Bugzilla – Full Text Bug Listing
| Summary: | [gtk3] 8 tests consistently fail on CentOS (disabled since 4.6 RC2) | ||
|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Dani Megert <daniel_megert> |
| Component: | UI | Assignee: | Sravan Kumar Lakkimsetti <sravankumarl> |
| Status: | VERIFIED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | akurtakov, arunkumar.thondapu, daniel_megert, ericwill, lbullen, mistria, nobody |
| Version: | 4.6 | Flags: | daniel_megert:
review+
mistria: review+ |
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | CentOS | ||
| See Also: |
https://git.eclipse.org/r/73047 https://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=a2f14eca2c5e0d30c1ee0cdc179ddc3adbb874b6 https://git.eclipse.org/r/73137 https://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=0687c87d2afb09d6d90cb6a62f5ce57634f03d48 https://git.eclipse.org/r/115651 https://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=44130248a27b8c658e1445287aa4f630cf543a33 |
||
| Whiteboard: | |||
| Bug Depends on: | 461354 | ||
| Bug Blocks: | 497330 | ||
|
Description
Dani Megert
This can be reproduced using gtk3 machine not necessary to have CentOS. There two problems in this particular case. 1. SWT.SetData event is getting delayed. atleast 3 tests gets cancelled because this event is not received in time for the test case to pass. 2. There is a delay in processing setSelection call. The delay is on the gtk side. This is causing the topindex calls to return wrong values. I will need some more to properly analyse these test failures Moving to RC3 as there is not much time. @Dani is this really needed for Neon? these problems exists in GTK3 implementations quite long time. (In reply to Sravan Kumar Lakkimsetti from comment #2) > Moving to RC3 as there is not much time. > > @Dani is this really needed for Neon? these problems exists in GTK3 > implementations quite long time. Then disable the tests for now, since others running those tests will also see them fail. New Gerrit change created: https://git.eclipse.org/r/73047 Gerrit change https://git.eclipse.org/r/73047 was merged to [master]. Commit: http://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=a2f14eca2c5e0d30c1ee0cdc179ddc3adbb874b6 Sravan, this causes 790 test failures on all non-Linux platforms, e.g. http://download.eclipse.org/eclipse/downloads/drops4/I20160518-2000/testresults/html/org.eclipse.ui.tests_ep46I-unit-mac64_macosx.cocoa.x86_64_8.0.html New Gerrit change created: https://git.eclipse.org/r/73131 New Gerrit change created: https://git.eclipse.org/r/73137 Hi Dani, Please review https://git.eclipse.org/r/73137. This has better implementation of verifying the GTK3 version. This uses system property to check for GTK3 instead of OS.GTK3. Thanks Sravan Gerrit change https://git.eclipse.org/r/73137 was merged to [master]. Commit: http://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=0687c87d2afb09d6d90cb6a62f5ce57634f03d48 With the latest patch, the test are skipped on a my GTK3-based Fedora. Reopening to investigate the failures during 4.7. I did some preliminary investigation: right off the bat I can see that 3 of the tests are failing due to List.getTopIndex(). Those are: testNoSelectionRefresh testSelectionRefresh testSelectionRevealBug177619 As for the others, I can't yet say if they are caused by getTopIndex() issues. Those test cases have more JFace in them so it will take me a bit longer to dig to the bottom of it. (In reply to Eric Williams from comment #13) > testSelectionRevealBug177619 This bug stems from List.getTopIndex(), but isn't the same problem outlined in bug 461354. This particular test failure is caused by gtk_tree_view_get_path_at_pos() returning false, which makes List.getTopIndex() to return 0. > testNoSelectionRefresh > testSelectionRefresh These two bugs are definitely caused by bug 461354. Working on a fix for that at the moment. (In reply to Eric Williams from comment #14) > This bug stems from List.getTopIndex(), but isn't the same problem outlined > in bug 461354. This particular test failure is caused by > gtk_tree_view_get_path_at_pos() returning false, which makes > List.getTopIndex() to return 0. I've actually included a fix for this in the patch for bug 461354 as well, since they are similar in nature. What is the status here Eric? Are the tests reenabled and fixed now? (In reply to Alexander Kurtakov from comment #16) > What is the status here Eric? Are the tests reenabled and fixed now? Will check, IIRC there are still 2 or 3 failures. (In reply to Eric Williams from comment #17) > Will check, IIRC there are still 2 or 3 failures. Tests are disabled, if I re-enable them on my local machine there are the old failures and a few new ones. Still work to be done then as part of bug 497330. Lucas, would you please take a look at these tests. New Gerrit change created: https://git.eclipse.org/r/115651 Gerrit change https://git.eclipse.org/r/115651 was merged to [master]. Commit: http://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=44130248a27b8c658e1445287aa4f630cf543a33 All tests running and passing. Verified in I20180121-2000. |