Thanks Ceki, I'll look read over the reference you gave and come back if I have any input/ideas.
On Thu, Nov 3, 2011 at 13:37, ceki <[email protected]> wrote: > Hi Chad, > > There was a lengthy discussion about this subject a while back. You > can already pass a map in a log call. For example, > > logger.debug("{}", map); > > the above is not pretty and somewhat lame but it will work. > > There was a lengthy discussion about this issue a while back with > Joern Huxhorn and Ralph Goers. Search for "SLF4J Message interface". > Joern Huxhorn came up with the idea of the Message interface [1] which > allows for a cleaner API. I think the Message interface will be added > into SLF4J some time in the future. > -- > Ceki > http://twitter.com/#!/ceki > > [1] > https://github.com/huxi/slf4j/tree/slf4j-redesign/slf4j-n-api/src/main/java/org/slf4j/n > > On 03/11/2011 2:59 PM, Chad La Joie wrote: >> >> We have a couple cases now where a facility like Apache's HTTPD >> logging mechanism would be nice. Specifically where, when an "event" >> is logged, what is provided is just a map of values and no format >> string. Then the the individual configuring logging configures a >> formatting string that controls what actually gets written out to the >> log file. >> >> From an SLF4J API perspective, I think this probably just means >> trace/debug/info/warn/error methods that take a map instead of a >> format and set of objects. It may even possible to do the actual >> implementation by just creating an MDC, add all the map information, >> calling the current logging methods with an empty format string, and >> then clearing out MDC, or something like that. >> >> Thoughts? >> > _______________________________________________ > slf4j-user mailing list > [email protected] > http://mailman.qos.ch/mailman/listinfo/slf4j-user > -- Chad La Joie www.itumi.biz trusted identities, delivered _______________________________________________ slf4j-user mailing list [email protected] http://mailman.qos.ch/mailman/listinfo/slf4j-user
