+print join " •\n", (
+ map {
+ join " ·\n", pairmap {
+ optionlink($b || ucfirst $a, '/charset'.($a && "/$a?compare"), $a eq $Request);
+ } @{$_}
+ }
+ [
+ iso => 'ISO',
+ win => 'Windows',
+ dos => 'DOS',
+ mac => 'Apple',
+ ebcdic => 'EBCDIC',
+ $tablist[0] eq 'default' ? () : ('' => 'common'),
+ ],
+ [
+ westeur => 'West',
+ centeur => 'Central',
+ norteur => 'North European',
+ turkish => 0,
+ greek => 0,
+ cyrillic => 0,
+ hebrew => 0,
+ ],
+);
+:>.
+</p>
+
+<:
+use POSIX qw( ceil );
+use Shiar_Sheet::FormatChar;
+my $glyphs = Shiar_Sheet::FormatChar->new;
+my @request;
+
+my $charsets = do 'charset-encoding.inc.pl'
+ or Alert('Encoding metadata could not be read', $@ || $!);
+
+sub tabinput {
+ # generate character table(s)
+ my $input = shift or return;
+ my $params = $input =~ s/[+](.*)\z// ? $1 : undef;
+ my $charset = $charsets->{lc $input} || {};
+
+ if (ref $charset ne 'HASH') {
+ $params and Alert("Parameters ignored for $input",
+ "Cannot apply <q>$params</q> to multiple charsets.",
+ );
+ tabinput($_) for ref $charset ? @{$charset} : $charset;
+ return;
+ }
+
+ state $visible = {'' => 1}; # all present tables
+ my %row = (offset => 0, cols => 16);
+
+ if (not defined $params) {
+ my @parents = @{ $charset->{inherit} || [] };
+
+ if (my ($parent, $part) = pairfirst { defined $visible->{$a} } @parents) {
+ $row{parent} = $parent;
+ $params = $part;
+ $params = 80 unless $visible->{$parent}
+ or ($input eq 'MacCroatian' and defined $visible->{MacRomanian});
+ }
+ elsif (defined $visible->{ascii}) {
+ $row{parent} = $parents[0];
+ $params = $parents[1] // 80;
+ $params = 80 if hex $params >= 0x80; # ascii offset at most
+ }
+ elsif (@parents) {
+ $row{parent} = $parents[0];
+ $params = $parents[1] if hex $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 (?<start> \p{AHex}+) (?: [-] (?<end> \p{AHex}+) )? \z }x) {
+ if (defined $row{endpoint}) {
+ # extend earlier range
+ my $skip = int(($row{endpoint} || $row{startpoint}) / $row{cols});
+ for ($skip + 1 .. (hex($+{start}) / $row{cols}) - 1) {
+ $row{skip}->{ $_ * $row{cols} - $row{startpoint} }++;
+ }
+ }
+ else {
+ $row{startpoint} = hex $+{start};
+ }
+ $row{endpoint} = hex($+{end} || 0);
+ }
+ else {
+ Alert("Unknown option <q>$param</q> for charset $input");
+ }
+ }
+
+ if ($charset->{setup}) {
+ eval { $charset->{setup}->(\%row) }
+ or Alert("Incomplete setup of $input", $@);
+ }
+ $row{endpoint} ||= 0xFF;
+
+ if ($row{set}) {}
+ elsif ($row{set} = Encode::resolve_alias($input)) {
+ $row{offset} = delete $row{startpoint};
+ if ($row{set} eq 'MacHebrew' or $row{set} eq 'MacThai') {
+ # array of possibly multiple characters per code point
+ $row{table} = [
+ map { Encode::decode($row{set}, pack 'C*', $_) } $row{offset} .. $row{endpoint}
+ ];
+ }
+ else {
+ # ~16x faster than decoding in loop;
+ # substr strings is twice as fast as splitting to an array
+ $row{table} = Encode::decode($row{set}, pack 'C*', $row{offset} .. $row{endpoint});
+ }
+
+ $row{endpoint} -= $row{offset};
+
+ $visible->{ascii} = # assume common base
+ $visible->{ $row{set} } = 1;
+ }
+ else {
+ Alert("Encoding <q>$input</q> unknown");
+ return;
+ }
+
+ if (my $replace = $charset->{replace}) {
+ while (my ($offset, $sub) = each %{$replace}) {
+ $offset -= $row{offset};
+
+ if (ref $row{table} eq 'ARRAY') {
+ $row{table}->[$offset] = $sub
+ if $offset >= 0 and $offset <= $row{endpoint};
+ next;
+ }
+
+ my $length = length $sub;
+
+ if ($offset < 0) {
+ $offset > -$length or next; # at least one character after start
+ # trim leftmost part to start at offset
+ substr($sub, 0, -$offset) = '';
+ $length += $offset;
+ $offset = 0;
+ }
+
+ if ((my $excess = $row{endpoint} - $offset - $length + 1) < 0) {
+ $excess > -$length or next;
+ # trim rightmost part to prevent overflow
+ substr($sub, $excess) = '';
+ $length += $excess;
+ }
+
+ substr($row{table}, $offset, $length) = $sub;
+ }
+ }
+
+ push @request, \%row;
+}
+tabinput($_) for @tablist;
+
+my $NOCHAR = chr 0xFFFD;
+
+sub range_cell {
+ my ($info, $offset) = @_;
+ my $table = $info->{cell} or return;
+ my $def = $table->{$offset} or return;
+ my ($len, $class, $name, $title) = @{$def};
+
+ my $cols = $info->{cols};
+ my $colsize = $table->{colsize} || 1;
+ my $attr = '';
+ $len /= $colsize;
+ $name //= $len <= 2 ? 'res' : 'reserved';
+
+ if (my $part = ($offset/$colsize - $info->{startpoint}) % $cols) {
+ # continued row
+ my $rest = $cols - $part; # remaining
+ $rest = $len if $len < $rest; #TODO: optimise
+ if ($len -= $rest) {
+ # continued on new row
+ my @next = ($len * $colsize, "$class joinu");
+ my $separate = $cols - $len > $rest; # columns not on next row
+ if ($len > $rest) {
+ # minority remains
+ push @next, $name, $title;
+ $title ||= $name;
+ $name = $separate && '…';
+ }
+ else {
+ # minority on next row
+ push @next, $separate && '"', $title || $name;
+ }
+ $table->{$offset + $colsize*$rest} //= \@next;
+ $class .= ' joind';
+ }
+ $len = $rest;
+ }
+ elsif (my $rows = int($len / $cols)) {
+ # multiple full rows
+ my $rowsize = $colsize * $cols;
+ if ($len -= $rows * $cols) {
+ # partial row remains
+ $table->{$offset + $rowsize * $rows} //= [$len*$colsize, "$class joinu", '', $title];
+ $class .= ' joind';
+ }
+
+ unless ($info->{realsize}) {
+ # coalesce multiple rows
+ while ($rows > 3) {
+ $info->{skip}->{$offset += $rowsize}++;
+ $rows--;
+ }
+ if ($rows > 2) {
+ $info->{skip}->{$offset += $rowsize} = 0;
+ }
+ }
+
+ $attr .= sprintf ' rowspan=%d', $rows;
+ $len = $cols;
+ }
+
+ $attr .= sprintf ' colspan=%d', $len unless $len == 1;
+ $attr .= $1 if $class and $class =~ s/( \w+="[^"]*")//;
+ $attr .= sprintf ' class="%s"', $class if $class;
+ $attr .= sprintf ' title="%s"', EscapeHTML($title) if $title;
+ return "<td$attr>$name\n";