ceabd83dcf
AFM metrics always have 1000 font units to a scaled PostScript point, but TTF/OTF doesn't have to. The easiest way (and the one which best avoids unnecessary rounding) is to store the scale in the metrics, and change pswidth.ph to return the width in PostScript points instead of font units. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
63 lines
1.4 KiB
Perl
63 lines
1.4 KiB
Perl
#!/usr/bin/perl
|
|
|
|
use Font::TTF::Font;
|
|
use Font::TTF::Head;
|
|
use Font::TTF::Hmtx;
|
|
use Font::TTF::Cmap;
|
|
use Font::TTF::Maxp;
|
|
use Font::TTF::PSNames;
|
|
use Font::TTF::Post;
|
|
|
|
use strict;
|
|
|
|
sub parse_ttf_file($) {
|
|
my($filename) = @_;
|
|
|
|
my $fontdata = {
|
|
widths => {},
|
|
kern => {}
|
|
};
|
|
|
|
my $f = Font::TTF::Font->open($filename);
|
|
|
|
return undef if (!defined($f));
|
|
|
|
$fontdata->{file} = $filename;
|
|
$fontdata->{type} = defined($f->{' CFF'}) ? 'otf' : 'ttf';
|
|
|
|
$f->{head}->read();
|
|
$fontdata->{scale} = $f->{head}{unitsPerEm};
|
|
|
|
$f->{maxp}->read();
|
|
my $glyphs = $f->{maxp}{numGlyphs};
|
|
|
|
$f->{cmap}->read();
|
|
$f->{hmtx}->read();
|
|
$f->{name}->read();
|
|
$fontdata->{name} = $f->{name}->find_name(6); # PostScript name
|
|
$f->{post}->read();
|
|
my $psglyphs = 0;
|
|
my $psmap = $f->{post}->{VAL};
|
|
$psmap = [] if (!defined($psmap));
|
|
#printf "Glyphs with PostScript names: %d\n", scalar(@$psmap);
|
|
|
|
# Can be done as an array of arrays in case of multiple unicodes to
|
|
# one glyph...
|
|
my @unimap = $f->{cmap}->reverse();
|
|
|
|
for (my $i = 0; $i < $glyphs; $i++) {
|
|
my $width = $f->{hmtx}->{advance}[$i];
|
|
my $psname = $psmap->[$i];
|
|
if (!defined($psname)) {
|
|
$psname = Font::TTF::PSNames::lookup($unimap[$i]);
|
|
}
|
|
next if (!defined($psname) || ($psname eq '.notdef'));
|
|
$fontdata->{widths}{$psname} = $f->{hmtx}->{advance}[$i];
|
|
}
|
|
|
|
$f->release;
|
|
|
|
return $fontdata;
|
|
}
|
|
|
|
1;
|