Thanks for fixing bug 786.
If you are interested, here are some other inconsistencies I found in enum.spec.
1. The PixelFormat and PixelInternalFormat enums use values from SGIX_icc_texture, but these values are commented out because the extension was incomplete. Suggestion: remove the uses from PixelFormat and PixelInternalFormat. See also http://oss.sgi.com/archives/ogl-sample/2001-05/msg00006.html.
2. The EXT_vertex_attrib_64bit enum uses DOUBLE from VERSION_1_1, whereas it is defined in DataType. Suggestion: change to use it from DataType.
3. The ARB_gpu_shader_5 enum uses MAX_VERTEX_STREAMS from ARB_texture_multisample, but it is actually defined in ARB_transform_feedback3. Suggestion: change to use from ARB_transform_feedback3.
4. NV_fbo_color_attachments uses COLOR_ATTACHMENT0 from NV_draw_buffers, but it comes from VERSION_3_0. Suggest to use COLOR_ATTACHMENT0_NV, which comes from NV_draw_buffers and has the same value. Same for the following 1 through 15. A tricky case, though.
That's all I found.
Khronos has moved to a new XML API Registry and headers generated
from that; see
We don't plan to maintain the old .spec files going forward. Some of
these issues have been fixed in the XML, and some can't be fixed
because it doesn't include the corresponding information (for example,
the XML format doesn't say that an enum is defined by a particular
extension - it defines the enum as a standalone thing and then
every extension which needs it will reference that enum).