Check all gl calls manually. Some of them are missing constants - e.g. glTexEnvi values.
Can we do anything about return values? Only seem to be used for checking the framebuffer status.
Can we do anything about the GL_TEXTURE0 + n pattern?
Related to the above: ternaries - i.e. blah ? GL_XXX : GL_YYY
What is the mysterious glTexImage2D call with an internalformat of 4 - is that a bug?
The code for inserting extra GL constants in order seems to be broken.
Use upper case for the hex literals we create in the GL class.
I guess it's a moot point now, but something else I wondered was whether we need to worry about the group->enums mapping at all. I think we do though.
Some of the extension methods in GlConstantTransformer might be useful more generally, or we might even want to move them to GlRegistry. (I think it'd be somewhat nice if GlRegistry didn't know about JavaParser types though.)
Should we use DI to pass the GlRegistry to the GlConstantTransformer?
Remaining items:
* [x] Check all gl calls manually. Some of them are missing constants - e.g. glTexEnvi values.
* [x] Can we do anything about return values? Only seem to be used for checking the framebuffer status.
* [x] Can we do anything about the `GL_TEXTURE0 + n` pattern?
* [x] Related to the above: ternaries - i.e. `blah ? GL_XXX : GL_YYY`
* [ ] What is the mysterious `glTexImage2D` call with an `internalformat` of 4 - is that a bug?
* [x] The code for inserting extra GL constants in order seems to be broken.
* [x] Use upper case for the hex literals we create in the GL class.
I guess it's a moot point now, but something else I wondered was whether we need to worry about the group->enums mapping at all. I think we do though.
Some of the extension methods in GlConstantTransformer might be useful more generally, or we might even want to move them to GlRegistry. (I think it'd be somewhat nice if GlRegistry didn't know about JavaParser types though.)
Should we use DI to pass the GlRegistry to the GlConstantTransformer?
The transformer currenty adds some constants with silly vendor suffixes, like _SGIS. We should check these exhaustively and fix the groups in the registry where appropriate.
* [x] The transformer currenty adds some constants with silly vendor suffixes, like `_SGIS`. We should check these exhaustively and fix the groups in the registry where appropriate.
https://raw.githubusercontent.com/KhronosGroup/OpenGL-Registry/master/xml/gl.xml
Remaining items:
GL_TEXTURE0 + n
pattern?blah ? GL_XXX : GL_YYY
glTexImage2D
call with aninternalformat
of 4 - is that a bug?I guess it's a moot point now, but something else I wondered was whether we need to worry about the group->enums mapping at all. I think we do though.
Some of the extension methods in GlConstantTransformer might be useful more generally, or we might even want to move them to GlRegistry. (I think it'd be somewhat nice if GlRegistry didn't know about JavaParser types though.)
Should we use DI to pass the GlRegistry to the GlConstantTransformer?
_SGIS
. We should check these exhaustively and fix the groups in the registry where appropriate.Everything apart from the mysterious
internalformat
of 4 has been completed, so I'm closing this.