[BUG] v0.1.7 failed to fetch when reading epub on the web app

[sudogitclone]

Description of the Bug

Hi,

On version 0.1.7 When reading an epub on the web app I get a failed to fetch error and the epub doesn't load properly like in the screenshot below on Edge

on Firefox

This happens when I access Stump behind a reverse proxy with https address e.g. https://stump.mydomain.com. If I access directly by the server IP like http://10.0.0.10:10801/books/44ba7eb7-473f-4700-a25d-a47fef547ca3/epub-reader the epub loads but the cover image is missing

I checked the network inspector on the non https page and it seems to make requests for ReadiumCSS-before.css ReadiumCSS-default.css and ReadiumCSS-afer.css as well as a lot of blob requests like blob:http://10.0.0.10:10801/9c627ca2-3bb8-419b-9c16-a626d727704f however those are not requested on the https page.

Thanks for checking and let me know if you need any more details!

Expected Behavior

As described above

Steps To Reproduce

As described above

Where is this issue happening?

• Web App
• Mobile App
• Server

Server & Client Details

Server Details:

• Release Channel: [latest]
• Server OS: [Debian Bookworm]
• Using Docker: [Yes]
• Server Version: [0.1.7]

Client Details:

• Device: [PC]
• OS: [Windows]
• Browser: [latest Edge, Firefox] (applicable only for Web App)

Additional context

Edit: PDF files load normally

[sudogitclone]

After checking the the docs, it seems adding https://www.stumpapp.dev/docs/guides/configuration/server-config#stump_trust_proxy_headers set to true fixed the issue and epubs are loading properly with the cover image and all behind a reverse proxy at https://stump.mydomain.com, this setting was not required on v0.1.6. However the cover image is still missing on the http page.

[xelab04]

I second the fix on https being setting trust proxy headers to true

[aaronleopold]

I've been able to repro this on my own instance and after some debugging believe I have a tentative fix prepared, however it would be against the upstream readium library. When accessing via ip, or really when accessing on an insecure protocol (i.e., http), the navigator upgrades the insecure connection to https which breaks unless you've actually configured your proxy to handle that.