Re: [AD] X11 unresponsiveness

[Elias Pschernig <elias@xxxxxxxxxx>]

On Mon, 2004-07-12 at 18:51 +0200, Evert Glebbeek wrote:
> On Monday 12 July 2004 18:38, Elias Pschernig wrote:
> > [Resent, sorry if it appears double now]
> 
> I tried sending messages to the mailing list several times during the 
> day... so far, they've only showed up once, so I think it will be ok ;)
> 
> > The attached patch makes all X11 programs (even expal, or the demo with
> > page flipping) completely responsive. And in the demo, I see no change
> > in FPS compared to before.
> 
> Sweet! I tested it locally and it works wonderful here too.
> 
> > There's a call to XLOCK protecting the input code, so there is no
> > problem with thread safety after removing that if. My guess is, the
> > removed line should maybe prohibit re-entrancy of the event handler
> > itself, so a lot of input wouldn't cause lots of XLOCKs. If that's true,
> > and it makes sense, we could add code for that again, but it should only
> > block on the function itself then, not on all XLOCKs. I'm not sure at 
> > all though that's what it was for though.
> 
> Have you checked with the sigalrm version instead of the pthreads version 
> of the library? That may make a difference, but I'm not really sure.
> 

No, but I will test before committing. AFAICS the sigalarm driver should
have the same problem, since it also calls _xwin_handle_input - just
from the signal handler and not the thread. But _xwin_handle_input also
just returns there, not processing input.

And I remember, Vincent complaint the most about bad input response, so
I guess it was even worse with the signals driver, and should all work
now :)

-- 
Elias Pschernig