OpenCL 2.0 compute shaders in OpenGL 4.6

Allow capabilities of OpenCL 2.0 in OpenGL compute/OpenCL shaders.

If someone wants to use such functionality in such a way, they then can.
Not everyone might want to rewrite their application to OpenGL NG right away.

Not everybody must rewrite their application right away.

This is the same fallacy that we occasionally see relating to OpenGL 3.x+. Many people using older OpenGL code (immediate mode, glTexEnv calls, etc) behave as though the very existence of OpenGL 3.x+ was an affront.

But if you have a legacy codebase using older OpenGL calls you always have the option of just keeping that legacy codebase. Don’t upgrade it, don’t port it, it will still work. Newer versions of OpenGL prove an option, they’re not mandatory.

Then there is no problem with integrating OpenCL 2.0 in the next minor version release of OpenGL for the situations you describe.

Don’t integrate SPIR(OpenCL 2.0) with OpenGL 4.6.
Is not good feature to integrate with current OpenGL 4.x.
Add OpenCL 2.0 compute shaders without the SPIR(OpenCL) IR in OpenGL 4.6.

Do specify SPIR (OpenCL) only for OpenGL NG/newer.
This feature (SPIR) fits very well with the suggested IR representation feature of OpenGL NG.