[
https://issues.apache.org/jira/browse/XALANJ-1844?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Joe Kesselman updated XALANJ-1844:
----------------------------------
Description:
Considering a highly multithreaded system, where threads have a short living
time, and some of the threads perform XSLT transformation, but only once, the
JVM soon reaches memory saturation, since XMLReaders are never re-used.
I would suggest maybe to keep the ThreadLocal mechanism instead of the map +
ThreadLocal duet which appears redundant to me, but maybe i am missing some
point ?
André
was:
Considering a hightly multithreaded system, where threads have a short living
time, and some of the threads perform XSLT transformation, but only once, the
JVM soons reaches memory saturation, since XMLReaders are never re-used.
I would suggest maybe to keep the ThreadLocal mecanism instead of the map +
ThreadLocal duet which appears redondant to me, but maybe i am missing some
point ?
André
> XMLReaderManager provides neither a way to clear its cache nor does it clear
> its cache by itself
> ------------------------------------------------------------------------------------------------
>
> Key: XALANJ-1844
> URL: https://issues.apache.org/jira/browse/XALANJ-1844
> Project: XalanJ2
> Issue Type: Bug
> Security Level: No security risk; visible to anyone(Ordinary problems in
> Xalan projects. Anybody can view the issue.)
> Components: Other
> Affects Versions: 2.7.1
> Environment: Operating System: All
> Platform: All
> Reporter: Andre Doherty
> Assignee: Henry Zongaro
> Priority: Blocker
> Fix For: 2.7
>
> Attachments: ASF.LICENSE.NOT.GRANTED--28082.patch,
> ASF.LICENSE.NOT.GRANTED--XMLReaderManager.java,
> ASF.LICENSE.NOT.GRANTED--XMLReaderManager.java
>
>
> Considering a highly multithreaded system, where threads have a short living
> time, and some of the threads perform XSLT transformation, but only once, the
> JVM soon reaches memory saturation, since XMLReaders are never re-used.
> I would suggest maybe to keep the ThreadLocal mechanism instead of the map +
> ThreadLocal duet which appears redundant to me, but maybe i am missing some
> point ?
> André
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
