X-Git-Url: http://git.shiar.nl/sheet.git/blobdiff_plain/dd5b69ff6ef39a7b8521bac32f2a1aa8ad327729..8521aa8e17b8e10fd0d9450d474d8ab06a8c2b73:/tools/mkcharinfo

diff --git a/tools/mkcharinfo b/tools/mkcharinfo
index b673105..988d7b1 100755
--- a/tools/mkcharinfo
+++ b/tools/mkcharinfo
@@ -8,12 +8,16 @@ no if $] >= 5.018, warnings => 'experimental::smartmatch';
 use open OUT => ':utf8', ':std';
 use Data::Dump 'pp';
 
-our $VERSION = '1.00';
+our $VERSION = '1.01';
 
 my %info = (
+	# prepare presentational string for some control(lish) entries
 	"\xAD"     => {string => '-'},
 	"\x{200E}" => {string => '→'},
 	"\x{200F}" => {string => '←'},
+	"\x{200B}" => {string => '␣'}, # nbsp: ~ in TeX
+	"\x{200C}" => {string => '|'}, # ISO-9995-7-081 lookalike (alt: ∣ ⊺ ⟙)
+	"\x{200D}" => {string => '⁀'}, # join (alt: ∤ |ͯ ⨝)
 );
 $info{chr $_} //= {} for 32 .. 126;
 
@@ -43,24 +47,32 @@ eval {
 } or warn "Failed importing html entities: $@";
 
 my %diinc = (
-	'digraphs.inc.pl' => 'u-di',
+	'data/digraphs-rfc.inc.pl' => 'u-di',
+	'data/digraphs-shiar.inc.pl' => 'u-prop ex',
+	'data/digraphs-vim.inc.pl' => 'u-prop',
 );
-for (keys %diinc) {
+for (sort keys %diinc) {
 	-e $_ or next;
 	my $di = do $_ or die "Error reading digraphs file $_: ", $@ || $!;
-	while (my ($mnem, $cp) = each %$di) {
+	for my $mnem (sort keys %{$di}) {
+		my $cp = $di->{$mnem};
 		length $mnem == 2 or next;  # limit to digraphs
 		my $class = $diinc{$_};
-		if (ref $cp) {
-			# old style array
-			$class = 'u-prop' if $cp->[2] and $cp->[2] =~ m/\bXz\b/;
-			$cp = chr $cp->[0];
-		}
 		$info{$cp}->{di} //= $mnem;
 		$info{$cp}->{class}->{$class}++;
 	}
 }
 
+eval {
+	# read introducing unicode versions for known characters
+	my $agemap = do 'unicode-age.inc.pl' or die $@ || $!;
+	for my $chr (keys %info) {
+		my $version = $agemap->{ord $chr} or next;
+		$info{$chr}->{class}->{'u-v'.$version}++
+	}
+	1;
+} or warn "Failed including unicode version data $@";
+
 for my $chr (keys %info) {
 	my $cp = ord $chr;
 	# attempt to get unicode character information
@@ -70,7 +82,11 @@ for my $chr (keys %info) {
 			|| { block => '?', category => 'Xn', name => '', script => '' }
 	} or next;
 
-	$info->{$_} = $info{$chr}->{$_} for qw(di html class string);
+	$info->{$_} = $info{$chr}->{$_} for keys %{ $info{$chr} };
+
+	# official digraphs either lose vim flag or gain experimental
+	delete $info->{class}->{'u-prop'} or $info->{class}->{ex}++
+		if $info->{class}->{'u-di'};
 
 	# categorise by unicode types and writing script
 	$info->{class}->{$_}++ for $info->{category};
@@ -80,22 +96,24 @@ for my $chr (keys %info) {
 	$info->{class}->{Xa}++ if $info->{block} eq 'Basic Latin';
 	$info->{class}->{Xl}++ if $info->{block} eq 'Latin-1 Supplement';
 
-	given ($cp) {
-		when ($info->{string}) {
+	{
+		if ($info->{string}) {
 			# keep predefined presentational string
 		}
-		when ($info->{combining}) {
+		elsif ($info->{combining}) {
 			# overlay combining accents
 			$info->{string} = chr(9676) . $chr;
 		}
-		when (($cp & ~0b1001_1111) == 0 or $cp == 127) {
+		elsif (($cp & ~0b1001_1111) == 0 or $cp == 127) {
 			# control characters (first 32 chars from 0 and 128)
 			# rename to something more descriptive
 			$info->{name} = $info->{unicode10}
 				? '<'.$info->{unicode10}.'>'  # the old name was much more useful
 				: sprintf('<control U+%04X>', $cp);  # at least identify by value
 			# show descriptive symbols instead of control chars themselves
-			$info->{string} = $cp < 32 ? chr($cp + 0x2400) : chr(0xFFFD);
+			$info->{string} = $cp < 32   ? chr($cp + 0x2400) :
+			                  $cp == 127 ? chr(0x2421) :
+			                               chr(0xFFFD);
 		}
 	}
 
@@ -103,6 +121,7 @@ for my $chr (keys %info) {
 }
 
 # output perl code of hash
+say "# automatically generated by $0";
 say 'use utf8;';
 say '+{';
 for my $cp (sort keys %info) {