#dev 2020-08-31

2020-08-31 UTC
cambridgeport90, [tw2113], beko, jonnybarnes, KartikPrabhu, [chrisaldrich] and nickodd joined the channel
#
@JesperBylund
↩️ Sounds lovely. I really like the indieweb trend with /now and webmentions. Feels retro, but also personal, and adds something to the network.
(twitter.com/_/status/1300305338122547200)
jonnybarnes and KartikPrabhu joined the channel
#
hirusi[m]
[Zegnat]: thanks! :)
#
hirusi[m]
After 5 whole months of slow-and-frustrating-work (lol), I'm finally onboard the IndieWeb train! https://rusingh.com/micro/
superkuh, moppy, jonnybarnes and [jeremycherfas] joined the channel
#
[jeremycherfas]
!tell [manton] [Zegnat] I adjusted the template to include `read-of` and `read-status`. The parsed mf2 looks OK to me but I would be very grateful if you would take a quick look. Thanks.
#
Loqi
Ok, I'll tell them that when I see them next
jeremy, jeremy-, [James_Gallaghe], jonnybarnes, fuzzbomb, [grantcodes], [Murray] and [tb] joined the channel
#
[tb]
Uh oh I have some debugging to do on my IndieAuth server today — it's returning an `unsupported_response_type` error to IndieLogin.com for some reason
#
[tb]
Hrm I think my coercion of `response_type=id` requests to `code` under the hood isn't working as intended
KartikPrabhu joined the channel
#
GWG
I'm still waiting for the spec revision for that
#
GWG
Only because I have other things
twomanytacos joined the channel
#
aaronpk
hey cool github put rel=me on the twitter link on your profile
#
[tb]
Part of my really wants to shave the yak more on this now that I've grafted IndieAuth onto Doorkeeper and just extract it all into an IndieAuth-only engine for Rails
#
[tb]
me*
dckc joined the channel
#
[tb]
You might see a bunch of requests from my dev site on IndieLogin.com this morning aaronpk
#
aaronpk
i don't really look at it :)
geoffo, [tantek], nickodd, MrHyde, [jgmac1106], Zegnat and jonnybarnes joined the channel
#
[tb]
Ah yay I fixed it
#
[tb]
Was a stupid copy paste error when I was redoing my authorization views for the vanilla OAuth2 provider I have alongside my IndieAuth provider — I copied the updated views to the IndieAuth ones and forgot to change the post URL on them to the IndieAuth one, so they were all hitting my POST /oauth2/authorize endpoint haha
#
[tb]
Oh btw [jgarber] I wrote a full endpoint discovery thing for webmentions this weekend (including an automated spec suite against all the discovery tests from webmention.rocks) and used your gem again https://github.com/craftyphotons/singulus/blob/main/app/services/discover_webmention_endpoint.rb
[fluffy] and [manton] joined the channel
#
[manton]
[jeremycherfas] Looks good. 👍
#
[manton]
I wonder if most read-of posts should also use p-uid for the ISBN (maybe as a hidden span) but that shouldn’t be required, and might be difficult to get widespread implementation since it will be a little fiddly.
#
aaronpk
i wouldn't overload uid like that
#
aaronpk
uid is meant to be the reference to the post
[jeremycherfas] joined the channel
#
[jeremycherfas]
Am I wrong to use it on the link in the h-cite?
cambridgeport90, [tw2113], jonnybarnes and [James_Gallaghe] joined the channel
#
[manton]
[aaronpk] Hmm. I was just copying that from e.g. this post: https://gregorlove.com/2018/06/want-to-read-welcoming-the/
#
Loqi
[gRegor Morrill] Want to read: Welcoming the Stranger: Justice, Compassion & Truth in the Immigration Debate by Matthew Soerens, Jenny Yang (ISBN 9780830845392)
#
aaronpk
oh i see
#
aaronpk
it's in the h-cite hmm
#
aaronpk
i missed that, thought you were talking about on thte h-entry review
#
aaronpk
only slightly strange thing about that is that there's no indication that it's an ISBN
#
aaronpk
according to http://microformats.org/wiki/h-cite `uid` needs to be a URL or URI
#
Loqi
[Tantek Çelik] Test #IndieWeb : Parvenir à « POSSEr » un article vers Twitter
#
aaronpk
i guess this is the reason for URNs
#
aaronpk
"urn:isbn:9780830845392"
#
[manton]
Interesting.
[schmarty] joined the channel
#
aaronpk
lol [schmarty]
#
[schmarty]
i used to think that "URL" was a very silly term / word but years of web development practice normalized it for me.
#
[schmarty]
"urn" is still very silly to me 😂
#
[jeremycherfas]
I saw that URN on bookstore.org but thought it would be more useful to link to the book itself rather than give the isbn/urn
#
aaronpk
linking to the book sure, but "uid" provides a stable identifier for it
#
[manton]
Are there any other real-world examples that use p-uid? I don’t think I notice it very often.
#
aaronpk
there's some mention of it in /authorship
[chrisaldrich] and [snarfed] joined the channel
#
[snarfed]
bridgy puts silo post ids (etc) in it
[Ana_Rodrigues] joined the channel
#
[tantek]
it's a use-case, which URNs never actually interoperably solved
#
[tantek]
disconnect between admission of a real problem / use-case, and the leap to an unproven proposal (happens a lot in standards work)
#
aaronpk
isn't this a perfect use for them though?
#
aaronpk
there isn't a real URL for an ISBN, but ISBNs exist in a single ISBN namespace
#
aaronpk
so if you want to find matching h-cites of a matching ISBN, you can't use the `url` property for that since what people actually link to for a given ISBN isn't consistent
#
[jeremycherfas]
Sorry, bookshop.org
#
[jeremycherfas]
And though it provides `EAN/UPC` That isn’t linked to anything. If you search for the number though, Amazon.com is generally the top hit. I prefer not to link directly to Amazon, so I provide and affiliate link to bookshop.org
jonnybarnes and strugee joined the channel
#
jacky
URNs could also be used to anchor to Web3-y stuff (ssb URLs, dat links etc)
#
jacky
a URN could be more effective than leaning on something like a http-to-$protocol hub too
jonnybarnes joined the channel
#
[manton]
The “don’t just link to Amazon” problem is why I ended up creating special book URLs on Micro.blog, like this: https://micro.blog/books/9781501173219
#
[manton]
They have no branding or any links to Micro.blog, so they are hopefully more useful to link to from any blog.
nickodd left the channel
#
jmac
[manton]: I didn't know about that! Cool
[tw2113] joined the channel
#
jmac
WorldCat++
#
Loqi
WorldCat has 1 karma over the last year
#
[jeremycherfas]
That’s also a good option.
jonnybarnes joined the channel
#
jacky
ooh nice
#
jacky
this is something I feel like I was needing to do on my own site if I ever did link to books
#
jacky
tbf, if there was some sort of thing that I could hit with a ISBN and a mf2 property, it'd update my post with the proper mf2 for a post, I'd be super interested in that
[tb], [chrisaldrich], jonnybarnes, [Ana_Rodrigues], [tw2113] and geoffo joined the channel
#
[tb]
Yay I think all of my test webmentions that I just sent out to the discovery tests on webmention.rocks worked
[snarfed], jonnybarnes and [jgarber] joined the channel
#
[jgarber]
[tb] Nice! Glad to see that gem getting some use. Let me know if you’ve got any ideas or feedback.
#
[jgarber]
Will give your code a closer look later, too.
[chrisaldrich] and KartikPrabhu joined the channel
#
[tb]
[jgarber] Worked like a charm again! One thing I noticed was that it didn't seem to handle the case of a `<>; rel=webmention` value, which came from me coercing all the `<a>` and `<link>` `href`s into link header format so I could parse everything the same. Not sure though if RFC5988/8288 actually allow for that anyway though so it could just be me abusing it 🙂
#
[jgarber]
I suspect that’s invalid, but I’d have to double-check those specs.
#
[tb]
Yeah I just inject the target URL webmention into it in that case as according to the webmention spec so it works either way
#
[tb]
But I was like, why write my own separate relative URL logic for the `a` and `link` tags when the link header parser already does exactly what I need!