Actually I though of something; the *second* byte in the UTF-8
sequence for Georgian letters is always 0x83 (which, by the way, you’re
not testing for), so you could do something like
\def^^e1#1#2{%
\ifx#1^^83% Second UTF-8 byte for Georgian (range U+10C0 - U+10FF)
\ifx#2^^ad^^e1 (E1)\else% ჭ U+10ED GEORGIAN LETTER CHAR
\errmessage{Hyphenation pattern file corrupted or T8M encoding not
supported!}%
\fi
\else#1#2% The current sequence is not UTF-8, but in some 8-bit encoding;
output the plain bytes
\fi}
This of course assumes that ^^83 is not used in T8M or T8K, but that’s
highly likely, and in any case we’re not chancing the system much more
than we’re already doing.
Arthur
