[Xcb] Problem with xcb-test

[Anni Kotilainen]

My processor is Athlon 9700 pro, distro is up to date Gentoo (2004.1). I 
can't give the xdpyinfo (at least right now), because at the moment I'm 
in the process of reinstalling my Gentoo due to unrelated reasons. The 
DISPLAY environmental variable was ":0.0", I actually had to set it just 
to stop getting the error that xcb couldn't connect to the server - it was 
empty before. I did try rendertest too but got no window.

In case it's relevant, I also had to fight Gentoo over autoconf and 
automake a bit because it insisted on using versions that weren't recent 
enough. After that I had to run autogen.sh twice (getting a lot of 
warnings over missing "'s or something on the way), but in the end it 
seemed to install okay.

I hope this helps, I'll be able to play with it more once my OS is up and 
running again.

--
Anni Kotilainen

On Mon, 14 Jun 2004, Jamey Sharp wrote:

> I can at least tell you that only the first two messages are
> interesting, and the rest are because the first failed. It appears to
> hang because it's looping, waiting for events: it's supposed to do that.
> The only way to end xcb-test under any circumstances is to kill it with
> a signal like ctrl-c.
> 
> These things would help: your processor type and OS variant
> (distribution and version, etc.); the output of the xdpyinfo shipped
> with your XFree86 bits; the value of your DISPLAY environment variable;
> and whether rendertest successfully puts a window on the screen.
> 
> The second error message below seems odd. That says that a GetGeometry
> request on the root window failed. xdpyinfo will help diagnose that, a
> little: maybe XCB and Xlib will disagree about which window is the root
> window, giving clues to a bug in XCB. (Unfortunately the XCB version of
> xdpyinfo doesn't output the root window ID, but that info is in the
> output you provided below.) I guess the BadWindow error would be due to
> the wrong root window being provided.
> 
> It's also quite possible that xcb-test itself is buggy. (I know of one
> bug that you don't seem to have run into; you'd see a BadMatch on seqnum
> 3 if you hit that.) I have a lot more faith in the XCB library than in
> any of the demos, mostly because the demos rely on the semantics of X,
> while the library only has to get the protocol right. I'm not very good
> at understanding X semantics yet... The rendertest application provides
> a good datapoint because it was contributed by Andy Howe and so doesn't
> have exactly the same semantic confusions as my code.
> 
> Thanks for trying this code out.
> 
> On Mon, Jun 14, 2004 at 09:06:11AM -0700, Bart Massey wrote:
> > I doubt it has to do with your server version.  Sounds more
> > like an outright XCB bug to me.  Platform, environment, etc?
> > Perhaps Jamey will have some insight...
> > 
> > In message <Pine.LNX.4.44.0406141130050.23635-100000 at melkinpaasi.cs.Helsinki.FI> you wrote:
> > > I'm trying to run xcb-test but I can't seem to get over the following
> > > error:
> > >  
> > > $ /usr/local/bin/xcb-test
> > > Error BadWindow on seqnum 3 (CreateWindow).
> > > Failed to get geometry for window 0x7d4029a4.
> > > Failed to get geometry for window 0x02400000.
> > > Error BadWindow on seqnum 6 (ChangeProperty).
> > > Error BadWindow on seqnum 7 (ChangeProperty).
> > > Error BadWindow on seqnum 8 (MapWindow).
> > > Error BadDrawable on seqnum 10 (GetGeometry).
> > > Error BadDrawable on seqnum 11 (GetGeometry).
> > > Failed to query tree for window 0x02400000.
> > > Error BadWindow on seqnum 12 (QueryTree).
> > > Failed to get attributes for window 0x02400000.
> > > Error BadWindow on seqnum 9 (GetWindowAttributes).
> > >  
> > > After that it just hangs until I abort it.
> > >  
> > > I'm currently using XFree 4.3.0, would it help if I installed XOrg
> > > instead?
> > > -- 
> > > Anni Kotilainen
>