#!/usr/bin/env perl

use strict;
use warnings;
use utf8;

use open OUT => ':utf8', ':std';

our $VERSION = '1.06';

# import and combine various digraph data
my $rfc = do 'data/digraphs-rfc.inc.pl'
	or die "error reading digraphs include: ", $@ // $!;
my $extra = do 'data/digraphs-shiar.inc.pl'
	or warn "could not include shiar proposals: ", $@ // $!;
my $vim = do 'data/digraphs-vim.inc.pl'
	or warn "could not include vim extensions ", $@ // $!;
my $vimold = do 'data/digraphs-vim-74.inc.pl'
	or warn "could not include vim compatibility ", $@ // $!;
my $di = { %{$vim // {}}, %{$rfc}, %{$extra // {}} };

# optionally get unicode character information
my $uninfo = do 'unicode-char.inc.pl'
	or warn "could not include unicode details: ", $@ // $!;

# 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(' ',
	grep { !defined $di->{$_} }
	map { substr($_, 1, 1).substr($_, 0, 1) } sort keys %{$di}
);
printf "q{%s}=>[%s],\n", s/(?=[\\}])/\\/gr, join(',',
	ord $di->{$_},   # original code point
	map {"'$_'"}
	$uninfo->{ $di->{$_} }->[1] // '',  # name
	join(' ',
		$rfc->{$_}
			? $vim->{$_} ? 'l4' : 'l1'  # vim+rfc or rfc only
			: $vim->{$_} ? 'l3' : 'l2', # vim only or neither
		$vimold && $vim->{$_} && !$vimold->{$_} ? 'ex' : (), # new vim feature
	),
	($uninfo->{ $di->{$_} }->[0] // '') =~ s/ u-di| u-prop| ex//gr,  # class
	$uninfo->{ $di->{$_} }->[4] // (),  # string
) for sort keys %{$di};
print "}\n";

__END__

=head1 NAME

mkdigraphlist - Output character list of combined digraph data

=head1 SYNOPSIS

    mkdigraphlist >digraphs.inc.pl
    perl -e'$di = do "digraphs.inc.pl"; print chr $di->{DO}->[0]'

=head1 DESCRIPTION

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
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, usage classes, unicode classes,
and replacement output string.
For example:

 +{
   AE => [198, 'LATIN CAPITAL LETTER AE', 'u-di', 'Latin Lu Xl u-v11'],
   EA => 'AE',
  }

=head1 AUTHOR

Mischa POSLAWSKY <perl@shiar.org>

=head1 LICENSE

Licensed under the GNU Affero General Public License version 3.