Khronos Public Bugzilla
Bug 277 - ProgramParameteri function is invalid
: ProgramParameteri function is invalid
Status: RESOLVED FIXED
Product: OpenGL
Registry
: 3.2
: All All
: P3 normal
: ---
Assigned To: Jon Leech
:
:
:
:
:
  Show dependency treegraph
 
Reported: 2010-03-14 17:44 CDT by Stephen A.
Modified: 2010-08-09 05:06 CDT (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Stephen A. 2010-03-14 17:44:00 CDT
Both gl.spec[1] and glext.h[2] list ProgramParameteri as a core 3.2 function.
This is incorrect: this function is part of ARB_geometry_shader4[3], which is
has not been promoted to core (its functionality has been folded into GLSL 1.5
geometry shaders instead) and the PDF spec holds no mention of it.

To be more specific, the .spec and .h files list 3 version of this function:
ProgramParameteri
ProgramParameteriARB
ProgramParameteriEXT

The last two are valid (and exposed in drivers(*)) but the first is invalid
(and not exposed in drivers). For this reason, ProgramParameteri should
probably be removed from the .spec and .h files.

Impact of the issue is minor, but it does confuse developers and introduces a
certain amount of support overhead[4][5].

(*) tested on Nvidia, status on AMD is as of yet unknown.

[1]http://www.opengl.org/registry/api/gl.spec
[2]http://www.opengl.org/registry/api/glext.h
[3]http://www.opengl.org/registry/specs/ARB/geometry_shader4.txt
[4]http://www.opentk.com/node/1637
[5]http://www.opengl.org/discussion_boards/ubbthreads.php?ubb=showflat&Number=273944
Comment 1 Tim Scheiter 2010-03-24 15:05:49 CDT
glProgramParameteri in OpenGL 3.2 Core Profile is valid on ATI 5700 serie, but
invalid on nvidia (both tested)
Comment 2 Jon Leech 2010-08-09 05:06:18 CDT
You are correct that ProgramParameteri is not a 3.2 core function; however
with the just-published updated to the .spec and headers for GL 4.1,
ProgramParameteri *does* exist in core 4.1 / ARB_get_program_binary, so
both ARB and non-ARB functions properly appear in the headers now. Thanks
for the bug report!