Re: glib event loop thrashing on solaris x86?

[Paul Davis <paul linuxaudiosystems com>]

>hi!
>   this is a somewhat vague question, which i apologize for.  i'm just 
>trying for suggestions of where i should look to find my glitch:
>
>i am running on a dual processor x86 solaris 8 box, and my gtkmm 
>application seems to spend a lot of time thrashing in the glib event 
>loop.  truss reveals lots of this:
>
>ioctl(4, FIONREAD, 0x0804657C)            = 0
>gettimeofday(0x08046584)            = 0
>poll(0x0832D400, 2, 0)                = 0
>gettimeofday(0x080465A4)            = 0
>ioctl(4, FIONREAD, 0x0804657C)            = 0
>gettimeofday(0x08046584)            = 0
>poll(0x0832D400, 2, 0)                = 0
>gettimeofday(0x080465A4)            = 0
>
>top reports my cpu usage at around 60%, which seems very high for the 
>amount of processing that is going on.  has anyone run into a similar 
>situation?  if so, can you suggest things i might be doing wrong?

without timestamps on that output, its meaningless. any X based
program works exactly like your trace shows.

--p