UbuntuIRC / 2008 /04 /18 /#upstart.txt
niansa
Initial commit
4aa5fce
[00:01] <Keybuk> http://people.ubuntu.com/~scott/dbus-1.1.20-exit.patch
[00:01] <Keybuk> http://people.ubuntu.com/~scott/dbus-1.1.20-bus-addresses-shutdown.patch
[00:01] <Keybuk> http://people.ubuntu.com/~scott/dbus-1.1.20-shared-connections-shutdown.patch
[00:06] <ion_> I see you use vim as well. :-)
[00:08] <Keybuk> vim?
[00:08] <Keybuk> oh, heh
[00:08] <Keybuk> I'm strange
[00:08] <Keybuk> I usually have both vim and emacs open at any one time
[00:10] <ion_> Don't they try to kill each other in the style of the animation "flash animation fights back"? :-)
[00:47] <Keybuk> meh
[00:47] <Keybuk> somewhere along the line, my test case failure for the first patch got lost
[00:47] <Keybuk> so now it fails by just exiting mid-tests
[00:52] <Keybuk> Testing nih_dbus_bus()
[00:52] <Keybuk> ...with session bus
[00:52] <Keybuk> ...with system bus
[00:52] <Keybuk> ...with disconnection before registration
[00:52] <Keybuk> BAD: unexpected exit(), unpatched D-Bus?
[00:52] <Keybuk> at tests/test_dbus.c:369 (test_bus).
[00:52] <Keybuk> better
[02:04] <Keybuk> http://bazaar.launchpad.net/~keybuk/libnih/trunk/revision/scott%40netsplit.com-20080418010152-l1gnavvxqqej62jt?start_revid=scott%40netsplit.com-20080418010152-l1gnavvxqqej62jt
[02:05] <Keybuk> ^ that's the "glue" bit of the D-Bus API committed
[11:51] <Keybuk> grr
[11:51] <Keybuk> D-Bus is winning again today
[11:53] <Keybuk> I don't _quite_ understand what these guids are for yet
[13:38] <Keybuk> and still have icky valgrind traces from it :p
[17:01] <Gadi> hi, all. Is there a way to specify in the event.d file what vt to output to (ie, chvt 3 and exxecutes some program)?
[17:01] <Gadi> if I do "console output" and then chvt in the script that gets executed, the output still goes to vt1
[17:10] <Keybuk> Gadi: use a program like openvt
[17:21] <Gadi> Keybuk: I actually found a way to get what I want by suppressing all rcS/rc2 console msgs with "console none" - I love oneliners
[17:21] <Gadi> :)
[18:25] <Keybuk> :-)(
[18:26] <Keybuk> 0.5 jobs can now have
[18:26] <Keybuk> oom killer
[18:26] <Keybuk> err
[18:26] <Keybuk> oom never
[18:26] <Keybuk> in their definition
[18:30] <ion_> Great
[18:31] <Keybuk> pls anything from oom -16 (almost never kill)
[18:31] <Keybuk> thru to oom +15 (ooh! ooh! kill me! kill me!)
[18:32] <ion_> How about setting niceness and ioniceness in the job definition as well? (I can't remember whether any of this is implemented already.)
[18:32] <Keybuk> nice is in there
[18:32] <Keybuk> ionice isn't though, how do you set that?
[18:35] <ion_> Well, there’s ionice(1), but i don’t know what it calls.
[18:35] <Keybuk> ioprio_set
[18:36] <ion_> -c The scheduling class. 1 for real time, 2 for best-effort, 3 for idle.
[18:36] <ion_> -n The scheduling class data. This defines the class data, if the class accepts an argument. For real time and best-effort, 0-7 is valid data.
[18:36] <ion_> I was thinking of equivalent values in the job description,
[18:37] <Keybuk> ionice realtime 4
[18:37] <ion_> Yeah
[18:40] <Keybuk> of course, there's no prototype for that one yet
[18:40] <Keybuk> I'll leave that one for now, can you file a wishlist bug to remind me to get to it later?
[18:41] <ion_> Will do.
[19:17] <Keybuk> http://rafb.net/p/jkcW3x33.html
[19:17] <Keybuk> ugh
[19:25] <Keybuk> heh
[19:25] <Keybuk> dbus-daemon does it too
[19:26] <Keybuk> so I'm going to stop caring :p
[19:28] <ion_> :-)
[19:34] <Keybuk> I have better things to do than valgrind dbus-daemon and fix their bugs as well as my own :p
[19:34] <ion_> Heh
[20:28] <elventear> Is there some source where I can look at all the events that upstart recognizes natively?
[20:30] <Keybuk> elventear: events.h
[20:30] <Keybuk> it's not very many
[20:31] <elventear> thx
[20:55] * Keybuk finishes fixing D-Bus anyway
[21:02] <Keybuk> (adding another patch to the list)
[21:08] <Keybuk> in other news, I passed PPL Meteorology today \o/
[21:09] <ion_> Congrats