[
https://issues.apache.org/jira/browse/POOL-178?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12926617#action_12926617
]
Phil Steitz commented on POOL-178:
----------------------------------
I am -1 on this patch, for the following reasons:
1. numActive and numIdle have a different meanings for GKOP vs GOP and need to
be documented separately and likely renamed for GKOP.
2. what is left as the top level interface does not encapsulate a "pool".
3. The only useful code in the new "Pool" class is "assertOpen" and
"isClosed".
I think it is better to keep separate base classes for keyed and non-keyed
pools as the core pool methods have different signatures and semantics. If we
think the small duplication of the assertOpen / isClosed code in BaseObjectPool
and BaseKeyedObjectPool is worth creating a common superclass, we can create a
common parent for them, called something like "Container" or somesuch; but I
don't at this point see the wisdom in getting rid of them.
> Common interfaces and abstract classes for BOP and BKOP
> -------------------------------------------------------
>
> Key: POOL-178
> URL: https://issues.apache.org/jira/browse/POOL-178
> Project: Commons Pool
> Issue Type: Improvement
> Affects Versions: Nightly Builds
> Reporter: Gary Gregory
> Fix For: 2.0
>
> Attachments: POOL-178.diff
>
>
> I have created a patch that IMO should make it easier for us to further
> refactor code.
> In brief, OP/BOP and KOP/BKOP do not have a common parent interface/class.
> This patch remedies this by pulling up the
> close/clear/getNumActive/getNumIdle methods into a new interface and class.
> This seems not very controversial (to me at least), so I would like to commit
> it (unless riots spread from Paris to this list.)
> Thoughts?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
