AtUri.http should return an actual http string or be removed

[gabeklavans]

atproto/packages/atproto_core/uri/uri.py

Lines 81 to 84 in 5bce23e

	@property
	def http(self) -> str:
	"""Convert instance to HTTP URI."""
	return str(self)

Right now it just returns the str representation, which seems to always return a string prefixed by at://

I think this property is a bit misleading, unless I'm totally missing something. I'd expect the return of this to be an http(s) url using the DidDocument resolver to get the proper PID base.

Let me know if this is meant for something else

[MarshalX]

Hello, I totally agree that at://... is not HTTP at all. And should be names like uri or at least url, BUT. The reason why it is named like this is pretty easy. That's how it's named in the official SDK :) https://github.com/bluesky-social/atproto/blob/371145432178b6c8c411f1289c266314cc7ec592/packages/syntax/src/aturi.ts#L131-L150

So I would prefer to ask them in the first place. Probably that's just a historical moment that should be accepted. The conversion using DidDocument goes beyond the scope of the AtUri helper, but having this utility somewhere in the SDK potentially could be useful

[gabeklavans]

That's how it's named in the official SDK :) https://github.com/bluesky-social/atproto/blob/371145432178b6c8c411f1289c266314cc7ec592/packages/syntax/src/aturi.ts#L131-L150

I may be mistaken, but it looks like that says href not http. href would make a bit more sense, tho it still seems a bit redundant.

The conversion using DidDocument goes beyond the scope of the AtUri helper, but having this utility somewhere in the SDK potentially could be useful

I totally agree on that, I've since learned about the IdResolver, which is the more appropriate place to be doing that resolution.

[MarshalX]

Oh wow. It's really not http at all xD. Probably now we now who and how did this. I literally misread it many times and even this time