use warnings;
use utf8;
-use open OUT => ':utf8', ':std';
-use Data::Dumper;
+use open OUT => ':encoding(utf-8)', ':std';
+use JSON ();
-our $VERSION = '1.03';
+our $VERSION = '1.07';
-# create a hash of desired input
-my $di = do 'data/digraphs-rfc.inc.pl'
+# import and combine various digraph data
+push @INC, 'data';
+my $rfc = do 'digraphs-rfc.inc.pl'
or die "error reading digraphs include: ", $@ // $!;
-
-# personal addendums
-my $extra = do 'data/digraphs-shiar.inc.pl'
+my $extra = do 'digraphs-shiar.inc.pl'
or warn "could not include shiar proposals: ", $@ // $!;
-$di = { %{$di}, %{$extra // {}} };
-
-$di->{chr $_} = $_ for 32 .. 126;
-$di->{'\\'.$_} = delete $di->{$_} for '{', '}', '\\';
+my $vim = do 'digraphs-vim.inc.pl'
+ or warn "could not include vim extensions ", $@ // $!;
+my $vimold = do 'digraphs-vim-74.inc.pl'
+ or warn "could not include vim compatibility ", $@ // $!;
+my $di = { %{$vim // {}}, %{$rfc}, %{$extra // {}} };
# optionally get unicode character information
-my %info = eval {
- require Unicode::UCD;
- map {
- $_ => Unicode::UCD::charinfo($di->{$_})
- || { block => '?', category => 'Xn', name => '', script => '' }
- } keys %{$di};
-};
-
-# add custom categories for certain blocks
-for (values %info) {
- $_->{category} .= ' Xa' if $_->{block} eq 'Basic Latin';
- $_->{category} .= ' Xl' if $_->{block} eq 'Latin-1 Supplement';
-}
-
-# mark unofficial extras as such
-$info{$_}->{category} .= ' Xz' for keys %{$extra};
-
-for (keys %{$di}) {
- $info{$_}->{string} = chr(9676) . chr($di->{$_}) if $info{$_}->{combining};
- # find control characters (first 32 chars from 0 and 128)
- next unless ($di->{$_} & ~0b1001_1111) == 0 or $di->{$_} == 127;
- # rename to something more descriptive
- $info{$_}->{name} = $info{$_}->{unicode10}
- ? '<'.$info{$_}->{unicode10}.'>' # the old name was much more useful
- : sprintf('<control U+%04X>', $di->{$_}); # at least identify by value
- # show descriptive symbols instead of control chars themselves
- $info{$_}->{string} = $di->{$_} < 32 ? chr($di->{$_} + 0x2400) : chr(0xFFFD);
-}
-# presentational string for some control(lish) entries
-$info{$_}->{string} = '-' for grep { $di->{$_} == 0x00AD } keys %{$di};
-$info{$_}->{string} = '␣' for grep { $di->{$_} == 0x200B } keys %{$di};
-$info{$_}->{string} = '|' for grep { $di->{$_} == 0x200C } keys %{$di};
-$info{$_}->{string} = '⁀' for grep { $di->{$_} == 0x200D } keys %{$di};
-$info{$_}->{string} = '→' for grep { $di->{$_} == 0x200E } keys %{$di};
-$info{$_}->{string} = '←' for grep { $di->{$_} == 0x200F } keys %{$di};
-
-# convert info hashes into arrays of strings to output in display order
-for my $row (values %info) {
- $row = [ map { $row->{$_} } qw/name category script string/ ];
- # strip off trailing missing values (especially string may be unknown)
- defined $row->[-1] ? last : pop @$row for 1 .. @$row;
-}
-
-# output perl code of hash
-# (assume no backslashes or curlies, so we can just q{} w/o escaping)
-print "# automatically generated by $0\n";
-print "use utf8;\n";
-print "+{\n";
-printf '(map {$_=>0} qw{%s}),'."\n", join(' ',
+my $uninfo = do 'unicode-char.inc.pl'
+ or warn "could not include unicode details: ", $@ // $!;
+
+# output json map of character info
+my %table;
+$table{$_} = 0 for (
+ grep { !defined $di->{$_} }
map { substr($_, 1, 1).substr($_, 0, 1) } sort keys %{$di}
);
-printf "q{%s}=>[%s],\n", $_, join(',',
- $di->{$_}, # original code point
- $info{$_} # optional additional arguments
- ? map {"'$_'"} @{ $info{$_} }
- : ()
-) for sort keys %{$di};
-print "}\n";
+$table{$_} = [
+ ord $di->{$_}, # original code point
+ $uninfo->{ $di->{$_} }->[1] // '', # name
+ (
+ $rfc->{$_}
+ ? $vim->{$_} ? 'l5' : 'l1' # vim+rfc or rfc only
+ : $vimold && $vimold->{$_} ? 'l4' # compat vim if known
+ : $vim->{$_} ? 'l3' : 'l2', # vim only or neither
+ ),
+ ($uninfo->{ $di->{$_} }->[0] // '') =~ s/ u-di| u-prop//gr, # class
+ $uninfo->{ $di->{$_} }->[4] // (), # string
+] for sort keys %{$di};
+
+print JSON->new->ascii->canonical->encode({
+ title => 'RFC-1345',
+ key => \%table,
+ intro => join("\n",
+ 'Character mnemonics following compose key ⎄:',
+ 'i^k in <a href="/vi">Vim</a>,',
+ '^u^\ in <a href="/readline">Emacs</a>,',
+ '^a^v in <a href="/screen">Screen</a>.',
+ 'Similar but different from <a href="/digraphs/xorg">X.Org</a>.',
+ 'Also see <a href="/unicode">common Unicode</a>.</p>',
+ '<p class="aside">Unofficial <span class="u-l2">proposals</span>',
+ 'are available as <a href="/digraphs.vim">ex commands</a>.',
+ ),
+ flag => {
+ l5 => 'full support',
+ l4 => 'vim extension',
+ l3 => 'vim v8.0',
+ l2 => 'proposal',
+ l1 => 'not in vim',
+ },
+ flagclass => {
+ l5 => '', # common
+ l3 => 'u-l5', # rare
+ },
+});
__END__
=head1 SYNOPSIS
- mkdigraphlist >digraphs.inc.pl
- perl -e'$di = do "digraphs.inc.pl"; print chr $di->{DO}->[0]'
+ mkdigraphlist | jq -r '.key."DO"[0]' | perl -nE 'say chr' # $
=head1 DESCRIPTION
-Parses the official RFC-1345 document, searching the
-'character mnemonic table' for all digraph definitions.
-If successful, Perl code is output resulting in a hash
-with character data keyed by digraph.
-Any errors and warnings are given at STDERR.
+Combines precompiled digraph includes of rfc (1345), vim, and shiar
+and outputs a complete map including character details and usage classes.
-The value can either be a scalar string containing another
+The C<key> values can either be a scalar string containing another
digraph which can be considered identical (usually inverted),
or an array ref containing at least the resulting character's
Unicode code point value. If available, the following UCD data
-is appended: character name, category, script, and output string.
+is appended: character name, usage classes, unicode classes,
+and replacement output string.
For example:
- +{
- AE => [198, 'LATIN CAPITAL LETTER AE', 'Lu Xl', 'Latin'],
- EA => 'AE',
+ {
+ "key": {
+ "AE" => [198, "LATIN CAPITAL LETTER AE", "u-di", "Latin Lu Xl u-v11"],
+ "EA" => "AE",
+ }
}
=head1 AUTHOR
git.shiar.nl / sheet.git / blobdiff