Hi Ceki,
I was only speaking for myself and giving feedback on what I have done for
Cactus ! I don't know what is the opinion of others on the subject and maybe
they have a better solution ! :-)
Thanks
-Vincent
----- Original Message -----
From: "Ceki Gülcü" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, August 01, 2001 10:26 PM
Subject: Re: [httpclient] log4j redux
Vincent,
I see that you already have a solution. I won't waste my time working on the
log4jME then.
Regards, Ceki
At 20:47 01.08.2001 +0100, Vincent Massol wrote:
>----- Original Message -----
>From: "Scott Sanders" <[EMAIL PROTECTED]>
>To: <[EMAIL PROTECTED]>
>Sent: Wednesday, August 01, 2001 7:57 PM
>Subject: Re: [httpclient] log4j redux
>
>
>> RE: [httpclient] log4j reduxIs it possible to allow httpclient to be
>passed
>> a Category that can override an internal Category-like class like Rodney
>> mentioned which just does System.out/err? This would require log4j at
>> compile time, but only optionally at runtime. Full control for the
log4jh
>> fanatic, without REQUIRING log4j.jar.
>>
>> I would see this as a great way to incorporate full-blown logging for
>those
>> who want it, while not requiring it at all if you do not need it.
>>
>
>This is almost exactly what I have done in Cactus. I have a very very thin
>wrapper and I use reflection to test if log4j classes are on the classpath,
>and if not, provide a dummy implementation that does nothing (it can be
>easily change to use stdout and stdee). If it is there then it uses log4j.
I
>like it this way.
>
>> I am a huge fan of log4j type logging, but I personally do not want to
use
>> it inside black-box type components within my system such as httpclient,
>> digester, collections, beanutils, etc.
>>
>> Scott Sanders
>-Vincent
--
Ceki Gülcü - http://qos.ch
