Community
Participate
Working Groups
On Windows XP using I20030923, "Ctrl+Space" gives two key down events. This key binding is used by Eclipse for code assist. Affects keys preference page, and causes a flicker as the code assist is executed twice. If the key is bound to any other command, it can make the command completely unusable. Problem does not occur on Linux-GTK. This is required for M4. DOWN: stateMask=0x0, keyCode=0x40000, character=0x0 '\0' DOWN: stateMask=0x40000 CTRL, keyCode=0x20, character=0x20 ' ' DOWN: stateMask=0x40000 CTRL, keyCode=0x20, character=0x20 ' ' UP : stateMask=0x40000 CTRL, keyCode=0x20, character=0x20 ' ' UP : stateMask=0x40000 CTRL, keyCode=0x40000, character=0x0 '\0'
I thought I already fixed this. Perhaps I fixed the bug and didn't mark the PR fixed? I can't seem to find a dup. Fixed > 20030925
*** Bug 43874 has been marked as a duplicate of this bug. ***
*** Bug 44055 has been marked as a duplicate of this bug. ***
Our test suite for this issue is currently failing on the Mac in 3.3.1. See http://eclipsebuildserv.ottawa.ibm.com/downloads/master/downloads/drops/M20070921-1145/testresults/html/org.eclipse.ui.tests_macosx.carbon.ppc_5.0.html
Kim, I could not find any change in either SWT or Platform UI on the maintenance branch that could explain the test failure. Did the Mac on which the tests are run change in any way, like an update to the OS?
I just checked the mac and I don't see any software updates.
Running Snippet25 does not show the problem on my Mac (for what it's worth). The problem could be in Display.post(). We're looking into it ...
There is nothing for us to debug. It works when we sit in front of the machine. Please open a new bug with steps, more information etc.
Sorry, I forgot to update this bug. We had the following two results: Build M20070918-1350: All tests pass. Build M20070921-1145: Test failures in org.eclipse.ui.tests_macosx.carbon.ppc_5.0 I found no relevant code changes between M20070918-1350 and M20070921-1145, i.e. no change at all in SWT, an unrelated change in Platform UI, and no changes in any other components. Kim Moir said that when she ran the tests from M20070918-1350 again, there was a failure. So it is either an intermittent failure or something changed on the Mac we use for the builds. Given that we cannot reproduce this on other Macs, I agree that there is nothing we can do right now.
I reran the tests locally on the machine and they passed. Usually they are invoked over ssh in the test framework. I don't know why this would make a difference.