seth vidal wrote: >So here's an interesting tidbit. Michael Stenner wrote a version of >logger that was syslog-capable. In fact, I think he still has it. I'd be >curious to know how he implemented it. > >Josko, could you take a look at: >http://linux.duke.edu/projects/mini/logger/ > >and see if it is still there? > > Actually it is. Sad thing is that I actually checked the link referenced in the logger.py distributed with yum and the version in that location is older one, nothing changed. I guess I should've also googled it. Anyway, taking a look at it now. Thanks for info. >If not, I'll take another look at your patch and see what can be done. > >I think syslogging yum output is a good idea. Care to write an epylog >module (http://linux.duke.edu/epylog/) to go along with it? > > That was my intention - was busy until today so I'll give it a try after trying to use newer logger instead of my changes. >Would it be a worthwhile idea to change around the stuff that gets >syslogged so it syslogs more important things and/or a slightly >different layout for easier parsing? > > It may be useful to log also, in case of an upgrade, what was the old version. You could parse that or not later but it's nice to have the info. Other than that probably not. Maybe a verbosity config that would make it possible to log to file and syslog also some warnings typically given by rpm. Format of the messages themselves is already easy enough to parse (epylog deals with *much* worse). Josko P.
