You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@felix.apache.org by "Andie Similon (Created) (JIRA)" <ji...@apache.org> on 2011/10/09 20:25:29 UTC
[jira] [Created] (FELIX-3151) Activating bundles in seperate thread
Activating bundles in seperate thread
-------------------------------------
Key: FELIX-3151
URL: https://issues.apache.org/jira/browse/FELIX-3151
Project: Felix
Issue Type: Improvement
Components: Framework
Reporter: Andie Similon
Priority: Minor
I was wondering about the following thing and I'd like one of the developers opinion about it. Is there a reason why bundles are not started in a seperate thread. I was wondering this because if a bundle developer creates a bundle that ends up in an infinite loop or something in the bundle start method, it will block the entire system. Is there a reason why the bundles aren't started in a seperate thread?
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (FELIX-3151) Activating bundles in seperate
thread
Posted by "Richard S. Hall (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/FELIX-3151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13124175#comment-13124175 ]
Richard S. Hall commented on FELIX-3151:
----------------------------------------
There is no particular reason. It could be done, but it would potentially complicate concurrency issues even more since we'd have to block the calling thread, call into the bundle with another thread, who could then in turn call back into the framework on another thread, which means we'd have to be extra careful with locking. This is really complicated by things like "refresh" that have to hold locks and operate on multiple bundles.
Further, the implied solution here is to have a timeout and try to kill the activator thread if it is taking too long, but we already know that killing threads in Java is unreliable and not recommended. Further still, this is just one failure mode associated with bundles and there are tons of other failure modes that we can't do anything about, since Java doesn't provide any way to isolate or limit resource consumption. So, in the end, if a bundle is not well behaved, it is probably better to learn about it quicker and stop using it.
> Activating bundles in seperate thread
> -------------------------------------
>
> Key: FELIX-3151
> URL: https://issues.apache.org/jira/browse/FELIX-3151
> Project: Felix
> Issue Type: Improvement
> Components: Framework
> Reporter: Andie Similon
> Priority: Minor
>
> I was wondering about the following thing and I'd like one of the developers opinion about it. Is there a reason why bundles are not started in a seperate thread. I was wondering this because if a bundle developer creates a bundle that ends up in an infinite loop or something in the bundle start method, it will block the entire system. Is there a reason why the bundles aren't started in a seperate thread?
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira