Bug#627007: Severe "Processor Leak" when Pointer is on GTK/GDK Window

[Elliott Mitchell]

>From: Josselin Mouette <joss at debian.org>
> Le mardi 17 mai 2011 ? 16:04 -0700, Elliott Mitchell a ?crit : 
> > Not much needed to reproduce. Start one of the many afflicted programs,
> > whenever you move the pointer into its window and see the processor usage
> > hit 100% whenever you move the pointer out of the window processor usage
> > will return to idle.
> 
> Guess what. I have hundreds of GTK-based applications installed and none
> of them exhibits the behavior you describe.
> 
> > I'm guessing some configuration setting(s) might be involved, but I've
> > only got wild guesses as to what combination is involved. The extra bits
> > that come to mind are, I've currently got metacity 1:2.30.1-3 as window
> > manager and have focus follows mouse set. There are also two monitors
> > attached to this system. I'll also theorize this may be effecting a fair
> > number of systems, but it likely isn't as noticable on multiprocessor
> > systems.
> 
> I?m guessing this only affects your system since this is the first time
> I ever heard of this bug.
> 
> What is your graphics hardware and which driver are you using?

I'm rather doubtful the graphics driver would effect this as no visual
glitches of any sort are present. In that unlikely event, nVIDIA hardware
and using the closed-source driver.

OTOH, I may have discounted the involvement of the window manager
(metacity) too early. The WM could be doing something odd/unusual that
tweaks GTK in the wrong way.


-- 
(\___(\___(\______          --=> 8-) EHM <=--          ______/)___/)___/)
 \BS (    |         EHeM at gremlin.m5p.com PGP F6B23DE0         |    )   /
  \_CS\   |  _____  -O #include <stddisclaimer.h> O-   _____  |   /  _/
2477\___\_|_/DC21 03A0 5D61 985B <-PGP-> F2BE 6526 ABD2 F6B2\_|_/___/3DE0