Bugzilla – Full Text Bug Listing
| Summary: | [Webkit2] Occasional crash on libjavascriptcoregtk-4.0.so during gtk_widget_destroy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Platform | Reporter: | Leo Ufimtsev <lufimtse> | ||||||||
| Component: | SWT | Assignee: | Leo Ufimtsev <lufimtse> | ||||||||
| Status: | RESOLVED FIXED | QA Contact: | |||||||||
| Severity: | normal | ||||||||||
| Priority: | P3 | CC: | benken, mike | ||||||||
| Version: | 4.7 | ||||||||||
| Target Milestone: | 4.7 M7 | ||||||||||
| Hardware: | PC | ||||||||||
| OS: | Linux | ||||||||||
| See Also: |
https://git.eclipse.org/r/93420 https://git.eclipse.org/c/platform/eclipse.platform.swt.git/commit/?id=95aee4d875873e0a962f582c45b052bd8ef741c5 https://bugs.eclipse.org/bugs/show_bug.cgi?id=522733 |
||||||||||
| Whiteboard: | |||||||||||
| Bug Depends on: | |||||||||||
| Bug Blocks: | 441568 | ||||||||||
| Attachments: |
|
||||||||||
Created attachment 267326 [details]
Webkit2_BrowserTest5_crash.log
Created attachment 267327 [details]
Webkit2_Btest_BrowserFunction_callback_with_integer_crash.log
Created attachment 267328 [details]
Webkit2_test_BrowserFunction_callback_with_multipleValues.log
New Gerrit change created: https://git.eclipse.org/r/93420 Gerrit change https://git.eclipse.org/r/93420 was merged to [master]. Commit: http://git.eclipse.org/c/platform/eclipse.platform.swt.git/commit/?id=95aee4d875873e0a962f582c45b052bd8ef741c5 (In reply to Eclipse Genie from comment #5) > Gerrit change https://git.eclipse.org/r/93420 was merged to [master]. > Commit: > http://git.eclipse.org/c/platform/eclipse.platform.swt.git/commit/ > ?id=95aee4d875873e0a962f582c45b052bd8ef741c5 Pending to see if crash is fixed by this. Will review in a week or so. Haven't observed crash since patch. Closing for now. Please re-open if you do find similar stack trace. If you hit this one from a Web Search, then see more recent Bug 522733 instead. (In reply to Michael Vorburger from comment #8) > If you hit this one from a Web Search, then see more recent Bug 522733 > instead. Thanks for linking these. |
*Very rarely*, but sometimes I catch Webkit2 crashing during jUnit test runs. It seems to happen once or twice a week. I don't have a way to reproduce it in a repeatable manner. Sometimes if I run AllBrowserTests in a loop for many hours, *sometimes* it also occurs. The pattern seems that the parent shell is 'closed', which eventually triggers OS.gtk_widget_destroy() on the parent. During this time it seems a javascript call is made and there is a crash deep down in libjavascriptcoregtk. My educated guess is that running 'evaluate()' or a callback from BrowserFunction while Browser is being disposed could be related as they would invoke libjavascriptcoregtk library at some point, which might run into some nullpointer. Attached are two crash dumps. One that occurred during BrowserTest5, the other in test_BrowserFunction_callback_with_integer. Both have a simmilar stack trace: Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code) > C [libjavascriptcoregtk-4.0.so.18+0xcbae5c] WTFCrash+0x1c > > Java frames: (J=compiled Java code, j=interpreted, Vv=VM code) > J 1621 org.eclipse.swt.internal.gtk.OS._gtk_widget_destroy(J)V (0 bytes) @ 0x00007f037921c481 [0x00007f037921c440+0x41] > J 1393 C1 org.eclipse.swt.widgets.Widget.release(Z)V (90 bytes) @ 0x00007f0379485824 [0x00007f0379484fe0+0x844] I will submit some patches aimed at trying to fix this. This would entail doing less evaluation()/execute() in the browser dispose mechanism. If someone catches more of these crashes, please attach crash dumps and let me know what was going on during this crash.