PDF/Word-Export is broken

[Web master]

Hello,

the exports are broken.

PDF-Export: The page style is awful. Headlines are exported like normal text (same size). The headline structure is for the trash in the export.

Impossible to read a well structured text.

The Word export is broken to. Cannot read the content. When I open it I see the technical content (XML).

[Nic Brough -Adaptavist-]

I think the first question has to be what you are doing differently to the rest of us.

When I grab a pdf of a page, I get several different levels of emphasis on header rows, depending on the level of header, all perfectly readable.

With Word, I get, well, a word document.  I know it's xml deep down, but it looks fine as a doc, I don't see xml, and save-as doc saves it fine.

[Andy Hoebeke]

Same problem for me. I am on Ubuntu. Tried with Firefox and Chrome. Have never experienced issued with downloading/exporting Word Doc export files from other platforms.

It looks like the doc is interpreted as a regular text file despite the file extension, neither Google Docs nor OpenOffice/LibreOffice are able to read it (no problem with documents exported from other platforms). It looks like an email header with mime details and then followed by HTML XML data.

So the question is, what is Confluence doing differently than other platforms when generating and exporting a Doc file?

[Nic Brough -Adaptavist-]

Nothing. 

The question is what your system is doing differently - as I said before, I (and most of the rest of us) get the pdf or doc we all expect to get.

[Web master]

I created a test case for you:

The simple text with some simple images in Confluence

Exported as PDF (see where the images are shifted and see the little headlines)

[Web master]

Exported as Word file opened with LibreOffice 7.1 on Windows 10

[Web master]

btw:

When I want to save this ".doc" file, then LibreOffice wants to safe it as txt file.

[Daniel Jakob John]

Hello,

I can confirm the behaviour reported by @Web master .

This behaviour can be reproduced in every cloud instance which is in the early access program for developer of us.

It looks like the pdf export is not using the default APIs. When I look into our logs I can confirm that e.g. dynamic content macros are not working properly because the renderModes "pdf" endpoint is not called. Instead it looks like the page is rendered in a webview an stored as pdf somehow.

The word export is totally broken like @Web master has described below.

[Simon Tost _TNG_]

Regarding the Word Export:

From this question: You might be running into the problem that "Word" means "Microsoft Word".

-> CONFCLOUD-26113  "Export Page to Word should support Office, Open Office and Google Docs"

(Server: CONFSERVER-26113)