On Fri, Jul 11, 2008 at 10:28 AM, Jeremy Haile <[EMAIL PROTECTED]> wrote:
> So, how exactly are you proposing I use SLF4J with the new code you checked
> in?  I don't see any support for it...

I haven't implemented that yet - it will take me about 5 minutes, I
just didn't have time yesterday

>
> It sounds like you are talking about doing detection of whether or not the
> jar is in the classpath - which would result in the same classloading
> hassles that is making everyone switch away from commons-logging to slf4j.
>  If we followed slf4js approach, then everyone would have to include
> jsecurity-slf4j.jar in their classpath, plus slf4j.jar plus slf4j-log4j.jar.
>  This seems like a needless hassle.

I am talking about that, but you're mistaken in that this would cause
the same classloading problems seen in Commons Logging.  Those CL
issues are related to using a class object to initialize a Log
instance.  I'm not doing that.  What I'm talking about is the log
framework detection exactly in the same manner that we detect if
Ehcache is available.  Same technique - unrelated to the Commons
Logging CL problems.

Reply via email to