#dev 2021-11-25
2021-11-25 UTC
tetov-irc, superkuh, akevinhuang2 and jjuran joined the channel
# micahrl[m] I got posting from Quill working with Interpersonal 😎 I didn't realize that the client sending the "me" value was optional and my code expected it. Ended up having to make some substantial changes bc Interpersonal supports multiple blogs at once, and I didn't have a way to determine the right blog for a given request right away, but now I do.
# micahrl[m] Needs a bunch more grit and polish, but I'm pretty happy that I got to see something new work today
kimberlyhirsh[d] and kogepan joined the channel
# jamietanna[m] <micahrl[m]> "I got posting from Quill working..." <- Nice one! That requirement was only added last year, when there were some significant changes to the spec, and some clients haven't yet updated, so that's likely why ☺
omz13_, omz13, rommudoh[m], nertzy__, tetov-irc, schmudde, [schmarty], chenghiz_, akevinhuang, ranuzz, squarepants and noobranu joined the channel; omz13_ left the channel
# capjamesg[d] How should one go about limiting resource access in accordance with Ticket Auth?
jamietanna joined the channel
# jamietanna[m] I think this would be based on Resource Indicators so you'd issue a ticket for a given set of resources
# capjamesg[d] I am at the "Redeem the ticket for an access token" part of the Ticket Auth spec.
# capjamesg[d] I am going to save a "resource" value that I can decode in the JWT token.
# capjamesg[d] When someone actually uses the token, how can I know that they are using it for the intended purpose?
# capjamesg[d] Yeah. I'm a bit confused though. I can verify what the resource is but how do I know that the client is requesting a resource that is permitted by the "resource" value and not something else?
# capjamesg[d] Oh, of course!'
# capjamesg[d] Ah, that was a silly question now I think about it.
# capjamesg[d] Thank you!
# capjamesg[d] GWG++
# capjamesg[d] That might be useful to add to the spec.
# capjamesg[d] I am still in the early stages of implementing it. Overall, I understand what is being said but I do think more clarity on what I mentioned earlier re: verifying a token is allowed to access something would be useful. I know the IndieAuth spec doesn't lay out how to issue resource-limited tokens but I'd love to see more on that. I do need to look at the GitHub issue linked in the wiki page though. Again, I'm not far into this yet.
# capjamesg[d] Are there any other explored use cases outside of private feed fetching? That's my primary goal for supporting this spec.
# capjamesg[d] Very true aaronpk.
# capjamesg[d] An accompanying "how to get started with ticket auth" blog post would be useful, separate from the spec... But that's a whole other thing haha.
# capjamesg[d] GWG I am going to use metadata, I think. Any pros / cons of metadata vs. rel?
# jamietanna capjamesg[d] remember that `resource` is an array of values :)
nertzy_, schmudde and angelo joined the channel
# capjamesg[d] I agree re: rel.
Seirdy joined the channel
# Seb[d] I wrote a lot of tests for my IndieAuth endpoint today
# Seb[d] one more box to tick (PKCE) and then I'll look into Ticket Auth
schmudde joined the channel
# [snarfed]1 Seb++ testing++
edburns[d] and tetov-irc joined the channel