Download files come out with wrong filename (without .pdf file extension)

[Popolechien]

Reported by user (Kiwix for Windows, 2.3.1) with the Milne textbooks zim where PDF files can be downloaded:

And here note the filename:

Adding .pdf at the end does not work, one needs to fully rename the file (e.g. fubar.pdf) for it to be downloaded successfully.

I know this isn't exactly a bug but since this is not expected behaviour I'm still flagging it as such (but feel free to remove tag).

[benoit74]

The problem is that online website resort on a special response header on the download request:

content-disposition: attachment; filename*=UTF-8''ACElectricalCircuitAnalysis.pdf;

Visible at https://milneopentextbooks.org/ac-electrical-circuit-analysis-a-practical-approach/

Currently we cannot store such headers in the ZIM, or replay them. This is typically a very good illustration of a case where we need to store and replay headers.

Not something going to happen soon without proper funding (need modifications of ZIM specification / libzim, probably readers, and warc2zim). But very important to keep track / open.

This is anyway a warc2zim issue, so I'm transferring there.