#dev 2021-01-12

2021-01-12 UTC
KartikPrabhu, [KevinMarks], [KevinMarks]1, ShadowKyogre, digiSal, gRegorLove and [schmarty] joined the channel; ShadowKyogre left the channel
#
Ruxton
GWG: have a code review :)
swentel joined the channel
#
swentel
sknebel, fun stuff isn't it ;)
#
swentel
It's a pity I didn't take a screenshot from my reader, it was an awesome view :)
#
sknebel
I think it's a misunderstanding on what a microformats parser does
#
sknebel
it does not work on text
#
sknebel
it works on HTML, and the HTML it outputs is equivalent to the input under HTML language rules
#
sknebel
HTML with or without the closing tag describe the same document, since if its missing a parser will insert it
#
sknebel
and nothing promises that you'll get the same bytesequence back at the output that you put in
#
swentel
true, and you're right, in case it would be an actual scrip tag, the page is broken anyway
#
swentel
I just noticed it because the input was wrong from granary
#
swentel
it does seem to try and correct tags though, which maybe it shouldn't do ?
#
swentel
but I don't have a strong opinion on that
#
sknebel
it doesn't "correct" tags
#
sknebel
it parses HTML to a DOM, works on it and serializes parts of it back to text
#
sknebel
it never sees how you formatted your tags on the input
#
swentel
so the dom php adds the closing </script> tag back
#
sknebel
because the entire thing is defined to work on HTML documents, not text
#
swentel
got it
#
swentel
yeah, nothing much to do about that then I guess :)
#
sknebel
and if you leave closing tag in or out, the result is the same document, and serializing doesn't know what your input was like (and is not supposed to know)
#
swentel
ok fwiw, the python parser gives the same 'wrong' result :)
#
swentel
so this is by design
#
swentel
just wanted to raise the issue so folks are aware
#
sknebel
yes. if a parser does something different, that's probably because its HTML parser is not as good
#
swentel
and it's good that it's fixed in granary
#
sknebel
(although I do wonder if script tags shouldn't be stripped from the html output too... I thought that was a rule. will need to look at that at some point)
#
sknebel
-> afk, work
#
swentel
well, I think input can be anything (that's at least how we deal with in in Drupal)
#
swentel
sanitizing should happen on output then
#
swentel
so, at least in my opinion, the behavior is fine
strugee and schmudde joined the channel
#
@nicklewis
↩️ Jekyll is superb too. Now got webmentions working for this setup, so if you revisit https://nicklewis.blog/notes/2021-01-11-second/ you will see the conversation replicated to the static site content. Various developers have used this in a number of creative ways.
(twitter.com/_/status/1348937207252455425)
swentel, myfreeweb, strugee, djmoch_ and jeremych_ joined the channel
hoschi joined the channel
#
@heydonworks
Got webmentions going on http://briefs.video now. Thanks to @mxbck for his writing on the subject. https://briefs.video/videos/is-progressive-enhancement-dead-yet/. I'm running free crons to keep them up to date too (http://cron-job.org).
(twitter.com/_/status/1349001768303259648)
ShadowKyogre, [tantek], [fluffy], schmudde and [manton] joined the channel; ShadowKyogre left the channel
#
@mxbck
↩️ well @zachleat did something like this - running sentiment analysis on webmentions: https://www.zachleat.com/web/snarky/ there are also blocklists for known WM spam sites, like this one by @swyx https://github.com/sw-yx/domainblocklist
(twitter.com/_/status/1349016851221766144)
alex11 and [schmarty] joined the channel; ShadowKyogre left the channel
BudaDude[m] and KartikPrabhu joined the channel
#
aaronpk
in case anyone wants to join the GNAP spec meeting it's starting in just a minute https://events.oauth.net/2021/01/gnap-interim-meeting-Z7Nw3hc8AIOM
#
Loqi
LIVE NOW: GNAP Interim Meeting
gxt joined the channel
#
[schmarty]
what is GNAP?
#
Loqi
It looks like we don't have a page for "GNAP" yet. Would you like to create it? (Or just say "GNAP is ____", a sentence describing the term)
#
[schmarty]
uh... asking for a friend?
#
aaronpk
aka next gen OAuth
#
sknebel
[schmarty]: the next generation of OAuth, with varying amounts of politics about if its the successor or different or :P
#
swentel
to be fair, it's an easier acronym to pronounce, +1 :P
#
aaronpk
i want to make sure that GNAP can support the indieauth use case more or less out of the box, so i'd love to get more folks here involved too
#
sknebel
noted, but this is a bit short notice :D
#
aaronpk
yes sorry, i will post more in advance next time :)
#
aaronpk
i'm also planning on doing a writeup and hopefully prototype of what that would look like so that there's something more concrete to review
#
aaronpk
throughout this process i'm also realizing how actually relatively complicated our use case is between having micropub servers that aren't necessarily coupled to the authorization server and such, so hopefully there will be some mechanisms in this that make that lot clearer
#
@frankmeeuwsen
↩️ Gebeurt allemaal automatisch dankzij het indieweb. Leuk toch! Zie je ook de likes en mentions? Dat kan met de Webmention plugin. Ik zal er eens een betere how-to over schrijven, gericht op het koppelen van tweets aan je blogposts.
(twitter.com/_/status/1349050901814800386)
[Murray] and gRegorLove joined the channel
#
[schmarty]
aaronpk: I'm not sure I understand the complications you're referring to? What might be simplified by having a micropub server coupled to the authorization server?
#
[schmarty]
Scopes, maybe?
#
aaronpk
i think they're more complicated gettting into the microsub/reader side
#
[schmarty]
Like for a authenticated friends-only feeds?
ShadowKyogre joined the channel; ShadowKyogre left the channel
#
aaronpk
yea that's where it gets really messy
#
[tantek]
reading friends-only posts is a hard problem! no one has a good federated solution to that!
#
aaronpk
in oauth terms it means each person's feed is a resource server, so now you've got access tokens that need to be used at many different resource servers issued by many different authorization servers
#
aaronpk
and those resource servers can't be tightly coupled to the authorization server by design
#
[tantek]
that's a good thing for encouraging things to be more distributed/decentralized, instead of biasing an architecture of centralization
#
aaronpk
but this is my point, that this kind of design isn't the default that most people are using oauth/gnap for, so i want to make sure that gets baked in
ShadowKyogre joined the channel
#
[tantek]
you're doing amazing work for actual decentralization aaronpk++
#
Loqi
aaronpk has 73 karma in this channel over the last year (212 in all channels)
#
[tantek]
this type of deep systems analysis & spec design to allow for and preferably even bias towards decentralization absolutely needs to be a tech talk for the Internet Archive's "DWeb Summit" crowd that often are stuck in the weeds of (re)inventing a new technology without performing such analysis
nickodd and KartikPrabhu joined the channel
#
@voxpelli
↩️ I'm an indie developer who wants to sign up to try my hands at creating a subscription mechanism for a SaaS-service. I submitted my https://webmention.herokuapp.com/ as the SaaS-service I wanted to try adding something to, but I'm getting too detailed questions about what exactly I will do
(twitter.com/_/status/1349067240415571974)
#
voxpelli
About the above, I haven’t decided whether to add something to that particular service, but it’s a good reference for trying something out
swentel, [Raphael_Luckom], [fluffy], [schmarty], [Rose] and leg joined the channel; ShadowKyogre left the channel
#
[Raphael_Luckom]
[aaronpk] I was intrigued by the conversation about gnap that I missed earlier. Is this draft spec a reasonably up-to-date thing? https://datatracker.ietf.org/doc/draft-ietf-gnap-core-protocol/?include_text=1
#
aaronpk
yep that's the latest published version, there's also a working draft on github and that's where more of the discussion is too
#
[Raphael_Luckom]
ah, this one: https://github.com/ietf-wg-gnap/gnap-core-protocol . thanks!
#
@hmans
I was building a decentralized, WebMention-based social network (way before Mastodon/ActivityPub), and if anyone had cared back then, it would be a nazi darling today and that just sucks.
(twitter.com/_/status/1349105824992538624)
[grantcodes] joined the channel
#
[grantcodes]
Don't think you need a whole protocol to negotiate with me
[snarfed] joined the channel; nickodd left the channel
#
[snarfed]
hey, true! pants! hmans was active here while he was building that
#
aaronpk
haha [grantcodes]
schmudde, jeremycherfas, [fluffy], hoschi and [KevinMarks] joined the channel
#
@ima_crayon
I'm playing around with implementing webmentions on my website. Does anybody out there have any recommended libraries? I haven't found one I like so I might end up building my own.
(twitter.com/_/status/1349137866056728576)
jeremycherfas joined the channel
#
@kevinmarks
↩️ What language are you using? There are several listed here: https://indieweb.org/Webmention-developer#Libraries
(twitter.com/_/status/1349141058161143810)
petermolnar and xdecimal joined the channel; petermolnar left the channel