[openal] OpenAL Soft string encodings

[Chris Robinson]

A sort-of announcement and sort-of query.

Recently with the Windows version of OpenAL Soft, I started handling the 
various device name strings by taking the wide-char representation (as 
given by the various *W functions), and converting them to UTF-8. This 
means that, on Windows, the strings now use UTF-8 encoding rather than 
the active code-page (ACP), and can effectively represent all Unicode 
characters. The reasoning for this is that the active code-page may not 
be able to properly represent any possible string identifier that the 
system may have.

Further, on other non-Windows systems the current assumption is that the 
system is set to use UTF-8 strings. The idea is to make OpenAL Soft use 
UTF-8 encoding for all strings that get passed in to and out from the 
lib, on as many systems as is feasible.

So the questions I have is:

Is this change on Windows (making strings use UTF-8 instead of ACP) 
going to be a problem?

Is there anything I can realistically do on non-Windows systems aside 
from assuming char strings are UTF-8?