#dev 2022-11-22

2022-11-22 UTC
#
IWDiscordRelay <j​acky#7226> Thank you! Now to make it work (lol)
#
IWDiscordRelay <j​acky#7226> I have the code for it in tests, but I still need to play with UX
gxt joined the channel
#
[snarfed] kongaloosh just fyi, your follow mf2 might have a bug, the u-url seems to have a Python dict embedded in it: https://pin13.net/mf2/?url=https://kongaloosh.com/following/nikolai@mstdn.ca
#
Loqi nikolai@mstdn.ca
#
[snarfed] hey gRegor could you link that post that you thought was missing from your BF feed?
#
gRegor [snarfed], https://indieweb.social/@tchambers/109383220637101265 or other posts by Tim today
#
gRegor Feed page is still empty. Is it only replies/mentions/likes/ etc that show up on the feed?
#
Loqi [Tim Chambers] #Admin #Indieweb.socialThis instance now has 9,347 people with this as their home base...With being so busy on admin work, I missed us crossing that line.From instnaces.social: this makes us 0.15% of total 6,150,381users they scan on the whole #Fediv...
#
[snarfed] Feed is primarily posts from your following. Equivalent to a silo (eg Twitter) home timeline
#
gRegor Ok, that's what I was thinking. I followed Tim yesterday
#
gRegor I see Tim's repost on the feed now. And yay, first repost of a federated post!
#
Loqi 😄
geoffo joined the channel
#
[tantek] barnaby, $t is either a universal CASSIS time variable across PHP & JS, or it's an arch-capitalist version of my identity.
#
gRegor lol
#
barnaby in a parallel universe, $t got rich off making apps which use gdata
#
[tantek] barnaby, there's probably a darkweb social media site somewhere, Griftr, where $ is used for mentions instead of @
#
barnaby griftr.money
#
GWG I think I need to look into some filtering of my feed into ActivityPub. The default is a bit messy looking
#
GWG Anyone have a link to a good simple representation of ActivityPub objects?
#
GWG Or whatever they call their post types?
#
[snarfed] lots in the spec, https://www.w3.org/TR/activitypub/
#
[snarfed] ActivityStreams 2
#
[snarfed] more in that spec, https://www.w3.org/TR/activitystreams-core/
#
GWG Okay. I'm reading [pfefferle]'s code, and as always, he left me some hooks to improve post rendering
#
Loqi [James M Snell] Activity Streams 2.0
#
GWG I think I need https://www.w3.org/TR/activitystreams-vocabulary
#
Loqi [James M Snell] Activity Vocabulary
#
GWG Considering the biggest audience is Mastodon, anyone have a link to what Mastodon supports?
#
[snarfed] they support AP and AS2
#
[snarfed] there are minor quirks, like content warnings, but very few
#
[snarfed] well, few
#
GWG This is going to require a lot of reading, I guess, to figure out how to map my Indieweb Post Types to the right vocabulary.
#
GWG Oh well, Thanksgiving project
#
[snarfed] granary does it, feel free to steal from there
#
GWG [snarfed]: Okay.
#
GWG Maybe I should have gone Bridgy Fed, but I thought this would give me more control
#
GWG I still haven't figured out how I want to discover and incorporate people's feeds from those sites, but I wasn't reading Twitter much, so I figure baby steps.
gRegorLove_ joined the channel
#
[tantek] snarfed, my understanding is that Mastodon supports a subset of AP/AS2, I have a feeling that's what GWG is asking
#
[tantek] what's the subset
#
GWG Exactly.
#
[snarfed] ah
#
GWG I want to just do a quick adjustment for display purposes
#
barnaby didn’t aaronpk post some stats on what AS2 object types his endpoint had seen? or was that activities?
#
barnaby could be useful for getting a sense of what’s widely used
#
aaronpk those were actions, not objects lol
#
barnaby lol yeah that indirection still irks me a little
#
aaronpk yeah def don't bother implementing all the AS2 vocab, nobody does 😂
#
aaronpk i think everything is a Note
#
aaronpk photos are included as attachments to the note
#
barnaby are mastodon’s polls also a note? I think AS2 has a Question type for polls
#
aaronpk good question
#
barnaby and is an answer an activity or an object 🙈
#
GWG I'm looking for the subset that maps to standard Indieweb post types.
#
GWG I'll have to research it, I guess. Thought someone might have taken notes.
#
GWG https://docs.joinmastodon.org/spec/activitypub/#supported-activities-for-statuses - Not bad.
#
aaronpk Payloads is what you're looking for
#
aaronpk Note and Question
#
aaronpk interesting that they map other object types to those
#
barnaby wonder if that’s left over from the switch from ostatus to AP/AS2
#
barnaby or if it’s just a consequence of mastodon’t focus on microblogging
#
barnaby oops
#
barnaby mastodon’s
#
aaronpk mastodo or mastodon't?
[chrisaldrich] joined the channel
#
[chrisaldrich] That is the question. Whether 'tis nobler in the mind....
#
[chrisaldrich] Twitter << A Twitter archive project for [[Eleventy]] https://github.com/tweetback/tweetback
#
Loqi ok, I added "A Twitter archive project for [[Eleventy]] https://github.com/tweetback/tweetback" to the "See Also" section of /Twitter https://indieweb.org/wiki/index.php?diff=84561&oldid=84514
#
[iambismark] Anyone ever written/come across a puppeteer script for dumping Slack messages? It’s for a restricted workspace so unlikely I could use the API.. not trying to do anything nefarious, I just want a way of grabbing all the tweets I’ve ever shared for archival purposes…
#
[iambismark] If not I’ll have to whip one up!
Nuve, Seirdy and joshproehl joined the channel
#
[iambismark] meh who needs puppeteer when you’ve got the dev console…
#
[iambismark] ```Array.from(document.querySelectorAll(".c-search__container .c-virtual_list__item")).filter(el => el.id.startsWith("messages_")).forEach(elem => {
#
[iambismark] let id = elem.id.split("_")[1];
#
[iambismark] let text = elem.querySelector('.p-rich_text_section').textContent;
#
[iambismark] let ts = elem.querySelector('.c-search_message__gutter a').attributes['data-ts'].value.split(".")[0];
#
[iambismark] messages[id] = { ts, text };
#
[iambismark] });
#
[iambismark] document.querySelector(".c-search .c-scrollbar__hider").scrollBy({top: 500});```
#
[iambismark] (please don’t judge my js)
mro, gxt, barnaby, Kaja, Nuve, sciamp[m], AramZ-S[m], ash[m], kandr3s[m], h4kor[m], Kat[m], Zagura[m], gRegorLove_, imachange, [chrisaldrich] and [jeremycherfas] joined the channel
#
[aciccarello] I know many people find the www unnecessary, but for sites that redirect the apex domain to a www subdomain, would there be a way in tooling like indieauth and bridgy fed to drop that part at least in UIs?
gxt and mro joined the channel
#
IWDiscordRelay <j​acky#7226> As someone who's going to transition to that (because I want to host other things), I'm curious about this too
#
barnaby why do you need to transition your site to www. to host other things? so you can use CNAME records?
#
[snarfed] [aciccarello] yeah that's obviously possible. I deliberately chose not to in BF because the exact domain is so central to your experience using it elsewhere, eg Mastodon
#
[snarfed] So I wanted to avoid confusion
#
[snarfed] Good idea otherwise though! Eg Bridgy non-Fed doesn't have that UX concern, so it does aggressively hide common subdomains like www, m, mobile
#
[aciccarello] I think it makes sense to use the exact domain by default. I'm curious what is so important about the exact domain in mastodon if the top level domain redirects. I know some people with a self hosted mastodon subdomain have it appear in mastodon as the top level domain (@person@social.personal.example is configured to show as @person@personal.example)
#
[snarfed] Ah, so are you asking, when BF sees www, could it special case that and drop it and make you @ciccarello.me@ciccarello.me instead?
#
[aciccarello] I'm not sure how you'd configure it (maybe a name attribute on <link rel="authorization_endpoint">?) but yes being able to drop the subdomain on the username. For a mastodon example, I think Zach does this on https://fediverse.zachleat.com/@zachleat which appears as @zachleat@zachleat.com
#
sknebel indieauth can auth you as the root
#
sknebel thats not a problem
#
sknebel and I guess bridgy fed would need you to put the webfinger bits at the root?
#
sknebel which also would work in many setups
#
[snarfed] Strictly speaking, I don't know that www and apex are controlled by the same person, eg on hosts like micro.blog, wordpress.com, etc. But realistically www is a special case, so I could maybe do it for that one specifically
#
[snarfed] [aciccarello] feel free to file an issue! I definitely agree that it would be nicer UX
#
sknebel what is bridgy fed doing to pick the domain it presents?
#
Loqi It looks like we don't have a page for "bridgy fed doing to pick the domain it presents" yet. Would you like to create it? (Or just say "bridgy fed doing to pick the domain it presents is ____", a sentence describing the term)
#
[snarfed] And if apex rejects to www, that's probably good enough of a check
#
aaronpk [aciccarello]: there's details in the IndieAuth spec for doing this, you have to use a particular HTTP redirect code for consumers to not treat the redirect as permanent
#
[aciccarello] Oh interesting
#
sknebel huh, not anymore
#
[snarfed] sknebel on the wm -> AP side, wherever the HTTP fetch ends up after redirects
#
sknebel aaronpk didnt we change that? doesnt it rely on the endpoint to tell the url, and then verifying the endpoint is present there?
#
sknebel I think you describe the old method
#
[aciccarello] Are you talking about https://www.w3.org/TR/indieauth/#redirect-examples
#
sknebel thats the old spec
#
sknebel in the current state the endpoint returns at the end who it authenticated
#
sknebel and the app verifies it was allowed to represent that url
#
sknebel https://indieauth.spec.indieweb.org/
#
[aciccarello] The only thing I'm seeing in the new spec is about 301 redirects. See "Basic Redirect" in this section https://indieauth.spec.indieweb.org/#authorization-server-confirmation
#
sknebel because its not using redirects anymore
#
sknebel what aaron said was the old way of doing it
#
barnaby https://indieauth.spec.indieweb.org/#email-like-identifier would be the perfect path for mastodon to adopt indieauth
#
barnaby btw aaronpk the paragraph fragment links to those examples are broken, they all link to https://indieauth.spec.indieweb.org/#authorization-server-confirmation
#
sknebel [aciccarello]: https://indieauth.spec.indieweb.org/#access-token-response - "me (required) - the canonical user profile URL for the user this access token corresponds to."
#
sknebel endpoint gives whatever url it wants as your profile url - i.e. your root domain
#
sknebel and then https://indieauth.spec.indieweb.org/#authorization-server-confirmation - client verifies that this url declares the endpoint as its endpoint
#
[aciccarello] Yeah, I saw that. So theoretically I should be able to drop the www from the h-card on my domain to use that as the canonical?
#
sknebel h-card?
#
aaronpk Oh you're right the redirect stuff is old
#
[snarfed] took a big step just now, switched the Mastdon button on the https://brid.gy/ home page to point to https://fed.brid.gy/, to surface the Fed option more visibly to new users
#
[snarfed] (there's still a UI path back to normal Mastodon Bridgy signup)
#
barnaby sounds good
#
barnaby are there supposed to be two identical mastodon buttons on the homepage now though?
#
[snarfed] hah no, bug, fixing
#
[aciccarello] sknebel, oh I think I read the "me value" part wrong. Sounds like that's coming from the authorization endpoint
#
sknebel yes
#
[snarfed] https://appdot.net/ 😂
#
aaronpk lol
#
sknebel tony finch still wins IMHO: https://dotat.at/ (look at the email address in the footer :D)
#
[snarfed] 💯
#
[snarfed] can't imagine trying to tell someone that address verbally
#
barnaby oh no
#
barnaby their blog URL adds yet more confusion https://dotat.at/@/
#
[manton] That URL! Yikes. But love how fast the site loads.
geoffo joined the channel
#
aaronpk sometimes i see these nice sites and want to just strip out all the fluff on my own site
#
[aciccarello] I'd love to drop all the material design styles from my site and strip it down to a simple stylesheet
#
barnaby with caches disabled, dotat.at vs aaronparecki.com total load time is 0.6s vs 4.3s
#
barnaby I’m a little surprised that a stylesheet is the largest individual piece of content fetched
#
[jacky] anyone have particular headers around caching that they use for their token response endpoint?
#
[jacky] going to peek around at the oauth site
#
[jacky] [aciccarello] I've been thinking of doing the same myself (though, I need to trim down on my selectors and lean on more native elements)
#
barnaby [jacky]: here’s what I use https://github.com/Taproot/indieauth/blob/main/src/Server.php#L913
#
[jacky] https://mblode.github.io/marx/ re: CSS strip down, was thinking of forking, moving to CSS vars where I can and go from there
#
[jacky] thank you barnaby!
#
[jacky] Do you have particular reasoning for those values?
#
barnaby [jacky]: there are some additional headers for the auth form, too https://github.com/Taproot/indieauth/blob/main/src/Server.php#L770
#
barnaby cache-control: no-store tells browsers and intermediary caches to never cache the content (which is different from cache-control: no-cache!)
#
barnaby Pragma: no-cache is an older version of the same thing as backcompat
#
[aciccarello] Yeah, I'd like to have more element selectors. I have a whole bunch of ideas for css but haven't prioritized it yet. https://github.com/aciccarello/ciccarello.me/issues/105
#
Loqi [aciccarello] #105 Simplify CSS
#
barnaby I copied both from the OAuth2 spec, and afaik they’re required, officially
#
[jacky] _very_ good to know, thank you!
#
aaronpk fyi Pragma is dropped from OAuth 2.1 https://datatracker.ietf.org/doc/html/draft-ietf-oauth-v2-1-07
#
[jacky] noted
#
[jacky] going to opt for the bleeding edge as much as I can
#
barnaby ah good to know, I assume due to it not being necessary any more? is there any good reason to remove it?
#
aaronpk it was a mistake
#
aaronpk https://www.rfc-editor.org/rfc/rfc7234#section-5.4 "Because the meaning of "Pragma: no-cache" in responses is
#
aaronpk not specified, it does not provide a reliable replacement for
#
aaronpk "Cache-Control: no-cache" in them."
#
[jacky] oh man [aciccarello] this ticket is going to inspire me to rework my site 🙂
#
aaronpk https://mailarchive.ietf.org/arch/msg/oauth/9DdkE2P0RrUZMeZAbdf3NrMfy0w/
#
barnaby ah, got it (although cache-control: no-cache is also a mistake afaik? the spec correctly states no-store)
#
[aciccarello] [jakewilkins] Your welcome? Or I'm sorry? I'm not sure which 😆
#
[jacky] lol it's a good thing 🙂
#
[aciccarello] Oops wrong tag. Thanks slack autocomplete..
#
[jacky] 😛
geoffo joined the channel
#
barnaby doesn’t seem like a big problem but I’ll open an issue on taproot/indieauth to consider removing that header
#
[jacky] this _kinda_ reminds me to do a quick lookover things for OWASP
#
[jacky] heh now I'd want to move some of these suggested headers and slide them in by default
#
[jacky] (by default when working with this project being Sele)
#
barnaby I referred to OWASP quite a bit when building taproot/indieauth, and should probably look over it again at some point. I built the bundled stateless auth form CSRF protection based on their example
#
[jacky] kinda like that ':add_secure_headers' thing in Rails
#
[jacky] ooh nice
#
barnaby ha ha [jacky] I already made an issue about those headers on the indieauth spec https://github.com/indieweb/indieauth/issues/117
#
Loqi [barnabywalters] #117 Include Cache-Control and Pragma headers in token exchange response example
#
[jacky] oh nice lol!
#
barnaby I guess this is a good opportunity to amend that to dropping the Pragma header from the suggestion
#
[jacky] yeah see I def was following the spec blindly
#
[jacky] yerp
#
barnaby as I assume bleeding-edge indieauth would prefer compatibility with OAuth 2.1 rather than strict compat with 2.0 (which would require the Pragma header)
#
[jacky] loose with 2.0, stricter with 2.1
#
barnaby finding problems like this is good, it means we can improve the indieauth spec to the point that it’s possible to make a solid implemention just by blindly following that one spec
#
[jacky] exactly!
jeremycherfas joined the channel
#
aaronpk yeah we should aim for OAuth 2.1
#
aaronpk fwiw i'm trying to get 2.1 to "last call" by about June next year
#
barnaby I have some downtime now while I wait for work stuff to run, aaronpk what’s the process for making changes to indieauth? just change public/source/index.php and send a PR? or is there a build process I need to run?
#
barnaby I have a couple of low-effort improvements lined up now I think, so once I know the process I can get them done quickly
#
[jacky] would be down to help
#
aaronpk yeah edit the file in the source folder, i'll build the compiled version when there's enough to release
#
barnaby great, will do
mro joined the channel
#
barnaby should the profile URL response also have a no-store header? it doesn’t contain sensitive tokens, but is otherwise similar https://indieauth.spec.indieweb.org/#profile-url-response
jeremycherfas joined the channel
#
barnaby I’ll add it for the moment for consistency and wait for a review
#
sknebel I think in general if it is accessed used a token a cache should know not to store, but telling it that explicitly again doesnt hurt?
#
barnaby I just checked for the token introspection endpoint and clients are allowed to cache those responses, no Cache-Control headers are specified https://datatracker.ietf.org/doc/html/rfc7662#section-2.2
gRegor joined the channel
#
GWG barnaby++
#
Loqi barnaby has 38 karma in this channel over the last year (56 in all channels)
mro joined the channel
#
GWG barnaby: Your logic is the reason I wanted refresh tokens in there, even if we cited the OAuth spec for more info
#
GWG I remember aaronpk saying there was no reason they weren't compatible with Indieauth, but it just wasn't in people's minds
#
barnaby yeah I’m definitely in favour of at least mentioning as many relevant things as possible in the spec and providing fleshed-out examples, even if they are non-normative and just link to some of the other OAuth specs. Fragmentation has always been a criticism of OAuth2, and though 2.1 does a lot to fix that, I think indieauth does a good job of being an example of how OAuth extension specs can be written to be clear and mostly self-contained
#
[jacky] Bingo
#
[KevinMarks] Anil Dash's site used to be something like dot-anil.dashes.com
[Jeremy_Keith] joined the channel
#
[Jeremy_Keith] In the interests of getting many eyeballs on a bug, can anyone figure out why webmention.io returns a status of `no_link_found` here? https://webmention.io/roytang.net/webmention/1e0dOomp9Ms1JcMbvSUh Here’s the source document, which has the target link in the HTML, all marked up in an h-entry: view-source:https://adactio.com/links/19669 What am I doing wrong?
#
sknebel looks fine to me too
#
sknebel have you retried it?
#
sknebel (not 100% if webmention.io reports if it gets an error while trying to fetch)
#
sknebel (if you retry, you could check the server logs if it went through cleanly o nthat part)
#
aaronpk hmm looks like webmention.io is getting an HTTP 406 error from your site when it fetches it
#
[Jeremy_Keith] That’s odd!
#
[Jeremy_Keith] [aaronpk] That’s probably why I’m getting the same response when I try to send you a webmention too: https://webmention.io/aaronpk/webmention/FkItiZSBlphkNua2pCsV
#
aaronpk it sends a tricky set of Accept headers because of mastodon content negotiation stuff
#
[snarfed] conneg--
#
Loqi conneg has -12 karma in this channel over the last year (-15 in all channels)
#
aaronpk heh
#
aaronpk i swear [snarfed] has a keyword alert set up for that
#
[Jeremy_Keith] Well, I guess I won’t be sending any webmentions through webmention.io then.
#
[snarfed] good idea
#
[snarfed] (that was to aaronpk)
#
sknebel aaronpk: because it uses xray which always does that? or some reason for wm.io to do it?
#
aaronpk yes via xray, webmention.io doesn't do its own parsing anymore
#
[KevinMarks] I tried with mention.tech and it worked:
#
[KevinMarks] 32 seconds ago adactio.com/links/19669
#
[KevinMarks] mentioned roytang.net/2022/11/word-web-persons/ ✅ 6 seconds ago mention sent '201'
#
aaronpk (the ruby mf2 parser was out of date for a while, so i switched webmention.io to use xray for parsing)
#
sknebel hm, that feels off for webmention.io to do somehow
#
[tantek] [Jeremy_Keith] do you have some special code setup to return a 406 under certain conditions?!?
#
[tantek] what is 406
#
Loqi It looks like we don't have a page for "406" yet. Would you like to create it? (Or just say "406 is ____", a sentence describing the term)
#
[Jeremy_Keith] [tantek] Nope, it’s a standard PHP page returning some HTML.
#
aaronpk must be apache doing it for some reason
#
aaronpk weird, this fails `curl -H "Accept: text/html" https://adactio.com/links/19669`
#
aaronpk i guess i should add */* at the end of the accept header, i don't think that will break other things
#
sknebel "Content-Type text/html; charset=utf-8"
#
sknebel but text/html is not accepted?!
#
sknebel thats just broken
#
aaronpk oh darn, that does break mastodon
#
barnaby conneg--
#
Loqi conneg has -13 karma in this channel over the last year (-16 in all channels)
#
aaronpk weird, mastodon is ignoring the number in the accept header too
#
[tantek] conneg--
#
Loqi conneg has -14 karma in this channel over the last year (-17 in all channels)
#
sknebel Available variants: <li><a href="links.php">links.php</a> , type application/x-httpd-php</li>
#
sknebel just no
tbbrown joined the channel
#
[tantek] you know, if only conneg had a test suite 🙄
#
aaronpk this seems wrong, mastodon returns HTML for this: `Accept: application/mf2+json, application/activity+json;q=1, text/html;q=0.9, application/json, application/xml, text/xml, */*;q=0.8`
#
aaronpk well i could add that `application/x-httpd-php` into it 😂
#
sknebel thats really apache thinking too much
#
[Jeremy_Keith] Hmm …I wonder if this is because Apache is configured with `MultiviewsMatch` for PHP files (so that I can use `/links/...`` in URLs when the actual file is `/links.php/...` (so it’s doing something behind the scenes that’s messing with the accept headers)
#
[tantek] what is MultiviewsMatch
#
Loqi It looks like we don't have a page for "MultiviewsMatch" yet. Would you like to create it? (Or just say "MultiviewsMatch is ____", a sentence describing the term)
#
aaronpk my main question is whether this is something that might be common on other peoples' sites too, in which case i should add that to the header. or if this is some obscure configuration that [Jeremy_Keith] has that's fixable
#
[KevinMarks] if the conneg barfs, could you retry with a plain GET?
#
[Jeremy_Keith] Is it because I have this in my Apache config? `AddType application/x-httpd-php .php` (I think I need that so that .php files are parsed as PHP)
#
sknebel thats wrong
#
sknebel AddHandler, not AddType
#
[Jeremy_Keith] A-ha!
#
sknebel (which you probably already have somewhere, otherwise it'd probably have offered your php files for download)
#
[Jeremy_Keith] sknebel++
#
Loqi sknebel has 19 karma in this channel over the last year (41 in all channels)
#
[Jeremy_Keith] By jove, I think that may have fixed it!
#
aaronpk hey cool, xray is parsing the pages now! https://xray.p3k.app/parse?url=https%3A%2F%2Fadactio.com%2Flinks%2F19669&pretty=true
#
[tantek] interesting sknebel. I just tried `AddType application/x-httpd-php .php` in two different htaccess validator services and neither found any errors with it.
#
[Jeremy_Keith] Yay! https://webmention.io/aaronpk/webmention/9UCsNKul-DD1zp7uwwqy
#
barnaby 🎉
#
aaronpk it's not *wrong* it just does something different
#
Loqi does a happy dance!
#
[Jeremy_Keith] Thanks for the eyeballs, folks!
#
[tantek] so I should be looking for an htaccess linter then?
#
sknebel aaronpk: I'd say its wrong because it never does something you want. (even if you planned to offer php files for download, thats not the MIME-type to give them)
#
aaronpk fair
#
[Jeremy_Keith] Yeah, it was definitely a screw-up on my part. It may be something I copied and pasted from a well-meaning but incorrect tutorial about setting up Apache and PHP.
#
[tantek] stackoverflow-- (that doesn't fix an overflow does it?)
#
Loqi stackoverflow has -1 karma over the last year
#
[tantek] copypasta--
#
Loqi copypasta has -1 karma over the last year
#
GWG [Jeremy_Keith]: A sighting...hi
#
[Jeremy_Keith] 👋
mro, [manton], [jacky], [Jeremy_Keith], [fluffy] and vilhalmer joined the channel
#
IWDiscordRelay <j​acky#7226> might be interesting https://github.com/reiver/greatape
#
Loqi [reiver] greatape: Social audio & video app
mro and bterry joined the channel
#
[tantek] better than a bored ape I suppose
#
[snarfed] discovered this just now 😆 🤔 https://socialhub.activitypub.rocks/t/about-the-fediverse-enhancement-proposals/1168
#
[tantek] saw that and didn't see any of the folks who have been deeply involved with AP so far (e.g. editors, or folks like nightpool who are active on W3C SocialCG IRC #/social)
#
aaronpk fep is such a terrible name
#
[tantek] it's a weird branch of standards culture, the only other place I've seen that pattern are the XMPP XEPs
#
[tantek] https://xmpp.org/extensions/
#
gRegor Getting likes and replies back from Mastodon so easily is really cool, snarfed++ for Bridgy Fed
#
[snarfed] thanks and welcome!
#
[snarfed] Python PEPs also
#
[snarfed] I think I've seen it elsewhere too. definitely a fork of standards culture, but has caught on in some places
#
[tantek] snarfed, except "fediverse" is not a standard, unlike XMPP, and Python
#
[tantek] so it's a bit brash/bold/presumptuous to declare something "FEPs" like that
jamietanna joined the channel
#
jamietanna tantek I ended up writing https://gitlab.com/jamietanna/content-negotiation-test-cases because I wanted consistent (and spec-compliant) test cases for my content negotiation pieces
#
[snarfed] oh sure. the "fediverse" part of FEP is weird and wrong ish, agreed
#
[snarfed] jamietanna noooo don't do it you just give it more power
#
[snarfed] #deplatformconneg
#
jamietanna obligatory conneg++
#
Loqi conneg has -13 karma in this channel over the last year (-16 in all channels)
#
jamietanna :P
#
[tantek] hahaha was waiting for that 😆
#
Loqi hehe
#
[snarfed] now you've done it
#
[snarfed] ...
#
[snarfed] C͎̦͖̬̼͝O̜̲N̻͖̘̫͕͖̼N̷̟̬͈͚̭͚̺E̷͍͈̜̜̭ͅG̰͡-̥͇̬̣̘́-̜͈͙
#
jamietanna zalgo++
#
Loqi zalgo has 1 karma over the last year
#
[tantek] jamietanna, "the standard" should link to it
#
jamietanna That is very weird that aaronpk's `accept` above didn't negotiate to activity+json, unless they've also got a preference internally to show HTML over `application/activity+json` / `application/json`
#
aaronpk i tried reading the spec for the Accept header at one point and i'm pretty sure most people don't actually implement it all properly
#
[tantek] ^ so that would actually be "hilarious" to have a test suite that strictly tests what the spec says, and then use that to demonstrate actual lack of compliant implementations / interop, which maybe would make snarfed happier 😉
#
[snarfed] conneg is a troll, don't feed the trolls
#
[jamietanna] Yeah I've definitely had that before, and part of me building that suite was to start seeing if other common implementations match my own usage and reading of it
gxt joined the channel
#
aaronpk lol weird bit of reporting
#
aaronpk "Automattic, we should note, already has some experience with ActivityPub. ... Automattic runs the blogging service WordPress.com and contributes to the open source WordPress blogging software platform. The latter today offers an ActivityPub plug-in for self-hosted WordPress sites."
#
aaronpk https://techcrunch.com/2022/11/21/tumblr-to-add-support-for-activitypub-the-social-protocol-powering-mastodon-and-other-apps/
#
aaronpk but hey cool that [pfefferle]'s plugin was linked to from a techcrunch article!
#
[aciccarello] I'd like to watch this later but it seems like it'd be relevant for some of the people here working on ActivityPub interop https://www.youtube.com/watch?v=vnciCz83t70
#
[tantek] [pfefferle]++ yay! You've been linked from TechCrunch! 😂
#
Loqi [pfefferle] has 2 karma in this channel over the last year (13 in all channels)
#
Loqi 😊
gRegorLove_ joined the channel
#
[tantek] snarfed, there's also this: https://socialhub.activitypub.rocks/c/fediversity/fediverse-futures/58
gRegorLove__ joined the channel
#
[tantek] hey has anyone tried this? https://github.com/mwt/apfollow/
#
[tantek] like does this work? https://apfollow.mwt.me/?user=tantek.com&instance=tantek.com
#
[tantek] all I see is "Couldn't find user"
#
[aciccarello] It works when I put in a mastodon username
[pfefferle] joined the channel
#
[pfefferle] oh nice 😍
#
[tantek] ^ snarfed, I wonder if you are seeing any "odd" Bridgy Fed requests from that link
#
[snarfed] I'll look
#
[aciccarello] Sounds related to https://github.com/snarfed/bridgy-fed/issues/60
#
Loqi [petermolnar] #60 add a "remote follow" page
#
[aciccarello] Also, I opened https://github.com/snarfed/bridgy-fed/issues/314 for the option to drop subdomains.
#
Loqi [aciccarello] #314 Option to hide www subdomain from account identifier
#
[snarfed] [tantek] I'm not sure I'd see apfollow.mwt.me anywhere in my logs, eg referrers, since I don't think it would redirect to BF at all
#
[snarfed] I also don't see its IP in my logs, but looks like it's hosted on https://bunny.net/, and it's CNAMEd to their CDN, so I doubt that IP is static or the same for its server side outbound traffic
[aciccarello]1 and [snarfed]1 joined the channel
#
[tantek] snarfed, I’d expect that that site would be doing some form of webfinger discovery on those params and thus end up querying BF
sknebel_, lanodan_ and walkah_ joined the channel
#
[snarfed]1 right. but as I said ^, I can't know the IP that traffic came from, and it doesn't include its host anywhere
#
IWDiscordRelay <j​acky#7226> what are webrings
#
IWDiscordRelay <j​acky#7226> hm
#
IWDiscordRelay <j​acky#7226> might add https://lib.rs/crates/faery-ring to that page
aaronpk, superkuh, angelo, AramZ-S[m] and Zagura[m] joined the channel
#
[KevinMarks] https://octodon.social/@cwebber/109389760204927183
nertzy and ash[m] joined the channel
#
Loqi [Christine Lemmer-Webber] If anyone wanted to look at it, the ActivityPub test suite (now down, I know. anyone want to take it over?) was built on Pubstrate, and it also included a kind of interesting set of experimental ideas on which I used to prototype ActivityPub things ... https://assets.octodon.social/media_attachments/files/109/389/753/657/487/322/original/7a4636d42dccd7cf.png
#
[snarfed] "clever" 😱
gxt joined the channel