[
https://issues.apache.org/jira/browse/DERBY-4700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12887358#action_12887358
]
Kristian Waagan commented on DERBY-4700:
----------------------------------------
A comment from Dag, posted to derby-dev:
-----
> Do we need a bogus port for an embedded TC?
Well, I used one in the BootLockTest, for example (although Kathey has
later rewritten that test to work without ports), so I wouldn't rule
it a priori. I wasn't aware that we created two concurrent TCs either,
presumably they are used in sequence, so we don't *really* need separate
port assignements for them?
> What about a JMX port?
I think JMX is useful also for an embedded Derby, so we would want to
be able to test it.
Thanks,
Dag
-----
> Add method to obtain a bogus port in TestConfiguration
> ------------------------------------------------------
>
> Key: DERBY-4700
> URL: https://issues.apache.org/jira/browse/DERBY-4700
> Project: Derby
> Issue Type: Improvement
> Components: Test
> Affects Versions: 10.7.0.0
> Reporter: Kristian Waagan
> Assignee: Kristian Waagan
> Priority: Minor
> Attachments: derby-4700-1a.diff, derby-4700-1b.diff
>
>
> In some cases one needs to obtain a port number on which there is no Derby
> network server.
> Today this is solved in an ad-hoc way, for instance by using the current port
> minus one. When running tests in parallel, there is a chance that a port
> where there actually is a Derby network server listening is picked.
> As a start, I suggest that the bogus port is allocated to the last port in
> the port range configured for the run: baseport + MAX_PORTS_USED -1.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
