Text with images pasted from Word 2016 doesn't preserve images.

[jswiderski]

📝 Provide detailed reproduction steps (if any)

1. Got to https://ckeditor.com/docs/ckeditor5/latest/features/pasting/paste-from-word.html
2. Extract attached file press Ctrl+A, Ctrl+C and paste its content into editor.
3. You can also select some text with an image. It doesn't have to be the whole document
   copy_paste_image_+_text.docx.zip

✔️ Expected result

Image with text gets pasted.

❌ Actual result

Only text gets pasted.

❓ Possible solution

If you have ideas, you can list them here. Otherwise, you can delete this section.

📃 Other details

From what we have checked, this MS Word file has been created with MS Word 2016 - https://superuser.com/a/1144401

• Browser: Any
• OS: Any
• First affected CKEditor version: N/A
• Installed CKEditor plugins: paste-from-office

---

If you'd like to see this fixed sooner, add a 👍 reaction to this post.

[Reinmar]

Is it a regression?

Is there a workaround (e.g. converting to a different format)?

Is it perhaps related to #2493 and #11921?

[jswiderski]

Is it a regression?

I don't see it as a regression. This is rather specific to this document of that type of documents. I based that on client comments:

Don’t see a issue if we create new document.
...
Even if I copy paste contents from old word file to new one – Don’t see a issue. But Save As old documents to new doesn’t help.

--

Is there a workaround (e.g. converting to a different format)?

According to the client "Even if I copy paste contents from old word file to new one – Don’t see a issue" but this is a weak workaround as not everyone has multiple MS word versions and the online one is not yet supported.

I personally haven't found any "trick" to get this working.

--

Is it perhaps related to #2493 and #11921?

It can be related to #2493 because from what we have checked, this file has been created with MS Word 2016 as well.

[arkflpc]

This is the content of the pasted clipboard from attached document: old.txt
This is the one after pasting it to empty document and copying it again: new.txt

The new one encodes images as:

<!--[if gte vml 1]>... <v:shape ...><v:imagedata src="file:///Path.jpg" /></v:shape><![endif]-->
<![if !vml]><img src="file:///Path.jpg"><![endif]>

v:shape is ignored during processing, and what the paste-from-office plugin actually supports is <img>

The old format is just:

<v:shape ...><v:imagedata src="file:///Path.jpg" /></v:shape>

[nileshawtade]

@mdabkowska88 - Any update on this ?

[Reinmar]

• The support for these images (v:imagedata) will need to land probably in https://github.com/ckeditor/ckeditor5/blob/d40dbad5f599c736e741466dd67d7c9d20656c00/packages/ckeditor5-paste-from-office/src/filters/image.js#L21-L22.
• The ultimate goal is to replace an entire <v:shape> with <img>.
• There's a slight risk that some images will get duplicated after doing the above. In the payloads mentioned by @arkflpc, the duped images are marked with a <!--[if gte vml 1]> but no one can predict what mess would Word produce in other cases.
• We're unsure whether we'll not run into issues with identifying which DOM node representing an image is represented by which image in the RTF format (from which we extract image data).

[mabryl]

Closed as mentioned in #2493 (comment)