Claire
Claire
Looks like a duplicate of #19004! Ideally we should also better validate the domains at domain block creation time because we actually don't support this kind of patterns.
Do you have an example of an HTTP URL that fails to be recognized? They should definitely work.
Yeah, this is because it's an IP address, not because it's using HTTP rather than HTTPS. EDIT: I edited the issue's title to reflect that, but I'll let you edit...
I'm not sure we want to repeat the `Group` mistake.
Everyone associating different semantics to `Group` actors and us having to deal with current behavior because we accepted them with old behavior.
I believe your issue is that your `Date` header is not a RFC2616 date. Related: #21617 should give better error messages around that.
> I checked those statuses.. There's no one mentioned, no hashtag etc. I suspect it's something to do with signatures - Trying to pull the public key of the requester...
Oh, good catch. We probably want to run the stoplight on more error types, although we need to be careful to report that in a non-confusing way so that implementors...
@Gargron I can't merge it because you have requested changes and not reviewed/accepted them.
Just a heads up that you need to also update the `Gemfile.lock` accordingly! EDIT: as described by https://github.com/k0kubun/hamlit/commit/8c659c28cf5585f2bfd7ea935d4b7ea1f4e41cfc, `haml` 6 is basically the successor of `hamlit`, so it should indeed...