common: bump version to 1.11
[sheet.git] / charset.plp
index 5668225..f9c50d8 100644 (file)
@@ -1,8 +1,11 @@
 <(common.inc.plp)><:
 
+my $mode = exists $get{compare};
+my @tablist = split m{/+}, $Request || 'default';
+
 Html({
        title => 'charset cheat sheet',
-       version => '1.0',
+       version => '1.2',
        description => [
                "Reference sheet with all glyphs in common character encoding tables,",
                "and an overview of Unicode ranges and UTF-8 bytes.",
@@ -12,13 +15,62 @@ Html({
                reference common overview table
        '],
        stylesheet => [qw'light'],
-       data => [qw'charset-unicode.inc.pl charset-utf8.inc.pl'],
+       data => [qw(
+               charset-encoding.inc.pl
+               charset-unicode.inc.pl charset-ucplanes.inc.pl charset-utf8.inc.pl
+       )],
 });
 
-my @tablist = split /[^\w-]+/, $Request || 'default';
+use List::Util qw( first pairmap pairfirst pairs );
 
 :>
-<h1>Character encoding</h1>
+<h1>Character encodings</h1>
+
+<p>
+<:
+if ($tablist[0] eq 'unicode') {
+       say "Detailed allocation of Unicode blocks.";
+       print "See ", showlink("charsets", '/charset'), " to compare specific encodings";
+}
+else {
+
+if ($tablist[0] eq 'default') {
+       say "Overview of Unicode ", showlink("allocation", '/charset/unicode');
+       say "and common latin code pages.";
+       say "Compare alternate charsets:";
+}
+else {
+       say "Charset comparison:";
+}
+
+print join " •\n", (
+       map {
+               join " ·\n", pairmap {
+                       showlink($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 );
@@ -26,198 +78,136 @@ 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} || {};
 
-       state $ALIAS = {
-               default    => [qw( unicode+0-192 utf-8 iso-8859-1 iso-8859-15 cp1252 cp437 cp850 )],
-               us         => [qw( cp437 cp863 AdobeStandardEncoding gsm0338 )],
-               ebcdic     => [qw( cp37 cp500 cp875 cp1026 cp1047 posix-bc )],
-               iso        => [map {"iso-8859-$_"} 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 )],
-
-               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 )],
-               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 )], # MacUkrainian is broken
-               arabic     => [qw( iso-8859-6 cp1006 cp864 cp1256 MacArabic )], # MacFarsi same as MacArabic?
-               greek      => [qw( iso-8859-7 cp1253 MacGreek cp737 cp869 )],
-               hebrew     => [qw( iso-8859-8 cp1255 MacHebrew cp862 )],
-               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;
+       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 = {};  # all present tables
+       state $visible = {'' => 1};  # all present tables
        my %row = (offset => 0, cols => 16);
-       my $endpoint = 255;
-       my $params = $input =~ s/[+](.*)\z// ? $1 : undef;
 
        if (not defined $params) {
-               use List::Util qw( first pairfirst pairs );
-
-               state $INHERIT = {
-                       'cp437'       => ['cp850' => 0], # ascii range overridden later
-                       'gsm0338'     => ['ascii' => '0-127'],
-
-                       'iso-8859-2'  => ['iso-8859-1' => '160'],
-                       'iso-8859-3'  => ['iso-8859-1' => '160'],
-                       'iso-8859-4'  => ['iso-8859-2' => '160'],
-                       'iso-8859-5'  => ['iso-8859-1' => '160'],
-                       'iso-8859-6'  => ['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'],
-                       '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'],
-                       'cp1254'      => ['iso-8859-9' => '128-159'],
-                       'cp874'       => ['iso-8859-11' => '128-159'], # windows-874 actually cp1162
-                       'cp1257'      => ['iso-8859-13' => '128'],
-                       'cp1251'      => ['cp1252' => '128'],
-                       'cp1253'      => ['cp1252' => '128'],
-                       'cp1255'      => ['iso-8859-8' => '128-223', 'cp1252' => '128'],
-                       'cp1256'      => ['cp1252' => '128'],
-
-                       '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' => '128-239', 'cp437' => '128'],
-                       'cp775'       => ['cp850' => '128'],  # partial cp437
-                       'cp866'       => ['cp437' => '128'],
-                       'cp855'       => ['cp437' => '128'],
-                       'cp1006'      => ['cp437' => '128'],
-                       'cp737'       => ['cp437' => '128'],
-                       'cp869'       => ['cp437' => '128'],
-                       'cp862'       => ['cp437' => '128'],
-                       'cp864'       => ['cp437' => '128'],
-
-                       'koi8-u'      => ['koi8-r' => 128],
-                       'koi8-f'      => ['koi8-u' => 128],
-
-                       'MacRomanian' => ['MacRoman' => '160'],
-                       'MacCroatian' => ['MacRoman' => '160'],
-                       'MacCentralEurRoman' => ['MacRoman' => '128'],
-                       'MacTurkish'  => ['MacRoman' => '208-223'], # F5 is unassigned
-                       'MacCyrillic' => ['MacRoman' => '128'],
-                       'MacHebrew'   => ['MacRoman' => '128'],
-               };
-
-               my @parents = @{ $INHERIT->{$input} || [] };
+               my @parents = @{ $charset->{inherit} || [] };
 
                if (my ($parent, $part) = pairfirst { defined $visible->{$a} } @parents) {
                        $row{parent} = $parent;
-                       $params = $visible->{$parent} ? $part : 128;
+                       $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] || 128;
-                       $params = 128 if $params >= 128;  # ascii offset at most
+                       $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};
+               $visible->{$_} //= 0 for $row{parent} || ();
        }
 
-       if (defined $params and $params =~ m/^ (\d+) (-\d+)? /x) {
-               $row{offset} = $1;
-               $endpoint = -$2 if $2;
+       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 ($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{len} = length $row{table};
-               $row{set} = sprintf 'Unicode block U+%02Xxx', $start >> 8;
+       if ($charset->{setup}) {
+               eval { $charset->{setup}->(\%row) }
+                       or Alert("Incomplete setup of $input", $@);
        }
-       elsif ($input eq 'U') {
-               $row{len} = 1024;
-               $row{set} = 'Unicode planes';
-               $row{cell} = do 'charset-ucplanes.inc.pl'
-                       or Alert('Table data could not be read', $@ || $!);
-               $row{cols} *= 2;
+       $row{endpoint} ||= 0xFF;
+
+       if (defined $row{table} or defined $row{cell}) {
+               $row{set} //= $input;
        }
        elsif ($row{set} = Encode::resolve_alias($input)) {
-               if ($row{set} eq 'Internal') {
-                       $row{len} = $endpoint < 255 ? 640 : 8192;
-                       $row{set} = 'Unicode BMP';
-                       $row{cell} = do 'charset-unicode.inc.pl'
-                               or Alert('Table data could not be read', $@ || $!);
-               }
-               elsif ($row{set} eq 'utf-8-strict') {
-                       $row{len} = 256;
-                       $row{set} = 'UTF-8';
-                       $row{cell} = do 'charset-utf8.inc.pl'
-                               or Alert('Table data could not be read', $@ || $!);
+               $row{offset} = delete $row{startpoint};
+               if ($charset->{varchar}) {
+                       # array of possibly multiple characters per code point
+                       $row{table} = [
+                               map { Encode::decode($row{set}, pack 'C*', $_) } $row{offset} .. $row{endpoint}
+                       ];
                }
                else {
-                       if ($row{set} eq 'MacHebrew') {
-                               # array of possibly multiple characters per code point
-                               $row{table} = [
-                                       map { Encode::decode($row{set}, pack 'C*', $_) } $row{offset} .. $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} .. $endpoint);
-                       }
-                       $row{len} = $endpoint - $row{offset};
-
-                       if ($row{set} eq 'cp437' and !$row{offset}) {
-                               substr($row{table}, 237, 1) = pack 'U*', 0x3D5; # phi sign
-                               substr($row{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
-                               );
-                       }
-
-                       $visible->{ascii} =  # assume common base
-                       $visible->{ $row{set} } = 1;
+                       # ~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};
+               $visible->{ascii}++;  # assume common base
        }
        else {
                Alert("Encoding <q>$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;
+       $visible->{ $row{set} } = 1 if $row{table};
 }
 tabinput($_) for @tablist;
 
@@ -235,22 +225,23 @@ sub range_cell {
        $len /= $colsize;
        $name //= $len <= 2 ? 'res' : 'reserved';
 
-       if (my $part = $offset/$colsize % $cols) {
+       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 = '';
+                               $name = $separate && '…';
                        }
                        else {
                                # minority on next row
-                               push @next, '"', $title || $name;
+                               push @next, $separate && '"', $title || $name;
                        }
                        $table->{$offset + $colsize*$rest} //= \@next;
                        $class .= ' joind';
@@ -259,11 +250,24 @@ sub range_cell {
        }
        elsif (my $rows = int($len / $cols)) {
                # multiple full rows
+               my $rowsize = $colsize * $cols;
                if ($len -= $rows * $cols) {
                        # partial row remains
-                       $table->{$offset + $colsize*$rows * $cols} //= [$len*$colsize, "$class joinu", '', $title];
+                       $table->{$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;
        }
@@ -272,7 +276,7 @@ sub range_cell {
        $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>$name";
+       return "<td$attr>$name\n";
 }
 
 for my $row (@request) {
@@ -281,13 +285,12 @@ for my $row (@request) {
        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 = 0;
-       my $endpoint = $offset + $row->{len} * $colsize;
+       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 grep { $_ ne 'iso-8859-1' } $row->{parent} // ();
+               for $row->{parent} || ();
        printf '<caption>%s</caption>', $title;
        print '<col>' x ($cols + 1);
        for my $section (qw{thead}) {
@@ -295,10 +298,19 @@ for my $row (@request) {
                printf '<th>%0*X', $coldigits, $_ * $colsize for 0 .. $cols - 1;
                print "\n";
        }
+
        print '<tbody>';
-       while ($offset < $endpoint - 1) {
+       while ($offset <= $row->{endpoint} * $colsize) {
+               if ($row->{skip}->{$offset}) {
+                       $offset += $cols * $colsize;
+                       next;
+               }
+
                print '<tr><th>';
-               {
+               if (defined $row->{skip}->{$offset}) {
+                       print '⋮';
+               }
+               else {
                        if (my $rowmod = $offset % $rowdiv) {
                                # offset in column units
                                printf '<small>+%X</small>', $rowmod;
@@ -308,25 +320,39 @@ for my $row (@request) {
                                printf '%X', ($offset + $row->{offset}) / $rowdiv;
                        }
                }
+               say '';
+
                for (1 .. $cols) {
                        if ($row->{cell}) {
                                print range_cell($row, $offset);
                                next;
                        }
 
+                       my $cp = $offset + $row->{offset};
                        my $glyph = ref $row->{table} eq 'ARRAY' ? $row->{table}->[$offset] :
                                substr $row->{table}, $offset, 1;
-                       if ($glyph eq $NOCHAR) {
-                               print '<td>';
-                               next;
+                       my ($cell, $name, $class) = !defined $glyph || $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}++;
                        }
 
-                       print "\n".$glyphs->glyph_cell($glyph);
+                       say sprintf $class ? '<td title="%s" class="X %s">%s' : '<td title="%s">',
+                               $name, $class, $cell;
                }
                continue {
                        $offset += $colsize;
                }
-               print "\n";
        }
        say '</table></div>';
 }
@@ -335,7 +361,13 @@ for my $row (@request) {
 <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>
@@ -367,11 +399,11 @@ for my $row (@request) {
        </table>
 
        <table class="glyphs"><tr>
-       <td class="X">unicode 7.0
+       <td class="X">unicode 10.0
        <td class="X Xr">proposed
        <td class="X Xd">deprecated
        <td class="">unassigned
        <td class="X Xi">invalid
-       </table>
+<: } :>        </table>
 </div>