CVS Wine doesn't build on NVidia system (2004-05-01)

[Saulius Krasuckas]

Am I the only NVidia owner who has this problem?  I did 'make clean' and
'cvs -PAd' and now am getting following:

make[2]: Entering directory `/mnt/antras/usr/src/wine-cvs/wine/dlls/uuid'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/mnt/antras/usr/src/wine-cvs/wine/dlls/uuid'
make[2]: Entering directory `/mnt/antras/usr/src/wine-cvs/wine/dlls/glu32'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/mnt/antras/usr/src/wine-cvs/wine/dlls/glu32'
make[2]: Entering directory `/mnt/antras/usr/src/wine-cvs/wine/dlls/d3d8'
gcc -c -I. -I. -I../../include -I../../include -I/usr/X11R6/include -D__WINESRC__  -D_REENTRANT -fPIC -Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wpointer-arith  -g -O2 -o utils.o utils.c
utils.c: In function `set_tex_op':
utils.c:1605: `GL_MODULATE_ADD_ATI_EXT' undeclared (first use in this function)
utils.c:1605: (Each undeclared identifier is reported only once
utils.c:1605: for each function it appears in.)
make[2]: *** [utils.o] Error 1
make[2]: Leaving directory `/mnt/antras/usr/src/wine-cvs/wine/dlls/d3d8'
make[1]: *** [d3d8] Error 2
make[1]: Leaving directory `/mnt/antras/usr/src/wine-cvs/wine/dlls'
make: *** [dlls] Error 2


Patch of cause is http://cvs.winehq.org/patch.py?id=12484 .  Nothing
changed in my NVidia drivers setup, nor in GL headers on my system AFAIK.

# lsmod | grep nv
nvidia               2123744   6  (autoclean)

# strings /lib/modules/`uname -r`/kernel/drivers/video/nvidia.o | grep id:
nvidia id: NVIDIA Linux x86 NVIDIA Kernel Module  1.0-5336  Wed Jan 14 18:29:26 PST 2004

Help :)

[Lionel Ulmer]

> gcc -c -I. -I. -I../../include -I../../include -I/usr/X11R6/include -D__WINESRC__  -D_REENTRANT -fPIC -Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wpointer-arith  -g -O2 -o utils.o utils.c
> utils.c: In function `set_tex_op':
> utils.c:1605: `GL_MODULATE_ADD_ATI_EXT' undeclared (first use in this function)
> utils.c:1605: (Each undeclared identifier is reported only once
> utils.c:1605: for each function it appears in.)

Well, the driver you use to get OpenGL running has nothing to do with the
way you build it.

Basically, NVIDIA should only ship the GL library (.so), the kernel module
and the XFree module. The rest is 'standard' accross all GL libraries (ie
the set of headers files should be the same for ATI, NVIDIA, DRI, ...).

Anyway, I still think that this is a bug in the code as, from what I can see
in the reference glext.h file (http://oss.sgi.com/projects/ogl-sample/ABI/glext.h),
one has the following constants:

#ifndef GL_ATI_texture_env_combine3
#define GL_MODULATE_ADD_ATI               0x8744
#define GL_MODULATE_SIGNED_ADD_ATI        0x8745
#define GL_MODULATE_SUBTRACT_ATI          0x8746
#endif

So I do not see why Christian used the (horrible :-) ) 'useext' macro here
as it should only be used for the other ARB / EXT texture combination code
(GL_ARB_texture_env_combine / GL_EXT_texture_env_combine) and NOT for the
ATI path.

Before Christian wakes-up, could you just remove the 'useext()' around the
GL_MODULATE_ADD_ATI constant in line 1605 and tell us if it builds ?

Thanks,
       Lionel

-- 
		 Lionel Ulmer - http://www.bbrox.org/

[Christian Costa]

Lionel Ulmer wrote:
> > gcc -c -I. -I. -I../../include -I../../include -I/usr/X11R6/include -D__WINESRC__  -D_REENTRANT -fPIC -Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wpointer-arith  -g -O2 -o utils.o utils.c
> > utils.c: In function `set_tex_op':
> > utils.c:1605: `GL_MODULATE_ADD_ATI_EXT' undeclared (first use in this function)
> > utils.c:1605: (Each undeclared identifier is reported only once
> > utils.c:1605: for each function it appears in.)
>
> Well, the driver you use to get OpenGL running has nothing to do with the
> way you build it.
>
> Basically, NVIDIA should only ship the GL library (.so), the kernel module
> and the XFree module. The rest is 'standard' accross all GL libraries (ie
> the set of headers files should be the same for ATI, NVIDIA, DRI, ...).
>
> Anyway, I still think that this is a bug in the code as, from what I can see
> in the reference glext.h file (http://oss.sgi.com/projects/ogl-sample/ABI/glext.h),
> one has the following constants:
>
> #ifndef GL_ATI_texture_env_combine3
> #define GL_MODULATE_ADD_ATI               0x8744
> #define GL_MODULATE_SIGNED_ADD_ATI        0x8745
> #define GL_MODULATE_SUBTRACT_ATI          0x8746
> #endif
>
> So I do not see why Christian used the (horrible :-) ) 'useext' macro here
> as it should only be used for the other ARB / EXT texture combination code
> (GL_ARB_texture_env_combine / GL_EXT_texture_env_combine) and NOT for the
> ATI path.

I don't see it too. ;-)
> Before Christian wakes-up, could you just remove the 'useext()' around the
> GL_MODULATE_ADD_ATI constant in line 1605 and tell us if it builds ?

Raphael has sent a patch for that.
And, I'm already awake. :-)
> Thanks,
>       Lionel

[Raphael]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 01 June 2004 15:46, Saulius Krasuckas wrote:
> Am I the only NVidia owner who has this problem?  I did 'make clean' and
> 'cvs -PAd' and now am getting following:

<snip>
> utils.c:1605: `GL_MODULATE_ADD_ATI_EXT' undeclared (first use in this
> function) utils.c:1605: (Each undeclared identifier is reported only once
> <snip>
>
> Patch of cause is http://cvs.winehq.org/patch.py?id=12484 .  Nothing
> changed in my NVidia drivers setup, nor in GL headers on my system AFAIK.

<snip>
> Help :)
> Hi,

You can use attached patch 
Christian think that all people sould have ATI or Mesa headers :)

Regards,
Raphael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAvODZp7NA3AmQTU4RAlSJAJ4kfj9jnn4jdVvlVLmg69AWD7+F+wCgkbWm
786R2YpxTd84ZhuJoip48AU=
=sD2U
-----END PGP SIGNATURE-----

[Christian Costa]

Raphael wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Tuesday 01 June 2004 15:46, Saulius Krasuckas wrote:
> > Am I the only NVidia owner who has this problem?  I did 'make clean' and
> > 'cvs -PAd' and now am getting following:
>
> <snip>
> > utils.c:1605: `GL_MODULATE_ADD_ATI_EXT' undeclared (first use in this
> > function) utils.c:1605: (Each undeclared identifier is reported only once
>
> <snip>
> > Patch of cause is http://cvs.winehq.org/patch.py?id=12484 .  Nothing
> > changed in my NVidia drivers setup, nor in GL headers on my system AFAIK.
>
> <snip>
> > Help :)
>
> Hi,
>
> You can use attached patch

Thanks for the patch.
> Christian think that all people sould have ATI or Mesa headers :)

:-)
This has nothing to do with ATI or Mesa headers.
I've mistakenly used the useext macro (curse on me) that do not add any 
suffixe on by box because my headers are OpenGL 1.3 compliants.
For other headers _EXT or _ARB may be added resulting in a compilation 
error.

Christian