Duncan

> I look at clause 7 and my English is not good enough to comprehend the way it is worded. > > I tested your suggestion with godbolt, and mutable...

You could at least add a note next to the user of it saying that it does not conform to std::erase_if's contract then

Instead of having only some using the correct size types I propose we just switch the codebase in full one type at a time it will lead to more consistent...

We already use git for source control we have absolutely no need to add unusable code branches. Just delete the code.

Why, we have already investigated and concluded is not used. Have a comment that there used to be special win9x code if you want. If in the future you can...

I would still like to see the dead code removed as it has already been investigated and if anyone for some reason wanted to view the original they can use...

Suggested to use https://github.com/TheAssemblyArmada/Thyme/blob/69b10921b42c6848fcd118c2fa257ec4f51c962d/src/w3d/renderer/dx8vertexbuffer.h#L45 Sadly this isn't the same this is for Vector4. Could still use the same naming scheme though.

We have the following vertex formats defined: `VertexFormatXYZDUV2` within w3dwater.h `VertexFormatXYZNDUV2` within dx8vertexbuffer.h `VertexFormatXYZDUV1` within dx8vertexbuffer.h This PR adds: `VertexFormatXYZWDUV1` within dx8wrapper.h `VertexFormatXYZWDUV2` within dx8wrapper.h

Yes that is possible. It will take a little while to get in the state where this could be enforced. This GitHub action could be used https://github.com/platisd/clang-tidy-pr-comments

> Hmm. To be fair it actually looks nicer to look at with the macros. The duplication of types and array sizes is quite a strain when typing it out...