Allowed types for specialization constants

The GLSL extension `GL_KHR_vulkan_glsl` says that specialization constants can only be used with scalar `int`, `float` or `bool`.

On the other hand, glslang compiles shaders with specialization constants whose type is any unsigned/signed integer variant of any size, or double.

Which types are actually allowed?

Start by reading the GL_KHR_vulkan_glsl specialization-constant rule and comparing it with the types glslang currently compiles. Determine which behavior is intended, then identify whether the resolution requires aligning the compiler or clarifying the extension; done means the allowed types and resulting behavior are unambiguous.

Written by the indexing model from the issue text.