Re: [dev] dwm does something bad to X.org 1.7.6

On Tue, Apr 13, 2010 at 11:58:58PM +0100, Anselm R Garbe wrote:
> On 13 April 2010 23:29, Benoit T <benoit.triquet_AT_gmail.com> wrote:
> > On Tue, Apr 13, 2010 at 03:55:46PM +0600, mikhail maluyk wrote:
> >> I have the same problem. After latest update on development version of
> >> debian, Xorg behaviour is insane. I'm not sure dwm is causing the trouble
> >
> > insane how? cpu consomption? event lagging? those are what i saw.
> >
> >> though.
> >
> > neither am i, i have no evidence of that, all i know is that dwm is
> > triggering something in xserver-xorg, that wmii seems to as well and
> > the other WMs i tried don't, including awesome (which has code from dwm)
> >
> > at any rate either dwm is doing unjustified amounts of requests - i
> > don't believe so - or xserver-xorg is compelled into doing something
> > stupid and that would be an xorg bug plain and simple, no matter what
> > dwm does exactly.
>
> Can you use hg tip dwm and redirect its output to some file and send
> me the output, perhaps that gives us an idea what is going on in this
> new xorg server.

I always use hg tip (as long as i don't see anything frightening in hg
incoming of course) plus my patch queue - nothing spectacular and nothing
that I believe could cause breakage. Anyway as mentioned in my first
message, I tested a pristine hg tip build and observed the phenomenon as
well.

When I discovered that tip was logging multi-megabytes to
.xsession-errors, I disabled the trace. I did not care to leave it
enabled because I did not see anything interesting (with my limited
knowledge of X) in that log of X events. As there are no timestamps, I
did nothing to compute their rate of arrival.

I'll reenable the trace and send you a log. I'd be glad if you come up
with things to try, because i have run out of ideas. I'll summarize my
most recent observations anyway, will try to get that done tomorrow
evening.

cheers

-- 
Benoit Triquet <benoit.triquet at gmail.com>
 .''`.
: :' :      We are debian.org. Lower your prices, surrender your code.
`. `'       We will add your hardware and software distinctiveness to
  `-        our own. Resistance is futile.