logging-log4j-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ceki Gülcü <c...@qos.ch>
Subject Re: [POLL] Moving UGLI outside Apache?
Date Wed, 06 Apr 2005 17:26:29 GMT


Thank you for your informative input which is highly appreciated. I agree 
with what you wrote so won't add anything except extend a firm invitation 
for your participation in the new project when it comes to being.

At 06:36 PM 4/6/2005, Simon Kitching wrote:
>Hi All,
>The set of people actually interested in working on JCL seems to have 
>dwindled back to two again (at least for the moment): Robert Donkin and 
>myself. And I´m still on holiday until this time next week; I´m following 
>the email discussions but can´t do any active work on things now. Note, by 
>the way, that neither of us is hugely interested in JCL per se; we happen 
>to work on projects that use it so need to get it working right in 
>self-defense! There are of course a number of additional people observing 
>but not actively contributing right now.
>There´s still no firm decision made on what will happen with JCL. Robert 
>has a list of scenarios that supposedly show limitations in UGLI 
>"discovery", and he is at least *interested* in exploring alternatives to 
>UGLI. I have been investigating slightly different issues recently, 
>related to singletons and studying up on j2ee classloading approaches and 
>implications which are also relevant to logging. I intend to have a good 
>look at UGLI and Robert´s concerns/alternatives when I get back from 
>holiday. Hopefully we can then have a big discussion/debate involving 
>everybody who is interested in this topic and finally select an approach - 
>which may or may not be UGLI.
>I think it would be healthy for apache to have a single logging wrapper if 
>we can. There´s no point in splitting the developer effort, or in 
>confusing potential users.
>Having some real code/unit-tests to demonstrate the pros and cons of 
>different approaches is much healthier than abstract theorising, and will 
>hopefully be able to make it crystal clear which scenarios are better 
>handled by which approach. Ceki posted some very useful demo code to the 
>commons-dev list a month or so ago. Robert has recently posted some code 
>too, which I hope you can find time to look at; see the mail archives for 
>details. I haven´t unfortunately been able to try it out yet but will next 
>week. With my new understanding of classloaders I hopefully can provide 
>some additional unit tests to demonstrate the pros/cons of various approaches.
>All this will of course take some time; that´s why my email of a week ago 
>said that it would likely be a few months before any concensus was reached 
>on the future of JCL.
>Re Ceki´s concerns on licensing: yes, it may be worth thinking about 
>hosting UGLI (or whatever) on an external site so that a pure BSD license 
>can be applied. I think this would be a bad idea from a community point of 
>view, but may be the only solution if there really are projects that are 
>avoiding the Apache 2.0 license.
>Jacob Kjome wrote:
>>Will this allow us to continue to include UGLI in Log4j.jar or will it be
>>required to be a separate jar?  Also, what are our plans for UGLI?  I haven't
>>followed the commons-dev list.  Is there any new information being provided
>>there as to JCL's plans?  Is UGLI part of it?  If we continue with UGLI 
>>and JCL
>>doesn't go with it, are we prepared to continue with UGLI or are we going to
>>drop it for whatever JCL comes up with?  Is JCL dead-set on developing 
>>JCL 2.0
>>instead of moving to an agnostic (external to Apache) logging API or is 
>>there a
>>possibility of discontinuing it?  Are we (Log4j and JCL) prepared to 
>>promote two
>>"competing" logging API's?
>>Moving UGLI outside Apache might help it become more accepted, and it is 
>>a good
>>point that GPL and LGPL projects probably need to depend on a logging API 
>>a  non-Apache license.  But there are also these issues with JCL that need to
>>be worked out.  Do we have official answers for these questions?  If not, 
>>the move is fine for UGLI but maybe not so fine (or maybe just 
>>irrelevant) for
>>Quoting Ceki Gülcü <ceki@qos.ch>:
>>>Greetings to all,
>>>As a result of discussions held on this list and on commons-dev, I
>>>think starting a new project based on or similar to UGLI outside
>>>Apache Logging Services would benefit the wider OS community.
>>>For one, many open source projects based on GPL or LGPL licenses
>>>refuse to adopt log4j because of its Apache license which they deem
>>>incompatible with their own. Whether the Apache Software license is
>>>incompatible with GPL or LGPL is a matter of debate beyond the scope
>>>of this email note. I am merely relaying the position adopted by the
>>>Free Software Foundation, whose opinion carries considerable weight in
>>>the open source world.
>>>Given  UGLI's tiny  size,  I  think we  should  consider changing  its
>>>license to BSD or MIT, in  order to facilitate its wider adoption.  As
>>>an immediate consequence,  we would need to move  UGLI outside the ASF
>>>because the foundation will not host software not Apache licensed.
>>>Just as importantly, I think the lessons learned from the JCL
>>>experience warrant a new approach. A new start should allow for
>>>participants from the commons-dev community as well as Apache Logging
>>>Services to share their knowledge in more open and cordial atmosphere.
>>>Do you think starting a new project, based or similar to UGLI outside
>>>the ASF is an option worth exploring?  Other comments? Opinions?
>To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
>For additional commands, e-mail: log4j-dev-help@logging.apache.org

Ceki Gülcü

   The complete log4j manual: http://www.qos.ch/log4j/

To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org

View raw message