[
https://issues.apache.org/jira/browse/DBCP-591?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17738305#comment-17738305
]
Phil Steitz commented on DBCP-591:
----------------------------------
Acknowledged. It is strange that intrinsic locks work differently, but I
understand that at this point at least they do. I am tempted to suggest
removing the sync entirely here; but we have not been that consistent in
stating the expectation that concurrent access to PoolableConnections is not a
good thing, so it is probably best to leave it in, but replace the intrinsic
lock with an a Reeentrant lock. Patches / better ideas welcome.
> Avoid synchronized in PoolableConnection.close
> ----------------------------------------------
>
> Key: DBCP-591
> URL: https://issues.apache.org/jira/browse/DBCP-591
> Project: Commons DBCP
> Issue Type: Improvement
> Reporter: Attila Kelemen
> Priority: Minor
>
> org.apache.commons.dbcp2.PoolableConnection.close uses an intrinsic lock to
> synchronize its close method. However, this can be an issue since this pins
> the carrier thread of a virtual thread, needlessly degrading performance.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
