#dev 2023-06-21

2023-06-21 UTC
moose333, gRegorLove_ and btrem joined the channel
#
btrem I seem to recall reading a post a long time ago arguing against using microformats classnames as css selectors. So long ago that I
#
btrem ...I'm not sure if I even remember correctly. Any thoughts on why I shouldn't use mf class names in my css?
#
[tantek] Yeah it was a hard learned lesson. Pretty sure we wrote it up in the FAQ
Zegnat joined the channel
#
epoch_ which FAQ?
#
epoch_ https://microformats.org/wiki/faq is my bet
#
epoch_ eh, maybe not.
#
epoch_ "Experience has shown that the best practice is for authors to use their own class names for styling, and only use microformats class names for expressing microformats semantics." touched on in there
#
[timothy_chambe] Regarding Test Suite for ActivityPub: Found this recent discussion on the topic of an ActivityPub test suite: https://lists.w3.org/Archives/Public/public-swicg/2023Mar/0215.html
#
btrem It seems like strange advice. I'm basically declaring all body elements with class='h-recipe' to be a class.
#
btrem Sorry, hit enter too early.
#
btrem I think of class as categorizing elements. And styles can then be applied to those groups, e.g., `a.ext {}` is a way to style all links that are external. And I think of `h-recipe` as a way to categorize a group of `<article>`s or `<body>`s (that is, pages).
#
aaronpk the main reason not to style microformats classes is so that you can later refactor your styling css independently of your data structure css
#
btrem Hmm, I guess. Seems like styling a group of articles that are h-recipe is not very different from styling a bunch of links with a certain class, or a group of lists, or what have you.
#
vladimyr Same like if you bind your js behaviors to the elements via css classes you are supposed to use some sort of prefix like js-carousel or something like that
#
vladimyr Same reasoning
#
aaronpk just saying that everyone I know who started by styling their microformats classes ended up regretting it
#
aaronpk so don't be surprised if you end up with the same conclusion after trying it out
#
btrem Same counter argument: if I'm binding behavior to a specific group of elements, and I want to style that same group of elements for the same reason (here are figures with images that I want to look a certain way and behave a certain way), why not use the same class name?
#
btrem Haha, yeah ok.
#
btrem I'm not trying to be argumentative.
#
btrem I just notice I'm doing things like `<article class="recipe h-recipe">`. Which seems to violate DRY.
#
aaronpk i wish we had more concrete documentation of this problem other than "experience has shown"
#
vladimyr I can fully atest what aaronpk already said and in my case it wasn't mixing mf2 with styling it was mixing behaviour binding classes with styling
#
vladimyr Once you go down that route you are destined to find yourself in a giant mess further down the line
#
btrem me /shrugs I guess I'll continue with "recipe h-recipe". Sure seems odd though.
#
vladimyr I mean, as with any advice, you can kinda ignore it if it is like 2 divs but as soon as numbers go up you'll wish you didn't do that
#
btrem Or I could throw caution to the wind and then some day regret it. One can never have enough regrets! :-D
#
aaronpk how about `<article class="h-recipe i-was-told-not-to-do-this-and-might-regret-it-later">` :D
#
aaronpk it's like a CSS comment
#
btrem Hahaha. That's a really funny idea!
#
Loqi hehe
#
btrem I'm so tempted to do that.
#
vladimyr I can haz data attrs -> `<article class="h-recipe" data-warning="folks-warned-me-not-to-style-mf2-classes" />`
#
vladimyr Then you just need bookmarklet that grabs all data-warnings and shows them as ::before content in red :D
#
aaronpk hey now that's actually useful
[fluffy] joined the channel
#
[fluffy] *[data-warning]::before { color: red; font-weight: bold; content: attr(‘data-warning’); }
#
[fluffy] oh wait just attr(data-warning), no quotes
#
vladimyr And you thought I wasn't serious :O
#
vladimyr ^ here is your gold medal 🥇
tiim, bret, jan6, oodani and [pfefferle] joined the channel
#
[pfefferle] ↩️ It's weird, but I see your reply only if I visit the instance of https://hachyderm.io/@renegadejade I can not see it when I click on her post in my timeline 🤔
[Ana_R] joined the channel
#
[KevinMarks] Neat font trick https://front-end.social/@kizu/110581379745089361
#
Loqi [preview] [Roma Komarov] What if I will tell you how we could solve fit-to-width text with pure #CSS without any hardcoded parameters? Curiously, scroll-driven animations will allow us to do just that!Welcome my new article — “Fit-to-Width Text” — where I continue e... [video]
nertzy joined the channel
#
[tantek] hmm, I tried using mention.tech to send webmentions to fed.bridgy for my most recent post (and the 100days post before it), and it said received successfully, but neither showed up in the Bridgy Fed dashboard. I tried again a few minutes later, and same result (received successfully, no sign of anything in the dashboard)
#
[tantek] [snarfed] is there an actual bug here or are "a few minutes" (so far) within the expected delay between webmention and showing up in the dashboard? if there is such a "expected delay" it might help to add a short sentence about it to https://fed.brid.gy/docs#troubleshooting
#
[tantek] just tried resending again (3rd webmention) from my recent post to Bridgy Fed
Xe, holiday_medley, [jacky], tei_ and btrem joined the channel
#
[snarfed] [tantek] it's a bug. sorry! still lots of moving pieces in the refactoring getting ready to support multiple protocols, looks like this bug was introduced just last night. will fix.
tei_ joined the channel
#
[snarfed] (you can tell because it responded 500 to your webmentions 😳)
t0nic joined the channel
#
[snarfed] ↩️ that's normal fediverse, right? replies get delivered to the original post's instance, but not necesssarily to the replier's followers or other instances
tei_, [jamietanna] and [aciccarello] joined the channel
#
[tantek] That's two bugs then, http://mention.tech said succeeded when it should have reported the 500. [KevinMarks] ?
#
[tantek] I’m still rewriting my publishing user flow (including sending webmentions) so as you say, moving pieces in refactoring
#
[KevinMarks] does bridgy return a 500 after the 202?
#
[KevinMarks] 202 mean it got the webmention, but hasn't processed it yet, right?
#
[snarfed] [KevinMarks] ohh sorry you're absolutely right, the external facing webmention 202s
#
[snarfed] forgot and fell victim to the same async problem I complained about just weeks ago, https://fed.brid.gy/docs#error-handling
#
[KevinMarks] I don't look for a Location on 201 or 202 at the moment
#
[snarfed] yeah and BF doesn't have wm status pages anyway
#
[snarfed] filed https://github.com/snarfed/bridgy-fed/issues/557 , thx for the nudge
#
Loqi [preview] [snarfed] #557 Status pages for object deliveries
#
[tantek] clear we need test cases on webmention.rocks for async. I hear those are hard 🙂
btrem, benji and tei_ joined the channel
#
[tantek] on a positive Webmention note, it appears Lemmy is sending Webmentions, or at least there’s an example in the wild! e.g. if you open up the Interactions with this post on https://www.jvt.me/posts/2019/10/20/indieweb-talk/ you can see "http://defed.lemmy.tf: (mention)" near the bottom of the mentions
#
Loqi [preview] [Jamie Tanna] The IndieWeb Movement: Owning Your Data and Being the Change You Want to See in the Web
#
[tantek] is anyone using or (cross-)posting to Lemmy, or has setup an instance?
jonnybarnes joined the channel
#
[pfefferle] ↩️ But why do I see other comments (not sent by http://brid.gy)???
#
[pfefferle] ↩️ What do you send in to and/or cc?
#
[tantek] looking into seeing if I can find the explicit webmention support in the source
[marksuth] and trwnh joined the channel
#
[timothy_chambe] Hi Everyone, LMK if this should be only here in Dev or also in the main Indieweb, but my team just launched this research:
#
[timothy_chambe] https://indieweb.social/@tchambers/110583050008843476
#
[timothy_chambe] And on Twitter: https://twitter.com/tchambers/status/1671547617895907337
#
Loqi [preview] [Tim Chambers] New research from my team on the #TwitterMigration: This is our 3rd quarterly update on which platforms are growing, new entries such as #BlueSky, #Substack Notes, #Nostr & all public data on Meta's #project92 Includes latest on the #RedditMigrat... https://cdn.masto.host/indiewebsocial/media_attachments/files/110/583/043/945/734/983/original/1106f5b2f703c80b.png
#
[tantek] [timothy_chambe] I'm interested in the methodology behind how accounts are categorized in the chart. E.g. is my Twitter bio mention of @tantek.com@tantek.com syntactically enough to be included as having "moved" to Mastodon or not? And there is the labeling issue of does Mastodon ~= ActivityPub/Fedi in these charts or ... ?
#
[tantek] And yeah, these kinds of details are dev-like enough to keep here for now
holiday_medley joined the channel
#
[timothy_chambe] Well: we don't claim that anyone doing the act of adding an alternative site in their bio has "fully migrated." But we do mark it as a key signal that they have commited enough to alternative social spaces to do this and drive friends to them on those places. We did find reserach on deactivation rates, and after a spike around Elon takeover those boomed, but then settled into a slow steady rate since.
#
[timothy_chambe] And [tantek] - for that chart we did list "mastodon" as for this we specifically scanned for several hundred Mastodon server domain names. So for instance that doesn't even include "cackley.social" non-Mastodon microbogging servers, and it does not include wordpress servers with ActivtyPub plugin active. We do list overall "Fediverse" numbers later in the report, and those came from http://Fedidb.org which to our eyes is the
#
[timothy_chambe] best datasource on that right now.
#
[timothy_chambe] For those in the ActivityPub Test Suite thread. We are getting some wilingness to build one out here: https://indieweb.social/@robz@toot.robzazueta.com/110583730695258922
#
Loqi [preview] [Rob Zazueta] @OpinionatedGeek @tchambers @evan @cwebber @pfefferle @gabek I can kind of kick it off. I'm building some tooling that exchanges ActibityPub messages. I need to test the secure digest signature part as well as ensure my messages are well formed. But ...
#
epoch_ kek. while testing my own stuff, I triggered the hypothetical loop I'd noticed a while ago
#
epoch_ when a message gets sent, the receiving end tries to grab the public key of the sender by making a signed request to the sender's actor url. this second request tries to get verified too, but needs to grab the public key first...
#
epoch_ except in this case, the loop got tighter than two actors
#
epoch_ my server is asking itself what my server's public key is using a signed message
#
epoch_ looks at all of the open connections from 127.0.0.1 to 127.0.0.1:443
#
epoch_ oh, key distribution problems
#
[tantek] on the Lemmy subject, I did find this issue which has some webmention related pointers: https://github.com/LemmyNet/lemmy/issues/1982
#
Loqi [preview] [dessalines] #1982 Make webmention a background task.
#
[tantek] nothing on IndieAuth or Micropub though, maybe worth filing feature requests: https://github.com/LemmyNet/lemmy/issues
tei_ joined the channel
#
rubenwardy Are there any indieweb standards for pages about interests and such? (music, food, books, etc)
#
rubenwardy well, *conventions
#
rubenwardy like how there is for /appearances
#
rubenwardy I guess it's /about ?
#
[tantek] rubenwardy, I think you can ask about each of those subjects for specific efforts on each
#
[tantek] e.g.
#
[tantek] what is food
#
Loqi Food is a post type that represents eating or drinking particular food or drink https://indieweb.org/food
#
[snarfed] [pfefferle] re https://chat.indieweb.org/dev/2023-06-21#t1687369562357200 (moving back to top level since Slack threads don't translate well to IRC etc)...you're right, reply visibility in the fediverse, ie "why do I see some replies to this remote post on my instance, but not others?" is afaik complicated at best
#
Loqi [preview] [[pfefferle]] ↩️ But why do I see other comments (not sent by http://brid.gy)???
#
[snarfed] your instance can learn about/fetch a reply (and thus show it under the original post) for a variety of reasons: someone on your instance liked or reposted or quoted it, they searched for it, etc
#
[snarfed] I'm far from an expert on this, but I think lots of people have looked at it and written about it in detail
#
[tantek] [snarfed] is it possible the receiving/displays of likes/reposts are also broken on Bridgy Fed? Usually I see at least a few on posts fairly quickly and have seen none on my most recent post (though maybe no one did care about it 😉 )
#
[snarfed] [tantek] yup very possible, will look
#
[tantek] this reminds me I need to use Bridgy Fed more frequently to help with noticing changes in behavior, possible regressions etc. 🙂
tei_ joined the channel
#
[tantek] separately, we discussed a nascent need for a "simple" (UI etc.) way to setup & edit a static website (where the content is mostly static, not about the plumbing, SSG etc.), that's easier to use than GH pages
#
[tantek] setting up GH pages for a new domain is still enough of a pain as to be a barrier for casual sideproject use IMO. i.e. got an idea, register a new domain for it, setup some static HTML+CSS exploring the idea
#
[snarfed] yeah right now regressions are more likely due to how deeply I'm refactoring, even with good test coverage, but that should be temporary
#
[tantek] (not looking to attempt to reinvent "visual editing" site builders either, that's a different use-case/audience)
btrem joined the channel
#
[snarfed] [pfefferle] re that reply's to/cc, it had "to": ["https://www.w3.org/ns/activitystreams#Public"], "cc": ["https://hachyderm.io/users/renegadejade", "https://hachyderm.io/users/renegadejade/followers", "https://www.w3.org/ns/activitystreams#Public"]. more details in the log: https://fed.brid.gy/log?path=%2F_ah%2Fqueue%2Fwebmention&start_time=1687220468&key=https%3A%2F%2Ftantek.com%2F2023%2F170%2Ft1%2F
#
[snarfed] hey [manton] I think I found a conneg bug in http://micro.blog's AP support. want me to file it somewhere? or describe here? or...?
#
[snarfed] conneg--
#
Loqi conneg has -22 karma in this channel over the last year (-30 in all channels)
#
[snarfed] filed https://github.com/microdotblog/issues/issues/246
#
Loqi [preview] [snarfed] #246 Content negotiation fails when Accept request header includes multiple types
eitilt1, holiday_medley, angelo and [manton] joined the channel
#
[manton] [snarfed] Thanks! I’ll check it out.
nertzy, IWSlackGateway, tei_, holiday_medley, [jacky] and [tantek] joined the channel
#
[tantek] jacky, I think it would be interesting to explore an IndieAuth-wall for posts by default and maybe only have an allowlist of certain hashtags for purely public posts
[snarfed], petermolnar and gRegor joined the channel
#
[jacky] oooh that sounds interestin
#
[jacky] *interesting
#
[jacky] are hashtags being suggested here for the ease of composition UX?
#
[jacky] I was thinking of leaning on /audience to help with this
holiday_medley joined the channel