Re: [Gnu-arch-users] Increasing the filename space (Or: begging for trouble?)

[Tom Lord]

    > From: Aaron Bentley <address@hidden>

    > I'm tempted to think that we should enforce utf-8 for log-files, so as
    > to standardize the format.  

Right, me too.   In patch logs and archives, logs should just be
UTF-8.

The other thing is that I think filenames should _also_ just be UTF-8 
and what I'm puzzling over concerning that is if there's an
incremental path to that.


    > The committer machine should know how to turn the local encoding
    > into utf-8, and the getter machine should know how to turn utf-8
    > into the local encoding.  But I'm not sure if we should assume
    > that the getter machine can turn the committer's local encoding
    > into utf-8.

I think we can assume with an escape clause.  Most people should agree
that Unicode is ample.   Many people will want to use a character set
with a lossy roundtrip to Unicode.   But as long as arch is tolerant
of private-use codepoints, this issue can reasonably be "not our problem".


-t