22

✏️ Editing: gs_ttf.ps

% Copyright (C) 2001-2019 Artifex Software, Inc.
% All Rights Reserved.
%
% This software is provided AS-IS with no warranty, either express or
% implied.
%
% This software is distributed under license and may not be copied,
% modified or distributed except as expressly authorized under the terms
% of the license contained in the file LICENSE in this distribution.
%
% Refer to licensing information at http://www.artifex.com or contact
% Artifex Software, Inc.,  1305 Grant Avenue - Suite 200, Novato,
% CA 94945, U.S.A., +1(415)492-9861, for further information.
%

% Support code for direct use of TrueType fonts.
% (Not needed for Type 42 fonts.)

% Note that if you want to use this file without including the ttfont.dev
% option when you built Ghostscript, you will need to load the following
% files before this one:
%	lib/gs_mgl_e.ps
%	lib/gs_mro_e.ps
%	lib/gs_wan_e.ps

% Thanks to B. Jackowski and GUST (the Polish TeX Users' Group) for
% the glyf-splitting code.

% ---------------- Font loading machinery ---------------- %

% Augment the FONTPATH machinery so it recognizes TrueType fonts.

/.scanfontheaders where {
  pop /.scanfontheaders [
   .scanfontheaders aload pop (\000\001\000\000*) (true*) (wOFF*)
  ] def
} if

% ---------------- Automatic Type 42 generation ---------------- %

% Load a TrueType font from a file as a Type 42 PostScript font.
% The thing that makes this really messy is the handling of encodings.
% There are 2 interacting tables that affect the encoding:
%       'cmap' provides multiple maps from character codes to glyph indices
%       'post' maps glyph indices to glyph names (if present)
% What we need to get out of this is:
%       Encoding mapping character codes to glyph names
%         (the composition of cmap and post)
%       CharStrings mapping glyph names to glyph indices
%         (the inverse of post)
% If the post table is missing, we have to take a guess based on the cmap
% table.

/.loadttfontdict 50 dict dup begin

/orgXUID AladdinEnterprisesXUID def
/maxstring 32764 def    % half the maximum length of a PostScript string,
                        % must be a multiple of 4 (for hmtx / loca / vmtx)

/.invert_encoding   % <array> invert_encoding <dict>
{ dup 256 dict exch
  0 exch 1 exch length 1 sub {     % [] <> i
    dup 3 index exch get           % [] <> i v
    dup /.notdef ne {
      exch 2 index 2 index .knownget {
        dup type /arraytype eq {
        [ exch aload pop counttomark 2 add -1 roll ]
        } {
        exch 2 array astore
        } ifelse
      } if 2 index 3 1 roll put
    } {
      pop pop
    } ifelse
  } for
  exch pop
} bind def

% Make /MacRomanEncodingForTrueType including additional
% characters in Mac OS Roman encoding, which is missing
% in /MacRomanEncoding. See PDF spec 1.7, p. 431 or
% "Inside Macintosh: Text" 1-55, Figure 1-36.
% This is useful to exploit the glyphs via Apple Roman
% TrueType cmap subtable, but not appropriate for information
% interchange.

/MacRomanEncodingForTrueType
  /MacRomanEncoding .findencoding
  dup length array copy
  <<
        /notequal	173
        /infinity	176
        /lessequal	178
        /greaterequal	179
        /partialdiff	182
        /summation	183
        /product	184
        /pi		185
        /integral	186
        /Omega		189
        /radical	195
        /approxequal	197
        /Delta		198
        /lozenge	215
        /Euro		219
        /apple		240
  >> {
    TTFDEBUG { (Extend MacRomanEncodingForTrueType for TrueType: ) =only } if

% check /.notdef to avoid overwriting
    2 index 1 index get dup /.notdef eq {
      TTFDEBUG { (insert /) =only 2 index =only ( @ ) =only 1 index //== exec } if
      pop
      exch 2 index 3 1 roll put
    } {
      TTFDEBUG { (cannot insert /) =only 2 index =only ( @ ) =only 1 index =only ( used for ) =only dup //== exec } if
      pop
      pop pop
    } ifelse
  } forall
  aload pop
  256 packedarray
  5 1 index .registerencoding
.defineencoding

% Define the Macintosh standard mapping from characters to glyph indices.
/MacRomanEncoding dup .findencoding def
/MacGlyphEncoding dup .findencoding def
/MacRomanEncodingForTrueType dup .findencoding def

% Invert the MacRomanEncoding.
/.romanmacdict MacRomanEncodingForTrueType .invert_encoding def
/.latin1isodict ISOLatin1Encoding .invert_encoding def

% Define remapping for misnamed glyphs in TrueType 'post' tables.
% There are probably a lot more than this!
/postremap mark
  /Eoverdot /Edotaccent
  /eoverdot /edotaccent
.dicttomark readonly def

% Array used for fast pre-filling of cmap array
/.array1024z [ 1024 { 0 } repeat ] def

% ---- Utilities ---- %

% Define a serial number for creating unique XUIDs for TrueType fonts.
% We used to use the checkSumAdjustment value from the font, but this is
% not reliable, since some fonts don't set it correctly.
% Note that we must do this in a string to make it immune to save/restore.

% handle 64 bit integer objects
/xuidstring 16#ffffffff 0 gt
{
  <8000000000000000>
}
{
  <80000000>
} ifelse def
/curxuid {		% - curxuid <int>
  .getCPSImode
  //false .setCPSImode
  0 xuidstring { exch 8 bitshift exch add } forall
  % for tbe benefit of pdfwrite/ps2write we want the resulting XUID contents
  % to fit into a 32 bit value
  16#ffffffff 0 gt
  {
    16#ffffffff idiv
  } if
  exch .setCPSImode
} bind def
/nextxuid {		% - nextxuid -
  3 -1 0 {
    xuidstring 1 index 2 copy get dup 255 ne {
      1 add put pop exit
    } if pop 0 put pop
  } for
} bind def

% <string> <index> getu16 <integer>
/getu16 {
  2 copy get 8 bitshift 3 1 roll 1 add get add
} bind def

% <string> <index> gets16 <integer>
/gets16 {
  getu16 16#8000 xor 16#8000 sub
} bind def

% <string> <index> getu32 <integer>
/getu32 {
  2 copy getu16 16 bitshift 3 1 roll 2 add getu16 add
} bind def

% <string> <index> gets32 <integer>
/gets32 {
  2 copy gets16 16 bitshift 3 1 roll 2 add getu16 add
} bind def

% <string|array> <index> getu16 <integer>
/getu16a {
  2 getinterval_from_stringarray 0 getu16
} bind def

% <string|array> <index> gets16 <integer>
/gets16a {
  2 getinterval_from_stringarray 0 gets16
} bind def

% <string|array> <index> getu32a <integer>
/getu32a {
  4 getinterval_from_stringarray 0 getu32
} bind def

% <string|array> <index> gets32a <integer>
/gets32a {
  4 getinterval_from_stringarray 0 gets32
} bind def

16#ffffffff 0 gt {  % 64-bit sign extension
  { /gets32 /gets32a} {
    mark 1 index load aload pop { 16#80000000 xor 16#80000000 sub } aload pop
    //.packtomark exec cvx def
  } bind forall
} if

% <string> <index> <integer> putu16 -
/putu16 {
  3 copy -8 bitshift put
  exch 1 add exch 16#ff and put
} bind def

% <string> <index> <integer> putu32 -
/putu32 {
  3 copy -16 bitshift putu16
  exch 2 add exch 16#ffff and putu16
} bind def

% <nametable> <nameid> findname <string> true
% <nametable> <nameid> findname false
/findname {
  TTFDEBUG { (findname: ) print dup =only  } if
  //false 3 1 roll
  1 index length 0 gt {	% check for zero length name table
    0 1 3 index 2 getu16 1 sub {
                % Stack: false table id index
      12 mul 6 add 2 index exch 12 getinterval
      dup 6 getu16 2 index eq 1 index 8 getu16 0 ne and {
                  % We found the name we want.
        exch pop
                  % Stack: false table record
        dup 10 getu16 2 index 4 getu16 add
        1 index 8 getu16 4 -1 roll 3 1 roll
            3 copy add 1 index length
            le {
              pop
              getinterval exch
                  % Stack: false string record
                  % Check for 8- vs. 16-bit characters.
          is2byte { //true } { dup is2byte2 } ifelse { string2to1 } if //true //null 4 -1 roll exit
            } {
              pop pop pop pop
              //false
                  exit
            } ifelse
      } if pop
    } for
  } if
  pop pop
  TTFDEBUG {
    dup { ( = ) print 1 index //== exec } { ( not found) = } ifelse
  } if
} bind def

% <namerecord> is2byte <bool>
/is2byte {
  dup 0 getu16
  {
    { pop //true }		% Apple Unicode
    { pop //false }		% Macintosh Script manager
    { 1 getu16 1 eq }		% ISO
    { 1 getu16 1 eq }		% Microsoft
    { pop //false }		% default - *MUST* be last
  }
  % check for a valid Platform ID, if we don't get one, assume single byte encoding
  dup length 1 sub
  2 index lt {exch pop dup length 1 sub}{exch} ifelse
  get exec
} bind def

% <string> is2byte2 <bool>
/is2byte2 {
  dup length
  dup 2 mod 0 ne {
    pop pop //false
  } {            % s l
    //true exch  % s b l
    1 sub 0 exch 2 exch {
      2 index exch get 0 ne {
        pop //false exit
      } if
    } for
    exch pop
  } ifelse
} bind def

% <string2> string2to1 <string>
/string2to1 {
  dup length 2 idiv string dup
  0 1 3 index length 1 sub {
    3 index 1 index 2 mul 1 add get put dup
  } for pop exch pop
} bind def

% Each procedure in this dictionary is called as follows:
%       <encodingtable> proc <glyph dictionary>
/cmapformats mark
  0 {		% Apple standard 1-to-1 mapping.
    6 256 getinterval_from_stringarray
    mark 0 3 -1 roll
    { 1 index 1 add } forall pop
    .dicttomark
  } bind
  2 {		% Apple 16bit CJK (ShiftJIS etc)

% /sHK_sz		subHeaderKey_size	% 1 * uint16
    % /sH_sz		subHeader_size		% 4 * uint16
    % /sH_len		subHeader_length
    % /cmapf2_tblen	total table length
    % /cmapf2_lang	language code (not used)
    % /sHKs		subHeaderKeys

/sHK_sz 2 def
    /sH_sz 8 def
    dup 2 getu16a /cmapf2_tblen exch def

dup 4 getu16a /cmapf2_lang exch def

dup 6 256 sHK_sz mul getinterval_from_stringarray /sHKs exch def

0		% initialization value for /sH_len
    0 1 255 {
       sHKs exch
       2 mul getu16a
       1 index	% get current max
       1 index	% get current subHeaderKey
       lt {exch} if pop
    } for
    /sH_len exch def

dup 6 256 sHK_sz mul add
    cmapf2_tblen 1 index sub getinterval_from_stringarray
    /sH_gIA exch def

/cmapf2_glyph_array 0 dict def

/.cmapf2_putGID {
        /cmapf2_ch cmapf2_ch_hi 8 bitshift cmapf2_ch_lo add def
        firstCode cmapf2_ch_lo le
        cmapf2_ch_lo firstCode entryCount add lt
        and { % true: j is inside
            sH_offset idRangeOffset add		% offset to gI
            cmapf2_ch_lo firstCode sub 2 mul	% rel. pos. in range
            add 6 add				% offset in sH_gIA
            sH_gIA exch getu16a
            dup 0 gt { %
                idDelta add
                cmapf2_glyph_array exch cmapf2_ch exch put
            } {
                pop
                % cmapf2_glyph_array cmapf2_ch 0 put
            } ifelse
        } {   % false: j is outside
            % cmapf2_glyph_array cmapf2_ch 0 put
        } ifelse
    } def

16#00 1 16#ff { % hi_byte scan
        /cmapf2_ch_hi exch def
        sHKs cmapf2_ch_hi sHK_sz mul getu16a
        /sH_offset exch def
        sH_gIA sH_offset sH_sz getinterval
            dup 0 getu16a /firstCode exch def
            dup 2 getu16a /entryCount exch def
            dup 4 gets16a /idDelta exch def
            dup 6 getu16a /idRangeOffset exch def
        pop
        sH_offset 0 eq {
           /cmapf2_ch_lo cmapf2_ch_hi def
           /cmapf2_ch_hi 0 def
           .cmapf2_putGID
        } {
           16#00 1 16#ff { % lo_byte scan
               /cmapf2_ch_lo exch def
               .cmapf2_putGID
           } for
        } ifelse
     } for
     pop
     cmapf2_glyph_array
  } bind
  4 {		% Microsoft/Adobe segmented mapping.
    /etab exch def
    /nseg2 etab 6 getu16a def
    14 /endc etab 2 index nseg2 getinterval_from_stringarray def
                % The Apple TrueType documentation omits the 2-byte
                % 'reserved pad' that follows the endCount vector!
    2 add
    nseg2 add /startc etab 2 index nseg2 getinterval_from_stringarray def
    nseg2 add /iddelta etab 2 index nseg2 getinterval_from_stringarray def
    nseg2 add /idroff etab 2 index nseg2 getinterval_from_stringarray def
                % The following hack allows us to properly handle
                % idiosyncratic fonts that start at 0xf000:
    pop
    /firstcode startc 0 getu16a 16#ff00 and dup 16#f000 ne { pop 0 } if def
    /putglyph {
      glyphs code 3 -1 roll put /code code 1 add def
    } bind def
    /glyphs 0 dict def

/numcodes 0 def /code 0 def
    % neg2 is number of segments x 2. A format 4 cmap, with usable
    % mappings cannot (according to the spec) have only 1 segment,
    % since format 4 requires a closing segment with start and end
    % codes being 0xffff. Unfortunately, we have found badly subsetted
    % TTF fonts which have one segment with valid mappings, and no
    % closing segment.
    % In such a case, force the loop to work as though as we have two
    % segments - this works as we don't rely on the terminating segment
    % but use the loop counter to spot completion - effrectively, we
    % ignore the closing segment when it's present, anyway.
    0 2 nseg2 dup 4 lt {pop 4}if 3 sub {
      /i2 exch def
      /scode startc i2 getu16a def
      /ecode endc i2 getu16a def
      % Bug 693538: see above
      scode 0 eq ecode 0 eq and not {
        scode ecode 1 add gt { /ecode scode 1 add def } if % Bug 691326. See above.

numcodes scode firstcode sub
                  % Hack for fonts that have only 0x0000 and 0xf000 ranges
        %dup 16#e000 ge { 255 and } if
        % the previous line is obstructive to CJK fonts, so it was removed
        exch sub 0 .max dup /code exch code exch add def
        ecode scode sub 1 add add numcodes add /numcodes exch def
        /delta iddelta i2 gets16a def
        TTFDEBUG {
          (scode=) print scode =only
          ( ecode=) print ecode =only
          ( delta=) print delta =only
          ( droff=) print idroff i2 getu16 =
        } if
        idroff i2 getu16a dup 0 eq {
          pop scode delta add 65535 and 1 ecode delta add 65535 and
          { putglyph } for
        } {       % The +2 is for the 'reserved pad'.
          /gloff exch 14 nseg2 3 mul add 2 add i2 add add def
          0 1 ecode scode sub {
            2 mul gloff add etab exch getu16a
            dup 0 ne { delta add 65535 and } if putglyph
          } for
        } ifelse
      } if
    } for glyphs /glyphs //null def	% for GC
  } bind
  6 {		% Single interval lookup.
    dup 6 getu16a /firstcode exch def
    dup 8 getu16a /ng exch def
    ng dict
    dup
                % Stack: tab dict dict
                % Fill elements firstcode .. firstcode+nvalue-1 with glyph values
    0 1 ng 1 sub {
      dup firstcode add exch
      2 mul 10 add 4 index exch getu16a 3 copy put pop pop
    } for pop exch pop
  } bind
  12 {		% Microsoft/Adobe segmented mapping.
    /etab exch def
    /glyphs 0 dict def

0 1
    etab 12 getu32a % nGroups
    1 sub {
      /ind exch def
      /gid etab 12 ind mul 8 add 16 add getu32a def % startGlyphID
      etab 12 ind mul 16 add getu32a % startCharCode
      1
      etab 12 ind mul 4 add 16 add getu32a % endCharCode
      {
        gid glyphs 3 1 roll put
        /gid gid 1 add def
      } for
    } for
    glyphs /glyphs //null def	% for GC
  } bind
.dicttomark readonly def                % cmapformats

% <cmaptab> cmapdict <glyph dictionary>
/cmapdict {
  dup 0 getu16a cmapformats exch .knownget {
    TTFDEBUG {
      (cmap: format ) print 1 index 0 getu16a = flush
    } if exec
  } {
    (Can't handle format ) print 0 getu16a = flush
    mark 0 1 255 { dup } for .dicttomark
  } ifelse
  TTFDEBUG {
    (cmap: length=) print dup length = dup ===
  } if
} bind def

/get_from_stringarray % <array|string> <offset> get_from_stringarray <int>
{ 1 index type /stringtype eq {
    get
  } {
    exch {                % o ()
      2 copy length ge {
        length sub
      } {
        exch get exit
      } ifelse
    } forall
  } ifelse
} bind def

/getinterval_from_stringarray % <array|string> <offset> <length> getinterval_from_stringarray <string>
{ % May allocate a string in VM.
  2 index type /stringtype eq {
    getinterval
  } {
    dup 65535 le {
      string exch 0                                 % [] s o p
      4 3 roll {                                    % s o p Si
        dup length                                  % s o p Si lSi
        dup 4 index lt {
          3 index exch sub                          % s o p Si o'
          exch pop 3 1 roll exch pop                % s o' p
        } {                                         % s o p Si lSi
          dup 3 1 roll                              % s o p lSi Si lSi
          4 index sub                               % s o p lSi Si lSi-o
          5 index length 4 index sub                % s o p lSi Si lSi-o ls-p
          2 copy gt { exch } if pop                 % s o p lSi Si minl
          dup 3 1 roll                              % s o p lSi minl Si minl
          5 index exch getinterval                  % s o p lSi minl from
          5 index 4 index 3 index                   % s o p lSi minl from s p minl
          getinterval                               % s o p lSi minl from to
          copy pop                                  % s o p lSi minl
          exch pop add exch pop 0 exch              % s 0 p'
          dup 3 index length ge { exit } if
        } ifelse
      } forall
      pop pop                                       % s
    } {
      [ 4 1 roll
        {
          dup 0 eq {
            pop pop pop exit
          } if
          3 copy
          dup 65535 ge {
            pop 65535
          } if
          getinterval_from_stringarray 4 1 roll
          3 index length sub
          exch
          3 index length add
          exch
        } loop
      ]
    } ifelse
  } ifelse
} bind def

/string_array_size  % <array|string> string_array_size <int>
{ dup type /stringtype eq {
    length
  } {
    0 exch { length add } forall
  } ifelse
} bind def

% Each procedure in this dictionary is called as follows:
%       posttable <<proc>> glyphencoding
/postformats mark
  16#00010000  {	% 258 standard Macintosh glyphs.
    pop MacGlyphEncoding
  }
  16#00020000  {	% Detailed map, required by Microsoft fonts.
    dup dup type /arraytype  eq { 0 get } if length 36 lt {
      TTFDEBUG { (post format 2.0 invalid.) = flush } if
      pop [ ]
    } {
      /postglyphs exch def
      /post_first postglyphs dup type /arraytype eq { 0 get } if def
      post_first 32 getu16 /numglyphs exch def
      % Build names array in the order they occur in the 'post' table
      /postpos numglyphs 2 mul 34 add  def
      /total_length postglyphs //string_array_size exec def

% the number of names in a post table is not declared up front
      % and there are fonts with more names than glyph indices <sigh>
      % so we have to pre-process and find the highest index used.
      % We start with a base value of numglyphs as it's most common,
      % and we can safely have a array that's too large.
      % Rather than parse out the indices twice, we store them in an
      % array on the stack, which we then reuse, overwriting entries
      % as we go later on.
      numglyphs array numglyphs postglyphs
      0 1 numglyphs 1 sub {
        dup 2 mul 34 add 2 index exch 2 //getinterval_from_stringarray exec
        dup 0 get 8 bitshift exch 1 get add
        dup 5 index exch
        4 -1 roll exch put
        258 sub dup 3 index gt
        % if we find a reference to an index higher than the value
        % already on the stack, replace the value on the stack
        { 3 -1 roll pop 2 1 roll} {pop} ifelse
      } for
      pop
      1 add array 0 1 2 index length 1 sub {
        postpos total_length ge {
          % Fill the rest with .notdef
          1 2 index length 1 sub { 1 index exch /.notdef put } for
          exit
        } if
        % No name available, /postnames will be defined as an empty
        % array and the glyph won't get a name attached.
        postglyphs postpos //get_from_stringarray exec
        postglyphs postpos 1 add 2 index
        2 copy add total_length gt {
          TTFDEBUG { (post table ends in the middle of the entry.) print flush } if
          pop pop pop pop
          % Fill the rest with .notdef
          1 2 index length 1 sub { 1 index exch /.notdef put } for
          exit
        } if
        //getinterval_from_stringarray exec cvn
        exch postpos add 1 add /postpos exch def
        2 index 3 1 roll
        put
      } for
      /postnames exch def
        % Some TrueType fonts converted by "Windows Type 1 Installer" have
        % problematic post table including MacGlyphEncoding entries which
        % should be omitted. Such extra entries in the beginning of glyphName
        % array make /Encoding broken. If we find populary predefined
        % ISOLatin1Encoding glyph name in the post table, empty encoding is
        % returned. Some TrueType fonts for Microsoft Windows redefines
        % MacGlyphEncoding glyph name out of predefined range). To permit
        % such fonts, ISOLatin1Encoding is used to find broken post. Bug 689495.
      /.broken_post //false def
      tabdict /name .knownget {
        (Windows Type 1 Installer V1.0) search {
          pop pop pop
          .latin1isodict postnames {
            dup /.notdef ne {
              2 copy known {
                TTFDEBUG { (ignore post table that redefines ISOLatin1Encoding glyph name ) print dup //== exec flush } if
                /.broken_post //true def
                pop exit
              } if
            } if
            pop
          } forall
          pop
        } {
          pop
        } ifelse
      } if
      % Loop through the array of name indices we created above.
      % This loop replaces name index in the array with the name
      % it references.
      0 1 2 index length 1 sub {
        dup 2 index exch get dup 258 lt {
          MacGlyphEncoding exch get
        } {
          258 sub postnames exch get
          % At least some of Microsoft's TrueType fonts use incorrect
          % (Adobe-incompatible) names for some glyphs.
          % Correct for this here.
          postremap 1 index .knownget { exch pop } if
        } ifelse
        2 index 3 1 roll
        put
      } for

.broken_post {
        pop
        [ postnames aload length 1 roll ]
      } if
    }
    ifelse
  } bind
  16#00030000  {	% No map.
    pop [ ]
  } bind
.dicttomark readonly def                % postformats

/call.readtable
{ .readtable
} bind def
/call.readbigtable
{ .readbigtable
} bind def

% Each procedure in this dictionary is called as follows:
%	<file> <length> -proc- <string|array_of_strings>
% Note that each table must have an even length, because of a strange
% Adobe requirement that each sfnts entry have even length.
/readtables mark
        % Ordinary tables
  (head) //call.readtable
  (hhea) 1 index
  (maxp) 1 index
  (name) 1 index
  (OS/2) 1 index
  (post) //call.readbigtable
  (vhea) //call.readtable
        % Big tables
  (cmap) //call.readbigtable
  (GSUB) //call.readbigtable
  (glyf) //call.readbigtable
  (loca) 1 index
  (hmtx) 1 index
  (vmtx) 1 index
  (cvt ) //call.readtable
  (fpgm) 1 index
  (prep) 1 index
.dicttomark
% Normally there would be a 'readonly' here, but the ttf2pf utility wants
% to include the 'kern' table as well, so we leave the readtables dictionary
% writable.
def                % readtables

/readtables_stripped readtables dup length dict copy
dup (loca) { .skiptable } put
dup (glyf) { .skiptable } put
def

% Read a table as a single string.
% <file> <length> .skiptable <string>
/.skiptable {
  pop pop ()
} bind def

% Read a table as a single string.
% <file> <length> .readtable <string>
/.readtable {
  dup dup 1 and add string
                % Stack: f len str
  dup 0 4 -1 roll getinterval
                % Stack: f str str1
        % Because of the absurd PostScript specification that gives an
        % error for reading into an empty string, we have to check for
        % this explicitly here.
  3 -1 roll exch
  dup () ne { readstring } if pop pop
} bind def

% Read a big table (one that may exceed 64K).
% <file> <length> .readbigtable <string[s]>
/.readbigtable {
  dup 65400 lt {
    .readtable
  } {
    currentuserparams /VMReclaim get -2 vmreclaim
    [ 4 2 roll {
                % Stack: mark ... f left
      dup maxstring le { exit } if
      1 index maxstring string readstring pop 3 1 roll maxstring sub
    } loop .readtable ]
    exch vmreclaim
  } ifelse
} bind def

end readonly def                % .loadttfontdict

% <tab> .printtab -
/.printtab {
  dup 0 4 getinterval print ( ) print
  dup 8 getu32 =only ( ) print
  12 getu32 =
} bind def

% <file> <bool> <SubfontID> .loadttfonttables -
% Pushes .loadttfontdict & scratch dict on d-stack.
% Defines f, offsets, tables, tabdict, tabs, tthdrlen.
% Skips loca and glyf if <bool> is true.
/.loadttfonttables {
  //.loadttfontdict begin
  40 dict begin
  /SubfontID exch def
  /load_stripped exch def
  /f exch def
  /offsets f 12 string readstring pop def
  /tthdrlen offsets length def
  load_stripped { readtables_stripped } { readtables } ifelse /readtables_ exch def
  offsets 0 4 getinterval (ttcf) eq {
    offsets 8 getu32 /num_fonts exch def
    SubfontID num_fonts ge {
      QUIET not { (True Type collection contains insufficient fonts.) = } if
      /.loadttfonttables cvx /invalidfont signalerror
    } if
    f SubfontID 4 mul dup 0 ne {
      () /SubFileDecode filter flushfile
    } {
      pop pop
    } ifelse

f 4 string readstring pop
    0 getu32 /ttc_offset exch def

f ttc_offset SubfontID 4 mul sub 16 sub
    dup 0 ne {
      () /SubFileDecode filter flushfile
    } {
      pop pop
    } ifelse
    /offsets f 12 string readstring pop def
  } {
    SubfontID 0 gt {
      QUIET not { (SubfontID > 0 with a True Type file which is not a collection.) = } if
      /.loadttfonttables cvx /invalidfont signalerror
    } if
    /ttc_offset 0 def
  } ifelse
        % Peek because table counter may be incorrect.
  /tables f offsets 4 getu16 16 mul string .peekstring pop def
  /tabdict tables length 16 idiv dict def
        % tabs = tables we want to keep, sorted by file position.
  /tabs [ 0 16 tables length 1 sub {
    tables exch 16 getinterval
    TTFDEBUG { dup //.printtab exec } if
    dup 0 4 getinterval readtables_ 1 index known {
      % put all 0 length tables at 0 to avoid overlap
      1 index 12 getu32 0 eq { 1 index 8 0 putu32 } if
      tabdict exch 2 index put
    } {
      pop pop
    } ifelse
  } for ] {
    exch 8 getu32 exch 8 getu32 lt
  } .sort def
        % In certain malformed TrueType fonts, tables overlap.
        % Truncate tables if necessary.
  0 1 tabs length 2 sub {
    dup tabs exch get exch 1 add tabs exch get
    1 index 8 getu32 2 index 12 getu32 add
    1 index 8 getu32 gt {
      (**** Warning: ) print 1 index 0 4 getinterval print
      ( overlaps ) print dup 0 4 getinterval print
      (, truncating.) = flush
      dup 8 getu32 2 index 8 getu32 sub
      2 index 12 3 -1 roll putu32
    } if pop pop
  } for
} bind executeonly def

% <file> <bool> <SubfontID> .loadwofftables -
% Pushes .loadttfontdict & scratch dict on d-stack.
% Defines f, offsets, tables, tabdict, tabs, tthdrlen.
% Skips loca and glyf if <bool> is true.
/.loadwofftables {
  //.loadttfontdict begin
  40 dict begin
  /SubfontID exch def
  /load_stripped exch def
  /f exch def
  % loading WOFFs "stripped" only works for .findfontvalue - we cannot
  % (currently) use a stripped wOFF font
  load_stripped { readtables_stripped } { readtables } ifelse /readtables_ exch def
  /offsets 12 string def
  /woffhdr f 44 string readstring pop def
  /tthdrlen woffhdr length def
  /ttc_offset 0 def
  woffhdr 0 4 getinterval (wOFF) eq not
  % allow loading wOFFOTTO fonts "stripped" for .findfontvalue
  woffhdr 4 4 getinterval (OTTO) eq load_stripped not and or
  {/.loadwofftables cvx /invalidfont signalerror}
  {
    offsets 0 woffhdr 4 4 getinterval putinterval
    /tables f woffhdr 12 getu16 20 mul string .peekstring pop def
    offsets 4 tables length 20 idiv putu16
    /tabdict tables length 20 idiv dict def
    /tabs [ 0 20 tables length 1 sub
      {
        tables exch 20 getinterval
        TTFDEBUG { dup //.printtab exec } if
        dup  0 4 getinterval readtables_ 1 index known
        {
          % put all 0 length tables at 0 to avoid overlap
          1 index 12 getu32 0 eq { 1 index 8 0 putu32 } if
          tabdict exch 2 index put
         }
         {pop pop}
         ifelse
      }for
    ]
    { exch 4 getu32 exch 4 getu32 lt } .sort
    def
  } ifelse
} bind def

/.file_table_pos_names
mark
/glyf 0
/loca 0
.dicttomark readonly def

% - .readttdata -
% Read data.  Updates offsets, tabs; stores data in tabdict.
/.readttdata {
  /file_table_pos 10 dict def
  /fpos tthdrlen ttc_offset add def
  /sfpos offsets length tabs length 16 mul add def
  offsets 4 tabs length putu16
  0 1 tabs length 1 sub {
    dup tabs exch get
    dup 0 4 getinterval /tname exch def
    dup length 20 eq
    {
      dup 4 getu32 /tpos exch def
      dup 8 getu32 /tclen exch def
      dup 12 getu32 /tlen exch def
    }
    {
      dup 8 getu32 /tpos exch def
      dup 12 getu32 /tlen exch def
      /tclen tlen def
    } ifelse
    load_stripped //.file_table_pos_names tname known and {
        pop
        file_table_pos tname [tpos tlen tclen] put
        tabdict tname () put
        pop
    } {
      dup length 20 eq
      {
        0 16 getinterval dup 8 sfpos putu32
        tabs 3 -2 roll put
      }
      {8 sfpos putu32 pop}
      ifelse
        % Skip data between the end of the previous table and
        % the beginning of this one, if any.
      tpos fpos gt {
        load_stripped {
          % 'setfileposition' is faster for skipping a big data.
          f tpos setfileposition
        } {
          f tpos fpos sub () /SubFileDecode filter dup flushfile closefile
          /fpos tpos def
        } ifelse
      } if
      /ff
      tlen tclen eq not
      {f tclen () /SubFileDecode filter /FlateDecode filter}
      {f tlen  () /SubFileDecode filter}
      ifelse
      def
      ff tlen readtables_ tname get exec
      ff closefile
      tabdict tname 3 -1 roll put
        % Round up the table length to an even value.
      /sfpos sfpos tlen dup 1 and add add def
    } ifelse
    /fpos fpos tclen add def
  } for
} bind def

% Find the string in a list of strings that includes a given index.
% <strings> <index> .findseg <string> <index'>
/.findseg {
  exch {
    dup length 2 index gt { exch exit } if
    length sub
  } forall
} bind def

% - .makesfnts -
% Defines checksum, getloca, head, locatable, numloca, post, sfnts, upem
% Note that the 'loca' table may be out of order.
/.makesfnts {
  //.readttdata exec
  /head tabdict /head get def
  /post tabdict /post .knownget {
    dup 0 get /post_first_part exch def
  } {
    //null
  } ifelse def
  load_stripped not {
    /locatable tabdict /loca get def
    /numloca
      locatable dup type /stringtype eq
       { length }
       { 0 exch { length add } forall }
      ifelse	% no def yet
    locatable type /stringtype eq {
      /.indexloca {} def
    } {
      /.indexloca //.findseg def
    } ifelse
    head 50 getu16 0 ne {
      /getloca {
        2 bitshift locatable exch .indexloca getu32
      } def
      4 idiv 1 sub
    } {
      /getloca {
        dup add locatable exch .indexloca getu16 dup add
      } def
      2 idiv 1 sub
    } ifelse def		% numloca
  } {
    % We did not load loca, take the number of glyphs from maxp.
    /numloca tabdict /maxp get 4 getu16 def
  } ifelse
  /sfnts [
    offsets tabs { concatstrings } forall
    tabs {
      0 4 getinterval tabdict exch get
      dup type /stringtype ne { aload pop } if
    } forall
  ] def
} .bind executeonly odef

/first_post_string % - first_post_string <string>
{
  post dup type /arraytype eq { 0 get } if
} bind def

% - .getpost -
% Uses post, defines glyphencoding
/.getpost {
  /glyphencoding post //null eq {
    TTFDEBUG { (post missing) = flush } if [ ]
  } {
    postformats first_post_string 0 getu32 .knownget {
      TTFDEBUG {
        (post: format ) print
        first_post_string
        dup 0 getu16 =only (,) print 2 getu16 = flush
      } if
      post exch exec
    } {
      TTFDEBUG { (post: unknown format ) print post 0 getu32 = flush } if [ ]
    } ifelse
  } ifelse
  TTFDEBUG { (post=) print dup //== exec } if
  def
} bind def

% - .ttkeys <key> <value> ...
/.ttkeys {
  count /ttkeycount exch def
  /upem head 18 getu16 def
  /FontMatrix matrix
  /FontBBox [ 36 2 42 { head exch gets16 upem div } for ]
  nextxuid
  tabdict /name .knownget {
                % Find the names from the 'name' table.
    /names exch def
    /FontName names 6 findname not { names 4 findname not { curxuid 16#ffffffff and 16 32 string cvrs } if } if
      /fontname 1 index def
    /FontInfo mark
      names 0 findname { /Notice exch } if
      names 1 findname { /FamilyName exch } if
      names 4 findname { /FullName exch } if
      names 5 findname { /Version exch } if
  } {
                % No name table, fabricate a FontName.
    /FontName curxuid 16#ffffffff and 16 32 string cvrs
      /fontname 1 index def
    /FontInfo mark
  } ifelse
                % Stack: ... /FontInfo mark key1 value1 ...
  post //null ne {
    /ItalicAngle first_post_string 4 gets32 65536.0 div
    /isFixedPitch first_post_string 12 getu32 0 ne
    /UnderlinePosition first_post_string 8 gets16 upem div
    /UnderlineThickness first_post_string 10 gets16 upem div
  } if
  counttomark 0 ne { .dicttomark } { pop pop } ifelse
  /XUID [orgXUID 42 curxuid]
  TTFDEBUG {
    tabs { //.printtab exec } forall
    [ sfnts { length } forall ] //== exec
    count ttkeycount sub array astore dup { //== exec } forall aload pop
  } if
  /sfnts sfnts
} bind def

% ---------------- Standard TrueType font loading ---------------- %

% - .pickcmap_with_no_xlatmap -
% Defines cmapsub, cmaptab
/.pickcmap_with_no_xlatmap {
  tabdict /cmap get
                % The Apple cmap format is no help in determining the encoding.
                % Look for a Microsoft table.  If we can't find one,
                % just use the first table, whatever it is.
  dup 4 8 getinterval_from_stringarray exch  % () [] % the default
  0 1 2 index 2 getu16a 1 sub {              % () [] i
    8 mul 4 add 1 index exch 8 getinterval_from_stringarray % () [] ()
    TTFDEBUG {
      (cmap: platform ) print dup 0 getu16 =only
      ( encoding ) print dup 2 getu16 = flush
    } if
    dup 0 getu16 3 eq { exch 3 -1 roll pop exit } if pop
  } for
                % Stack: subentry table
  /cmapsub 2 index def  % () []
  exch 4 getu32 1 index string_array_size 1 index sub getinterval_from_stringarray
  /cmaptab exch def
} bind def

% - .pickcmap_with_xlatmap -
% Defines cmapsub, cmaptab
/.pickcmap_with_xlatmap {
  .xlatmap_dict /TrueType known not {
    (Emulating a CID font with a True Type file, ) print
    (the file gs/lib/xlatmap must contain /TrueType key.) =
    /.pickcmap_with_xlatmap cvx /configurationerror signalerror
  } if
  //false
  .xlatmap_dict /TrueType get
  dup length 2 sub 0 exch 2 exch {                       % bool [] i
    2 copy get                                           % bool [] i ()
    (.) search {                                         % bool [] i post match pre
      cvi exch pop exch cvi                              % bool [] i PlatID SpecID
    } {
      (gs/lib/xlatmap containg a record with an invalid (PlatformID.SpecificID)) =
      /.pickcmap_with_xlatmap cvx /configurationerror signalerror
    } ifelse
    TTFDEBUG {
      (Seeking a cmap for platform=) print 1 index =only ( encoding=) print dup =
    } if
    tabdict /cmap get                                    % bool [] i PlatID SpecID (cmap)
    dup /cmaptab exch def % temporary
    0 1 2 index 2 getu16a 1 sub {                        % bool [] i PlatID SpecID (cmap) j
      8 mul 4 add 1 index exch 8
      getinterval_from_stringarray                       % bool [] i PlatID SpecID (cmap) (cmapsub)
      TTFDEBUG {
        (cmap: platform ) print dup 0 getu16 =only
        ( encoding ) print dup 2 getu16 = flush
      } if
      dup 0 getu16 4 index eq {
        dup 2 getu16 3 index eq {                        % bool [] i PlatID SpecID (cmap) (cmapsub)
          TTFDEBUG {
            (Choosen a cmap for platform=) print 3 index =only
            ( encoding=) print 2 index =
          } if
          /cmapsub 1 index def
          dup 4 getu32                                   % bool [] i PlatID SpecID (cmap) (cmapsub) p
          cmaptab 1 index getu16a                        %     get cmap format
          8 lt {                                         %     length for traditional 16bit format 0, 2, 4, 6
            cmaptab 1 index 2 add getu16a
          } {                                            %     length for advanced 32bit format 8, 10, 12
            cmaptab 1 index 4 add getu32a
          } ifelse                                       % bool [] i PlatID SpecID (cmap) (cmapsub) p l
          cmaptab 3 1 roll getinterval_from_stringarray
          /cmaptab exch def                              % bool [] i PlatID SpecID (cmap) (cmapsub)
          5 index 5 index 1 add get                      % bool [] i PlatID SpecID (cmap) (cmapsub) /Decoding
          /Decoding exch def                             % bool [] i PlatID SpecID (cmap) (cmapsub)
          7 -1 roll pop //true 7 1 roll                  % true [] i PlatID SpecID (cmap) (cmapsub)
        } if
      } if
      pop                                                % true [] i PlatID SpecID (cmap)
      5 index { exit } if
    } for                                                % bool [] i PlatID SpecID (cmap)
    pop pop pop pop                                      % bool []
    1 index { exit } if
  } for                                                  % bool []
  pop                                                    % bool
  not {
    QUIET not { (True Type font doesn't contain a charset listed in gs/lib/xlatmap.) = } if
    /.pickcmap_with_xlatmap cvx /invalidfont signalerror
  } if                                                   %
} bind def

% - .pickcmap -
% Defines cmapsub, cmaptab
/.pickcmap {
  % Currently we use xlatmap only for emulation CIDFontType 2 with
  % a disk True Type font files, and use load_stripped
  % to check this regime. We would like to do so
  % while emulating a Type 42, but first the old code
  % about handling them to be changed
  % with adding a handling of a Decoding.
  % fixme : A correct way to fix this is to implenent
  % the Type 42 emulation with gs_fntem.ps .
  % Also note that PDF embedded fonts probably don't need a xlatmap -
  % see PDF spec, "Encodings for True Type fonts".
  load_stripped {
    //.pickcmap_with_xlatmap exec
  } {
    //.pickcmap_with_no_xlatmap exec
  } ifelse
} .bind executeonly odef
currentdict /.pickcmap_with_xlatmap .undef
currentdict /.pickcmap_with_no_xlatmap .undef

% <glyph> .nname <_name>
/.nname {
  =string cvs (_) exch concatstrings cvn
} bind def

% - .charkeys /CharStrings <charstrings> /Encoding <encoding>
% Resets glyphencoding
/.charkeys {
  TTFDEBUG {
    (glyphencoding: length=) print glyphencoding dup length = === flush
  } if
                % Hack: if there is no usable post table but the cmap uses
                % the Microsoft Unicode encoding, use ISOLatin1Encoding.
     % if 'post' presents, .charkeys computes (with dropping minor details) :
     %     CharStrings = glyphencoding^-1
     %     Encoding = cmap*glyphencoding
     % because 'post' maps glyph indices to glyph names.
     % Otherwise .charkeys must compute (with dropping same details) :
     %     CharStrings = glyphencoding^-1 * cmap
     %     Encoding = glyphencoding
     % because glyphencoding is stubbed with an encoding,
     % which maps char codes to glyph names.
  glyphencoding length 0 eq {
    /have_post //false def
    cmapsub 0 4 getinterval <00030001> eq {
    TTFDEBUG { (No post but have cmap 3.1, so use ISOLatin1Encoding) = } if
      /glyphencoding ISOLatin1Encoding dup length array copy def
    } {
      TTFDEBUG { (No encoding info, use .GS_extended_SymbolEncoding) = } if
      /glyphencoding /.GS_extended_SymbolEncoding findencoding dup length array copy def
    } ifelse
  } {
    /have_post //true def
  } ifelse
                % If necessary, fabricate additional glyphencoding entries
                % to cover all of loca
  glyphencoding length numloca lt {
    /glyphencoding numloca array
      glyphencoding length dup 1 sub 0 1 3 2 roll {
        dup glyphencoding exch get
        3 index 3 1 roll put
      } for
      % /glyphencoding <newarray> <glyphencoding length>
      1 numloca 1 sub {
        1 index exch dup //.nname exec put
      } for
    def
  } if

/cmapa cmaptab cmapdict def
                % Some badly designed Chinese fonts have a post table
                % in which all glyphs other than 0 are named .null.
                % Use CharStrings to keep track of the reverse map from
                % names to glyphs, and don't let any name be used for
                % more than one glyph.
  /CharStrings glyphencoding dup length 1 add dict	% +1 for .notdef
    0 1 3 index length 1 sub {
                % Stack: glyphencoding dict index
      dup 3 index 1 index get
      have_post not {
        exch
        cmapa exch .knownget not {
          0   % a stub for a while. Must skip the entry.
        } if
        exch
      } if
      3 index 1 index known
      { % The same name maps to more than one glyph. Change the name.
        % No special treatment for /.notdef glyph. Bug 689408.
        pop //.nname exec 3 index 2 index 2 index put
        2 index exch 3 -1 roll put
      } {
        3 index exch 3 -1 roll put	% unique name
        pop
      } ifelse
    } for exch pop
                % If there is no .notdef entry, map it to glyph 0.
  dup /.notdef known not { dup /.notdef 0 put } if

TTFDEBUG {
    (CharStrings:)= dup { exch =string cvs print ( ) print //== exec } forall
  } if

% Provide all known aliases for each glyph
  cmapsub 0 4 getinterval <00030001> eq % is the cmap table a unicode one?
  cmapa //ReverseAdobeGlyphList //AdobeGlyphList
  {                                  %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) gname(name) codep(integer)
    exch
    dup 6 index exch .knownget
    {                                %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) codep(integer) gname(name) gindex(integer)
      3 index 3 index .knownget
      {                              %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) codep(integer) gname(name) gindex(integer) gnames(array)
        {                            %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) codep(integer) gname(name) gindex(integer) gname(name)
          TTFDEBUG { (\n1 setting alias: ) print dup ==only
                ( to be the same as  ) print 2 index //== exec } if

7 index 2 index 3 -1 roll exch .forceput
        } forall
        pop pop pop
      }
      {
        pop pop
      } ifelse
    }
    {                                %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) gname(name) codep(integer)
      exch
      % Only do this for Unicode cmap table
      4 index
      {
        dup 4 index exch .knownget
        {                            %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) gname(name) codep(integer) gindex(integer)
          exch pop
          TTFDEBUG { (\n2 setting alias: ) print 1 index ==only
                     ( to use glyph index: ) print dup //== exec } if
          5 index 3 1 roll .forceput
          //false
        }
        {
          //true
        }ifelse
      }
      {
        //true
      } ifelse

{                        %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) gname(name) codep(integer)
        16 4 string cvrs dup length neg 7 add (uni0000) dup
        4 -2 roll exch putinterval cvn 3 index 1 index .knownget
        {                            %  CharStrings(dict) isunicode(boolean) cmap(dict) RAGL(dict) gname(name) codep(integer) gindex(integer)
          TTFDEBUG { (\3 nsetting alias: ) print 1 index ==only
                ( to be index: ) print dup //== exec } if
          exch pop 5 index 3 1 roll .forceput
        }
        {
          pop pop
        } ifelse
      } if
    } ifelse
  } forall
  pop pop pop

readonly

/Encoding [
      have_post {
        0 1 255 {
          cmapa exch .knownget not { 0 } if
          glyphencoding dup length
          2 index le {
            pop pop /.notdef
          } {
            exch get
          } ifelse
        } for
      } {
        glyphencoding  dup length 256 gt { 0 256 getinterval } if
        aload pop
        counttomark 256 exch sub { /.notdef } repeat
      } ifelse
    ]
  TTFDEBUG { (Encoding: ) print dup === flush } if
} .bind executeonly odef		% hides .forceput

% ---------------- CIDFontType 2 font loading ---------------- %

% Fill a string with sequential CIDs starting from the initial value.
% <string> <value> .fill_identity_cmap <string>
/.fill_identity_cmap {             % () v
  1 index length 2 sub          % () v n-2
  0 2 3 2 roll {                % () v 0 2 n-1
      3 copy exch               % () v i () i v
      -8 bitshift               % () v i () i v>>8
      put                       % () v i
      3 copy 1 add              % () v i () v i+1
      exch 255 and              % () v i () i+1 v&255
      put                       % () v i
      pop 1 add                 % () v+1
  } for
  pop
} bind def

% <CIDSystemInfo dict> <dict> .definettcidfont <font>
/.definettcidfont {
  dup begin
  /CIDFontName fontname def
  /CIDFontType 2 def
  /CIDSystemInfo 4 -1 roll def
  /CharStrings mark /.notdef 0 .dicttomark def
                % The cmap isn't of any use even if it is present.
                % Just construct an identity CIDMap covering all the glyphs.

/CIDCount numloca % Wrong if a CIDFontType2 embedded into PDF with a non-Identity CIDToGIDMap.
  def               % processCIDToGIDMap may replace.
  /CIDMap numloca maxstring le {
                % Use a single string.
      numloca 2 mul string 0 //.fill_identity_cmap exec
  } {
                % We must use 2 strings.
      maxstring 2 mul string 0 //.fill_identity_cmap exec
      numloca maxstring sub 2 mul string maxstring //.fill_identity_cmap exec
      2 array astore
  } ifelse
  def

/GDBytes 2 def
  end

end end dup /CIDFontName get exch /CIDFont defineresource
} bind def

% <CIDSystemInfo dict> <file> <Substite name> .loadttcidfont <cidtype2font>
/.loadttcidfont {
  exch
  //false 0 .loadttfonttables
  .makesfnts
        % CIDFontType2 fonts don't have a cmap: they are indexed by CID.
  mark
  //.ttkeys exec
  .dicttomark dup              % convert keys to dict and copy dict
  3 -1 roll                    % Bring PDF substitute name to top of stack

dup
  /fontname exch def
  /FontName exch put           % replace any definition of /FontName
  //.definettcidfont exec
} .bind executeonly odef

% <file> <SubfontID> .load_tt_font_stripped <font_data>
% The font_data includes sfnts, NumGlyphs, TT_cmap, file_table_pos, Decoding.
% CIDMap to be created later from TT_cmap.
/.load_tt_font_stripped {
  //true exch .loadttfonttables
  .makesfnts
  .pickcmap
  mark
  //.ttkeys exec
  /NumGlyphs numloca
  /TT_cmap cmaptab cmapdict
  /file_table_pos file_table_pos
  /Decoding Decoding
  .dicttomark
  end end
} .bind executeonly def

/.load_woff_for_cid {
  //false exch //.loadwofftables exec
  .makesfnts
  % we have to set load_stripped here so
  % .pickcmap will follow the right logic
  % for emulating a CIDFont
  /load_stripped //true def
  .pickcmap
  /load_stripped //false def
  mark
  //.ttkeys exec
  /NumGlyphs numloca
  /TT_cmap cmaptab cmapdict
%  /file_table_pos file_table_pos
  /Decoding Decoding
  .dicttomark
  end end
} .bind executeonly def

% ---------------- PDF TrueType font loading ---------------- %

% Strictly speaking, this code should be loaded only if we have a PDF
% interpreter, but it's so closely tied to the rest of the code in this
% file that we always include it.

% <plat+enc> .findcmap <subtable> true
% <plat+enc> .findcmap false
/.findcmap {
  //false exch tabdict /cmap get
                % Some fonts have multiple cmaps with the same platform and
                % encoding.  Use the first one we find.
  0 1 2 index 2 getu16a 1 sub {
                % Stack: false plat+enc cmap index
    8 mul 4 add 1 index exch 8 getinterval_from_stringarray
    dup 0 4 getinterval 3 index eq {
      4 getu32 1 index exch 1 index string_array_size 1 index sub getinterval_from_stringarray
      4 -1 roll not 4 2 roll exit
    } if pop
  } for
                % Stack: false plat+enc cmap || subtable true plat+enc cmap
  pop pop
} bind def

% Build .symbol_list for .pdfcharkeys .
% It is a dictionary containing all SymbolEncoding glyph names
% and random names for filling gaps in the character code range.
/.symbol_list 256 dict def
{
  =string 0 (x) 0 get put
  /SymbolEncoding .findencoding
  0 1 255 {
    dup 2 index exch get
    dup /.notdef eq {
      pop dup
      =string 1 3 getinterval cvs length 1 add
      =string exch 0 exch getinterval cvn
    } if
    exch //.symbol_list 3 1 roll put
  } for
  pop
} bind exec

% Create .GS_extended_SymbolEncoding as inverse of .symbol_list .
{
  /.GS_extended_SymbolEncoding 256 array
  //.symbol_list {
    exch 2 index 3 1 roll put
  } forall
  .defineencoding
} bind exec

/.hexdigits (0123456789ABCDEF) def

/.is_hex_digit { %   <int> .is_hex_digit <bool>
  dup 48 ge exch
  dup 57 le exch
  dup 65 ge exch
  70 le and
  3 1 roll
  and or
} bind def

/.popfex { pop //false exit } bind def
/.pop3ex { pop pop pop exit } bind def

/.addglyph {  % <dict> <name> <glyph#> .addglyph -
  1 index .namestring     % d n g s
  dup length 7 eq {
    % Bug688467.ps doesn't work if the uniXXXX mapping is allowed with any cmap.
    % Allow it with cmap 3.1 only.
    currentdict /.allow_uniXXXX_glyph_names .knownget not { //false } if
    } { //false
    } ifelse              % d n g s b
  {
    % An undocumented Adobe feature (not sure) :
    % if the name is uniXXXX, obtain Unicode code from it.
    % See bug 688946.
    TTFDEBUG { (Try uniXXXX:) print dup print } if
    { dup 0 get 117 ne //.popfex if % u
      dup 1 get 110 ne //.popfex if % n
      dup 2 get 105 ne //.popfex if % i
      dup 3 get //.is_hex_digit exec not //.popfex if
      dup 4 get //.is_hex_digit exec not //.popfex if
      dup 5 get //.is_hex_digit exec not //.popfex if
      dup 6 get //.is_hex_digit exec not //.popfex if
      dup 3 1 getinterval .hexdigits exch search pop length exch pop exch pop 12 bitshift exch
      dup 4 1 getinterval .hexdigits exch search pop length exch pop exch pop  8 bitshift exch
      dup 5 1 getinterval .hexdigits exch search pop length exch pop exch pop  4 bitshift exch
      dup 6 1 getinterval .hexdigits exch search pop length exch pop exch pop exch pop
      add add add

cmapencoding exch .knownget not { 0 } if
      dup 0 eq //.popfex if

TTFDEBUG { ( index=) print dup =string cvs print } if
      exch pop
      put //true exit
    } loop
    TTFDEBUG { ()= } if
  } {
    pop //false
  } ifelse
  not {                   % d n g
    {
      cmapencoding exch .knownget not { 0 } if
      dup 0 eq //.pop3ex if
      3 copy pop known //.pop3ex if
      put exit
    } loop
  } if
} bind def

% <chartoglyphmap> <subcmap> <AGL> .pdfmapchars /CharStrings <charstrings>
/.pdfmapchars {
  exch cmapdict /cmapencoding exch def     % c2g
  /CharStrings 0 dict                       % c2g /CS <<>>
  % Add glyphs of <AGL>*<subcmap> :
  3 2 roll {                               % /CS <<>> v
    dup type /arraytype eq {               % /CS <<>> /name []
      { 3 copy //.addglyph exec
        pop
      } forall
    } {
      3 copy //.addglyph exec pop
    } ifelse
    pop
  } forall
  % Add glyphs of 'post' with lower priority :
  0 1 glyphencoding length 1 sub {
    dup glyphencoding exch get exch
    dup 0 ne {
      3 copy pop known not {
        3 copy put
      } if
    } if
    pop pop
  } for

% Finally, add glyphs from the prebuilt_encoding with lowest priority
  % this is a catch-all for poorly matched font/PDF combos
  3 2 roll {                               % /CS <<>> v
    dup type /arraytype eq {               % /CS <<>> /name []
      { 3 copy //.addglyph exec
        pop
      } forall
    } {
      3 copy //.addglyph exec pop
    } ifelse
    pop
  } forall

dup /.notdef 0 put
} bind def

% <subtable> .pdfmapsymbolic  /Encoding [] /CharStrings <<>>
/.pdfmapsymbolic {
  //true /.render_notdef gput
  cmapdict
  dup length 1 add dict begin
  [ 64 { /.notdef /.notdef /.notdef /.notdef } repeat ]
  exch {                              % [] char glyph
    exch
    dup 255 gt {
      255 and
      2 index 1 index get /.notdef ne {
        pop pop
        //false
      } {
        //true
      } ifelse
    } {
      //true
    } ifelse {
      dup (01234567) cvs cvn          % glyph char /char
      3 1 roll                        % /char glyph char
      3 index exch                    % /char glyph [] char
      3 index put                     % /char glyph
      def                             % -
    } if
  } forall
  /.notdef 0 def
  /Encoding exch
  /CharStrings currentdict end
} bind def

% - .pdfcharkeys /CharStrings <charstrings> /Encoding <encoding>
/.pdfcharkeys {
        % The following algorithms are per the PDF 1.7 Reference.
  TTFDEBUG { (.pdfcharkeys beg) = } if

% if the flags say symbolic, *and* the font contains a symbol or UCS-2
  % cmap table, ignore Encoding if there is one, treat as symbolic.
  % If the flags say symbolic, and the font doesn't have a symbol cmap
  % and we have an Encoding, treat it as non-symbolic
  % if we don't have an Encoding, try the MacRoman cmap table.
  % Finally, if none of that works, remove the Encoding, and treat it
  % as non-symbolic.
  is_symbolic {
    <00030000> //.findcmap exec
    { //true }
    {
      prebuilt_encoding //null eq {
        <00030001> //.findcmap exec
        {//true}
        {<00010000> //.findcmap exec}
        ifelse
      }
      {//false}
      ifelse
    } ifelse

{
      //.pdfmapsymbolic exec
      /prebuilt_encoding //null def
      //false
    } {
      /prebuilt_encoding //null def
      //true
    } ifelse
  } {
    //true
  }
  ifelse
  {
    <00030001> //.findcmap exec {
      prebuilt_encoding //null ne {
        TTFDEBUG { (Using cmap 3.1 with prebuilt_encoding for non-symbolic.) = } if
        //true /.render_notdef gput
        % The character code in /prebuilt_encoding[] is not guaranteed to be
        % compatible with the Microsoft UCS2 TrueType cmap subtable.
        % If the glyphname is known by AdobeGlyphList, the charcode for the glyph
        % is replaced by UCS2 charcode from AdobeGlyphList. Otherwise it is left
        % as in /prebuilt_encoding[]. /.pdfmapchars should dredge them.

/.allow_uniXXXX_glyph_names //true def
        AdobeGlyphList prebuilt_encoding .invert_encoding dup {    % <<AGL>> <<pbe>> <glyphname> <pbecode>
          pop                                                      % <<AGL>> <<pbe>> <glyphname>
          TTFDEBUG { (check glypname /) print dup =only flush } if
          2 index 1 index .knownget {                              % <<AGL>> <<pbe>> <glyphname> <AGLcode>
            2 index 3 1 roll                                       % <<AGL>> <<pbe>> <<pbe>> <glyphname> <AGLcode>
            TTFDEBUG { ( redefine codepoint by AdobeGlyphList, ) print dup //== exec flush } if
            put                                                    % <<AGL>> <<pbe>>
          } {                                                      % <<AGL>> <<pbe>> <glyphname>
            TTFDEBUG { ( unknown glyphname, leave as it is ) = flush } if
            pop                                                    % <<AGL>> <<pbe>>
          } ifelse
        } forall
        exch pop                                                   % <<pbe>>
        TTFDEBUG { dup (<<\n) print { exch (    ) print =only (\t) print //== exec } forall (>>\n) print flush } if
        exch AdobeGlyphList
        //.pdfmapchars exec
        /Encoding prebuilt_encoding
      } {
        % Likely this branch is now obsolete.
        TTFDEBUG { (Using cmap 3.1 for non-symbolic.) = } if
        0 dict exch
        AdobeGlyphList //.pdfmapchars exec
        /Encoding /WinAnsiEncoding .findencoding
          % WinAnsiEncoding is just a stub here.
          % It will be replaced with one from font resource,
          % because PDF spec requires it.
      } ifelse
    } {
      <00010000> //.findcmap exec {
        TTFDEBUG { (Using cmap 1.0 for non-symbolic.) = } if
        prebuilt_encoding //null ne {
          prebuilt_encoding .invert_encoding exch .romanmacdict //.pdfmapchars exec
          prebuilt_encoding
        } {
          0 dict exch
          /MacRomanEncodingForTrueType .findencoding
        } ifelse
        /Encoding exch
      } {
        % Apply the default algorithm for using the 'post'.
        .charkeys
        % But use PDF encoding when available.
        prebuilt_encoding //null ne {
          pop prebuilt_encoding
        } if
      } ifelse
    } ifelse
  } if
  TTFDEBUG { (.pdfcharkeys end) = } if
} bind executeonly def

% <file> <is_symbolic> <Encoding|null> <FontName> .loadpdfttfont <type42font>
/.loadpdfttfont {
  TTFDEBUG { (.loadpdfttfont FontName=) print dup == } if
  /font_name gput
  TTFDEBUG { (.loadpdfttfont Encoding=) print dup //== exec } if
  /prebuilt_encoding gput     % for .pdfcharkeys
  /is_symbolic gput
  % init key for whether to display the TTF notdef
  //false /.render_notdef gput
  //false 0 .loadttfonttables
  .makesfnts
  tabdict /cmap known not {
    (   **** Warning: Embedded TT font lacks required cmap table. Using identity cmap.\n)
        pdfformaterror
    % Unknown (i.e. not 0 ) language ID brings in post table.
    % Fields: cmap_version, number_of_subtables,
    %  plarform_id, encoding_id, offset,
    %  format, subtable_length, language,
    %  256 bytes of identity mapping.
    tabdict /cmap
    < 0000 0001
      0001 D00D 0000000c
      0000 0106 0000
      00 01 02 03 04 05 06 07 08 09 0a 0d 0c 0d 0e 0f
      10 11 12 13 14 15 16 17 18 19 1a 1d 1c 1d 1e 1f
      20 21 22 23 24 25 26 27 28 29 2a 2d 2c 2d 2e 2f
      30 31 32 33 34 35 36 37 38 39 3a 3d 3c 3d 3e 3f
      40 41 42 43 44 45 46 47 48 49 4a 4d 4c 4d 4e 4f
      50 51 52 53 54 55 56 57 58 59 5a 5d 5c 5d 5e 5f
      60 61 62 63 64 65 66 67 68 69 6a 6d 6c 6d 6e 6f
      70 71 72 73 74 75 76 77 78 79 7a 7d 7c 7d 7e 7f
      80 81 82 83 84 85 86 87 88 89 8a 8d 8c 8d 8e 8f
      90 91 92 93 94 95 96 97 98 99 9a 9d 9c 9d 9e 9f
      a0 a1 a2 a3 a4 a5 a6 a7 a8 a9 aa ad ac ad ae af
      b0 b1 b2 b3 b4 b5 b6 b7 b8 b9 ba bd bc bd be bf
      c0 c1 c2 c3 c4 c5 c6 c7 c8 c9 ca cd cc cd ce cf
      d0 d1 d2 d3 d4 d5 d6 d7 d8 d9 da dd dc dd de df
      e0 e1 e2 e3 e4 e5 e6 e7 e8 e9 ea ed ec ed ee ef
      f0 f1 f2 f3 f4 f5 f6 f7 f8 f9 fa fd fc fd fe ff
    > readonly put
  } if
  .getpost
  .pickcmap
  mark
  //.pdfcharkeys exec
  //.ttkeys exec
  /FontType 42
  /PaintType 0
  /.render_notdef .render_notdef
  TTFDEBUG {
    (numloca=) print numloca =
  } if
  .dicttomark
  dup /FontName font_name put        % replace any definition of /FontName
  end end
 .completefont
} .bind executeonly odef

%--------- Define main type42font loading routines, after their dependencies ------------
% <file> .loadttfont <type42font>
/.loadttfont {
  TTFDEBUG { (.loadttfont) = } if
  //false
  % if this came from Fontmap, there may be a dictionary on the dict stack
  % with a SubfontID in it. Validate it *is* a Fontmap record by also
  % checking for a Path key
  /SubfontID where
  {
    dup /Path known
    {/SubfontID get}
    {pop 0} ifelse
  }
  {0} ifelse
  .loadttfonttables
  .makesfnts
  .getpost
  .pickcmap
  mark
  % if we're loading to substitute for TTF in a PDF,
  % load the font the "PDF way"
  /prebuilt_encoding where
  {pop //.pdfcharkeys exec}
  {.charkeys} ifelse
  //.ttkeys exec
  /FontType 42
  /PaintType 0
  TTFDEBUG {
    (numloca=) print numloca =
  } if
  .dicttomark
  end end dup /FontName get exch definefont
} .bind executeonly def

% <file> .loadwofffont <type42font>
/.loadwofffont
{
  TTFDEBUG { (.loadwofffont) = } if
  //false 0 //.loadwofftables exec
  .makesfnts
  .getpost
  .pickcmap
  mark
  .charkeys
  //.ttkeys exec
  /FontType 42
  /PaintType 0
  TTFDEBUG {
    (numloca=) print numloca =
  } if
  .dicttomark
  end end dup /FontName get exch
  definefont
} .bind executeonly def

% ----- Utilities for loading possible TrueType font ------
% <file> .findttfontname <fname> true
% <file> .findttfontname false
% Closes the file in either case.
/.findttfontname {
  //true 0 .loadttfonttables
  tabdict /name .knownget {
    dup 8 getu32 f exch setfileposition
    12 getu32
    dup 65535 gt { pop 65535 } if % protect against extremely large name
    string f exch readstring pop
    dup
    6 findname not {
      4 findname	% Try FullName
    } {
      exch pop //true
    }
    ifelse
  } {
    //false
  } ifelse
  f closefile end end
} bind def

/.findwoffname {
  //true 0 //.loadwofftables exec
  tabdict /name .knownget {
    dup /clen exch 8 getu32 def
    dup /olen exch 12 getu32 def
    4 getu32 f exch setfileposition
    olen dup 65535 gt { pop 65535 } if % protect against extremely large name
    /fs
    clen olen ne
    {f clen () /SubFileDecode filter /FlateDecode filter}
    {f clen () /SubFileDecode filter}
    ifelse def
    string fs exch readstring pop
    fs closefile
    dup
    6 findname not {
      4 findname	% Try FullName
    } {
      exch pop //true
    }
    ifelse
  } {
    //false
  } ifelse
  f closefile end end
} bind def

/tt_tag_dict << <00010000> 0 (true) 0 (typ1) 0 (ttcf) 0 >> readonly def
/ttf_otf_tag_dict << <00010000> 0 (true) 0 (typ1) 0 (ttcf) 0 (OTTO) 0>> readonly def
% The wOFF tag is immediately followed by the sfnt "flavour" so for simplicity
% combine them
/woff_tag_dict <<
  (wOFF) 0
  <774F464600010000> 0 % (wOFF)<00010000> .concatstrings
  (wOFFtrue) 0
  (wOFFtyp1) 0
>> readonly def

% <file> .is_ttf_or_otf <bool>
/.is_ttf_or_otf {
    dup 0 setfileposition (1234) .peekstring { //ttf_otf_tag_dict exch known } { //false } ifelse
} bind def

/.is_woff {
    dup 0 setfileposition (1234) .peekstring { //woff_tag_dict exch known } { //false } ifelse
} bind def

% <file> <key> .findfontvalue <value> true
% <file> <key> .findfontvalue false
% Closes the file in either case.
/.findnonttfontvalue /.findfontvalue load def
/.findfontvalue {
  1 index //.is_woff exec {
    dup /FontType eq {
      pop closefile 42 //true
    } {
      dup /FontName eq { pop //.findwoffname exec} { pop closefile //false } ifelse
    } ifelse
  }
  {
    1 index //.is_ttf_or_otf exec {
                  % If this is a font at all, it's a TrueType font.
        dup /FontType eq {
          pop closefile 42 //true
        } {
          dup /FontName eq { pop //.findttfontname exec} { pop closefile //false } ifelse
        } ifelse
    } {
                  % Not a TrueType font.
        .findnonttfontvalue
    } ifelse
  } ifelse
} bind executeonly def

% Load a font file that might be a TrueType font.
% <file> .loadfontfile -
/.loadnonttfontfile /.loadfontfile load def
/.loadfontfile {
  dup (12345678) .peekstring { //woff_tag_dict exch known } { //false } ifelse
  {
    //.loadwofffont exec pop
  }
  {
    dup (1234) .peekstring { //tt_tag_dict exch known } { //false } ifelse {
                  % If this is a font at all, it's a TrueType font.
      //.loadttfont exec pop
    } {
                  % Not a TrueType font.
      .loadnonttfontfile
    } ifelse
  } ifelse
} bind def

% Undef the local utility funcs
[
    /.findttfontname
    /.findwoffname
    /ttf_otf_tag_dict
    /tt_tag_dict
    /woff_tag_dict
    /.is_ttf_or_otf
    /.is_woff
]
{currentdict exch .undef} forall
% ----- END .loadfontfile that supports possible TrueType font ------

% Undef these, not needed outside this file
[
 /.addglyph
 /.definettcidfont
 /.fill_identity_cmap
 /.findseg
 /.loadttfontdict
 /.is_hex_digit
 /.pop3ex
 /.popfex
 /.ttkeys
 /.printtab
 /.readttdata
 /.nname
 /.loadttfont
 /.loadwofffont
 /.loadwofftables
 /.findcmap
 /.pdfmapsymbolic
 /.pdfcharkeys
 /.pdfmapchars
] {systemdict exch .forceundef} forall

MR SKB