[PATCH]: Fix GLX crashes on RISC cpus.

[David Miller]

From: Daniel Stone <daniel at fooishbar.org>
Date: Thu, 27 Dec 2007 09:56:43 +0000

> On Thu, Dec 27, 2007 at 12:34:16AM -0800, David Miller wrote:
> > As another example, half the input drivers in the tree don't even
> > build, and it's been like this for over two years.  There has been
> > some level of forward progress, but the regression is still mostly
> > there.
> 
> http://bugs.freedesktop.org/show_bug.cgi?id=10262 which has been
> RESOLVED/FIXED since August?

Look at how it's like pulling teeth to get someone to fix build
regressions they added, even downloading the drivers to do a test
build seems to be considered a chore.

Also note how many times that bug had to get closed and reopenned
before it really was fixed.  Finally, some of the input drivers
currently don't build:

xf86-input-elographics:

 gcc -DHAVE_CONFIG_H -I. -I. -I.. -g -O2 -I/opt/XORG/include/xorg -I/opt/XORG/include/pixman-1 -I/opt/XORG/include -I../src -MT xf86Elo.lo -MD -MP -MF .deps/xf86Elo.Tpo -c xf86Elo.c  -fPIC -DPIC -o .libs/xf86Elo.o
xf86Elo.c:68:24: error: xf86_ansic.h: No such file or directory
make[2]: *** [xf86Elo.lo] Error 1

xf86-input-magictouch:

clude/pixman-1 -I/opt/XORG/include -I../src -MT xf86MagicTouch.lo -MD -MP -MF .deps/xf86MagicTouch.Tpo -c xf86MagicTouch.c  -fPIC -DPIC -o .libs/xf86MagicTouch.o
xf86MagicTouch.c:25:24: error: xf86_ansic.h: No such file or directory
xf86MagicTouch.c: In function 'xf86MagicAllocate':
xf86MagicTouch.c:1019: warning: assignment makes pointer from integer without a cast
make[2]: *** [xf86MagicTouch.lo] Error 1

>      I'm not holding this up as an example of brilliant development,
>      but it's not a two-year regression, however you paint it.

Sure, it was a one-year regression.

But a known build regression that lasts for a year, sucks.