Cross-origin Link headers
Technical Discussion
3
Post
2
Autori
0
Visualizzazioni
-
evan@cosocial.ca question for you — is there any guidance in the spec about whether
idandurlfor a given AP object needs to be same-origin?js@podcastindex.social and I were recently discussing this in a related context (
Linkheaders specifically, for HTTP discovery) and I wasn't entirely sure whether this was a valid use-case. -
@julian @js @trwnh there's a whole section on verification in the discovery report.
https://swicg.github.io/activitypub-html-discovery/#verification
Feed RSS
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@julian @js @trwnh there's a whole section on verification in the discovery report.
https://swicg.github.io/activitypub-html-discovery/#verification
-
@julian @js @trwnh Absolutely not! Same-origin is useful for some verification tasks but otherwise is completely orthogonal to ActivityPub.
-
evan@cosocial.ca question for you — is there any guidance in the spec about whether id and url for a given AP object needs to be same-origin?
js@podcastindex.social and I were recently discussing this in a related context (Link headers specifically, for HTTP discovery) and I wasn't entirely sure whether this was a valid use-case.
-
rimu@piefed.social Do you send the Undo(Announce(Create)) as well for microblog compatibility?
-
@julian Well, that's good to know 😀 That said, the other aspect: I believe the biggest thing ActivityPub brings to the world is networking existing communities using existing platforms. They're not "legacy", they're first class. It's completely normal to build a platform with a specific schema for your own needs, and then hook up to AP later. To the rest of the network, it looks like you're extending JSON-LD in exciting innovative ways, while to you it just looks like dumping rows into JSON with a weird @context element.
-
@sanity Almost all AP implementations have an internal database with a fixed schema. When JSON-LD arrives, they take pieces they recognise and force them into that schema. They throw away anything they don't recognise. It's not really a case of "supporting" RDF. They support it just fine, provided you only use exactly the structure they expect. This is definitely not how AP was conceived, and I would be very interested in an implementation that stored and forwarded JSON-LD as it was intended.
-
Another tricky aspect of this is different clients have all kinds of different designs and colors so depending on the situation the tags could look garish and crazy. But on the other hand a bit of crazy is fun.
-
@rimu Thank you! 🙂
I took a look at the source code too and it seems to expect CSS hex notation (either 6-digit RGB or compacted 3-digit form).
One of the things I've been agonizing over is how to allow richer CSS colors, HDR etc. It'd be a short spec to be like “just do anything https://drafts.csswg.org/css-color/#color-syntax allows” but I'd want to give non-browser-based clients some sort of 8-bit RGB fallback. And all that without exploding the complexity. 😓 Will keep thinking on it.
