[Pidgin] #2002: idle reporting doesn't work after coming back from idle

[Pidgin]

#2002: idle reporting doesn't work after coming back from idle
--------------------+-------------------------------------------------------
 Reporter:  zarvok  |        Owner:  seanegan    
     Type:  defect  |       Status:  new         
Milestone:          |    Component:  pidgin (gtk)
  Version:  2.4.1   |   Resolution:              
 Keywords:  idle    |  
--------------------+-------------------------------------------------------

Comment(by darkrain42):

 Hmm, I could have sworn I searched for tickets before I filed a bug. I
 filed #7839, which describes (as best as I can tell) largely this same
 issue -- namely, how to wedge the idle reporting. Copying my bug
 description in here, since I did a little bit of gdb work to diagnose the
 issue.


 It's possible to wedge the purple idle time reporting into never reporting
 me idle again. Also, by my reading of the code, the "mins_before_away"
 setting in the preferences should be grouped with "Report Idle time" in
 the preferences because it is always in effect (if idle time reporting is
 enabled), irrespective of the auto-away settings.

 Anyway, for wedging the idle time reporting:

 Report Idle Time: From last sent message ("purple")
 Change status when idle: Not checked
 Minutes before becoming idle: 1

 Steps to reproduce:

    1. Restart pidgin
    2. Allow 1 minute to pass, so that you go idle.
    3. Send a message to someone (I use myself) to go unidle
    4. Wait 1 more minute -- you will not go idle.

 From putting watches on idle_timer and time_until_next_idle_event , one
 can see that after going idle, the idle timer is removed and is never set
 again (check_idleness_timer is never called).

 I believe this happens because even though purple_idle_touch is called
 when sending an IM, no_away is TRUE. Additionally, the idle time is never
 repolled so, when first going idle, the idle_recheck_interval in
 check_idleness is set to 0 ("don't poll when using purple idle reporting")

 I'd submit a patch to fix it, but the idle code drives me nuts because I
 still haven't figured out what the no_away variable does.

-- 
Ticket URL: <http://developer.pidgin.im/ticket/2002#comment:47>
Pidgin <http://pidgin.im>
Pidgin