[ 
https://issues.apache.org/jira/browse/LOG4J2-3691?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17889196#comment-17889196
 ] 

Piotr Karwasz commented on LOG4J2-3691:
---------------------------------------

{quote}The `toBuilder()` method sounds interesting I hope something happens 
there!  It would probably only refect the interpreted configuration and not the 
original XML representation right?  By that I mean some builders return `null` 
if the element is misconfigured or assume defaults.{quote}

I am afraid nothing will happen automatically, since our backlog of new 
features is enough for a lifetime. If you find the idea useful, you should 
start a thread on [dev@logging mailing 
list|https://logging.apache.org/support.html#discussions-maintainer]. 
Personally I would like to see the project go into that direction, but I don't 
have the time to work on it now. From what I understand you are starting a big 
migration task from Log4j 1 to Log4j 2. Contributing some of that work back to 
the Log4j Project might save you some time. Not all new features are accepted 
in the Log4j codebase, but the `toBuilder()` feature probably would be accepted.

PS: Any configuration file that comes from a runtime configuration, will only 
contained the resolved value of the {{${...}}} property substitution 
placeholders. Maybe a better way to get the current configuration is to get the 
original configuration file through JMX or 
[Configuration.getConfigurationSource()|https://logging.apache.org/log4j/2.x/javadoc/log4j-core/org/apache/logging/log4j/core/config/Configuration.html#getConfigurationSource()].

> Documentation: CompositeTriggeringPolicy - nested <Policies> element?
> ---------------------------------------------------------------------
>
>                 Key: LOG4J2-3691
>                 URL: https://issues.apache.org/jira/browse/LOG4J2-3691
>             Project: Log4j 2
>          Issue Type: Bug
>          Components: Configuration, Documentation
>    Affects Versions: 2.24.0
>            Reporter: Jeff Thomas
>            Priority: Minor
>
> According to my JetBrains AI Assistant :):
> "According to the Log4j 2 configuration guidelines, nesting a {{Policies}} 
> element within another {{Policies}} element is not supported. Each 
> {{RollingFile}} appender should have one {{Policies}} element, which in turn, 
> directly contains the individual policies."
> Example:
> {code:java}
> <RollingFile name="FILE"
>              fileName="app.log"
>              filePattern="app.%d{yyyy-MM-dd}.%i.log">
>   <JsonTemplateLayout/>
>   <Policies>
>     <OnStartupTriggeringPolicy/>
>     <Policies>
>       <SizeBasedTriggeringPolicy/>
>       <TimeBasedTriggeringPolicy/>
>     </Policies>
>   </Policies>
> </RollingFile> {code}
> I could not find an explicit statement regarding this in the new Log4j 2.x 
> documentation.
> Also in the code of the `CompositeTriggeringPolicy` class it seems that there 
> is no validation check to ensure that this does not happen.  
> If this is in fact, undesirable maybe the documentation should state this and 
> also enforce it in code (or alternatively aggregate the policies - flatten 
> them to the top-level).
> Side note: the documentation and implementation don't mention adding multiple 
> policies of the same type to a composite-policy (i.e. two 
> "CronTriggeringPolicy" elements) - whether this is supported or actively 
> discouraged.
> NOTE: The same question could be applied to the `CompositeFilter`constructor 
> not checking if one of the provided filters is also a `CompositeFilter`.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to