[
https://issues.apache.org/jira/browse/GERONIMO-3889?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sangjin Lee updated GERONIMO-3889:
----------------------------------
Attachment: GERONIMO-3889.patch
A suggested fix.
Essentially it is adopting the same ownership model as that for the main I/O
thread pool. If the caller passes in a scheduled executor service, the
lifecycle management is up to the caller, i.e. HttpIoHandler does not attempt
to shut it down. For the case where one was not provided, we create a
single-threaded *daemon* executor so that we don't need to explicitly clear it.
I think it would lead to a most straightforward approach to this...
> HttpIoHandler shuts down the scheduled executor service even if it is passed
> in by caller
> -----------------------------------------------------------------------------------------
>
> Key: GERONIMO-3889
> URL: https://issues.apache.org/jira/browse/GERONIMO-3889
> Project: Geronimo
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: AsyncHttpClient
> Affects Versions: 1.x
> Reporter: Sangjin Lee
> Assignee: Rick McGuire
> Priority: Minor
> Attachments: GERONIMO-3889.patch
>
>
> HttpIoHandler uses a scheduled executor service to handle timing out
> requests. It can either take one from the caller, or it will create one by
> itself. Therefore, the ownership becomes confusing if the caller passes in
> one.
> This effectively prevents multiple instances of AsyncHttpClient from sharing
> a single scheduled executor service. It should allow them to share a
> scheduled executor service.
> I'll see if I can come up with a good way to address this...
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
