> From: "peter_jacobi.rm"
> - no correct uppercasing, only ASCII chars are uppercased at all
>
> - character length limit not enforced (can store 'ABC' into char(1)
> character set UNICODE_FSS)
>
> - no non-trivial collations defined for UNICODE_FSS
>
> - UNICODE_FSS is unaware of the astral planes, but you get way with
> this by simply treating UNICODE_FSS as CESU-9 in your code (if you
> need astral planes at all)

It could be perhaps easier to list features UNICODE_FSS _does_ support:

- It is possible to transliterate characters between UNICODE_FSS
and other character sets.

Is that all or did I forget something ?
Most people wanting Unicode support may as well use NONE character set
without loosing any functionality, especially because using more character sets
in one database is ... problematic enyway, thanks to _forcing_ client
to choose single connection character set, and enforcing transliteration
even without explicit request from client (and I do not like software
that tries to please me by doing things without me asking for it).

Ivan