+use strict;
+use warnings;
use Unicode::Normalize qw( NFKD );
use Text::Unidecode qw( unidecode );
+our %rows;
+my %g; # present group classes
{
rows => [0, 1],
def => {
my $v = $rows{$_};
my $class = (
!defined $v || $_ eq $v ? 1 # identical
- : $v =~ /\A\p{Mn}+\z/ ? 8 # combining accent
- : NFKD($v) =~ $_ ? 2 # decomposed equivalent
- : $v =~ /^\p{Latin}/ ? 4 # latin script
- : unidecode($v) =~ /^\W*\Q$_/ ? 5 # transliterated
- : 7
+ : $v =~ /\A\p{Mn}+\z/ ? 9 # combining accent
+ : $v =~ /\A[\p{Sk}\p{Lm}]+\z/ ? 8 # modifier symbol
+ : $v =~ /\A[\pM\pP]+\z/ ? 7 # mark
+ : NFKD($v) =~ /\Q$_/ ? 2 # decomposed equivalent
+ : unidecode($v) =~ /\Q$_\E+/i ? 4 # transliterated
+ : $v =~ /^\p{Latin}/ ? 5 # latin script
+ : 6
);
+ $g{$class} = 1;
$_ => "g$class"
} keys %rows
},
flag => {
g1 => ['unaltered', "same results as without modifier"],
g2 => ['accented', "decomposes to the original letter with a combining accent"],
- g4 => ['latin', "a different (accented) latin letter"],
- g5 => ['similar', "transliterates (mostly) into the unmodified letter"],
- g7 => ['other', "symbol not directly deducible from key"],
- g8 => ['combining', "diacritical mark to be combined with a following character"],
+ g4 => ['similar', "transliterates (mostly) into the unmodified letter"],
+ g5 => ['latin', "a different (accented) latin letter"],
+ g6 => ['symbol', "other character not directly deducible from key"],
+ g7 => ['punctuation', "(punctuation) mark"],
+ !$g{9} ? () : (
+ g8 => ['mark', "modifier letter or mark (spacing diacritic)"],
+ g9 => ['combining', "diacritical mark to be combined with a following character"],
+ ),
},
}