-my @nibble = (0..9, 'A'..'F');
-my $nibsize = 1;
-
-use Encode qw(decode resolve_alias);
-# generate character table(s)
-# (~16x faster than decoding in loop;
-# substr strings is twice as fast as splitting to an array)
-my %ALIAS = (
-# default => [qw(unicode utf-8 iso-8859-1 cp437 -cp1252- --iso-8859-15- -koi8-f)],
- default => [qw(unicode- utf-8 iso-8859-1 -cp1252- --iso-8859-15- cp437 -cp850)],
- 0 => [qw(cp437 -cp863)],
- 1 => [qw(iso-8859-1 -cp1252 -MacRoman -cp850)],
- 2 => [qw(iso-8859-2 -cp1250 -cp852 -MacCentralEurRoman -MacCroatian -MacRumanian)],
- 5 => [qw(koi8-f -iso-8859-5 -cp1251 -MacCyrillic -cp855 -cp866)],
- 7 => [qw(iso-8859-7 -cp1253 -MacGreek -cp737 -cp869)],
- 8 => [qw(iso-8859-8 -cp1255 -MacHebrew -cp862)],
-);
-my @request = map {
- if (my $input = $_) {
- my %row = (offset => 0);
- my $endpoint = 255;
- if ($input =~ s/^--//) {
- $row{offset} = $endpoint > 160 ? 160 : 48;
- }
- elsif ($input =~ s/^-//) {
- $row{offset} = $endpoint > 128 ? 128 : 32;
- }
- if ($input =~ s/-$//) {
- $endpoint = $row{offset} ? $row{offset} < 160 ? 159 : 191 : 127;
- }
- if ($row{offset}) {
- $row{setnote} = 'over cp437' if $input eq 'cp850';
- $row{setnote} = 'over iso-8859-1' if $input =~ /^iso-8859-|^cp125/;
- }
-
- if ($input =~ /^U([0-9a-f]+)(?:-([0-9a-f]+))?/) {
- my $start = hex($1) << ($2 ? 4 : 8);
- my $end = $2 ? hex($2) << 4 : $start + 240;
- $row{table} = join '', map { chr } $start .. $end+15;
- utf8::upgrade($row{table}); # prevent latin1 output
- $row{set} = sprintf 'Unicode block U+%02Xxx', $start >> 8;
- }
- elsif ($input eq 'U') {
- $row{table} = ' ' x 1024;
- $row{set} = 'Unicode planes';
- $row{cell} = do 'charset-ucplanes.inc.pl'
- or printf "<p class=error>Table data could not be read: <em>%s</em>.</p>\n", $@ || $!;
- @nibble = (map { $_.0, $_.8 } 0 .. 7);
- $nibsize = 8;
- }
- elsif ($row{set} = resolve_alias($input)) {
- if ($row{set} eq 'Internal') {
- $row{table} = ' ' x ($endpoint < 255 ? 640 : 8192);
- $row{set} = 'Unicode BMP';
- $row{cell} = do 'charset-unicode.inc.pl'
- or printf "<p class=error>Table data could not be read: <em>%s</em>.</p>\n", $@ || $!;
- }
- elsif ($row{set} eq 'utf-8-strict') {
- $row{table} = undef;
- $row{set} = 'UTF-8';
- $row{cell} = do 'charset-utf8.inc.pl'
- or printf "<p class=error>Table data could not be read: <em>%s</em>.</p>\n", $@ || $!;
+my @request;
+
+sub tabinput {
+ # generate character table(s)
+ my $input = shift or return;
+
+ state $ALIAS = {
+ default => [qw( u+0-639 utf-8+realsize iso-8859-1 iso-8859-15 cp1252 cp437 cp850 )],
+ unicode => [qw( uu+cols=32+realsize u+0-4095 u+4096-6319 u+6320-8191 )],
+ us => [qw( cp437 cp863 gsm0338 AdobeStandardEncoding )],
+ ebcdic => [qw( cp37 cp500 cp1047 posix-bc cp1026 cp875 )],
+ iso => [map {"iso-8859-$_"} 1 .. 11, 13 .. 16],
+ dos => [qw( cp437 cp865 cp861 cp860 cp863 cp850 cp857 cp852 cp775
+ cp737 cp869 cp866 cp855 cp862 cp864 )],
+ aix => [qw( cp1006 )],
+ win => [qw( cp1252 cp1250 cp1254 cp1257 cp1258 cp1253 cp1251 cp1255 cp1256 cp874 )],
+ mac => [qw( MacRoman MacRomanian MacRumanian MacCroatian MacCentralEurRoman MacTurkish MacIcelandic MacSami
+ MacGreek MacCyrillic MacHebrew MacArabic MacFarsi MacThai )],
+
+ # languages
+ westeur => [qw( iso-8859-1 iso-8859-15 cp1252 iso-8859-14 cp850 hp-roman8 nextstep MacRoman )],
+ centeur => [qw( iso-8859-2 iso-8859-16 cp1250 cp852 MacRomanian MacCroatian MacCentralEurRoman )], # MacRumanian only for DB
+ turkish => [qw( iso-8859-9 iso-8859-3 cp1254 cp857 MacTurkish )],
+ norteur => [qw( baltic nordic )],
+ baltic => [qw( iso-8859-4 iso-8859-13 cp1257 cp775 )],
+ nordic => [qw( iso-8859-10 cp865 cp861 MacIcelandic MacSami )],
+ cyrillic => [qw( koi8-r koi8-u koi8-f iso-8859-5 cp1251 MacCyrillic cp866 cp855
+ U4 U2DE-2DF UA64-A69 U50-52 )], # MacUkrainian is broken
+ arabic => [qw( iso-8859-6 cp1256 MacArabic cp864 cp1006 MacFarsi
+ U6 U8A-8F+0-31+64 U75-77 )],
+ greek => [qw( iso-8859-7 cp1253 MacGreek cp737 cp869 U37-3F U1F )],
+ hebrew => [qw( iso-8859-8 cp1255 MacHebrew cp862 U59-5F )],
+ thai => [qw( iso-8859-11 cp874 MacThai )],
+ vietnamese => [qw( viscii cp1258 MacVietnamese )],
+ symbol => [qw( symbol dingbats MacDingbats AdobeZdingbat AdobeSymbol )],
+
+ # iso-code shorthand
+ 1 => 'westeur',
+ 2 => 'centeur',
+ 3 => 'turkish',
+ 4 => 'baltic',
+ 5 => 'cyrillic',
+ 6 => 'arabic',
+ 7 => 'greek',
+ 8 => 'hebrew',
+ 9 => 'turkish',
+ 10 => 'nordic',
+ 11 => 'thai',
+ };
+ if (my $follow = $ALIAS->{$input}) {
+ tabinput($_) for ref $follow ? @{$follow} : $follow;
+ return;
+ }
+
+ state $visible = {}; # all present tables
+ my %row = (offset => 0, cols => 16);
+ my $params = $input =~ s/[+](.*)\z// ? $1 : undef;
+
+ if (not defined $params) {
+ state $INHERIT = {
+ 'cp437' => ['cp850' => 0, 'ascii' => '0-31+128'], # ascii range overridden later
+ 'gsm0338' => ['ascii' => '0-127'],
+
+ 'iso-8859-2' => ['iso-8859-1' => '160'],
+ 'iso-8859-3' => ['iso-8859-1' => '160'], #TODO: also apply to iso-8859-9
+ 'iso-8859-4' => ['iso-8859-2' => '160'],
+ 'iso-8859-5' => ['iso-8859-1' => '160'],
+ 'iso-8859-6' => ['cp1256' => '128', 'iso-8859-1' => '160'],
+ 'iso-8859-7' => ['iso-8859-1' => '160'],
+ 'iso-8859-8' => ['iso-8859-1' => '160'],
+ 'iso-8859-9' => ['iso-8859-1' => '208-223+240'],
+ 'iso-8859-10' => ['iso-8859-4' => '160'],
+ 'iso-8859-11' => ['iso-8859-1' => '160'],
+ 'iso-8859-13' => ['iso-8859-4' => '160'],
+ 'iso-8859-14' => ['iso-8859-1' => '160'],
+ 'iso-8859-15' => ['iso-8859-1' => '160-191'],
+ 'iso-8859-16' => ['iso-8859-2' => '160'],
+ 'hp-roman8' => ['iso-8859-1' => '160'],
+
+ 'cp1252' => ['iso-8859-1' => '128-159'],
+ 'cp1250' => ['iso-8859-2' => '128-191', 'cp1252' => '128'],
+ 'cp1254' => ['iso-8859-9' => '128-159', 'cp1252' => '128-159+208'],
+ 'cp874' => ['iso-8859-11' => '128-159', 'cp1252' => '128'], # windows-874 actually cp1162
+ 'cp1257' => ['iso-8859-13' => '128-159+255', 'cp1252' => '128'],
+ 'cp1251' => ['cp1252' => '128'],
+ 'cp1253' => ['cp1252' => '128'],
+ 'cp1255' => ['iso-8859-8' => '128-223', 'cp1252' => '128'],
+ 'cp1256' => ['cp1252' => '128'],
+ 'cp1258' => ['cp1252' => '128-159+192'],
+
+ 'cp850' => ['cp437' => '144'],
+ 'cp860' => ['cp437' => '128-175'],
+ 'cp861' => ['cp865' => '128-175'],
+ 'cp863' => ['cp437' => '128-175'],
+ 'cp865' => ['cp437' => '144-175'],
+ 'cp852' => ['cp850' => '128', 'cp437' => '128'],
+ 'cp857' => ['cp850' => '128-175+208-239', 'cp437' => '128'],
+ 'cp775' => ['cp850' => '128'], # partial cp437
+ 'cp866' => ['cp437' => '128-175+224'],
+ 'cp855' => ['cp437' => '128'],
+ 'cp1006' => ['iso-8859-6' => '160', 'cp437' => '128'],
+ 'cp737' => ['cp437' => '128-175+224'],
+ 'cp869' => ['cp437' => '128'],
+ 'cp862' => ['cp437' => '128-159'],
+ 'cp864' => ['MacArabic' => '128', 'iso-8859-6' => '128', 'cp437' => '128'], #TODO: compare form variants
+
+ 'koi8-u' => ['koi8-r' => '144-191'],
+ 'koi8-f' => ['koi8-u' => '144-191'],
+
+ 'MacRomanian' => ['MacRoman' => '160-191+208-223'],
+ 'MacRumanian' => ['MacRomanian' => '160-191+208-223', 'MacRoman' => '160-191+208-223'],
+ 'MacCroatian' => ['MacRoman' => '160'],
+ 'MacCentralEurRoman' => ['MacRoman' => '128'],
+ 'MacIcelandic'=> ['MacRoman' => '160-175+208-239'],
+ 'MacTurkish' => ['MacRoman' => '208-223'], # F5 is unassigned
+ 'MacSami' => ['MacIcelandic' => '144-191+208-223+240', 'MacRoman' => '144'],
+ 'MacGreek' => ['MacRoman' => '128'],
+ 'MacCyrillic' => ['MacRoman' => '128'],
+ 'MacHebrew' => ['iso-8859-8' => '128', 'MacRoman' => '128-143+160'], # partial ascii
+ 'MacArabic' => ['iso-8859-6' => '128', 'cp864' => '128', 'MacRoman' => '128'], #TODO: multiple parents
+ 'MacFarsi' => ['MacArabic' => '176-191', 'MacRoman' => '128'],
+
+ 'cp37' => ['posix-bc' => '0'],
+ 'posix-bc' => ['cp1047' => '64'],
+ 'cp500' => ['cp37' => '64-95+176-191'],
+ 'cp1047' => ['cp37' => '16-47+80-95+160-191'],
+ 'cp1026' => ['cp37' => '64'],
+ 'cp875' => ['cp37' => '48'],
+
+ };
+
+ my @parents = @{ $INHERIT->{$input} || [] };
+
+ if (my ($parent, $part) = pairfirst { defined $visible->{$a} } @parents) {
+ $row{parent} = $parent;
+ $params = $part;
+ $params = 128 unless $visible->{$parent}
+ or ($input eq 'MacCroatian' and defined $visible->{MacRomanian});
+ }
+ elsif (defined $visible->{ascii}) {
+ $row{parent} = $parents[0];
+ $params = $parents[1] // 128;
+ $params = 128 if $params >= 128; # ascii offset at most
+ }
+ elsif (@parents) {
+ $row{parent} = $parents[0];
+ $params = $parents[1] if $parents[1] == 0; # apply ascii end
+ }
+ $visible->{$_} //= 0 for $row{parent} || ();
+ }
+
+ for my $param (split /[+]+/, $params // '') {
+ if ($param eq 'realsize') {
+ $row{realsize}++;
+ }
+ elsif ($param =~ m{ \A cols = (\d+) \z }x) {
+ $row{cols} = $1;
+ }
+ elsif ($param =~ m{ \A (?<offset> \d+) (?: [-] (?<endpoint> \d+) )? \z }x) {
+ if (defined $row{endpoint}) {
+ # extend earlier range
+ my $skip = int(($row{endpoint} || $row{offset}) / $row{cols});
+ for ($skip + 1 .. ($+{offset} / $row{cols}) - 1) {
+ $row{skip}->{ $_ * $row{cols} - $row{offset} }++;
+ }