X-Git-Url: http://git.shiar.nl/sheet.git/blobdiff_plain/9a4d89685edcb4399aa500d38de84b0101ad09b2..2f6bd0f228e04949e561c9e5d7713a0baef77cd5:/charset.plp

diff --git a/charset.plp b/charset.plp
index e84ae82..ed495fe 100644
--- a/charset.plp
+++ b/charset.plp
@@ -1,8 +1,11 @@
 <(common.inc.plp)><:
 
+my $mode = exists $get{compare};
+my @tablist = split m{/+}, $Request || 'default';
+
 Html({
 	title => 'charset cheat sheet',
-	version => 'v1.0',
+	version => '1.0',
 	description => [
 		"Reference sheet with all glyphs in common character encoding tables,",
 		"and an overview of Unicode ranges and UTF-8 bytes.",
@@ -12,146 +15,352 @@ Html({
 		reference common overview table
 	'],
 	stylesheet => [qw'light'],
+	data => [qw(
+		charset-encoding.inc.pl
+		charset-unicode.inc.pl charset-ucplanes.inc.pl charset-utf8.inc.pl
+	)],
 });
 
+use List::Util qw( first pairmap pairfirst pairs );
+
 :>
-<h1>Character encoding</h1>
+<h1>Character encodings</h1>
 
+<p>
 <:
-my $diinfo = do 'digraphs.inc.pl';
-my %di = map { $diinfo->{$_}->[0] => $_ } grep { ref $diinfo->{$_} }
-	keys %$diinfo;
-
-use Encode qw(decode resolve_alias);
-# generate character table(s)
-# (~16x faster than decoding in loop;
-#  substr strings is twice as fast as splitting to an array)
-my %ALIAS = (
-#	default => [qw(unicode utf-8 iso-8859-1 cp437 -cp1252- --iso-8859-15- -koi8-f)],
-	default => [qw(unicode- utf-8 iso-8859-1 -cp1252- --iso-8859-15- cp437 -cp850)],
-	0 => [qw(cp437 cp863)],
-	1 => [qw(iso-8859-1 cp1252 MacRoman cp850)],
-	2 => [qw(iso-8859-2 cp1250 cp852 MacCentralEurRoman MacCroatian MacRumanian)],
-	5 => [qw(koi8-f iso-8859-5 cp1251 MacCyrillic cp855 cp866)],
-	7 => [qw(iso-8859-7 cp1253 MacGreek cp737 cp869)],
-	8 => [qw(iso-8859-8 cp1255 MacHebrew cp862)],
+if ($tablist[0] eq 'default') {
+	say "Overview of Unicode allocation and common latin code pages.";
+	say "Compare alternate charsets:";
+}
+else {
+	say "Charset comparison:";
+}
+
+sub optionlink {
+	my ($title, $href, $selected) = @_;
+	return sprintf(
+		$selected ? '<strong>%s</strong>' : '<a href="%2$s">%s</a>',
+		EscapeHTML($title), $href
+	);
+}
+
+print join " •\n", (
+	map {
+		join " ·\n", pairmap {
+			optionlink($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,
+	],
 );
-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} ? $row{offset} < 160 ? 159 : 191 : 127;
-		}
-
-		if ($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 512;
-			$row{set} = 'Unicode planes';
-			$row{cell} = do 'charset-ucplanes.inc.pl';
-		}
-		elsif ($row{set} = resolve_alias($input)) {
-			if ($row{set} eq 'Internal') {
-				$row{table} = ' ' x ($endpoint < 255 ? 640 : 4096);
-				$row{set} = 'Unicode BMP';
-				$row{cell} = do 'charset-unicode.inc.pl';
-			}
-			elsif ($row{set} eq 'utf-8-strict') {
-				$row{table} = undef;
-				$row{set} = 'UTF-8';
-				$row{cell} = do 'charset-utf8.inc.pl';
+:>.
+</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 {
-				$row{table} = decode($row{set}, pack 'C*', $row{offset} .. $endpoint);
+				$row{startpoint} = hex $+{start};
 			}
+			$row{endpoint} = hex($+{end} || 0);
 		}
 		else {
-			print "<p>Encoding $input unknown</p>\n";
+			Alert("Unknown option <q>$param</q> for charset $input");
+		}
+	}
+
+	if ($charset->{setup}) {
+		eval { $charset->{setup}->(\%row) }
+			or Alert("Incomplete setup of $input", $@);
+	}
+
+	if ($row{set}) {}
+	elsif ($row{set} = Encode::resolve_alias($input)) {
+		$row{offset} = delete $row{startpoint};
+		$row{endpoint} ||= 0xFF;
+		if ($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} .. $row{endpoint});
+		}
+
+		if ($row{set} eq 'cp437') {
+			if ($row{offset} <= 0xED and $row{endpoint} >= 0xED) {
+				# replace phi glyph
+				substr($row{table}, 0xED - $row{offset}, 1) = 'ϕ';
+			}
+			if ($row{offset} < 0x20) {
+				# replace control characters by visible variants
+				my $sub = substr ' ☺☻♥♦♣♠•◘○◙♂♀♪♫☼►◄↕‼¶§▬↨↑↓→←∟↔▲▼', $row{offset};
+				substr($row{table}, 0, length $sub) = $sub;
+			}
+		}
+		elsif ($row{set} eq 'symbol') {
+			if ($row{offset} <= 0x60 and $row{endpoint} >= 0x60) {
+				# replace radical extender by closest unicode equivalent
+				substr($row{table}, 0x60 - $row{offset},  1) = '│';
+			}
+			if ($row{offset} <= 0xBD and $row{endpoint} >= 0xFF) {
+				substr($row{table}, 0xBD - $row{offset},  2) = '⏐⎯'; # arrow extenders
+				substr($row{table}, 0xD2 - $row{offset},  3) = '®©™'; # serif variants
+				substr($row{table}, 0xE0 - $row{offset},  1) = '◊'; # replace lookalike, should match AdobeSymbol
+				substr($row{table}, 0xE2 - $row{offset},  3) = '®©™'; # sans-serif variants
+				substr($row{table}, 0xE6 - $row{offset}, 10) = '⎛⎜⎝⎡⎢⎣⎧⎨⎩⎪';
+				substr($row{table}, 0xF0 - $row{offset},  1) = '€';
+				substr($row{table}, 0xF4 - $row{offset}, 11) = '⎮⌡⎞⎟⎠⎤⎥⎦⎫⎬⎭';
+			}
 		}
-		\%row;
+
+		$row{endpoint} -= $row{offset};
+
+		$visible->{ascii} =  # assume common base
+		$visible->{ $row{set} } = 1;
 	}
 	else {
-		();
+		Alert("Encoding <q>$input</q> unknown");
+		return;
 	}
-} map { defined $ALIAS{$_} ? @{ $ALIAS{$_} } : $_ }
-	$ENV{PATH_INFO} =~ /\w/ ? split(m{[/+\s]}, $ENV{PATH_INFO}) : 'default';
+	push @request, \%row;
+}
+tabinput($_) for @tablist;
+
 my $NOCHAR = chr 0xFFFD;
 
-for my $cp437 (grep {$request[$_]->{set} eq 'cp437'} 0 .. $#request) {
-	substr($request[$cp437]->{table}, 237, 1) = pack 'U*', 0x3D5; # phi sign
-	substr($request[$cp437]->{table}, 0, 32) = pack 'U*', map {hex} qw(
-		2007 263A 263B 2665 2666 2663 2660 2022 25D8 25CB 25D9 2642 2640 266A 266B 263C
-		25BA 25C4 2195 203C 00B6 00A7 25AC 21A8 2191 2193 2192 2190 221F 2194 25B2 25BC
-	);
+sub range_cell {
+	my ($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";
 }
 
-my @nibble = (0..9, 'A'..'F');
 for my $row (@request) {
-	printf '<section><table class="glyphs%s">', !$row->{cell} && ' charmap';
-	printf '<caption>%s</caption>', $row->{set};
-	print '<col>' x 17;
+	my $cols = $row->{cols};
+	my $colsize = $row->{cell} && $row->{cell}->{colsize} || 1;
+	my $coldigits = ceil(log($colsize * $cols) / log(16));  # uniform length of hexadecimal header
+	my $rowdiv = 16 ** $coldigits;  # row divide for column digits
+	$rowdiv = 1 if $rowdiv != $cols * $colsize;  # divide only if all columns are matched
+	my $offset = $row->{startpoint} * $colsize || 0;
+
+	printf '<div class="section"><table class="glyphs%s">', !$row->{cell} && ' charmap';
+	my $title = $row->{set};
+	$title .= " <aside>(over $_)</aside>"
+		for $row->{parent} || ();
+	printf '<caption>%s</caption>', $title;
+	print '<col>' x ($cols + 1);
 	for my $section (qw{thead}) {
-		print "<$section><tr><th>↱";
-		print '<th>', $_ for @nibble;
+		print "<$section><tr><th>", $rowdiv == 1 ? '+' : '↱';
+		printf '<th>%0*X', $coldigits, $_ * $colsize for 0 .. $cols - 1;
 		print "\n";
 	}
-	print '<tbody>';
-	for my $msb (0 .. (length($row->{table}) || 256) - 1 >> 4) {
-		printf '<tr><th>%X', $msb + ($row->{offset} >> 4);
-		for my $lsb (0 .. $#nibble) {
-			if ($row->{cell}) {
-				print $row->{cell}->(($msb<<4) + $lsb);
-				next;
-			}
 
-			my $glyph = substr $row->{table}, ($msb<<4) + $lsb, 1;
-			if ($glyph eq $NOCHAR) {
-				print '<td>';
-				next;
-			}
+	print '<tbody>';
+	while ($offset <= $row->{endpoint} * $colsize) {
+		if ($row->{skip}->{$offset}) {
+			$offset += $cols * $colsize;
+			next;
+		}
 
-			my $info = [ord $glyph];
-			if (defined (my $mnem = $di{ord $glyph})) {
-				$info = $diinfo->{$mnem};
+		print '<tr><th>';
+		if (defined $row->{skip}->{$offset}) {
+			print '⋮';
+		}
+		else {
+			if (my $rowmod = $offset % $rowdiv) {
+				# offset in column units
+				printf '<small>+%X</small>', $rowmod;
 			}
 			else {
-				require Unicode::UCD;
-				my $fullinfo = Unicode::UCD::charinfo(ord $glyph);
-				$info = [@$fullinfo{qw/code name category script string/}] if $fullinfo;
+				# divided row offset
+				printf '%X', ($offset + $row->{offset}) / $rowdiv;
 			}
-			my ($codepoint, $name, $prop, $script, $string) = @$info;
+		}
+		say '';
 
-			$glyph = EscapeHTML($string || $glyph);
-			my $desc = sprintf 'U+%04X%s', $codepoint, $name && " ($name)";
-			my @class = ('X', grep {$_} $prop, $script);
+		for (1 .. $cols) {
+			if ($row->{cell}) {
+				print range_cell($row, $offset);
+				next;
+			}
 
-			$glyph = "<span>$glyph</span>" if $prop eq 'Zs';
+			my $cp = $offset + $row->{offset};
+			my $glyph = ref $row->{table} eq 'ARRAY' ? $row->{table}->[$offset] :
+				substr $row->{table}, $offset, 1;
+			my ($cell, $name, $class) = $glyph eq $NOCHAR ? () :
+				$glyphs->glyph_html($glyph);
+
+			if ($mode) {
+				state $visible = {};
+				$class = (
+					$cp == ord $glyph ? 'l4' :
+					$row->{parent} && $glyph eq
+						Encode::decode($row->{parent}, pack 'C', $cp) ? 'l3' :
+					!$class ? undef :
+					$visible->{$glyph} ? 'l2' :
+					'l1'
+				);
+				$visible->{$glyph}++;
+			}
 
-			printf "\n".'<td class="%s" title="%s">%s',
-				join(' ', @class), EscapeHTML($desc), $glyph;
+			say sprintf $class ? '<td title="%s" class="X %s">%s' : '<td title="%s">',
+				$name, $class, $cell;
+		}
+		continue {
+			$offset += $colsize;
 		}
-		print "\n";
 	}
-	print "</table></section>\n";
+	say '</table></div>';
 }
 
 :>
 <hr>
 
 <div class="legend">
-	<table class="glyphs"><tr>
+	<table class="glyphs"><tr><: if ($mode) { :>
+	<td class="X l4">unicode
+	<td class="X l3">inherited
+	<td class="X l2">existing
+	<td class="X l1">original
+	<td class="">unassigned
+<: } else { :>
 	<td class="X Cc">control
 	<td class="X Zs"><span>whitespace</span>
 	<td class="X Mn">diacritic<table class="glyphs"><tr>
@@ -183,11 +392,11 @@ for my $row (@request) {
 	</table>
 
 	<table class="glyphs"><tr>
-	<td class="X">unicode 5.0
+	<td class="X">unicode 7.0
 	<td class="X Xr">proposed
 	<td class="X Xd">deprecated
 	<td class="">unassigned
 	<td class="X Xi">invalid
-	</table>
+<: } :>	</table>
 </div>