#dev 2020-10-30

2020-10-30 UTC
deathrow1|absnt, [tantek], [schmarty], geoffo, [Denver_Prophit], ludovicchabant, themaxdavitt, jbove, [aaronpk], nickodd and prologic joined the channel
#
prologic
Heya all! I'm in the processing of _trying_ to add IndieAuth support to Twtxt (https://twtxt.net); and am running into some odd decisions I have to make. Section 4.2.1 (https://www.w3.org/TR/indieauth/#application-information) states that client's **SHOULD** have some kind of h-app landing page. But in the case of https://pin13.net/login/ which uses client_id=https://indielogin.com/ there is no such mf2 information at https://indielogin.co
#
prologic
m/ at all.
#
prologic
What _should_ I do in this case to present any useful information to the user on my side?
#
aaronpk
good question!
#
aaronpk
i fall back to showing the host name of the client ID by itself
#
GWG
aaronpk: Did we ever discuss that manifest thing?
#
aaronpk
GWG not in the latest round
#
aaronpk
we were focusing on bringing it in line with OAuth
#
prologic
Okay showing the hostname is probably a fair enough fallback
#
GWG
aaronpk: Maybe it's worth a proposal..I forget if we have an issue
#
prologic
I guess that's the thing about RFCs, nobody follows them :D
#
aaronpk
that’s one of the benefits of the client ID being a URL, at least that URL is going to be somewhat recognizable
#
prologic
Although I'm trying to (hence mty question) and implementing from scratch :P
#
prologic
aaronpk true
#
prologic
Thanks for the help!
#
prologic
I'll carry on then, I was reading that part of the spec, and implementing as I go along and useing real things to test against
#
prologic
and then went "wtf?!"
#
aaronpk
this is why feedback from people who are actually building stuff is so important! i’ve filed an issue so we can improve that part of the spec https://github.com/indieweb/indieauth/issues/64
#
prologic
Thank you! 🙇‍♂️
#
Loqi
[aaronpk] #64 Clarify what an AS should display if no app information was discovered at the client ID URL
#
jacky
came in here b/c of that ticket landing in his e-mail
#
jacky
and now I see more context
#
jacky
leaves
#
prologic
aaronpk++
#
Loqi
aaronpk has 63 karma in this channel over the last year (224 in all channels)
#
prologic
Thanks again, I'll be back if I run into any more issues :)
KartikPrabhu and [fluffy] joined the channel; prologic left the channel
#
aaronpk
I really need to make a dark mode for monocle
KartikPrabhu, [chrisbergr], DanC, jeremy, [Rose] and prologic joined the channel; nickodd left the channel
#
prologic
Hiya 👋 (again), just fyi https://github.com/jointwt/twtxt/pull/260 implements _very_ basic support for IndieAuth endpoint support for Twtxt.net and Twt.social pods in general (that use https://github.com/jointwt/twtxt). Thanks for the help! Its probably full of bugs and security holes, but the happy path works! 🤗
#
Loqi
[prologic] #260 Add support for IndieAuth endpoint
#
Zegnat
prologic: have you been implementing based on the W3C frozen version of IndieAuth or the living spec that is being tweaked to get ever close to OAuth 2.1?
#
prologic
the W3C
#
prologic
but also testing against a live client
#
Zegnat
https://indieauth.spec.indieweb.org/ has a few more changes trying to streamline it for standard OAuth client libraries and in general getting rid of more moving parts. And will hopefully continue to do so.
#
Zegnat
The W3C version is frozen. I think forever frozen unless the WG gets reactivated?
#
prologic
Not sure? :)
#
prologic
My implementation is pretty basic though so I may not have to change much, based on cide review can you spot what I'd have to adapt/change to be compliant with the newer spec? (reading is tedious for me being blind et all)
[capjamesg] joined the channel
#
Zegnat
I can have a look at it this weekend prologic! I am pretty booked up with the dayjob and other things today
#
Zegnat
Warning that I am not a Go programmer, so I can only try to spot for things in the IndieAuth related flow
#
GWG
Zegnat: I am hoping to close out my map failure and look at your notes this weekend too
#
GWG
Visvalingam–Whyatt didn't work better for me than Ramer–Douglas–Peucker and it is possible that I have no idea what I'm doing
#
Zegnat
prologic: I am ony not sure in as much as I am not sure what we need to do to be allowed to make any changes to the W3C published version :P I keep asking, and getting the answer, and then promptly forgetting it again.
#
Zegnat
So I would definitely refer to https://indieauth.spec.indieweb.org/ for the up-to-date spec
#
Zegnat
GWG: my notes on what?
geoffo joined the channel
#
aaronpk
I really need to make a test suite!
#
Zegnat
I have been thinking a lot about IndieAuth tests too, aaronpk. Now that I am writing an AS again. Been looking into how I best write up automatic integration tests. E.g. for testing if the right OAuth 2.0 errors are returned
#
Zegnat
It is a pretty long list of things I want to test for though, and that makes me not want to start at all *shame*
[chrisbergr] and jamietanna joined the channel
gRegorLove, [chrisaldrich] and [Denver_Prophit] joined the channel
#
GWG
Zegnat: This is why I have issues with tests
#
sknebel
it feels like at least parts of a testsuite could be made reusable
#
sknebel
not sure on whats best practices for that though
[KevinMarks], gbmor, kitt, [chrisbergr] and jamietanna joined the channel
#
jamietanna
Tools wise there are several choices, ie contract testing using something like Pact. Alternatively we could share the context not the implementation using Gherkin so although there will need to be language specific code, it follows the same steps
prologic and geoffo joined the channel
#
prologic
Zegnat thanks! appreicate some eyeballs on what I've done!
[fluffy] and [tantek] joined the channel
#
jacky
this is a nice Web Component
#
jacky
I still wonder if it's even possible to build something (like a set of Web Components) that could be used as a "drop-in" for comment feeds (powered by Webmentions? a h-feed that's polled every two minutes?)
#
KartikPrabhu
a webcomponent could pull-in webmetions from webmention.io right?