[
https://issues.apache.org/jira/browse/LOG4J2-1343?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15235609#comment-15235609
]
Remko Popma edited comment on LOG4J2-1343 at 4/11/16 6:02 PM:
--------------------------------------------------------------
The solution to the above deadlock problem (as currently in master) still has
issues because there are two objects that are locked on during Layout.encode():
the OutputStreamManager and the ByteBufferDestination. (This is not a problem
if {{Constants.ENABLE_DIRECT_ENCODERS}} is false.)
The cleanest solution is to collapse these two locks into one and let all
OutputStreamManagers implement the ByteBufferDestination interface. This
removes a lot of {{if (Constants.ENABLE_DIRECT_ENCODERS)...}} complexity from
the managers. Custom appender/manager subclasses then also no longer need to
worry if they can call the {{write(...)}} methods or if they first need to call
{{ByteBufferDestination.drain(...)}}: all writing would first go to the buffer,
then to the underlying output stream or randomAccessFile.
Work for this is in progress on branch
{{LOG4J2-1343-no-gc-os-appenders-bytebuffered}}.
was (Author: rem...@yahoo.com):
The solution to the above deadlock problem (as currently in master) still has
issues because there are two objects that are locked on during Layout.encode():
the OutputStreamManager and the ByteBufferDestination. (This is not a problem
if {{Constants.ENABLE_DIRECT_ENCODERS}} is false.)
The cleanest solution is to collapse these two locks into one and let all
OutputStreamManagers implement the ByteBufferDestination interface. This
removes a lot of {{if (Constants.ENABLE_DIRECT_ENCODERS)...}} complexity from
the managers. Custom appender/manager subclasses now also no longer need to
worry if they can call the {{write(...)}} methods or if they first need to call
{{ByteBufferDestination.drain(...)}}: all writing now goes first to the buffer,
then to the underlying output stream or randomAccessFile.
Work for this is in progress on branch
{{LOG4J2-1343-no-gc-os-appenders-bytebuffered}}.
> Update ConsoleAppender to utilize gc-free Layout method
> -------------------------------------------------------
>
> Key: LOG4J2-1343
> URL: https://issues.apache.org/jira/browse/LOG4J2-1343
> Project: Log4j 2
> Issue Type: Improvement
> Components: Appenders
> Affects Versions: 2.5
> Reporter: Remko Popma
> Assignee: Remko Popma
> Fix For: 2.6
>
>
> TBD: would we want to include this in the upcoming 2.6 release?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscr...@logging.apache.org
For additional commands, e-mail: log4j-dev-h...@logging.apache.org
