procyberian/emacs
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Further Unicode restrictive fixups

Browse source 
* puny.el (puny-highly-restrictive-p): Include the extra
identifier characters from table 3.
This commit is contained in:
Lars Ingebrigtsen 2015-12-29 17:46:00 +01:00
parent d181366dc3
commit d259328fb8
1 changed files with 27 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

31
lisp/net/puny.el
View file

@ -191,13 +191,36 @@ For instance \"xn--bcher-kva\" => \"bücher\"."
(buffer-string)))
;; http://www.unicode.org/reports/tr39/#Restriction_Level_Detection
;; http://www.unicode.org/reports/tr31/#Table_Candidate_Characters_for_Inclusion_in_Identifiers
(defun puny-highly-restrictive-p (string)
(let ((scripts
(seq-uniq
(seq-map (lambda (char)
(aref char-script-table char))
string))))
(delq
t
(seq-uniq
(seq-map (lambda (char)
(if (memq char
;; These characters are always allowed
;; in any string.
'(#x0027 ; APOSTROPHE
#x002D ; HYPHEN-MINUS
#x002E ; FULL STOP
#x003A ; COLON
#x00B7 ; MIDDLE DOT
#x058A ; ARMENIAN HYPHEN
#x05F3 ; HEBREW PUNCTUATION GERESH
#x05F4 ; HEBREW PUNCTUATION GERSHAYIM
#x0F0B ; IBETAN MARK INTERSYLLABIC TSHEG
#x200C ; ERO WIDTH NON-JOINER*
#x200D ; ERO WIDTH JOINER*
#x2010 ; YPHEN
#x2019 ; IGHT SINGLE QUOTATION MARK
#x2027 ; YPHENATION POINT
#x30A0 ; KATAKANA-HIRAGANA DOUBLE HYPHEN
#x30FB)) ; KATAKANA MIDDLE DOT
t
(aref char-script-table char)))
string)))))
(or
;; Every character uses the same script.
(= (length scripts) 1)