[
https://issues.apache.org/jira/browse/IGNITE-1006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14599320#comment-14599320
]
Artem Shutak commented on IGNITE-1006:
--------------------------------------
1. Reusing nodes().
Why you need any refactoring? Why you cannot just get all nodes, call
hostNames() for them and add it in one set?
2. Test.
My point is you have too easy test. For good test you need to have more then 1
node and test that ClusteGroup.hostNames() will return all hosts which will
return ClusterNode.hostNames() for all nodes in your cluster group.
And currently, your test can test nothing, because inputHostNames and
nodeHostNames can be empty - you do not check it .
Also, why inputHostNames.containsAll(nodeHostNames) ? Why contains? I don't
like this approach. It would be better expect equals of objects.
> Add hostnames() method to ClusterGroup to get all hostnames in current
> ClusterGroup
> -----------------------------------------------------------------------------------
>
> Key: IGNITE-1006
> URL: https://issues.apache.org/jira/browse/IGNITE-1006
> Project: Ignite
> Issue Type: Task
> Reporter: Atri Sharma
> Assignee: Atri Sharma
> Fix For: sprint-7
>
> Attachments: ignite-sprint-5_ignite-1006ver3.patch
>
>
> We should have method like nodes() in ClusterGroup to allow getting all
> hostnames currently in this ClusterGroup. Method should ensure handling for
> duplicates as well.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
