From 246253107a945a5276b9aaf1c32c391fd99ae052 Mon Sep 17 00:00:00 2001
From: John Whitington <john@coherentgraphics.co.uk>
Date: Sat, 8 Jun 2024 13:41:18 +0100
Subject: [PATCH] Matterhron 31_009

---
 cpdffont.ml  | 18 ++++++++++++------
 cpdffont.mli |  4 +++-
 cpdfua.ml    |  6 +++++-
 3 files changed, 20 insertions(+), 8 deletions(-)

diff --git a/cpdffont.ml b/cpdffont.ml
index 67bd829..cd52e1d 100644
--- a/cpdffont.ml
+++ b/cpdffont.ml
@@ -97,7 +97,7 @@ let is_missing pdf dict =
               | Some _ -> false
               | None -> true
 
-let missing_font pdf page (name, dict) =
+let missing_font ?l pdf page (name, dict) =
   if is_missing pdf dict then
     let subtype =
       match Pdf.lookup_direct pdf "/Subtype" dict with
@@ -110,12 +110,13 @@ let missing_font pdf page (name, dict) =
     and encoding =
      match Pdf.lookup_direct pdf "/Encoding" dict with
       | Some (Pdf.Name n) -> n
-      | _ -> ""
+      | _ -> "Built-in"
     in 
-      if Pdftext.standard_font_of_name basefont <> None then () else
-      Printf.printf "%i, %s, %s, %s, %s\n" page name subtype basefont encoding
+      match l with
+      | None -> Printf.printf "%i, %s, %s, %s, %s\n" page name subtype basefont encoding
+      | Some r -> r := Printf.sprintf "%i %s %s %s %s" page name subtype basefont encoding::!r
 
-let missing_fonts pdf range =
+let missing_fonts ?l pdf range =
   Cpdfpage.iter_pages
     (fun num page ->
        match Pdf.lookup_direct pdf "/Font" page.Pdfpage.resources with
@@ -130,11 +131,16 @@ let missing_fonts pdf range =
                      | _ -> [(name, dict)])
                   fontdict)
            in
-             iter (missing_font pdf num) name_dict_pairs
+             iter (missing_font ?l pdf num) name_dict_pairs
        | _ -> ())
     pdf
     range
 
+let missing_fonts_return pdf range =
+  let l = ref [] in
+    missing_fonts ~l pdf range;
+    !l
+
 let print_font_table pdf fontname pagenumber =
   let page = try List.nth (Pdfpage.pages_of_pagetree pdf) (pagenumber - 1) with e -> error "page not found" in
     match Pdf.lookup_direct pdf "/Font" page.Pdfpage.resources with
diff --git a/cpdffont.mli b/cpdffont.mli
index db23b3f..6ae10f9 100644
--- a/cpdffont.mli
+++ b/cpdffont.mli
@@ -20,7 +20,9 @@ val embed_missing_fonts : string -> bool -> string -> string -> unit
 val copy_font : Pdf.t -> string -> int -> int list -> Pdf.t -> Pdf.t
 
 (** Report missing fonts *)
-val missing_fonts : Pdf.t -> int list -> unit
+val missing_fonts : ?l:string list ref -> Pdf.t -> int list -> unit
+
+val missing_fonts_return : Pdf.t -> int list -> string list
 
 (** Print a font table to Standard Output. *)
 val print_font_table : Pdf.t -> string -> int -> unit
diff --git a/cpdfua.ml b/cpdfua.ml
index 59c2152..7668260 100644
--- a/cpdfua.ml
+++ b/cpdfua.ml
@@ -383,7 +383,11 @@ let matterhorn_31_008 pdf =
 
 (* For a font used by text intended to be rendered the font program is not
    embedded. *)
-let matterhorn_31_009 pdf = ()
+(* NB This, for now, reports all unembedded fonts. *)
+let matterhorn_31_009 pdf =
+  let l = Cpdffont.missing_fonts_return pdf (ilist 1 (Pdfpage.endpage pdf)) in
+    if l <> [] then
+      raise (MatterhornError (`List (map (fun x -> `String x) l)))
 
 (* For a font used by text the font program is embedded but it does not contain
    glyphs for all of the glyphs referenced by the text used for rendering. *)