Open Bug 514886 Opened 10 years ago
Consolidate perodic timers and sends
We currently have two intervals - AWAY_TIMEOUT (1m) and NOTIFY_TIMEOUT (5m) - and a number of less synchronised background operations - ALIVECHECK (sent when notify list is empty) and LAGTIMER (send when the server pings us). I think we can consolidate these a bit. The main issue I have with these is how unreliable LAGTIMER is - it only happens when the server pings *us*, which is usually only when we've been idle for 90s or so. What I'd like to do is have a single 1m interval cover all of these: - Send the LAGTIMER on every network (do it first, too). - Send the next away check for each network who's autoAwayPeriod has expired. - Send the notify checks for each network who's notifyPeriod (new pref, 5m default) has expired. As a little side note, an idle connection would actually use less bandwidth (I forget exactly, but I did calculate is - it's not a trivial saving) with the new arrangement. The important bit, however, is merging the functionality of ALIVECHECK and LAGCHECK, and making LAGCHECK more useful/reliable.
You need to log in before you can comment on or make changes to this bug.