From e1d01c391481cd8d1ad5ae28d9635628531a1676 Mon Sep 17 00:00:00 2001
From: John Whitington <john@coherentgraphics.co.uk>
Date: Mon, 10 Oct 2022 16:40:57 +0100
Subject: [PATCH] more

---
 cpdfembed.ml     | 23 +++++++++++++----------
 cpdftruetype.ml  | 44 +++++---------------------------------------
 cpdftruetype.mli |  2 +-
 3 files changed, 19 insertions(+), 50 deletions(-)

diff --git a/cpdfembed.ml b/cpdfembed.ml
index 45e8c07..28b8916 100644
--- a/cpdfembed.ml
+++ b/cpdfembed.ml
@@ -24,13 +24,7 @@ let basename () =
   incr fontnum;
   "AAAAA" ^ string_of_char (char_of_int (!fontnum + 65))
 
-let embed_truetype pdf ~fontfile ~fontname ~codepoints ~encoding =
-  let glyphlist_table = Pdfglyphlist.reverse_glyph_hashes () in 
-  let encoding_table = Pdftext.reverse_table_of_encoding encoding in
-  let accepted_unicodepoints =
-    calc_accepted_unicodepoints encoding_table glyphlist_table codepoints
-  in
-  let f = hd (Cpdftruetype.parse ~subset:accepted_unicodepoints fontfile encoding) in
+let make_single_font ~fontname ~encoding pdf f = 
   let name_1 = basename () in
   let module TT = Cpdftruetype in
   let fontfile =
@@ -50,7 +44,7 @@ let embed_truetype pdf ~fontfile ~fontname ~codepoints ~encoding =
       done;
       a
   in
-  [(codepoints,
+  (f.TT.subset,
    SimpleFont
     {fonttype = Truetype;
      basefont = Printf.sprintf "/%s+%s" name_1 fontname;
@@ -72,5 +66,14 @@ let embed_truetype pdf ~fontfile ~fontname ~codepoints ~encoding =
                     float_of_int f.TT.maxx, float_of_int f.TT.maxy);
         fontfile = Some (FontFile2 fontfile_num);
         charset = None;
-        tounicode = None};
-     encoding})]
+        tounicode = f.TT.tounicode};
+     encoding})
+
+let embed_truetype pdf ~fontfile ~fontname ~codepoints ~encoding =
+  let glyphlist_table = Pdfglyphlist.reverse_glyph_hashes () in 
+  let encoding_table = Pdftext.reverse_table_of_encoding encoding in
+  let accepted_unicodepoints =
+    calc_accepted_unicodepoints encoding_table glyphlist_table codepoints
+  in
+  let fs = Cpdftruetype.parse ~subset:accepted_unicodepoints fontfile encoding in
+    map (make_single_font ~fontname ~encoding pdf) fs
diff --git a/cpdftruetype.ml b/cpdftruetype.ml
index b3cbde3..fa7398c 100644
--- a/cpdftruetype.ml
+++ b/cpdftruetype.ml
@@ -21,44 +21,10 @@ type t =
    widths : int array;
    subset_fontfile : Pdfio.bytes;
    subset : int list;
-   tounicode : Pdfio.bytes option}
+   tounicode : (int, string) Hashtbl.t option}
 
 let dbg = ref false
 
-let tounicode_preamble =
-"/CIDInit /ProcSet findresource begin\n\
-12 dict begin\n\
-begincmap\n\
-/CIDSystemInfo <<\n\
-  /Registry (Adobe)\n\
-  /Ordering (UCS)\n\
-  /Supplement 0\n\
->> def\n\
-/CMapName /Adobe-Identity-UCS def\n\
-/CMapType 2 def\n\
-1 begincodespacerange\n\
-<00><FF>\n\
-endcodespacerange\n"
-
-let tounicode_postamble =
-"endbfrange\n\
-endcmap\n\
-CMapName currentdict /CMap defineresource pop\n\
-end\n\
-end\n"
-
-let tounicode_map s us =
-  let b = Buffer.create 1024 in
-  let s = ref s in
-  Buffer.add_string b tounicode_preamble;
-  Buffer.add_string b (Printf.sprintf "%i beginbfrange\n" (length us));
-  iter
-    (fun u -> Buffer.add_string b (Printf.sprintf "<%02x><%02x><%04x>" !s !s u);
-     s := !s + 1)
-    us;
-  Buffer.add_string b tounicode_postamble;
-  bytes_of_string (Buffer.contents b)
-
 let required_tables =
   ["head"; "hhea"; "loca"; "cmap"; "maxp"; "cvt "; "glyf"; "prep"; "hmtx"; "fpgm"]
 
@@ -563,13 +529,13 @@ let parse ?(subset=[]) data encoding =
               subset_font major minor !tables indexToLocFormat (if subset = [] then [] else [hd subset])
               encoding !glyphcodes loca mk_b glyfoffset data
             in
-            let subset_tounicode = 
+            (*let subset_tounicode = 
               if subset = [] then None else Some (tounicode_map 0 [hd subset])
-            in
-              begin match subset_tounicode with Some x -> Printf.printf "%S\n" (string_of_bytes x) | None -> () end;
+            in*)
+              (*begin match subset_tounicode with Some x -> Printf.printf "%S\n" (string_of_bytes x) | None -> () end;*)
               [{flags; minx; miny; maxx; maxy; italicangle; ascent; descent;
                 capheight; stemv; xheight; avgwidth; maxwidth; firstchar; lastchar;
                 widths; subset_fontfile = main_subset; subset = if subset = [] then [] else tl subset; tounicode = None};
                {flags; minx; miny; maxx; maxy; italicangle; ascent; descent;
                 capheight; stemv; xheight; avgwidth; maxwidth; firstchar; lastchar;
-                widths; subset_fontfile = second_subset; subset = if subset = [] then [] else [hd subset]; tounicode = subset_tounicode}]
+                widths; subset_fontfile = second_subset; subset = if subset = [] then [] else [hd subset]; tounicode = None}]
diff --git a/cpdftruetype.mli b/cpdftruetype.mli
index f15b75c..2235aa4 100644
--- a/cpdftruetype.mli
+++ b/cpdftruetype.mli
@@ -20,7 +20,7 @@ type t =
    widths : int array;
    subset_fontfile : Pdfio.bytes;
    subset : int list;
-   tounicode : Pdfio.bytes option}
+   tounicode : (int, string) Hashtbl.t option}
 
 (* Parse the given TrueType font file. It will return one or more fonts. The
    first, a plain Latin font in the given encoding. Others are for the