Archive.is redirecting HTTPS to HTTP

[ghost]

It seems archive.is has started redirecting HTTPS requests to HTTP (with an annoying 10-second delay). For example: https://archive.is/2VUZ (though I can't consistently reproduce this 100% of the time). Not sure how to disable a rule (just comment it out, or something else?) so I'm just opening an issue.

[fuglede]

I tried refreshing and waiting for 10 seconds quite a few times in both Firefox and Chromium and did not manage to reproduce the issue.

What goes wrong with respect to HTTPS Everywhere? If the site provided a single redirect, I would expect it to simple re-secure the connection (with an additional redirect).

Edit: And to answer your question, to disable a ruleset completely (which is often not completely necessary), the solution would to add default_off="explanation of why" to the ruleset root XML element.

[ghost]

Sorry, I forgot to mention that when it happens, it just shows a white page saying "redirecting to http://archive.is/2VUZ in 10 seconds", and when that elapses, HTTPS Everywhere redirects back to HTTPS and it starts over...

[ghost]

OK, it seems it only does this if the referrer is Wikipedia (telling RefControl to forge it as https://en.wikipedia.org/ works).

[ghost]

Or maybe not, it may be some weird A/B testing. After disabling the referrer spoofing I still got redirected a few times and then it stopped, and now different message: "Redirecting to https://archive.today/2VUZ in 10 seconds".

[fuglede]

Yeah, I see what you are saying:

$ curl -I http://archive.is/test
HTTP/1.1 200 OK
Refresh: 12;url=http://archive.today/test

$ curl -I https://archive.is/test
HTTP/1.1 200 OK
Refresh: 12;url=https://archive.today/test

$ curl -I http://archive.today/test
HTTP/1.1 302 Found
Location: //archive.is/test

$ curl -I https://archive.today/test
HTTP/1.1 302 Found
Location: //archive.is/test

Now I'm wondering why Firefox isn't picking this up.

In any case, as I'm reading those responses, you would also see that issue without HTTPS Everywhere -- what happens if you disable the ruleset?

[fuglede]

And it is pretty random indeed;

$ curl -Is https://archive.is/2VUZ | grep Refresh
$ curl -Is https://archive.is/2VUZ | grep Refresh
Refresh: 12;url=https://archive.today/2VUZ

[fuglede]

I contacted the webmaster of the site, and it seems that the behaviour above is due to a bot request throttling mechanism (which could also explain why I was only able to reproduce after testing for a while).