Magnet links don't work

[msxfm]

Issue by scstanton
Wednesday Nov 20, 2013 at 15:30 GMT
Originally opened as RequestPolicy/requestpolicy#410

---

After installing the beta, magnet: links no longer invoke my torrent client.

[msxfm]

Comment by cryptonot
Thursday Nov 21, 2013 at 22:22 GMT

---

Now that I've tested it, I have the same issue. Tried it on multiple FF profiles. Only way to use the links is copying them.

[nodiscc]

Reproducible every time for me. It looks like RP breaks all links that do not point to http:// or https:// URIs.

I use RP in blacklist mode (the default is to allow everything), so I don't think it's related the default allow/deny policy. I can confirm the addon at least breaks magnet: and mailto: links. Disable blocking seems to work around this problem.

Related bugs:

• conflict with RSS #395, SeaMonkey feeds not working with 1.0 RequestPolicy/requestpolicy#313, "Subscribe to This Page" browser feature breaks RequestPolicy/requestpolicy#386, Requestpolicy 1.0b3 prevent RSS icon subscription RequestPolicy/requestpolicy#372 (feed: URIs)
• GoToMeeting links cannot be allowed  #412 (comment) (UT2004: URIs)
• RequestPolicy blocks maito-links in Firefox RequestPolicy/requestpolicy#362 (mailto: URIs)
• Magnet links don't work RequestPolicy/requestpolicy#410 (magnet: URIs)

Should we rename this bug to something like Breaks all non-http/https links and use it as a tracking bug?

[myrdd]

Should we rename this bug to something like Breaks all non-http/https links and use it as a tracking bug?

In the long run I'd like to solve this by displaying blocked URI schemes (see nsIURI) in the menu, and maybe we might show a notification bar (like we do when a redirect is blocked).

Should this be solved until 1.0.0? At least the mailto link, of course. But I don't like to hardcode schemes like UT2004, and also I think it's a good idea to block cross-URI-scheme requests by default. But, we could allow them for link-clicks. What do you think?

Thanks for the putting those issues together, @nodiscc. But why did you link to the old repository sometimes? I'd suggest to always link to the new repository.

[myrdd]

I've created a new issue including a proposal to solve all those issues: #447

[myrdd]

@scstanton @cryptonot
version 1.0.beta8 contains a workaround against this bug, so I close this issue. For details take a look at issue #447, it aims to solve this problem cleanly.