Migrating templates to UTF-8?

[timeless]

David Miller wrote:
> I'm not sure offhand what CVS would do, I think it just treats it as 
> 8-bit data, and probably slightly sloppily if it matches something that 
> looks like an RCS keyword.  I'm cross-posting this to the developers 
> list as I don't think the localizers list has that big of an audience 
> (and was really only intended for distributing template-related security 
> information to the template authors prior to security releases).
> 
> We do have UTF-8 data in CVS currently, both in the Bugzilla source code 
> for contributors' names, as well as in the website source, and haven't 
> run into any problems with it, but we might just be lucky.

http://www.devguy.com/fp/cfgmgmt/cvs/

Unicode Files

UNICODE files must be marked BINARY; otherwise, CVS will corrupt them.

The latest CVSNT server can handle UCS-2 (NT UNICODE) and UTF-8 files. 
UCS-2 files "cvs add -ku"