package Font::TTF::Utils; =head1 NAME Font::TTF::Utils - Utility functions to save fingers =head1 DESCRIPTION Lots of useful functions to save my fingers, especially for trivial tables =head1 FUNCTIONS The following functions are exported =cut use strict; use vars qw(@ISA @EXPORT $VERSION @EXPORT_OK); require Exporter; @ISA = qw(Exporter); @EXPORT = qw(TTF_Init_Fields TTF_Read_Fields TTF_Out_Fields TTF_Pack TTF_Unpack TTF_word_utf8 TTF_utf8_word TTF_bininfo); @EXPORT_OK = (@EXPORT, qw(XML_hexdump)); $VERSION = 0.0001; =head2 ($val, $pos) = TTF_Init_Fields ($str, $pos) Given a field description from the C section, creates an absolute entry in the fields associative array for the class =cut sub TTF_Init_Fields { my ($str, $pos, $inval) = @_; my ($key, $val, $res, $len, $rel); $str =~ s/\r?\n$//o; if ($inval) { ($key, $val) = ($str, $inval); } else { ($key, $val) = split(',\s*', $str); } return (undef, undef, 0) unless (defined $key && $key ne ""); if ($val =~ m/^(\+?)(\d*)(\D+)(\d*)/oi) { $rel = $1; if ($rel eq "+") { $pos += $2; } elsif ($2 ne "") { $pos = $2; } $val = $3; $len = $4; } $len = "" unless defined $len; $pos = 0 if !defined $pos || $pos eq ""; $res = "$pos:$val:$len"; if ($val eq "f" || $val eq 'v' || $val =~ m/^[l]/oi) { $pos += 4 * ($len ne "" ? $len : 1); } elsif ($val eq "F" || $val =~ m/^[s]/oi) { $pos += 2 * ($len ne "" ? $len : 1); } else { $pos += 1 * ($len ne "" ? $len : 1); } ($key, $res, $pos); } =head2 TTF_Read_Fields($obj, $dat, $fields) Given a block of data large enough to account for all the fields in a table, processes the data block to convert to the values in the objects instance variables by name based on the list in the C block which has been run through C =cut sub TTF_Read_Fields { my ($self, $dat, $fields) = @_; my ($pos, $type, $res, $f, $arrlen, $arr, $frac); foreach $f (keys %{$fields}) { ($pos, $type, $arrlen) = split(':', $fields->{$f}); $pos = 0 if $pos eq ""; if ($arrlen ne "") { $self->{$f} = [TTF_Unpack("$type$arrlen", substr($dat, $pos))]; } else { $self->{$f} = TTF_Unpack("$type", substr($dat, $pos)); } } $self; } =head2 TTF_Unpack($fmt, $dat) A TrueType types equivalent of Perls C function. Thus $fmt consists of type followed by an optional number of elements to read including *. The type may be one of: c BYTE C CHAR f FIXED F F2DOT14 l LONG L ULONG s SHORT S USHORT v Version number (FIXED) Note that C, C and C are not data types but units. Returns array of scalar (first element) depending on context =cut sub TTF_Unpack { my ($fmt, $dat) = @_; my ($res, $frac, $i, $arrlen, $type, @res); while ($fmt =~ s/^([cflsv])(\d+|\*)?//oi) { $type = $1; $arrlen = $2; $arrlen = 1 if !defined $arrlen || $arrlen eq ""; $arrlen = -1 if $arrlen eq "*"; for ($i = 0; ($arrlen == -1 && $dat ne "") || $i < $arrlen; $i++) { if ($type eq "f") { ($res, $frac) = unpack("nn", $dat); substr($dat, 0, 4) = ""; $res -= 65536 if $res > 32767; $res += $frac / 65536.; } elsif ($type eq "v") { ($res, $frac) = unpack("nn", $dat); substr($dat, 0, 4) = ""; $res = sprintf("%d.%04X", $res, $frac); } elsif ($type eq "F") { $res = unpack("n", $dat); substr($dat, 0, 2) = ""; # $res -= 65536 if $res >= 32768; $frac = $res & 0x3fff; $res >>= 14; $res -= 4 if $res > 1; # $frac -= 16384 if $frac > 8191; $res += $frac / 16384.; } elsif ($type =~ m/^[l]/oi) { $res = unpack("N", $dat); substr($dat, 0, 4) = ""; $res -= (1 << 32) if ($type eq "l" && $res >= 1 << 31); } elsif ($type =~ m/^[s]/oi) { $res = unpack("n", $dat); substr($dat, 0, 2) = ""; $res -= 65536 if ($type eq "s" && $res >= 32768); } elsif ($type eq "c") { $res = unpack("c", $dat); substr($dat, 0, 1) = ""; } else { $res = unpack("C", $dat); substr($dat, 0, 1) = ""; } push (@res, $res); } } return wantarray ? @res : $res[0]; } =head2 $dat = TTF_Out_Fields($obj, $fields, $len) Given the fields table from C writes out the instance variables from the object to the filehandle in TTF binary form. =cut sub TTF_Out_Fields { my ($obj, $fields, $len) = @_; my ($dat) = "\000" x $len; my ($f, $pos, $type, $res, $arr, $arrlen, $frac); foreach $f (keys %{$fields}) { ($pos, $type, $arrlen) = split(':', $fields->{$f}); if ($arrlen ne "") { $res = TTF_Pack("$type$arrlen", @{$obj->{$f}}); } else { $res = TTF_Pack("$type", $obj->{$f}); } substr($dat, $pos, length($res)) = $res; } $dat; } =head2 $dat = TTF_Pack($fmt, @data) The TrueType equivalent to Perl's C function. See details of C for how to work the $fmt string. =cut sub TTF_Pack { my ($fmt, @obj) = @_; my ($type, $i, $arrlen, $dat, $res, $frac); $dat = ''; while ($fmt =~ s/^([flscv])(\d+|\*)?//oi) { $type = $1; $arrlen = $2 || ""; $arrlen = $#obj + 1 if $arrlen eq "*"; $arrlen = 1 if $arrlen eq ""; for ($i = 0; $i < $arrlen; $i++) { $res = shift(@obj) || 0; if ($type eq "f") { $frac = int(($res - int($res)) * 65536); $res = (int($res) << 16) + $frac; $dat .= pack("N", $res); } elsif ($type eq "v") { if ($res =~ s/\.([0-9a-f]+)$//oi) { $frac = $1; $frac .= "0" x (4 - length($frac)); } else { $frac = 0; } $dat .= pack('nn', $res, hex($frac)); } elsif ($type eq "F") { $frac = int(($res - int($res)) * 16384); $res = (int($res) << 14) + $frac; $dat .= pack("n", $res); } elsif ($type =~ m/^[l]/oi) { $res += 1 << 32 if ($type eq 'L' && $res < 0); $dat .= pack("N", $res); } elsif ($type =~ m/^[s]/oi) { $res += 1 << 16 if ($type eq 'S' && $res < 0); $dat .= pack("n", $res); } elsif ($type eq "c") { $dat .= pack("c", $res); } else { $dat .= pack("C", $res); } } } $dat; } =head2 ($num, $range, $select, $shift) = TTF_bininfo($num) Calculates binary search information from a number of elements =cut sub TTF_bininfo { my ($num, $block) = @_; my ($range, $select, $shift); $range = 1; for ($select = 0; $range <= $num; $select++) { $range *= 2; } $select--; $range /= 2; $range *= $block; $shift = $num * $block - $range; ($num, $range, $select, $shift); } =head2 TTF_word_utf8($str) Returns the UTF8 form of the 16 bit string, assumed to be in big endian order, including surrogate handling =cut sub TTF_word_utf8 { my ($str) = @_; my ($res, $i); my (@dat) = unpack("n*", $str); return pack("U*", @dat) if ($] >= 5.006); for ($i = 0; $i <= $#dat; $i++) { my ($dat) = $dat[$i]; if ($dat < 0x80) # Thanks to Gisle Aas for some of his old code { $res .= chr($dat); } elsif ($dat < 0x800) { $res .= chr(0xC0 | ($dat >> 6)) . chr(0x80 | ($dat & 0x3F)); } elsif ($dat >= 0xD800 && $dat < 0xDC00) { my ($dat1) = $dat[++$i]; my ($top) = (($dat & 0x3C0) >> 6) + 1; $res .= chr(0xF0 | ($top >> 2)) . chr(0x80 | (($top & 1) << 4) | (($dat & 0x3C) >> 2)) . chr(0x80 | (($dat & 0x3) << 4) | (($dat1 & 0x3C0) >> 6)) . chr(0x80 | ($dat1 & 0x3F)); } else { $res .= chr(0xE0 | ($dat >> 12)) . chr(0x80 | (($dat >> 6) & 0x3F)) . chr(0x80 | ($dat & 0x3F)); } } $res; } =head2 TTF_utf8_word($str) Returns the 16-bit form in big endian order of the UTF 8 string, including surrogate handling to Unicode. =cut sub TTF_utf8_word { my ($str) = @_; my ($res); return pack("n*", unpack("U*", $str)) if ($^V ge v5.6.0); $str = "$str"; # copy $str while (length($str)) # Thanks to Gisle Aas for some of his old code { $str =~ s/^[\x80-\xBF]+//o; if ($str =~ s/^([\x00-\x7F]+)//o) { $res .= pack("n*", unpack("C*", $1)); } elsif ($str =~ s/^([\xC0-\xDF])([\x80-\xBF])//o) { $res .= pack("n", ((ord($1) & 0x1F) << 6) | (ord($2) & 0x3F)); } elsif ($str =~ s/^([\0xE0-\xEF])([\x80-\xBF])([\x80-\xBF])//o) { $res .= pack("n", ((ord($1) & 0x0F) << 12) | ((ord($2) & 0x3F) << 6) | (ord($3) & 0x3F)); } elsif ($str =~ s/^([\xF0-\xF7])([\x80-\xBF])([\x80-\xBF])([\x80-\xBF])//o) { my ($b1, $b2, $b3, $b4) = (ord($1), ord($2), ord($3), ord($4)); $res .= pack("n", ((($b1 & 0x07) << 8) | (($b2 & 0x3F) << 2) | (($b3 & 0x30) >> 4)) + 0xD600); # account for offset $res .= pack("n", ((($b3 & 0x0F) << 6) | ($b4 & 0x3F)) + 0xDC00); } elsif ($str =~ s/^[\xF8-\xFF][\x80-\xBF]*//o) { } } $res; } =head2 XML_hexdump($context, $dat) Dumps out the given data as a sequence of blocks each 16 bytes wide =cut sub XML_hexdump { my ($context, $depth, $dat) = @_; my ($fh) = $context->{'fh'}; my ($i, $len, $out); $len = length($dat); for ($i = 0; $i < $len; $i += 16) { $out = join(' ', map {sprintf("%02X", ord($_))} (split('', substr($dat, $i, 16)))); $fh->printf("%s%s\n", $depth, $i, $out); } } =head2 XML_outhints Converts a binary string of hinting code into a textual representation =cut { my (@hints) = ( ['SVTCA[0]'], ['SVTCA[1]'], ['SPVTCA[0]'], ['SPVTCA[1]'], ['SFVTCA[0]'], ['SFVTCA[1]'], ['SPVTL[0]'], ['SPVTL[1]'], ['SFVTL[0]'], ['SFVTL[1]'], ['SPVFS'], ['SFVFS'], ['GPV'], ['GFV'], ['SVFTPV'], ['ISECT'], # 10 ['SRP0'], ['SRP1'], ['SRP2'], ['SZP0'], ['SZP1'], ['SZP2'], ['SZPS'], ['SLOOP'], ['RTG'], ['RTHG'], ['SMD'], ['ELSE'], ['JMPR'], ['SCVTCI'], ['SSWCI'], ['SSW'], # 20 ['DUP'], ['POP'], ['CLEAR'], ['SWAP'], ['DEPTH'], ['CINDEX'], ['MINDEX'], ['ALIGNPTS'], [], ['UTP'], ['LOOPCALL'], ['CALL'], ['FDEF'], ['ENDF'], ['MDAP[0]'], ['MDAP[1]'], # 30 ['IUP[0]'], ['IUP[1]'], ['SHP[0]'], ['SHP[1]'], ['SHC[0]'], ['SHC[1]'], ['SHZ[0]'], ['SHZ[1]'], ['SHPIX'], ['IP'], ['MSIRP[0]'], ['MSIRP[1]'], ['ALIGNRP'], ['RTDG'], ['MIAP[0]'], ['MIAP[1]'], # 40 ['NPUSHB', -1, 1], ['NPUSHW', -1, 2], ['WS', 0, 0], ['RS', 0, 0], ['WCVTP', 0, 0], ['RCVT', 0, 0], ['GC[0]'], ['GC[1]'], ['SCFS'], ['MD[0]'], ['MD[1]'], ['MPPEM'], ['MPS'], ['FLIPON'], ['FLIPOFF'], ['DEBUG'], # 50 ['LT'], ['LTEQ'], ['GT'], ['GTEQ'], ['EQ'], ['NEQ'], ['ODD'], ['EVEN'], ['IF'], ['EIF'], ['AND'], ['OR'], ['NOT'], ['DELTAP1'], ['SDB'], ['SDS'], # 60 ['ADD'], ['SUB'], ['DIV'], ['MULT'], ['ABS'], ['NEG'], ['FLOOR'], ['CEILING'], ['ROUND[0]'], ['ROUND[1]'], ['ROUND[2]'], ['ROUND[3]'], ['NROUND[0]'], ['NROUND[1]'], ['NROUND[2]'], ['NROUND[3]'], # 70 ['WCVTF'], ['DELTAP2'], ['DELTAP3'], ['DELTAC1'], ['DELTAC2'], ['DELTAC3'], ['SROUND'], ['S45ROUND'], ['JROT'], ['JROF'], ['ROFF'], [], ['RUTG'], ['RDTG'], ['SANGW'], [], # 80 ['FLIPPT'], ['FLIPRGON'], ['FLIPRGOFF'], [], [], ['SCANCTRL'], ['SDPVTL[0]'], ['SDPVTL[1]'], ['GETINFO'], ['IDEF'], ['ROLL'], ['MAX'], ['MIN'], ['SCANTYPE'], ['INSTCTRL'], [], # 90 [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], # A0 [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], # B0 ['PUSHB1', 1, 1], ['PUSHB2', 2, 1], ['PUSHB3', 3, 1], ['PUSHB4', 4, 1], ['PUSHB5', 5, 1], ['PUSHB6', 6, 1], ['PUSHB7', 7, 1], ['PUSHB8', 8, 1], ['PUSHW1', 1, 2], ['PUSHW2', 2, 2], ['PUSHW3', 3, 2], ['PUSHW4', 4, 2], ['PUSHW5', 5, 2], ['PUSHW6', 6, 2], ['PUSHW7', 7, 2], ['PUSHW8', 8, 2], # C0 ['MDRP[0]'], ['MDRP[1]'], ['MDRP[2]'], ['MDRP[3]'], ['MDRP[4]'], ['MDRP[5]'], ['MDRP[6]'], ['MDRP[7]'], ['MDRP[8]'], ['MDRP[9]'], ['MDRP[A]'], ['MDRP[B]'], ['MDRP[C]'], ['MDRP[D]'], ['MDRP[E]'], ['MDRP[F]'], # D0 ['MDRP[10]'], ['MDRP[11]'], ['MDRP[12]'], ['MDRP[13]'], ['MDRP[14]'], ['MDRP[15]'], ['MDRP[16]'], ['MDRP[17]'], ['MDRP[18]'], ['MDRP[19]'], ['MDRP[1A]'], ['MDRP[1B]'], ['MDRP[1C]'], ['MDRP[1D]'], ['MDRP[1E]'], ['MDRP[1F]'], # E0 ['MIRP[0]'], ['MIRP[1]'], ['MIRP[2]'], ['MIRP[3]'], ['MIRP[4]'], ['MIRP[5]'], ['MIRP[6]'], ['MIRP[7]'], ['MIRP[8]'], ['MIRP[9]'], ['MIRP[A]'], ['MIRP[B]'], ['MIRP[C]'], ['MIRP[D]'], ['MIRP[E]'], ['MIRP[F]'], # F0 ['MIRP[10]'], ['MIRP[11]'], ['MIRP[12]'], ['MIRP[13]'], ['MIRP[14]'], ['MIRP[15]'], ['MIRP[16]'], ['MIRP[17]'], ['MIRP[18]'], ['MIRP[19]'], ['MIRP[1A]'], ['MIRP[1B]'], ['MIRP[1C]'], ['MIRP[1D]'], ['MIRP[1E]'], ['MIRP[1F]']); my ($i); my (%hints) = map { $_->[0] => $i++ if (defined $_->[0]); } @hints; sub XML_binhint { my ($dat) = @_; my ($len) = length($dat); my ($res, $i, $text, $size, $num); for ($i = 0; $i < $len; $i++) { ($text, $num, $size) = @{$hints[ord(substr($dat, $i, 1))]}; $num = 0 unless (defined $num); $text = sprintf("UNK[%02X]", ord(substr($dat, $i, 1))) unless defined $text; $res .= $text; if ($num != 0) { if ($num < 0) { $i++; my ($nnum) = unpack($num == -1 ? 'C' : 'n', substr($dat, $i, -$num)); $i += -$num - 1; $num = $nnum; } $res .= "\t" . join(' ', unpack($size == 1 ? 'C*' : 'n*', substr($dat, $i + 1, $num * $size))); $i += $num * $size; } $res .= "\n"; } $res; } sub XML_hintbin { my ($dat) = @_; my ($l, $res, @words, $num); foreach $l (split(/\s*\n\s*/, $dat)) { @words = split(/\s*/, $l); next unless (defined $hints{$words[0]}); $num = $hints{$words[0]}; $res .= pack('C', $num); if ($hints[$num][1] < 0) { $res .= pack($hints[$num][1] == -1 ? 'C' : 'n', $#words); $res .= pack($hints[$num][2] == 1 ? 'C*' : 'n*', @words[1 .. $#words]); } elsif ($hints[$num][1] > 0) { $res .= pack($hints[$num][2] == 1 ? 'C*' : 'n*', @words[1 .. $hints[$num][1]]); } } $res; } } =head2 make_circle($f, $cmap, [$dia, $sb, $opts]) Adds a dotted circle to a font. This function is very configurable. The parameters passed in are: =over 4 =item $f Font to work with. This is required. =item $cmap A cmap table (not the 'val' sub-element of a cmap) to add the glyph too. Optional. =item $dia Optional diameter for the main circle. Defaults to 80% em =item $sb Side bearing. The left and right side-bearings are always the same. This value defaults to 10% em. =back There are various options to control all sorts of interesting aspects of the circle =over 4 =item numDots Number of dots in the circle =item numPoints Number of curve points to use to create each dot =item uid Unicode reference to store this glyph under in the cmap. Defaults to 0x25CC =item pname Postscript name to give the glyph. Defaults to uni25CC. =item -dRadius Radius of each dot. =back =cut sub make_circle { my ($font, $cmap, $dia, $sb, %opts) = @_; my ($upem) = $font->{'head'}{'unitsPerEm'}; my ($glyph) = Font::TTF::Glyph->new('PARENT' => $font, 'read' => 2, 'isDirty' => 1); my ($PI) = 3.1415926535; my ($R, $r, $xorg, $yorg); my ($i, $j, $numg, $maxp); my ($numc) = $opts{'-numDots'} || 16; my ($nump) = ($opts{'-numPoints'} * 2) || 8; my ($uid) = $opts{'-uid'} || 0x25CC; my ($pname) = $opts{'-pname'} || 'uni25CC'; $dia ||= $upem * .8; # .95 to fit exactly $sb ||= $upem * .1; $R = $dia / 2; $r = $opts{'-dRadius'} || ($R * .1); ($xorg, $yorg) = ($R + $r, $R); $xorg += $sb; $font->{'post'}->read; $font->{'glyf'}->read; for ($i = 0; $i < $numc; $i++) { my ($pxorg, $pyorg) = ($xorg + $R * cos(2 * $PI * $i / $numc), $yorg + $R * sin(2 * $PI * $i / $numc)); for ($j = 0; $j < $nump; $j++) { push (@{$glyph->{'x'}}, int ($pxorg + ($j & 1 ? 1/cos(2*$PI/$nump) : 1) * $r * cos(2 * $PI * $j / $nump))); push (@{$glyph->{'y'}}, int ($pyorg + ($j & 1 ? 1/cos(2*$PI/$nump) : 1) * $r * sin(2 * $PI * $j / $nump))); push (@{$glyph->{'flags'}}, $j & 1 ? 0 : 1); } push (@{$glyph->{'endPoints'}}, $#{$glyph->{'x'}}); } $glyph->{'numberOfContours'} = $#{$glyph->{'endPoints'}} + 1; $glyph->{'numPoints'} = $#{$glyph->{'x'}} + 1; $glyph->update; $numg = $font->{'maxp'}{'numGlyphs'}; $font->{'maxp'}{'numGlyphs'}++; $font->{'hmtx'}{'advance'}[$numg] = int($xorg + $R + $r + $sb + .5); $font->{'hmtx'}{'lsb'}[$numg] = int($xorg - $R - $r + .5); $font->{'loca'}{'glyphs'}[$numg] = $glyph; $cmap->{'val'}{$uid} = $numg if ($cmap); $font->{'post'}{'VAL'}[$numg] = $pname; delete $font->{'hdmx'}; delete $font->{'VDMX'}; delete $font->{'LTSH'}; $font->tables_do(sub {$_[0]->dirty;}); $font->update; return ($numg - 1); } 1; =head1 BUGS No known bugs =head1 AUTHOR Martin Hosken L. =head1 LICENSING Copyright (c) 1998-2016, SIL International (http://www.sil.org) This module is released under the terms of the Artistic License 2.0. For details, see the full text of the license in the file LICENSE. =cut