# new Unicode release, to make sure things haven't been changed by it.
my @properties = qw(
+ NONLATIN1_SIMPLE_FOLD
NONLATIN1_FOLD
ALPHANUMERIC
ALPHA
XDIGIT
VERTSPACE
IS_IN_SOME_FOLD
- BACKSLASH_FOO_LBRACE_IS_META
+ MNEMONIC_CNTRL
);
# Read in the case fold mappings.
my %folded_closure;
my @hex_non_final_folds;
+my @non_latin1_simple_folds;
my @folds;
use Unicode::UCD;
my $from = hex $hex_from;
- # Perl only deals with C and F folds
- next if $fold_type ne 'C' and $fold_type ne 'F';
+ # Perl only deals with S, C, and F folds
+ next if $fold_type ne 'C' and $fold_type ne 'F' and $fold_type ne 'S';
# Get each code point in the range that participates in this line's fold.
# The hash has keys of each code point in the range, and values of what it
push @{$folded_closure{$fold}}, $from if $fold < 256;
push @{$folded_closure{$from}}, $fold if $from < 256;
- if ($i < @folded-1
- && $fold < 256
- && ! grep { $_ eq $hex_fold } @hex_non_final_folds)
+ if (($fold_type eq 'C' || $fold_type eq 'S')
+ && ($fold < 256 != $from < 256))
+ {
+ # Fold is simple (hence can't be a non-final fold, so the 'if'
+ # above is mutualy exclusive from the 'if below) and crosses
+ # 255/256 boundary. We keep track of the Latin1 code points
+ # in such folds.
+ push @non_latin1_simple_folds, ($fold < 256)
+ ? $fold
+ : $from;
+ }
+ elsif ($i < @folded-1
+ && $fold < 256
+ && ! grep { $_ eq $hex_fold } @hex_non_final_folds)
{
push @hex_non_final_folds, $hex_fold;
push @{$folded_closure{$from}}, @{$folded_closure{$folded}};
}
}
+
+ # We have the single-character folds that cross the 255/256, like KELVIN
+ # SIGN => 'k', but we need the closure, so add like 'K' to it
+ foreach my $folded (@non_latin1_simple_folds) {
+ foreach my $fold (@{$folded_closure{$folded}}) {
+ if ($fold < 256 && ! grep { $fold == $_ } @non_latin1_simple_folds) {
+ push @non_latin1_simple_folds, $fold;
+ }
+ }
+ }
}
sub Is_Non_Latin1_Fold {
return join("\n", @return) . "\n";
}
+sub Is_Non_Latin1_Simple_Fold { # Latin1 code points that are folded to by
+ # non-Latin1 code points as single character
+ # folds
+ return join("\n", map { sprintf "%X", $_ } @non_latin1_simple_folds) . "\n";
+}
+
sub Is_Non_Final_Fold {
return join("\n", @hex_non_final_folds) . "\n";
}
$re = qr/\p{_Perl_Quotemeta}/;
} elsif ($name eq 'NONLATIN1_FOLD') {
$re = qr/\p{Is_Non_Latin1_Fold}/;
+ } elsif ($name eq 'NONLATIN1_SIMPLE_FOLD') {
+ $re = qr/\p{Is_Non_Latin1_Simple_Fold}/;
} elsif ($name eq 'NON_FINAL_FOLD') {
$re = qr/\p{Is_Non_Final_Fold}/;
} elsif ($name eq 'IS_IN_SOME_FOLD') {
$re = qr/\p{_Perl_Any_Folds}/;
- } elsif ($name eq 'BACKSLASH_FOO_LBRACE_IS_META') {
-
- # This is true for FOO where FOO is the varying character in:
- # \a{, \b{, \c{, ...
- # and the sequence has non-literal meaning to Perl; so it is true
- # for 'x' because \x{ is special, but not 'a' because \a{ isn't.
- $re = qr/[gkNopPx]/;
+ } elsif ($name eq 'MNEMONIC_CNTRL') {
+ # These are the control characters that there are mnemonics for
+ $re = qr/[\a\b\e\f\n\r\t]/;
} else { # The remainder have the same name and values as Unicode
$re = eval "qr/\\p{$name}/";
use Carp;
# Output the table using fairly short names for each char.
foreach my $charset (get_supported_code_pages()) {
- my @a2n = get_a2n($charset);
+ my @a2n = @{get_a2n($charset)};
my @out;
print $out_fh "\n" . get_conditional_compile_line_start($charset);