I was initially planning to phrase this as a question, but this is something I've bumped up against repeatedly while iterating on `vulkan.cppm`, and was wondering what the wider community thinks of this, which is quite a common error to stumble upon when working with an intermediate codebase that has both module imports and headers.
The standard as far as I can tell doesn't explicitly say anything about this, but de-facto compiler behaviour (GCC, MSVC) is to allow headers-before-modules, but disallow the reverse ordering.
I'd like to know what everyone thinks about disallowing any
#include statements after an import statement in the global module fragment (GMF)—effectively splitting it into two segments, which would also solve this problem.https://redd.it/1oub82l
@r_cpp
