[Xquartz-dev] 1.5 branch notes (was Re: 2.4.0_beta1)

[Jeremy Huddleston]

BTW, since pushing out 2.3.3, I've been focusing back on the 1.5 and  
1.6 branches of the server for 2.4.0.  The fullscreen crash mentioned  
below has been fixed in git.  You can test it out for yourself by  
installing 2.3.3.2 then dropping in the updated X11.bin from:
http://static.macosforge.org/xquartz/downloads/X11.bin-1.5.3-apple7.bz2

detailed instructions for the daring:
http://xquartz.macosforge.org/trac/wiki/Releases#xorg-server

for those with more patience, I hope to have a 2.4.0_beta2 available  
by the end of the week.  I will be away for approximately 10 days  
starting on the 7th, so I want to get some new toys out there for  
testing =)

The only major regression I am aware of with 1.5 has to do with  
keyboard mapping.  You'll probably run into problems if you have a  
~/.Xmodmap file present.  Hopefully I'll have a fix for that in time  
for the 2.4.0_beta2 release.

Thanks,
Jeremy

On Feb 23, 2009, at 15:40, William Davis wrote:

> crashed right after cmd-opt-a
>
> WD
> On Feb 23, 2009, at 1:52 PM, Jeremy Huddleston wrote:
>
>> Hi everyone.  I'm proud to announce that 2.4.0_beta1 of XQuartz is  
>> ready.  With this release, we've jumped to the 1.6 server branch,  
>> updated libX11, libxcb, pixman and various others from their rc  
>> versions to final release versions, and integrated massive  
>> improvements to our GLX support thanks to George's hard work.
>>
>> A list of all changes since 2.3.2.1 can be seen in our ChangeLog:
>> http://xquartz.macosforge.org/trac/wiki/ChangeLog
>>
>> And here's the dmg:
>> http://static.macosforge.org/xquartz/downloads/X11-2.4.0_beta1.dmg
>>
>> For those that are interested, the source code for our libGL is  
>> available in svn (see our websvn interface at http://xquartz.macosforge.org/trac/changeset) 
>>  or from http://xquartz.macosforge.org/downloads/src.
>>
>> The X11 server is seeing massive changes to its input handling to  
>> support multiple devices and multitouch.  As such, the 1.6 server  
>> branch has seen a fair number of changes as a stepping stone toward  
>> that end.  Unfortunately, all the kinks have not been worked out of  
>> input handling, so most of the known regressions are related to  
>> input:
>>
>> * capslock doesn't work
>> * a possible race condition can cause xmodmap to ruin your key  
>> mappings
>> * workaround:  move your ~/.Xmodmap file somewhere else and load it  
>> after X has already started
>> * sometimes X may stop responding to your mouse
>> * George has seen this, but I haven't.  If someone can help with  
>> reproducibility, I'd be grateful
>>
>> George also has some note regarding the GLX improvement (we now  
>> support GLX-1.4):
>>
>> AppleSGLX Release Notes
>>
>> o OpenGL Support
>>
>> AppleSGLX supports the same version of OpenGL as Leopard (OpenGL  
>> 2.1).
>> Many extensions from the OpenGL framework are now builtin.
>>
>> This adds support for GLSL, and a variety of other features.
>>
>> o Thread Support
>>
>> Thread support has been improved since the libGL in XQuartz 2.3.2.1.
>>
>> o GLX 1.4 Support
>>
>> The GLX 1.3 and 1.4 functions should all work with a few exceptions
>> as outlined in this document.
>>
>> o glXMakeContextCurrent (a GLX 1.3 feature)
>>
>> glXMakeContextCurrent should work with the readable drawable.  The
>> OpenGL functions: glReadPixels, glCopyPixels, and glCopyColorTable,
>> should use the readable drawable if it's different than the rendering
>> drawable.
>>
>> o glXGetProcAddress (a GLX 1.4 feature and ARB extension)
>>
>> glXGetProcAddress should work and allow getting the address of any
>> extension functions you may need from the X11 libGL, or OpenGL  
>> framework
>> libGL.  Previous versions of the X11 libGL didn't allow getting the  
>> newer
>> OpenGL framework addresses.
>>
>> o GLXPixmaps
>>
>> New support for GLXPixmaps works well with mixed X11 and OpenGL  
>> drawing
>> operations.  You can create them using glXCreateGLXPixmap or
>> glXCreatePixmap.
>>
>> o GLXPbuffers
>>
>> Support for GLXPbuffers has been added.  These are drawables that are
>> not possible to render to with X11, which is allowed by the spec.
>> A GLXPbuffer will never generate a clobber event, however
>> glXSelectEvent and glXGetSelectedEvent should operate normally.
>>
>> Clobber events are not generated due to low-level architectural
>> differences.  The contents of your pbuffers will not be clobbered.
>>
>> o Shared Contexts
>>
>> Due to basic low-level architectural differences the usage of shared
>> contexts requires a similar visual or GLXFBConfig be used in the
>> creation of a shared context.  It's best if you specify the same
>> visual.  This is due to a CGL design difference, and not something
>> that is easily worked around.
>>
>> o Indirect
>>
>> The X server supports indirect fairly well, so OpenGL applications
>> can be run remotely and displayed by XQuartz.  This means you can run
>> applications from a remote host on an XQuartz X server.
>>
>> AppleSGLX does not support indirect rendering.  Any indirect context
>> created will appear to glXIsDirect as an indirect context, but it
>> does not actually support indirect rendering to a remote X server.
>>
>> AppleSGLX supports GLXPixmaps and GLXPbuffers with direct and  
>> indirect
>> contexts, though they are all direct contexts by definition (see  
>> above).
>>
>>
>>
>> _______________________________________________
>> Xquartz-dev mailing list
>> Xquartz-dev at lists.macosforge.org
>> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev
>
>
>
> William Davis
> frstanATbellsouthDOTnet
> Mac OS X.5.6 Darwin 9.5.0
> XQuartz 2.3.2 (xorg-server 1.4.2-apple31)
> Mac Mini Intel Duo @ 1.86 GHz
>
> Mundus vult decepi, ego non
>
> _______________________________________________
> Xquartz-dev mailing list
> Xquartz-dev at lists.macosforge.org
> http://lists.macosforge.org/mailman/listinfo.cgi/xquartz-dev