I have certainly seen such code in the past, that reliedon developers working with the same code page, and
was essentially in-house code for developers with
non-English native languages.
If those APIs ever have to be shared elsewhere within
a larger organization, it is not unreasonable to imagine
the kind of mapping I suggest above as expediency,
rather than taking time to rename, refactor, and retest
all code using those APIs (the preferred solution).
AlisdairM
A more interesting question is whether we expect tools
to provide a translation layer using \uXXXX in identifiers
to alias more reasonable names in a user’s native
encoding?
Thinking of running a script to provide a translation of
a header with more usable names for my working
environment.
A yet more interesting question is whether characters outside of what is defined as the basic source character set appear in interfaces of libraries that are shared in environments that are not able to represent these characters.
I haven't seen such characters in identifiers in header files, ever.