[slf4j-user] Fail Silently on NOP implementation

[Phillip Lord]

Ceki Gulcu <ceki at qos.ch> writes:
>> I do acknowledge the bootstrap difficulty this causes for SLF4J, but
>> there are solutions. The issue of library vs. application only comes
>> into question when we discuss the problem with your solution (i.e bind
>> to nop).
>
> OK, admittedly, I am assuming that there is a distinction between
> library vs. application, a distinction although natural to me might not
> always be natural to others.
>
> FYI, I added the following paragraph to
> http://www.slf4j.org/codes.html#StaticLoggerBinder
>
>   If you are packaging an application and you do not care about logging,
>   then placing slf4j-nop.jar on the class path of your application will
>   get rid of this warning message. Note that embedded components such as
>   libraries or frameworks should not declare a dependency on any SLF4J
>   binding but only depend on slf4j-api. When a library declares a
>   compile-time dependency on a SLF4J binding, it imposes that binding on
>   the end-user, thus negating SLF4J's purpose.

This is good, although I would probably change the wording a bit, as it
conflicts with the maven terminology. Refering statically to a SLF4J
binding really would impose the choice (as you say). Adding a slf4j
binding as either a compile *or* runtime scoped dependency would imply a
choice, but not impose it, as in either case it can be over-ridden.


> In any case, thank you for initiating and valiantly pursuing this
> discussion. Such design questions are quite fundamental.

And thanks to you, and everyone else for replying so thoughtfully. I
shall leave the issue (and the list!) at this point.

Phil