| Summary: | Deadlock reported during startup of org.eclipse.tm.tcf.cdt.ui | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Tools] TCF | Reporter: | Anton Leherbauer <aleherb+eclipse> | ||||
| Component: | Debug | Assignee: | Project Inbox <tcf.debug-inbox> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | P3 | CC: | eugene | ||||
| Version: | unspecified | ||||||
| Target Milestone: | 0.4.0 | ||||||
| Hardware: | PC | ||||||
| OS: | Windows XP | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
Anton Leherbauer
Created attachment 189073 [details] Proposed patch Putting the initialization of the TCFBreakpointStatusListener last in the start method of the activator fixes it. The reason behind the deadlock is explained in https://bugs.eclipse.org/bugs/show_bug.cgi?id=86713 As understand, Equinox can dead lock if a thread, other then main, tries to load a class form a plugin being started. Looks like a serious platform bug to me. I wonder why the bug is closed with "Marking this bug as invalid because sufficient details were not provided". Anyway, I have committed the patch. Thanks! (In reply to comment #2) > As understand, Equinox can dead lock if a thread, other then main, tries to > load a class form a plugin being started. Looks like a serious platform bug to > me. I wonder why the bug is closed with "Marking this bug as invalid because > sufficient details were not provided". It's actually considered a Sun/Oracle JVM bug http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4670071 and the last comment is probably due to an automatic update. It was resolved as REMIND which is now an invalid state. |