Unicode Char Names: describe-char-unicode-data vs get-char-code-property

["T.V. Raman" <tv.raman.tv@gmail.com>]

I use describe-char-unicode-data to speak unicode chars in
emacspeak when navigating by character.

That function claims to be semi-obsolete -- and recommends
get-char-code-property instead; however the answers are not
consistent in the two solutions.

Try non-breaking space \240 (octal)

(setq char ? )
160 (#o240, #xa0, ? )
(describe-char-unicode-data char) returns
(("Name" "NO-BREAK SPACE") ("Category" "Separator, Space")
("Combining class" "Zs") ("Bidi category" "Zs") ("Decomposition"
"noBreak ") ("Decimal digit value" nil) ("Digit value" nil)
("Numeric value" nil) ("Mirrored" nil) ("Old name" "NON-BREAKING
SPACE") ("ISO 10646 comment" nil) ("Uppercase" nil) ...)
whereas
(get-char-code-property char 'name) returns
"COMMA NO-SHE IOTA ARABIC DIAERESIS"
also looking up char in ucs-names yields something different
(car (rassoc char (ucs-names))) returns
"NON-NEITHER UNDER CM COMMA"
Note that all of the above is with the unicdoedata file present
and correctly found by emacs.


Also, and perhaps related:

C-x 8 spc inserts non-breaking space correctly; however when you
use C-x 8 ret, non-breaking space doesn't show up in the
available completions -- either using it's "name" or "old-name"
--