[Chicken-users] Strange output from unit irregex

[Henry Hu]

Hello,

I am getting a strange result from unit irregex having to do with matching
character sets.

I recently upgraded to 4.13.0 to get the bug fix having to do with an extra
empty list in the SRE: https://github.com/ashinn/irregex/pull/18. I was
happy to find that "[]" bracketed character sets without "^" are working
beautifully! I am, however, observing strange things with the "^"
exclusion character.

The ⟀ character has three bytes and when displayed in byte form, looks like
`\342\276\200`:

INPUT:
(use irregex) ; Not doing (use utf8) because I want start-index and
end-index to function correctly
(irregex-match-substring (irregex-search (irregex "[^⟀]" 'utf8) "⟀⟀⟀"))

EXPECTED OUTPUT:
Considering a UTF-8 character as a single character anywhere it appears:
`#f`
Considering a UTF-8 character as a single character sometimes and a byte
string sometimes: `<the first byte of ⟀>` (displayed as `\342`), or #f
Considering a UTF-8 character as a byte string always: #f

OUTPUT:
`<the first byte of ⟀><the second byte of ⟀>` (looks like `\342\276`)

EVEN WORSE:
(irregex-match-substring (irregex-search (irregex "[^Ç]" 'utf8) "Ç")) --->
"Ç" ; A two-byte character

Am I doing something wrong? Is "^" not designed to be used with multibyte
characters? Why would it return two bytes and not 0, 1, or 3?

Thank you!