normal' workbench? How come we have to have the viewer so large / on top / etc?
*** Bug 414097 has been marked as a duplicate of this bug. ***
And another DNF: http://download.eclipse.org/eclipse/downloads/drops4/N20130816-2000/testResults.php
And Another DNF: http://download.eclipse.org/eclipse/downloads/drops4/I20140522-1330/testResults.php
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. 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. If the bug is still relevant, please remove the stalebug whiteboard tag.
