[jbosstools-issues] [JBoss JIRA] (JBIDE-17656) Forge crashed Eclipse

[Vineet Reynolds (JIRA)]

    [ https://issues.jboss.org/browse/JBIDE-17656?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13001955#comment-13001955 ] 

Vineet Reynolds commented on JBIDE-17656:
-----------------------------------------

While high GC activity is not the cause of the issue, it was definitely noticed and fixed. In fact, more operations must be performed (that the 4 steps outlined in the JIRA to see the high GC behavior). However, It is only tangentially related to the issue.

The root cause of the crash is a bug in the JDK where the JVM attempts to clean up a native buffer, without considering that it may be used and accessed by other threads. In our case of JBDS, one such thread would be of Forge, while another may be of Eclipse, or any other plugin in it. A patch was supplied in Furnace and Forge 2.9.0+ to delay clean of this buffer (this is not in repsin-a, but should be available in respin-b).

> Forge crashed Eclipse
> ---------------------
>
>                 Key: JBIDE-17656
>                 URL: https://issues.jboss.org/browse/JBIDE-17656
>             Project: Tools (JBoss Tools)
>          Issue Type: Bug
>          Components: forge
>    Affects Versions: 4.2.0.Beta3
>         Environment: Windows 7 64bits, JDK 8.0.0, JBDS 8.0.0.Beta2 
>            Reporter: Fred Bricon
>            Assignee: Vineet Reynolds
>            Priority: Critical
>              Labels: respin-b
>             Fix For: 4.2.0.CR1
>
>         Attachments: generate.fsh, hs_err_pid4536.log, hs_err_pid5320.log, hs_err_pid5372-psrna.log, hs_err_pid640.log, JVM-GCactivity-heap.png, JVM-GCactivity-Linux.png, JVM-GCactivity.png, psrna-forge-crash-win7.png
>
>
> I tried to reproduce the steps from JBIDE-17655 a 2nd time, Eclipse crashed while forge was scaffolding 



--
This message was sent by Atlassian JIRA
(v6.3.1#6329)