Currently, GnuPG libraries (libgpg-error, libassuan, etc.) have header file which is architecture dependent, generated at build time.
Ideally, if it offers same API, it's good if it can be a single unified header file among architectures.
If so, it will be possible to share the single unified header file among architectures which enables omitting --includedir option at configure time for cross build.
We have had this in the past but it led to subtle build and, worse, runtime problems. Thus the decision to provide architecture dependent files and have configure complain for wrong files. Right, you sometimes get false warnings for non-matching cpu-vendor-os strings but I consider this less severe than the old problem.
Of course we could split the header files in a generic and a system dependent file. This has the drawback that the osearch path for header files is not always robust and introduces other errors.
Also recall that the header files are created from configure test and thus we can't provide static files. The libgpg-error syscfg stuff is annoying enough - better don't add yet another one.
I think we should close this bug or re-purpose it to silence those warnings in common cases.