Community
Participate
Working Groups
http://download.eclipse.org/eclipse/downloads/drops4/S-4.2M7-201205031800/testresults/html/org.eclipse.jdt.text.tests_win32.win32.x86_7.0.html http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC1-201205182145/testresults/html/org.eclipse.jdt.text.tests_win32.win32.x86_7.0.html http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC2-201205242100/testresults/html/org.eclipse.jdt.text.tests_win32.win32.x86_7.0.html The failures are in org.eclipse.jdt.text.tests.contentassist.CodeCompletionTest org.eclipse.jdt.text.tests.SemanticHighlightingTest org.eclipse.jdt.text.tests.MarkOccurrenceTest The tests pass in the 3.8 builds http://download.eclipse.org/eclipse/downloads/drops/I20120523-2100/testResults.php http://download.eclipse.org/eclipse/downloads/drops/I20120527-2100/testResults.php The failures look a bit mysterious to me, especially the failures in CodeCompletionTest. I have not been able to reproduce the failures on my machine so far.
There is an exception in the console log (http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC2-201205242100/testresults/consolelogs/win7consolelog.txt) ----------------------------------------------------------------------------- [echo] Running org.eclipse.jdt.text.tests.JdtTextTestSuite. Result file: c:\hb\workspace\JUnit-win2\workarea\I20120524-2100\eclipse-testing/results/win32.win32.x86_7.0/org.eclipse.jdt.text.tests.JdtTextTestSuite.xml. [java] Exception in thread "Thread-388" java.util.ConcurrentModificationException [java] at java.util.HashMap$HashIterator.nextEntry(HashMap.java:806) [java] at java.util.HashMap$KeyIterator.next(HashMap.java:841) [java] at org.eclipse.jface.text.source.AnnotationModel.getAnnotationIterator(AnnotationModel.java:767) [java] at org.eclipse.jface.text.source.AnnotationModel.getAnnotationIterator(AnnotationModel.java:691) [java] at org.eclipse.jface.text.source.projection.ProjectionSummary.removeSummaries(ProjectionSummary.java:173) [java] at org.eclipse.jface.text.source.projection.ProjectionSummary.internalUpdateSummaries(ProjectionSummary.java:152) [java] at org.eclipse.jface.text.source.projection.ProjectionSummary.access$3(ProjectionSummary.java:147) [java] at org.eclipse.jface.text.source.projection.ProjectionSummary$Summarizer.run(ProjectionSummary.java:72) ----------------------------------------------------------------------------- but that also happens in 3.8 build - http://download.eclipse.org/eclipse/downloads/drops/I20120527-2100/testresults/consolelogs/win7consolelog.txt
I am clueless now.. - I cannot reproduce this by not having focus in the target workbench - Tried with different JDKs - 5,6,7 - I do not see any obvious differences in Hudson jobs for 3.8 and 4.2 here - https://hudson.eclipse.org/hudson/view/Eclipse%20and%20Equinox/
(In reply to comment #1) > There is an exception in the console log So, you can't say that this is related to the failure, right? Please try to trigger the failure on your machine.
> - I cannot reproduce this by not having focus in the target workbench It depends on the timing. In the IDE you can easily see that the occurrence marks go away when you remove the focus from the workbench window.
(In reply to comment #4) > > - I cannot reproduce this by not having focus in the target workbench > It depends on the timing. In the IDE you can easily see that the occurrence > marks go away when you remove the focus from the workbench window. But this behavior is same in 4.2 and 3.8. Furthermore, the failure in CodeCompletionTest is more puzzling because there instead of real parameter names we get arg0, arg1 etc even though everything is contained in 1 java file.
> Furthermore, the failure in CodeCompletionTest is more puzzling because there > instead of real parameter names we get arg0, arg1 etc even though everything > is > contained in 1 java file. You mean we get these for source? That would indeed be very strange.
(In reply to comment #6) > You mean we get these for source? That would indeed be very strange. Exactly! Which is why I said... :-) (In reply to comment #2) > I am clueless now..
A common pattern of these failures is that they need an up-to-date AST from ASTProvider or from SelectionListenerWithASTManager. If e.g. OverrideCompletionProposal#getRecoveredAST(IDocument, int, Document) returns a wrong AST (e.g. one for the previous editor), then StubUtility#suggestArgumentNames(IJavaProject, IMethodBinding) will fall into "catch (JavaModelException e)" and return "arg0" et al. However, I can also not reproduce these failures locally.
(In reply to comment #8) > A common pattern of these failures is that they need an up-to-date AST from > ASTProvider or from SelectionListenerWithASTManager. Interestingly ASTProviderTest also fails http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC2-201205242100/testresults/html/org.eclipse.jdt.ui.tests_win32.win32.x86_7.0.html http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC1-201205182145/testresults/html/org.eclipse.jdt.ui.tests_win32.win32.x86_7.0.html
And now a different set of tests fail - those in the LeakTestSuite http://download.eclipse.org/eclipse/downloads/drops4/I20120601-1900/testresults/html/org.eclipse.jdt.ui.tests_win32.win32.x86_7.0.html http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC3-201205311500/testresults/html/org.eclipse.jdt.ui.tests_win32.win32.x86_7.0.html
(In reply to comment #10) > And now a different set of tests fail - those in the LeakTestSuite > > http://download.eclipse.org/eclipse/downloads/drops4/I20120601-1900/testresults/html/org.eclipse.jdt.ui.tests_win32.win32.x86_7.0.html > > http://download.eclipse.org/eclipse/downloads/drops4/S-4.2RC3-201205311500/testresults/html/org.eclipse.jdt.ui.tests_win32.win32.x86_7.0.html Opened bug 381942 for these failures. CodeCompletionTest, SemanticHighlightingTest and MarkOccurrenceTest now pass in the build. Keeping this bug open to make sure that the test results are stable and reliable in the 4.2 build.
Things are a bit more stable now i.e. the almost the same tests fail, there are only one or two variations.
CodeCompletionTest failed again on Mac.
Marking as fixed: the tests were quite stable in the past weeks.