Is there some reason you can't do:
public static final Marker VERBOSE_MARKER = MarkerManager.getMarker("VERBOSE");
and then logger.info(VERBOSE_MARKER, "Reading Accounts");
then in your configuration you would do
<MarkerFilter marker="VERBOSE" onMatch="ACCEPT" onMismatch="DENY"/>
and finally, in your Appender you could use a PatternLayout that includes
pattern="%p %marker %msg"
You could also wrap the marker in a %replace so that you don't end up with two
spaces when there is not Marker.
The point is that the purpose of Markers is to fill the need for these kinds of
things without needing to create an endless number of logging levels. Remember
that every new Level that gets added causes 14 new methods to be added to the
Logger interface.
Another consideration is that if a new level was added between info and debug
when mapping the Log4j 2 API to SLF4J we would have to make an arbitrary choice
as to whether the events should be mapped to INFO or DEBUG since SLF4J (as well
as JCL) don't have a corresponding level. OTOH, JUL has a CONFIG level between
FINE and INFO, which seems to correlate to the examples you show.
On Oct 7, 2013, at 8:50 PM, Gary Gregory <[email protected]> wrote:
> On Mon, Oct 7, 2013 at 11:16 PM, Scott Deboy <[email protected]> wrote:
> If your examples included loggers I think it would show these added verbose
> entries would be another logger at INFO but not a separate severity.
>
> I think of levels in terms of relative numeric severity. DEBUG < INFO <WARN
> <ERROR etc. Chainsaw supports this kind of logic in filtering colorizing and
> searching for events.
>
> Where would verbose go? I would put it below DEBUG myself.
>
>
> I should not have proposed a name for this level since 'verbose' means
> different things to different people. As I stated in my original message,
> this new level would be between INFO and DEBUG.
>
> Gary
>
> Btw this advice to use another logger in most cases doesn't mean there aren't
> cases where additional levels aren't useful.
>
> Just something to think about.
>
> Scott
>
> On Oct 7, 2013 7:30 PM, "Gary Gregory" <[email protected]> wrote:
> Hi All:
>
> I've just come across the need for distinguishing log entries somewhat
> between what INFO and DEBUG provide.
>
> I think of the DEBUG level as a watermark where data is provided for
> developers and support to do deep debugging. INFO is for users. But I see the
> need now for providing an easy way to give user more information (hence
> VERBOSE or another word) which is not at the DEBUG level of detail.
>
> Ideally, I'd like this in INFO mode:
>
> INFO Reading configuration
> INFO Running server
>
> Then in a new VERBOSE mode:
>
> INFO Reading configuration
> VERBOSE Reading accounts
> VERBOSE Reading this config data
> VERBOSE Reading that config data
> VERBOSE Reading other config data
> INFO Running server
>
> The in DEBUG mode, you'd get all the things that are real debug information
> like
>
> DEBUG Configuration location was not provided on the command line.
> DEBUG Searching classpath for configuration: cp entries
> DEBUG Searching user home for configuration: c:\users\...
> DEBUG Found configuration at location xzy, last modified date: ...
> INFO Reading configuration
>
> Thoughts?
>
> --
> E-Mail: [email protected] | [email protected]
> Java Persistence with Hibernate, Second Edition
> JUnit in Action, Second Edition
> Spring Batch in Action
> Blog: http://garygregory.wordpress.com
> Home: http://garygregory.com/
> Tweet! http://twitter.com/GaryGregory
>
>
>
> --
> E-Mail: [email protected] | [email protected]
> Java Persistence with Hibernate, Second Edition
> JUnit in Action, Second Edition
> Spring Batch in Action
> Blog: http://garygregory.wordpress.com
> Home: http://garygregory.com/
> Tweet! http://twitter.com/GaryGregory
