From b6a65445e10d74dfe384763c37438338bd395372 Mon Sep 17 00:00:00 2001
From: mbrackeantidot <65160241+mbrackeantidot@users.noreply.github.com>
Date: Thu, 29 Apr 2021 18:11:44 +0200
Subject: Docx reader: add handling of vml image objects (jgm#4735) (#7257)

They represent images, the same way as other images in vml format.
---
 src/Text/Pandoc/Readers/Docx/Parse.hs |  11 +++++++++--
 test/Tests/Readers/Docx.hs            |   4 ++++
 test/docx/image_vml_as_object.docx    | Bin 0 -> 74199 bytes
 test/docx/image_vml_as_object.native  |   2 ++
 4 files changed, 15 insertions(+), 2 deletions(-)
 create mode 100644 test/docx/image_vml_as_object.docx
 create mode 100644 test/docx/image_vml_as_object.native

diff --git a/src/Text/Pandoc/Readers/Docx/Parse.hs b/src/Text/Pandoc/Readers/Docx/Parse.hs
index f8ed248d7..7325ff300 100644
--- a/src/Text/Pandoc/Readers/Docx/Parse.hs
+++ b/src/Text/Pandoc/Readers/Docx/Parse.hs
@@ -709,7 +709,8 @@ elemToParPart ns element
      case drawing of
        Just s -> expandDrawingId s >>= (\(fp, bs) -> return $ Drawing fp title alt bs $ elemToExtent drawingElem)
        Nothing -> throwError WrongElem
--- The below is an attempt to deal with images in deprecated vml format.
+-- The two cases below are an attempt to deal with images in deprecated vml format.
+-- Todo: check out title and attr for deprecated format.
 elemToParPart ns element
   | isElem ns "w" "r" element
   , Just _ <- findChildByName ns "w" "pict" element =
@@ -717,9 +718,15 @@ elemToParPart ns element
                   >>= findAttrByName ns "r" "id"
     in
      case drawing of
-       -- Todo: check out title and attr for deprecated format.
        Just s -> expandDrawingId s >>= (\(fp, bs) -> return $ Drawing fp "" "" bs Nothing)
        Nothing -> throwError WrongElem
+elemToParPart ns element
+  | isElem ns "w" "r" element
+  , Just objectElem <- findChildByName ns "w" "object" element
+  , Just shapeElem <- findChildByName ns "v" "shape" objectElem
+  , Just imagedataElem <- findChildByName ns "v" "imagedata" shapeElem
+  , Just drawingId <- findAttrByName ns "r" "id" imagedataElem
+  = expandDrawingId drawingId >>= (\(fp, bs) -> return $ Drawing fp "" "" bs Nothing)
 -- Chart
 elemToParPart ns element
   | isElem ns "w" "r" element
diff --git a/test/Tests/Readers/Docx.hs b/test/Tests/Readers/Docx.hs
index 263e04173..2cce70cc5 100644
--- a/test/Tests/Readers/Docx.hs
+++ b/test/Tests/Readers/Docx.hs
@@ -155,6 +155,10 @@ tests = [ testGroup "document"
             "VML image"
             "docx/image_vml.docx"
             "docx/image_vml.native"
+          , testCompare
+            "VML image as object"
+            "docx/image_vml_as_object.docx"
+            "docx/image_vml_as_object.native"
           , testCompare
             "inline image in links"
             "docx/inline_images.docx"
diff --git a/test/docx/image_vml_as_object.docx b/test/docx/image_vml_as_object.docx
new file mode 100644
index 000000000..7e1e4d2ca
Binary files /dev/null and b/test/docx/image_vml_as_object.docx differ
diff --git a/test/docx/image_vml_as_object.native b/test/docx/image_vml_as_object.native
new file mode 100644
index 000000000..6e689486a
--- /dev/null
+++ b/test/docx/image_vml_as_object.native
@@ -0,0 +1,2 @@
+[Para [Str "Test",Space,Str "with",Space,Str "object",Space,Str "as",Space,Str "image:"]
+,Para [Image ("",[],[]) [] ("media/image1.emf","")]]
-- 
cgit v1.2.3