On Jan 5, 2011, at 2:37am, Jukka Zitting wrote:
Hi,
From: Mattmann, Chris A (388J):
I think the intention is *to not do* any logging as you mentioned.
Yep. The last time we discussed this was in 2009 [1], and then the
consensus was to avoid all (or at least most) logging in Tika and
simply leave all upstream parsers to use their own logging mechanism.
The result is that currently we have transitive dependencies to both
commons-logging and slf4j, and I believe some of our parsers also
use java.util.logging. I wouldn't be surprised if we also ended up
with a dependency to log4j at some point.
It's then up to downstream projects to set up their logging
environment in a way that works for all of these libraries. For
example in Jackrabbit we use the slf4j bindings to map all commons-
logging and java.util.logging messages to slf4j, which we then
direct to Logback for writing out to a log file.
My question was motivated by wondering what to do when Tika itself is
"the upstream parser". E.g. if there's some code being added to Tika
that implements a parser (or parser-specific functionality). In these
situations logging becomes more interesting, and harder to avoid
completely.
It sounds like such code could use slf4j, commons-logging or
java.util.logging, but it probably shouldn't add a new dependency on
log4j.
Thanks,
-- Ken
--------------------------
Ken Krugler
+1 530-210-6378
http://bixolabs.com
e l a s t i c w e b m i n i n g
