[jira] [Commented] (VELOCITY-989) Multiple connection leaks in DataSourceResourceLoader
"Claude Brisson (Jira)" <[email protected]> Wed, 22 Apr 2026 17:36:00 +0000 (UTC)
| Newsgroups | gmane.comp.jakarta.velocity.devel |
|---|---|
| Message-ID | <[email protected]> |
[ https://issues.apache.org/jira/browse/VELOCITY-989?page=3Dcom.atlassi=
an.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D18=
075505#comment-18075505 ]=20
Claude Brisson commented on VELOCITY-989:
-----------------------------------------
I made a new proposal, changing the contract of the factory class (in fact =
deprecating the current one in favor of a new one, and instrumentalizing th=
e deprecated one so that it works with a class proxy...).
[~fschlimme] , [~michael-o] , let me know if the proposal looks okay to you=
: [PR 62|https://github.com/apache/velocity-engine/pull/62]
> Multiple connection leaks in DataSourceResourceLoader
> -----------------------------------------------------
>
> Key: VELOCITY-989
> URL: https://issues.apache.org/jira/browse/VELOCITY-989
> Project: Velocity
> Issue Type: Bug
> Reporter: Francis M Schlimmer
> Priority: Major
>
> The first case is in the implementation of FilterReader. It attempts to g=
et a statement from a result set after closing the result set. This always =
throws an exception as it is an invalid operation after closing the result =
set, which results in not closing the connection.
> The second is in the getResourceReader method. In the happy path a Filter=
Reader is returned which eventually closes the connection when close is cal=
led. However, if a template is not found, an exception is thrown without cl=
osing the connection.
> =C2=A0
> Pull request here [https://github.com/apache/velocity-engine/pull/57]
> =C2=A0
--
This message was sent by Atlassian Jira
(v8.20.10#820010)