This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
| Summary: | Preventing deadlock #65543: Do not call FolderLookup from inside DataObject operations! | ||
|---|---|---|---|
| Product: | javame | Reporter: | Lukas Hasik <lhasik> |
| Component: | Build System | Assignee: | Adam Sotona <asotona> |
| Status: | VERIFIED FIXED | ||
| Severity: | blocker | CC: | dkaspar |
| Priority: | P2 | Keywords: | THREAD |
| Version: | 6.x | ||
| Hardware: | All | ||
| OS: | All | ||
| Issue Type: | DEFECT | Exception Reporter: | |
| Attachments: |
stacktrace
exception exception |
||
|
Description
Lukas Hasik
2007-02-15 12:50:59 UTC
Created attachment 38535 [details]
stacktrace
I am not able to prevent LocalHistoryVCSInterceptor to request EditorSupport at the same time of DataObject creation. ur EditorSupport needs FolderLookup so the only place to cut this potetial deadlock is where LocalHistoryVCSInterceptor listens on DO creation. BTW asking for EditorSupport for each newly created DO is wrong even from performance point of view. Increasing priority because the probablility of deadlock is very high. Several similar cases were found and fixed last milestone. should be already fixed http://versioncontrol.netbeans.org/source/browse/versioncontrol/localhistory/src/org/netbeans/modules/localhistory/LocalHistoryVCSInterceptor.java?r1=1.8&r2=1.9 I'm able to reproduce with 070215 build of NB + 070216 of Mobility. Test case: Create 50 times new Visual MIDlet into a j2me project. The exception appeared 4 times. Test is available at /cvs/mobility/project/test/qa-functional/src/validation/NewProjectFileTest.java The exception is different. Seems that the localhistory isn't involved in it anymore. Tomas, please evaluate the exception and reassign to "mobility/build system". thanks. Created attachment 38610 [details]
exception
Created attachment 38611 [details]
exception
doesn't look like local history anymore reassigning back to Adam. This is something else, there is no more risk of deadlock in this thread dump. BTW: I don't see what is the exception about. closing the original issue. I will report the new exception as new issue. verified in 070321 |