Branch: refs/heads/master
Home: https://github.com/jenkinsci/remoting
Commit: 1083a97145b83f88d9eee0a920a9495e192cd480
https://github.com/jenkinsci/remoting/commit/1083a97145b83f88d9eee0a920a9495e192cd480
Author: Kohsuke Kawaguchi <[email protected]>
Date: 2014-09-05 (Fri, 05 Sep 2014)
Changed paths:
M src/main/java/org/jenkinsci/remoting/nio/NioChannelHub.java
Log Message:
-----------
[FIXED JENKINS-24050] don't let canceled keys kill the selector thread
In looking at the proposed PR #24
(https://github.com/jenkinsci/remoting/pull/24), I feel bit
uneasy to mask the problem like it does.
The code in question is looping through selected keys and processing it one by
one.
The only code that calls key.cancel() is done from the selector thread that
runs this loop.
So I don't understand how it is possible that the key picked up from selected
key set is
already cancelled here. I wonder if something more is going on.
Regardless, I agree that this shouldn't kill the selector thread, which breaks
all the slaves
in one go. This change flags and reports the problem, kill the connection
related to that key,
then continue to serve other connections.
--
You received this message because you are subscribed to the Google Groups
"Jenkins Commits" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
