Re: wm.log time stamp broken?

Endre Stølsvik <[email protected]> Thu, 6 Apr 2006 10:46:29 +0200 (CEST)
Newsgroups gmane.comp.java.webmacro.user
Message-ID <[email protected]>
| Actually, SLF4J doesn't support the trace() method.  From their FAQ:

I know.

jcl (clogging) has it. And have had to map it to log4j's debug. And JBoss 
have extended log4j to have it since beginning of time.

It's a stupid decision to not have it, in my opionion.

In _particular_ for libraries, the trace is much better than debug, since 
debug is more a "holistic" information than trace. Trace could display 
every load/entry of every little piece of information, and maybe even the 
merging process itself.

One shouldn't have to _configure_ one's system to shut up about 
unimportant information, either.

The debug/trace argument boils down to this: There are just _one_ 
"development-level" in the log4j system (and thus slf4j, due to Ceki), 
while there are (in log4j) _four_ for "production-level": fatal, error, 
warn and info. Thus you either get _absolutely every single idiotic 
debug/trace-line_, or _nothing_. This isn't good - it ruins the issue 
where you just need the "most important" contextual debug information - 
and trace would alleviates this considerably for most scenarios. And again 
- a library should really output much _debug_ information at all, sicen 
this is mostly up to the application using it, while it could output a 
good lot of _trace_ information, as that would help if you really had some 
strange problems going.
  In comparison, JSR 47 (JDK 1.4) have three levels that seems aimed at 
"development-levels": fine, finer, finest (in addition to method-entry and 
method-exit), in addition to their severe, warning, info and config 
"production" levels.

Ceki has always been very against the TRACE level, and have refused to add 
it to log4j for many years, although that has always been the one single 
feature that has been requested the most by users and other communities 
(e.g. jcl and Jboss). He was finally overruled by the rest of the 
committers, and log4j 1.2.13 have it now.

But forget this shit: my argument is still valid IMO - WebMacro is in need 
for some evaluation of its logging - just exchange "trace" with "debug" in 
my original post. The creation of a Context is e.g. not an "informational" 
issue. One shouldn't have to set the threshold to _warning_ to get WM to 
stop emitting absolutely non-essential trace/debug information.

Regards,
Endre.


-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642