[racket] DrRacket spiking CPU

From: Robby Findler (robby at eecs.northwestern.edu)
Date: Mon Aug 22 23:33:36 EDT 2011

That doesn't sound familiar to me, but you might try sending it a
break and seeing if you get a stack trace that would give more of a

Specifically, start up drracket from a command-line like this:

  racket -W info -l- drracket

and then when it is doing this bad thing, hit control-c. The output
might tell us something, maybe.

Also: does this happen with an empty buffer? does it happen with the
default preferences (ie, move the preferences file out of the way and
then start up drracket).


On Mon, Aug 22, 2011 at 8:52 PM, Don Blaheta <blahetadp at blahedo.org> wrote:
> I've installed DrRacket in an Ubuntu lab at school (with, admittedly,
> some fairly old systems), and while it is briefly responsive when I
> first fire it up, it then spikes to 100% CPU and is totally useless for
> a bit; if I take focus away it eventually registers the keystrokes I
> made, but the moment I give focus back (without even clicking or typing
> anything) it spikes again.  This is occurring under both 5.1.2 and
> 5.1.3.  If I run command-line racket it seems to be just fine.
> Running an strace indicates that during the spike there are a LOT of
> reads from file descriptors linked to unix sockets (6 and 10) that fail
> with an EAGAIN error.  What's especially puzzling to me is that the
> program runs at all if this is happening (and as I said, it does,
> eventually, come through with a response to the few keystrokes I'd
> queued up).
> I'm still triaging, but has anyone seen anything like this?  Anyone have
> ideas where I should look next?
> --
> -=-Don Blaheta-=-=-blahetadp at blahedo.org-=-=-=-<http://www.blahedo.org/>-=-
> "There's safety in numbers.  Large prime numbers."      --John Gilmore
> _________________________________________________
>  For list-related administrative tasks:
>  http://lists.racket-lang.org/listinfo/users

Posted on the users mailing list.