[09:27] <tjaalton> interesting.. tuxracer doesn't work with nvidia on jaunty
[09:27] <tjaalton> *** etracer error: Couldn't initialize video: X11 driver not configured with OpenGL (Resource temporarily unavailable)
[09:28] <tjaalton> sigh, and intel seems to crash on resume even without uxa
[11:13] <mnemo> tjaalton: I current have 2826M RES on compiz.real in top... how can I file a meaningful bug report on that?
[11:13] <mnemo> I mean can I enumerate the heap allocations and get the megabyte count for each allocation site?
[11:14] <mnemo> for example, i would want the stack to the malloc that created these big chunks
[11:14] <mnemo> i know valgrind can do something like this but only for apps you started from valgrind
[11:15] <mnemo> i wonder if there is a way to "attach" valgrind to an existing process and then check the memory or something
[12:35] <tjaalton> mnemo: sorry, I've no experience about that
[18:03] <jcristau> fwiw MAXDEVICES is a build-time constant. (set to 20, apparently)
[19:42] <tjaalton> ok, but not 6
[19:42] <tjaalton> hmm
[19:43] <tjaalton> but does the output device count? in that case "6 seats" might be the default limit (3 * (mouse, keyboard, output))
[19:48] <jcristau> that's input devices
[19:54] <tjaalton> ok..
[19:55] <jcristau> with evdev you get a bit more than that. there's the vck and vcp, and then one for each physical device
[20:00] <tjaalton> ok, so six sounds about right after all?
[20:00] <jcristau> yeah
[20:01] <tjaalton> hehe
[20:02] <tjaalton> it's just hard to argue with the middle-man
[20:02] <tjaalton> or easy, depending on how you see it