[
https://issues.apache.org/jira/browse/DISPATCH-111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14488166#comment-14488166
]
Ted Ross commented on DISPATCH-111:
-----------------------------------
The cause of this problem comes from the fact that there are two asynchronous
paths through the router used in link routing. The example program, once it is
given credit, sends a transfer immediately followed by a detach. Inside the
router, the detach actually passes the transfer and is sent to the broker in
front of the transfer. This is obviously incorrect.
The solution is to move the propagation of the detach onto the same internal
async FIFO that is used for delivery transfers and disposition updates.
> helloworld example hangs when run against a broker via link-routing
> -------------------------------------------------------------------
>
> Key: DISPATCH-111
> URL: https://issues.apache.org/jira/browse/DISPATCH-111
> Project: Qpid Dispatch
> Issue Type: Bug
> Components: Router Node
> Affects Versions: 0.4
> Reporter: Ted Ross
> Assignee: Ted Ross
> Fix For: 0.5
>
>
> The helloworld.py example (from proton/examples/engine/py) hangs when run
> against a router via Dispatch Router.
> Produced messages are not forwarded to the broker queue. The reason is that
> the sender detach is sent immediately following the message delivery and
> Dispatch is not properly handling the partial-closed status of the link.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
