> I'd especially like to know what guidance there might be on what to
> do for code points that don't have named characters/escapes in groff
> (ű, ő, ż, ă, ą, ā, ș, ć, č, etc.) -- basically, for Latin-2 I guess
> (or Unicode Latin extended-A).
In case you have proper fonts (or using a UTF-8 encoded TTY), just use
\[u...] entities.
> But for Eastern European languages, if the user has UTF-8 source and
> keeps the output encoding for the stylesheet set at its default
> value, any UTF-8 characters in the source that don't have named
> characters in groff are passed through as-is.
You might also try the `preconv' converter (to be activated with
groff's command line option -k) in the current CVS. Then the
conversion to \[u...] entities happens automatically.
Werner
