32 really isn't a valid depth though.  Maybe we should just make that do the same thing as x24+32 instead.

(In reply to comment #1)
> 32 really isn't a valid depth though.  Maybe we should just make that do the
> same thing as x24+32 instead.
> 
or just change the manpage to not give a non-working example :)

Besides the manpage bug, setting 24 bits even with -pixdepths 32 does not work (says only 'Xvfb failed to start') and my images with alpha I am creating save with no alpha channel.
I test with a python program called 'pyglet' that is a wrapper around OpenGL.


This is on a headless EC2 Ubuntu 9.04 box with Xvfb version 2:1.6.0-0ubuntu14 which I access with ssh and the '-x' option.
Here is my test: (using the Python 2.6.2 interactive prompt)
xvfb-run --server-args="-screen 0 1024x768x24 -pixdepths 32" /usr/bin/python-2.6

>>> import pyglet
>>> import pyglet.gl as gl
>>> window = pyglet.window.Window()
>>> config = pyglet.gl.Config()
>>> current_config = pyglet.window.Screen.get_best_config(window.screen)
>>> current_config
XlibCanvasConfig10([('double_buffer', 0),
 ('stereo', 0),
 ('buffer_size', 24),
 ('aux_buffers', 0),
 ('sample_buffers', 0),
 ('samples', 0),
 ('red_size', 8),
 ('green_size', 8),
 ('blue_size', 8),
 ('alpha_size', 0),
 ('depth_size', 0),
 ('stencil_size', 0),
 ('accum_red_size', 0),
 ('accum_green_size', 0),
 ('accum_blue_size', 0),
 ('accum_alpha_size', 0),
 ('major_version', None),
 ('minor_version', None),
 ('forward_compatible', None),
 ('debug', None)])
>>> config.alpha_size = 8
>>> config.buffer_size = 32
>>> window = pyglet.window.Window(config=config)
>>> pyglet.window.Screen.get_best_config(window.screen)
XlibCanvasConfig10([('double_buffer', 0),
 ('stereo', 0),
 ('buffer_size', 24),
 ('aux_buffers', 0),
 ('sample_buffers', 0),
 ('samples', 0),
 ('red_size', 8),
 ('green_size', 8),
 ('blue_size', 8),
 ('alpha_size', 0),
 ('depth_size', 0),
 ('stencil_size', 0),
 ('accum_red_size', 0),
 ('accum_green_size', 0),
 ('accum_blue_size', 0),
 ('accum_alpha_size', 0),
 ('major_version', None),
 ('minor_version', None),
 ('forward_compatible', None),
 ('debug', None)])

So as you can see, even explicitly setting the depth to 32 does not work.

I put the wrong command line.  In fact specifying '24+32' for the depth does not work either.

Is this still an issue?  Does the problem exist with Xfake or with the null 
video driver and Xorg DDX?

I tried your example, but python and I don't get along and I was unable to 
reproduce.  If you can provide a reduced teste case without python I'd be 
appreciative.

> --- Comment #5 from Jeremy Huddleston <jeremyhu@freedesktop.org> 2011-10-15 18:48:30 PDT ---
> Is this still an issue?  Does the problem exist with Xfake or with the null 
> video driver and Xorg DDX?
> 
the Xvfb manpage still has a broken example of depth 32, yes.

I stumbled upon this today. If 32 isn't a valid depth, then xvfb should exit with an appropriate error message. By the way, why isn't 32 valid? (I don't have much experience on those details yet)

(In reply to Maik Riechert from comment #7)
> I stumbled upon this today. If 32 isn't a valid depth, then xvfb should exit
> with an appropriate error message. By the way, why isn't 32 valid? (I don't
> have much experience on those details yet)

32 isn't a valid depth, in X11's opinion, because there is no pixel format where all 32 bits are used for r/g/b. Depth 24 is valid, and it will use 32 bits per pixel, but depth 32 is not a real thing.

Manual fixed now though:

commit d2d664df974ac5a55d5819f0379fcdac05d22fa3
Author: Adam Jackson <ajax@redhat.com>
Date:   Mon Apr 23 17:21:09 2018 -0400

    vfb: Fix man page in re depth
    
    32 is not a valid depth, and the default is now 24 not 8.
    
    Signed-off-by: Adam Jackson <ajax@redhat.com>
    Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>