On Thu, Jul 17, 2008 at 11:41 AM, Jeremy Haile <[EMAIL PROTECTED]> wrote:
> I don't have a strong preference on JCL vs SLF4J, as I haven't personally > experienced many CL issues using JCL. > > But I'm curious, if SLF4J is good enough for large open-source projects > like Spring (OSGI), Hibernate, and MINA to standardize on, why is it not > good enough for JSecurity? If there isn't a _need_ for it, why should we worry about it? If you deploy jcl-over-slf4j.jar with jsecurity instead of commons-logging.jar, you get the SLF4J CL behavior. It is a deployment-time decision - not something our code must force... > To mitigate the size of the download, you could always distribute a > jsecurity.zip and jsecurity-with-dependencies.zip just like Spring does. Why even worry about that while we still have beanutils as a dependency and commons-logging works fine, or if there are CL issues, jcl-over-slf4J.jar works fine. I don't think there is any debate over what is easier to deploy. Until an end-user request this and determines jcl-over-slf4j.jar is not sufficient, why spend more time on this? What is the *need* here? As Craig said, substituting hard dependency on a logging API for another doesn't seem worthwhile, especially when the end-user can determine the behavior they want at deployment time. And in keeping with previous comments, there might be more effort to document how to set up logging correctly in JSecurity with SLF4J: "Include jsecurity.jar, slf4j-api.jar, and then - go find the appropriate binding implementation for your logging subsystem, include that. Oh, and because commons-beanutils.jar uses JCL, you'll have to include jcl-over-slfj4.jar too". That seems like one of the major points people were trying to avoid with my solution - potential user confusion and us having to support it. Doesn't that argument swing both ways? On Jul 17, 2008, at 11:34 AM, Les Hazlewood wrote: Ah, I see what he/you were saying now, and yes, I agree with that. My POV > is that it is not worth arguing about any more until a *need* surfaces. > > On Thu, Jul 17, 2008 at 11:33 AM, Emmanuel Lecharny <[EMAIL PROTECTED]> > wrote: > > Les Hazlewood wrote: >> >> Maybe you didn't read the email thoroughly: >>> >>> >>> Alan just said that when you wrote : >> >> "I will be happy to re-visit this issue if end-users start to complain >> about JCL or express a definite *need* for SLF4J..." >> >> this is your POV, not the community POV. If we decide that we need to move >> to SLF4J, as a community, then we will do it, even if users don't complain >> about CL issues. If we need a formal vote about it, then let's vote. >> >> That being said, reverting the code to its previous state before injecting >> it to Apache was the best thing to do, IMO. >> >> Thanks ! >> >> >> -- >> -- >> cordialement, regards, >> Emmanuel Lécharny >> www.iktek.com >> directory.apache.org >> >> >> >>
