[jboss-jira] [JBoss JIRA] (LOGMGR-117) Fork Log4J ConosleAppender to simply handling of calls to System.out

[RH Bugzilla Integration (JIRA)]

    [ https://issues.jboss.org/browse/LOGMGR-117?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13082627#comment-13082627 ] 

RH Bugzilla Integration commented on LOGMGR-117:
------------------------------------------------

Nikoleta Ziakova <nziakova at redhat.com> changed the Status of [bug 1196228|https://bugzilla.redhat.com/show_bug.cgi?id=1196228] from ON_QA to VERIFIED

> Fork Log4J ConosleAppender to simply handling of calls to System.out
> --------------------------------------------------------------------
>
>                 Key: LOGMGR-117
>                 URL: https://issues.jboss.org/browse/LOGMGR-117
>             Project: JBoss Log Manager
>          Issue Type: Enhancement
>            Reporter: Kyle Lape
>            Assignee: Lin Gao
>             Fix For: 1.1.2.CR1
>
>
> While the use of {{ConsoleAppender}} in Log4J application logging configurations is discouraged, JBoss could probably do a better job of managing its use.
> Currently it's possible to encounter a deadlock if thread A invokes {{System.out.println()}} and thread B invokes {{Logger.info()}}.  Thread A will get the lock on {{System.out}} (a {{java.io.PrintWriter}}), but then thread B will get the lock on the {{org.apache.log4j.ConsoleAppender}}.  Then each thread is waiting on the other's lock.
> The suggested improvement (discussed with [~jamezp]) is to fork Log4J's {{ConsoleAppender}} so that instead of calling {{System.out.println}}, it calls the real {{stdout}}.
> *Bonus enhancement*: Drop a {{WARN}} message to the system root logger discouraging the use of {{ConsoleAppender}} in application configurations.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)