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.

Bug 31206 - Deadlock in NbJavaCvsFileSystem$JavaCvsAttributes
Summary: Deadlock in NbJavaCvsFileSystem$JavaCvsAttributes
Status: VERIFIED WONTFIX
Alias: None
Product: versioncontrol
Classification: Unclassified
Component: CVS (show other bugs)
Version: 3.x
Hardware: PC All
: P2 blocker (vote)
Assignee: issues@versioncontrol
URL:
Keywords: RANDOM, THREAD
Depends on:
Blocks:
 
Reported: 2003-02-19 11:16 UTC by Vitezslav Stejskal
Modified: 2007-01-04 17:14 UTC (History)
0 users

See Also:
Issue Type: DEFECT
Exception Reporter:

Attachments
full thread dump (18.44 KB, text/plain)
2003-02-19 11:18 UTC, Vitezslav Stejskal 		Details
FTD with DEADLOCK when doing copy&paste action on JavaCVS FS (24.28 KB, text/plain)
2003-05-05 19:02 UTC, dmladek 		Details
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Vitezslav Stejskal 2003-02-19 11:16:29 UTC 
There is deadlock in
NbJavaCvsFileSystem$JavaCvsAttributes which
occasianly happens during the project switch. IMO
it's caused by touching settings
(SharedClassObject) from
org.netbeans.modules.cvsclient.NbJavaCvsFileSystem$NbInfoImpl.outputStream(NbJavaCvsFileSystem.java:499)
which holds LOCK. The SCO uses lookup which waits
for Folder recognizer thread which is trying to
use the LOCK when reading some attrs from
NbJavaCVSFileSystem.

I will attach the thread-dump.
Comment 1 Vitezslav Stejskal 2003-02-19 11:18:10 UTC 
Created attachment 9019 [details]
full thread dump
Comment 2 Martin Entlicher 2003-02-20 14:16:29 UTC 
Deadlock needs to be fixed into 3.5
Comment 3 Tomas Zezula 2003-03-13 15:36:23 UTC 
Scheduled for next release, will be solved while refactoring settings.
Comment 4 dmladek 2003-05-05 14:45:02 UTC 
isn't it possible easy reproduce by copy & paste actions on JavaCVS FS?

I'll check it tomorrow. Increasing the priority to P2 (Deadlock has to
be always P2)
Comment 5 Martin Entlicher 2003-05-05 15:20:27 UTC 
If this is not much reproducible and does not occur during common
tasks I would like to defer this for NB 4.0. The reason is, that this
problem will disappear by itself by removing javacvs filesystem from
the IDE distribution. Only vcsgeneric (with added profile for cvslib)
will be present in NB 4.0.
Comment 6 Vitezslav Stejskal 2003-05-05 16:52:14 UTC 
Read the 'Bug Priority Guidelines':

P1 - reproducible, unavoidable crash or deadlock
P2 - intermittent crash or deadlock, happens fairly often

This happens occasionaly during the projects switch, I never saw it
when working with regular files. IMO it's not P2.
Comment 7 dmladek 2003-05-05 19:02:10 UTC 
Created attachment 10222 [details]
FTD with DEADLOCK when doing copy&paste action on JavaCVS FS
Comment 8 dmladek 2003-05-05 19:14:57 UTC 
well, there's only one thing which is common for your scenario and
mine scenarion. And it is the thread dump and wisible stacktrace. Also
even I'm not able to reproduce everytime my scenario. 

But it occures when:
====================
Using Nevada RC2, jdk1.4.2 rc #b21, rh80.
Mount JAvaCVS FS.
HAve already commited some dir structure + files in it.
Create a new dir.
Copy & Paste in it some already commited subpackages+files.

Then the DEADLOCK with almost identical threadDump occures.


So, then, I'm still for it is P2 ('cause DEADLOCK are serious) and it
is not only about projects, but because it is not 100%reproductable
I'm for "solving" it in 4.0.
Also the P2 "says" it won't go into 3.5 release/Nevada, right?
Comment 9 Martin Entlicher 2004-01-14 14:46:14 UTC 
As described at
http://www.netbeans.org/servlets/ReadMsg?msgId=655179&listName=nbusers
the development of javacvs module has been stopped.
Comment 10 Peter Pis 2005-07-20 15:53:39 UTC 
Verified.