Text is Not Rendered in Certain PDFs

[TechD123]

Since updating from 3.19 to the recently published 3.26 (via F-Droid), certain PDFs don't render any text. As my main use case for this app is its PDF reading feature, I have reverted to the previous version.

I'll share a link to a sample PDF once I find one that's affected. So far I've only seen this on PDFs that would leak PII if I shared them here... :P

[andiwand]

Thank you for reporting this @TechD123 ! If you don't want to share the files publicly you can send them to us via mail if that is an option. Otherwise we can wait for another file which has this issue

[TomTasche]

I've seen the same behavior, but can't share those PDFs publicly either. @andiwand @ViliusSutkus89 let me know if you are interested in taking a look.

[andiwand]

can you put it into the private testing repo? otherwise you can mail it to me @TomTasche

[ViliusSutkus89]

If I could reproduce it, I could tell if the problem is in upstream pdf2htmlEX or if it's on our side

[ViliusSutkus89]

@TomTasche tried those pdf's on pdf2htmlEX docker image available on docker hub and it gives the same result. Although that official pdf2htmlEX release docker image is 4 years old and using outdated Poppler and FontForge.

I've inspected the HTML DOM, all the text is there, but hidden by CSS. I assume it's a font issue, will try to debug

[TomTasche]

Since PDF is quite important for our users, I'd actually propose to set a bounty for this issue. @andiwand @ViliusSutkus89

I'd love to have a fix for this soon.

[ViliusSutkus89]

pdf2htmlEX-Android conan is nearly there. Then I can focus on debugging upstream pdf2htmlEX

[ViliusSutkus89]

Did some debugging, usually this error happens when fontforge errors out while trying to save a "malformed" font. Don't know how much of that malform is actually bad font and how much of that is bad parsing. Will see what I can do to workaround

[TomTasche]

A fix for that is rolling out now, thanks a lot @ViliusSutkus89 ! 🎉