> If you know of shortcomings of _GL_CONCAT, we need to determine whether we > should just document a limitation, or spend the necessary macro complexity > in order to fix it.
Sorry for the misunderstanding; _GL_CONCAT is fine because it takes a detour over _GL_CONCAT0. But not every macro-writing user who wants to reliably concatenate tokens may know this. That's why I said that it may make sense to present even simply macros like GL_CONCAT to the user. It may make user code more readable if they are used instead of ad-hoc named macros.
