Word doc (.docx) fails to convert to Markdown

[clfriede]

Hi all. I wanted to post this here before posting an issue. Seeing this problem when I'm trying to convert a word document to markdown (Actual filename replaced with to provide command I used to generate error):

pandoc --extract-media ./images -o .md .docx
couldn't unpack docx container: Did not find end of central directory signature

This seems to happen no matter what .docx file I provide. Tried without --extract-media as well to see if extracting images may have led to the issue. This is the version information for my pandoc:

pandoc -v
pandoc.exe 3.5
Features: +server +lua
Scripting engine: Lua 5.4
User data directory: C:\Users\clfriede\AppData\Roaming\pandoc
Copyright (C) 2006-2024 John MacFarlane. Web: https://pandoc.org
This is free software; see the source for copying conditions. There is no
warranty, not even for merchantability or fitness for a particular purpose.

Please let me know if there is something else I can try. If I don't hear anything or the recommendation is to file an issue, I will do that. Thanks.

[jgm]

Interesting. How did you install this? Was it from our release binaries?

[clfriede]

I tried two different ways. The first time I installed using the 3.5 MSI. When I got the error, I saw that there were other methods to install, so I uninstalled using the MSI installer and then reinstalled using the winget method. Both methods produced the same error.

Regarding the file, this happened to many different files I tried. I cannot provide those because they contain non-public information, but I could probably create a sample file, make sure it produces the same error and then provide that. I will post it here once I do.

[jgm]

(There was a bug in zip-archive that caused this but it got fixed a while back. Perhaps the pandoc you're using got compiled against an older version of zip-archive -- or maybe it's a different issue.)

[jgm]

If you can upload the docx (or email it to me) I could take a look.

[clfriede]

@jgm posting two files here. The first is a sample document that seems to work ok when I run pandoc against it with the arguments I posted above. It's called Document.docx.

The second file is one that fails that can be released publicly. It's called AvereOS_Upgrade_Template.docx.

Document.docx
AvereOS_Upgrade_Template.docx

What's strange to me is that the second file was, many moons ago, created by using the same template that I used to create the first file today.

If there's anything else I can provide to help, please let me know.

[jgm]

Thank you. I can reproduce the problem.

Note: not even the unzip utility can read this:

%  unzip -d Avero ~/Downloads/AvereOS_Upgrade_Template.docx 
Archive:  /Users/jgm/Downloads/AvereOS_Upgrade_Template.docx
  End-of-central-directory signature not found.  Either this file is not
  a zipfile, or it constitutes one disk of a multi-part archive.  In the
  latter case the central directory and zipfile comment will be found on
  the last disk(s) of this archive.
unzip:  cannot find zipfile directory in one of /Users/jgm/Downloads/AvereOS_Upgrade_Template.docx or
        /Users/jgm/Downloads/AvereOS_Upgrade_Template.docx.zip, and cannot find /Users/jgm/Downloads/AvereOS_Upgrade_Template.docx.ZIP, period.

Note that docx files are just zip containers, so unzip should work.

Second thing to note: when I try to open this in Word, I get

So, I'm guessing that Microsoft has put some kind of protection on this. It is not a zip file (doesn't have the right kind of header). There's no way pandoc can deal with the MS authentication, so I think we're at a dead end here.

[clfriede]

Ok, I think I may know what may be going on. What is strange though is that it doesn't work with my permissions as I can see those files without issue and can also edit them. I'll try a few things here and let you know.

[clfriede]

Figured it out. Recent security change caused the data classification for these documents to change in a way that prevents pandoc from being able to access them. Not entirely sure why, but when I changed it back to its old classification, it worked without issue. I think I'm ok from here. Thanks for all of the help @jgm!