wcsfc_s: add reorder? about safeclib HOT 1 CLOSED

Nope. Of all valid NFD sequences with >3 chars, if we manually reorder the last 2 (usually marks), there's no reordering, i.e. all the marks have the same Combining Class. Which is probably a Unicode ordering bug, as we cannot compare them, even if they are the same grapheme cluster, the sequence is different.

i.e.
\x1d5 => NFD=U\x308\x304, CC 308 + 304 = 230: NFD("U\x304\x308") => "U\x304\x308"

There is a problem with greek:

      FC        \x{41}\x{3B1}\x{345}\x{342} = \x{61}\x{3B1}\x{3B9}\x{342}
      NFD FC                           Aᾷ  = \x{61}\x{3B1}\x{3B9}\x{342}
      NFD       \x{41}\x{3B1}\x{345}\x{342} = \x{41}\x{3B1}\x{342}\x{345}
      FC NFD                           Aᾷ  = \x{61}\x{3B1}\x{342}\x{3B9} */

as fc 345 => 3b9. (iota)
Initially I thought the natural order would be to do NFD first, and then FC. only then reordering leads to improvement. CC 342=230, CC 345=240, CC 3b9=0!

but the iota must be followed by the 342mark, so fc alone is perfect. NFD reordering is not needed, FCD alone is enough.
Closing.

from safeclib.