[Groff] Re: preconv supported encodings

[Bruno Haible]

Werner LEMBERG wrote:
> > Big5 is not a good candidate to support here, because there are many
> > variants of Big5 and none of it is formally standardized.  [...]
>
> I know that too well, but the ETen variant is still the most favourite
> encoding in Taiwan (besides Unicode).

Are you sure about that?

That most Big5 variants include the Eten extensions, is a myth.
The Eten encoding maps
   0xC7F3 to U+0410 CYRILLIC CAPITAL LETTER A
   0xF9F0 to U+2565 BOX DRAWINGS DOWN DOUBLE AND HORIZONTAL SINGLE
   0xF9D6 to U+7881
The only mappings that do this are:
  - BIG5-HKSCS
  - ICU's MACOS-2566-10.2.TXT
  - ICU's IBM-1375_P100-2003.TXT

The reality is that many Big5 variants include _parts_ of Eten.
Some have the 0xC6..0xC7 rows; some have their contents swapped. Some have
the 0xF9 row, some don't.

Bruno