[slf4j-dev] svn commit: r158 - in slf4j/trunk/src/java/org/slf4j: .

[elkner at linofee.org]

'ceki at slf4j.org wrote:'
> 
> Author: ceki
> Date: Fri Aug 12 19:58:45 2005
> New Revision: 158

Hi ceki,

can you give us the log.trace() now? As metioned previously, log.trace()
is used by many other projects and I think, the success/adoption of slf4j
really depends on it (magnitudes more than on markers or child loggers). 

E.g., one may think: just for having a trace() I need to wrap e.g. nlog4j
again, so why should I switch ? Just to do all the work again ... 

Also Markers are by far NOT an appropriate replacement:
as mentioned before, markers are a completely different playground
(aka axis or orthogonal to log.xxx) - when I have a problem and enable
debug, I get a StackTrace, which gives me clearly a hint, on which class
I may configure a trace and that's it. A stacktrace gives NO CLUE at all,
what markers I've to use, to get a trace (or do you imply, that if one uses
markers, that he includes the markers in the getMessage() output or even
extend/wraps all exceptions to have a getMarkers() method ?) ...

So please add a trace() and make people happy ;-)

Regards,
jens.