I often find myself needing the full logs of a TP component (generally a CM). I usually manually restart the CM with, say, GABBLE_DEBUG=all GABBLE_LOGFILE=/tmp/gabble.log and then pray to be able to reproduce the issue I noticed. As a Telepathy developer I'm interested in most CMs and may find myself needing logs from basically as CM. I could tweak my session add define a *_DEBUG and *_LOGFILE variable for any CM but that's kinda lame and doesn't really scale. So I'm proposing to add a couple of generic TP variables: - TP_DEBUG : if the CM specific *_DEBUG variable is not present we use this one - TP_LOGDIR: if the CM specific *_LOGFILE variable is not present, we log the debug output to TP_LOGDIR/$cm-name.log That way I'd just have to define "TP_DEBUG=all TP_LOGDIR=/tmp/telepathy-logs/" and be done. Note that would also make debugging easier for bug reporters as they wouldn't have to play with a billion env variables any more. So, crack or not?
A global TP_LOGFILE means every app will output at the same place and messages will be mixed. Maybe TP_LOGDIR and each component will create there its own file? also only TP_DEBUG=all only really make sense, since if you want specific parts, they are different in each app. Maybe also TP_PERSIST ?
(In reply to comment #1) > A global TP_LOGFILE means every app will output at the same place and messages > will be mixed. Maybe TP_LOGDIR and each component will create there its own > file? Forget this, I just can't read a friday...
(In reply to comment #1) > A global TP_LOGFILE means every app will output at the same place and messages > will be mixed. Maybe TP_LOGDIR and each component will create there its own > file? Very good idea, it's almost like I suggested exactly that. :p > also only TP_DEBUG=all only really make sense, since if you want specific > parts, they are different in each app. Yeah, TP_DEBUG=1 maybe? > Maybe also TP_PERSIST ? This one seems a bit less useful to me (you generally use it to debug a specific component) but why not.
Sorry, what's wrong with the Debug interface?
(In reply to comment #4) > Sorry, what's wrong with the Debug interface? It's not exactly the same use case. The debug interface and its UI is useful when you know where to look and you notice the issue right away (as it's a ring buffer, the logs disappear after a while). But what I want is to be able to log *everything*, including extra stuffs for lower layer. Typically, I'd probably add WOCKY_DEBUG=all to see the raw xmpp data, which are not displayed in empathy-debugger.
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/telepathy/telepathy-spec/issues/164.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.