For the time being I'm going to NACK the PYTHONIOENCODING patch until
we have a good solution for modifying LANG. I don't think it's proper
to make a laundry list of interpreter-specific variables; it could get
out of control eventually.
What are our options? Wait for C.UTF-8 to be available everywhere?
Parse LANG to determine the actual language, and then remove the end
and add on ".utf-8", as in John's example?
This is pretty extensive here, but it's a bit complicated...
CGit mailing list