fenix/cpdf-source
1
0
Fork 0
mirror of https://github.com/johnwhitington/cpdf-source.git synced 2025-06-05 22:09:39 +02:00
Code Issues Packages Projects Releases Wiki Activity

First implementations for 01_003-5

Browse Source
This commit is contained in:
John Whitington
2025-01-20 13:44:14 +08:00
parent 3c1effa120
commit 161276ebb3
1 changed files with 26 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
cpdfua.ml
View File

@ -173,32 +173,42 @@ let all_ops pdf =
in in
form_xobject_ops @ page_ops form_xobject_ops @ page_ops
let rec artifact_in_content c a = function
| [] -> ()
(* Start content. *)
| Pdfops.Op_BDC _::t -> artifact_in_content true a t
(* Start artifact. If already in content, this is an error. *)
| Pdfops.Op_BMC "/Artifact"::t -> if c then merror () else artifact_in_content c true t
(* End artifact or content, whichever is true. If both true, an error. *)
| Pdfops.Op_EMC::t -> if c && a then merror () else if c then artifact_in_content false a t else artifact_in_content c false t
(* Anything else *)
| h::t -> artifact_in_content c a t
let rec content_in_artifact c a = function
| [] -> ()
(* Start content. If already in artifact, this is an error. *)
| Pdfops.Op_BDC _::t -> if a then merror () else content_in_artifact true a t
(* Start artifact. *)
| Pdfops.Op_BMC "/Artifact"::t -> content_in_artifact c true t
(* End artifact or content, whichever is true. If both true, an error. *)
| Pdfops.Op_EMC::t -> if c && a then merror () else if c then content_in_artifact false a t else content_in_artifact c false t
(* Anything else *)
| h::t -> content_in_artifact c a t
(* Content marked as Artifact is present inside tagged content. *) (* Content marked as Artifact is present inside tagged content. *)
let matterhorn_01_003 _ _ pdf = let matterhorn_01_003 _ _ pdf =
let artifact_in_content ops = iter (fun ops -> artifact_in_content false false ops) (all_ops pdf)
false
in
iter
(fun ops -> if artifact_in_content ops then merror ())
(all_ops pdf)
(* Tagged content is present inside content marked as Artifact. *) (* Tagged content is present inside content marked as Artifact. *)
let matterhorn_01_004 _ _ pdf = let matterhorn_01_004 _ _ pdf =
let content_in_artifact ops = iter (fun ops -> content_in_artifact false false ops) (all_ops pdf)
false
in
iter
(fun ops -> if content_in_artifact ops then merror ())
(all_ops pdf)
(* Content is neither marked as Artifact nor tagged as real content. *) (* Content is neither marked as Artifact nor tagged as real content. *)
let matterhorn_01_005 _ _ pdf = let matterhorn_01_005 _ _ pdf =
let untagged_content ops = let untagged_content ops =
false if Cpdftype.add_artifacts ops <> ops then merror ()
in in
iter iter (fun ops -> untagged_content ops) (all_ops pdf)
(fun ops -> if untagged_content ops then merror ())
(all_ops pdf)
(* Suspects entry has a value of true. *) (* Suspects entry has a value of true. *)
let matterhorn_01_007 _ _ pdf = let matterhorn_01_007 _ _ pdf =