#dev 2022-11-14

2022-11-14 UTC
[keithjgrant] joined the channel
#
[keithjgrant] Haha, yes! Got micropub working again. This time using Netlify functions instead of Heroku
#
[keithjgrant] This lib was storing the with token to the post’s frontmatter, which seems, uh… not good. But I was able to modify that out
#
barnaby [keithjgrant]++ congrats! yep storing the token is a fairly common mistake I think, especially when it’s sent as part of the request body. I remove it from the post body in my MP library to reduce how much that happens
#
Loqi [keithjgrant] has 1 karma over the last year
#
@indigitalcolor It looks like it’s time for me to dust off my webmentions which is why I originally added a rel="me" to my site. (twitter.com/_/status/1591967101216239616)
gxt, jjuran, slyduda, mro, barnaby, cambridgeport90 and geoffo joined the channel
#
@thejuanlam @schnarfed Hey! Really quick question, I'm trying to POSSE to Mastodon over webmentions with the Netlify Webmentions plugin, but I keep on getting the following error: https://brid.gy/log?start_time=1668434228&key=agdicmlkLWd5ck0LEg1QdWJsaXNoZWRQYWdlIiZodHRwczovL2p1YW5sYW0uY29tL25vdGVzLzE0MTEyMDIyNTEyLwwLEgdQdWJsaXNoGICAmLGQrLwKDA&module=default Any ideas? Bridgy is fully authorized on Mastodon. I don't know what's wrong! (twitter.com/_/status/1592175012773076995)
#
@thejuanlam @schnarfed Hey! Really quick question, I'm trying to POSSE to Mastodon over webmentions with the Netlify Webmentions plugin, but I keep on getting the following error: https://brid.gy/log?start_time=1668434228&key=agdicmlkLWd5ck0LEg1QdWJsaXNoZWRQYWdlIiZodHRwczovL2p1YW5sYW0uY29tL25vdGVzLzE0MTEyMDIyNTEyLwwLEgdQdWJsaXNoGICAmLGQrLwKDA&module=default Any ideas? Bridgy is fully authorized on Mastodon. I don't know what's wrong! (twitter.com/_/status/1592175012773076995)
geoffo and [benatwork] joined the channel
#
@schnarfed ↩️ Argh, this happens sometimes. Could you disable and re-enable publish on https://brid.gy/mastodon/@juanlam@indieweb.social and then try again? (twitter.com/_/status/1592181821177233413)
mro, [campegg], [timothy_chambe] and slyduda joined the channel
#
tommorris interesting post here by someone getting their blog indiewebified, but their h-entry syntax isn't getting picked up by indiwebify.me (which uses the php mf2 parser). https://www.lifelog.be/indiewebifying-my-blog-part-1 - any interest in helping them out or seeing if there's an issue with the parser
#
tommorris thought it might be because it's react/JSX but SSR is working, so dunno
vilhalmer joined the channel
#
sknebel <html lang="en" dir="ltr" class="h-full <- that looks potentially problematic
#
sknebel !mf2 https://www.lifelog.be/indiewebifying-my-blog-part-1
#
Kaja https://php.microformats.io/?url=https%3A%2F%2Fwww.lifelog.be%2Findiewebifying-my-blog-part-1
#
sknebel yep
#
sknebel tommorris: ^^^
#
slyduda hmmm so tailwind heights conflict with h-entry?
#
tommorris siiiigh. :(
#
tommorris sknebel++
#
Loqi sknebel has 15 karma in this channel over the last year (35 in all channels)
#
tommorris sknebel: will email the author to tell him :)
pharalia joined the channel
#
tommorris given tailwind probably won't go away, might not be a bad idea to hack together a way for uf2 parsers to selectively exclude likely overlaps. from my reading of the tailwind docs, it looks like h-[0-9], h-[0-9]/[0-9] and a few keywords could easily be ignored. guess it might be time to document it on the wiki
[aciccarello] joined the channel
#
[aciccarello] [tommorris] I added a section to the tailwind page about microformats. Feel free to update as you see fit
mro joined the channel
#
aaronpk We did already update the parsing spec for things like that
#
aaronpk quite a while ago
#
sknebel yes, but only ones with numbers
#
sknebel h-<word> is a bit much
#
tommorris mf2py + mf2util handles it correctly. it's been a while since I looked at the spec (or the parsers)
#
aaronpk Yes I believe the update was ignoring numbers and words with caps
#
slyduda that's a good catch, as someone who also uses tailwind i ditched h-<word>s right before i started optimizing for microformats!
#
tommorris I've filed a mf2py issue - https://github.com/microformats/mf2py/issues/170
#
Loqi [tommorris] #170 Ignore Tailwind's h- prefix values
yarrie joined the channel
#
[snarfed] tommorris++ W3C's priority of constituencies is new to me and interesting: https://www.w3.org/TR/html-design-principles/#priority-of-constituencies
#
Loqi tommorris has 3 karma in this channel over the last year (6 in all channels)
#
[snarfed] reminds me a bit of Bryan Cantrill's eng values idea: https://vimeo.com/230142234
#
[schmarty] snarfed: adactio touts the w3c priority of constituencies quite a bit in his talks about the web, progressive enhancement, making sure things fail well, etc.
#
[schmarty] it's a good one!
#
Zegnat If Tailwind keeps growing as it has, it might be time to look at options outside of classes again? https://github.com/microformats/php-mf2/pull/204 / https://github.com/microformats/microformats2-parsing/issues/43 comes to mind.
#
Loqi [aaronpk] #204 parse microformats classes from the property attribute
#
aaronpk oh this will end well https://twitter.com/elonmusk/status/1592177471654604800
#
@elonmusk @sampullara Part of today will be turning off the “microservices” bloatware. Less than 20% are actually needed for Twitter to work! (twitter.com/_/status/1592177471654604800)
#
sknebel that reminds me of the old "doom as a process manager" patches :P
#
yarrie why does he feel the need to make product announcements within replies
#
yarrie means you have to stalk his profile to figure anything out
#
barnaby yarrie: there’s your answer right there. he craves attention, and making product announcements from his personal account brings that in bucketloads
#
yarrie barnaby: that explanation seems insane but it wouldnt surprise me
#
barnaby I’ve had him blocked for years, but I think he did something very similar with spacex, live-tweeting updates and analysis from his personal account
#
barnaby it’s also similar to how Tr*mp used twitter
#
yarrie live tweeting important, mostly bs, decisions is one thing, but doing it within replies when the twitter UI for viewing replies is so awful is just insulting
#
yarrie at least trump (lol) kept it in the feed
#
[jacky] has anyone heard about https://castopod.org/
#
barnaby nope, looks interesting
#
Zegnat I would not neccessarily label that tweet as any sort of announcement that even needs to be public.
#
Zegnat [jacky]: I am still not clear on what Podcasting 2.0 is supposed to mean, though I do keep bumping into that term
#
Zegnat What is Podcasting 2.0?
#
Loqi It looks like we don't have a page for "Podcasting 2.0" yet. Would you like to create it?_ (Or just say "Podcasting 2.0 is ____", a sentence describing the term)
#
barnaby at least (in this case) it doesn’t look like it has anything to do with blockchains
#
yarrie am i being cynical or does podcasting 2.0 seem like a term being pushed by a single org
#
[jacky] yeah
#
[jacky] it looks like it
#
aaronpk yes, i would not define it on the wiki yet
#
aaronpk oh maybe not https://theaudacitytopodcast.com/what-is-podcasting-2-0-and-why-does-it-matter/
#
aaronpk or maybe it is still?
#
yarrie sometimes i can't tell bcs open standards orgs often behave like startups today
#
yarrie im still pretty sure its just a single org (non-profit or not) pushing it though, at least has adam curry for credibility
[chrisaldrich] joined the channel
#
[jacky] yeah and it being funded by nlnet is interesting too
#
[jacky] b/c their focus _is_ open tech
#
[jacky] I can see (barely) a case to put a mention to this _under_ any page on podcasting but also be like "here be dragons"
#
[jacky] this is the only thing that's probably worth keeping on the wiki (if needed) https://podcastnamespace.org/
#
[jacky] from that link you shared, aaronpk
#
aaronpk nlnet has been throwing small-ish grants at a lot of things, so i wouldn't necessarily give a lot of weight to that either
#
[jacky] good to know
#
[jacky] https://github.com/Podcastindex-org/podcast-namespace/blob/main/docs/1.0.md#locked ugh
#
barnaby well that’s one approach to limited audience content
#
[jacky] right lol
#
barnaby brb I’ll go propose p-original-content-do-not-steal=yes on for h-entry
#
[jacky] LOL
#
barnaby and the opposite property, p-goes-so-hard=feel-free-to-screenshot
gRegor joined the channel
#
gRegor [snarfed], For Bridgy Fed, if I want to syndicate some notes, do I need to have the Atom feed advertised on the homepage? I have my articles Atom feed there already, so will it look through both of them?
#
gRegor (I haven't started with BF yet, just a preliminary q)
vilhalmer joined the channel
#
slyduda [jacky]: that looks sick!!!
vilhalmer joined the channel
#
gRegor Hm, the Tailwind h-* stuff isn't ideal for the parsed results, but IWM does still show the first h-entry it finds even if it's nested. I suppose readers would probably fail, though. Monocle preview doesn't show the post.
#
gRegor https://indiewebify.me/validate-h-entry/?url=https%3A%2F%2Fwww.lifelog.be%2Fninety-days-in-a-new-country
#
aaronpk there's a million edge cases for things like this in XRay already, and adding one more may break others
#
aaronpk it looks simple at first glance but when you realize all the different combinations of things people try to do with their markup it gets complicated
#
aaronpk one other problem with that page is it's redirecting no-www to www, but including no-www links in the properties
#
gRegor Yeah, this is more #microformats I guess, but not sure about going down the route of keeping a list of h-* to ignore
#
gRegor Maybe in helper libs, but probably not in the parsing spec
#
aaronpk the other way to go here is encourage Tailwind to switch to css classes that don't interfere with microformats :)
#
aaronpk h-Full for example
#
gRegor Their blog post said "is not detecting the microformat tags" though so I'm wondering what they're seeing different.
#
[jacky] I have a vendatta against tailwind (I think people need to learn modern css! it's actually better) and would mark this as a WontFix
#
gRegor I might reach out to get more info
#
gRegor Haha, definitely no fan of Tailwind either
#
barnaby hmm I thought we raised an issue on their gh before, but I don’t see anything there so perhaps now
#
barnaby *not
#
[jacky] they won't
#
slyduda also not a fan but i use it constantly haha
#
[jacky] is going to brew a post called "Tailwind Considered Harmful"
#
slyduda not advocating for it to be fixed^ i would be fine change the name
#
barnaby from what I’ve seen the creator write, I don’t imagine they’d be interested in changing it for mf2 compat, but it might be worth a try
#
aaronpk the one thing a consumer can do with that extra h-full wrapper object is to look for an h-entry with a u-url that matches the URL of the page it's fetched from
#
aaronpk but that would fail on that example anyway because of the www issue
#
aaronpk https://pin13.net/mf2/?id=20221114213355491
#
barnaby or search the tree and take the first h-entry it finds
#
aaronpk that is worse than it sounds
#
aaronpk that's where the edge cases come in
#
barnaby that’s what I’ve been doing for ages and have never had any problems in practise
#
barnaby provided you only apply it when you’re parsing a page where you expect a single h-entry (e.g. when handling an incoming webmention)
#
aaronpk it works in some cases, but not if you want to parse arbitrary URLs to recognize what's on the page
#
gRegor Yeah, the mf-cleaner findByMicroformat method works well in my experience, though that's outside XRay
#
aaronpk if you know what you're looking for then it's fine
#
aaronpk but that's only a small part of consuming cases
#
gRegor yeah
#
aaronpk this is why there are so many test files in xray https://github.com/aaronpk/XRay/blob/main/tests/ParseTest.php
#
barnaby I’d be up for adding a function to mf-cleaner which removes any microformats with these false-positive root classnames from the tree, re-sorting any children back into place
#
aaronpk how do you know if it's a false positive? hard-code the list of known object names?
#
barnaby yeah, document them on a case-by-case basis and hard code a list of suggestions, but allow the caller to override it if they want
#
aaronpk seems reasonable
#
barnaby not an ideal solution, but IMO mf-cleaner is a good place for stuff like this which is useful but definitely doesn’t belong in the spec
#
[jacky] here's my idea of keeping the peace (in a opt-in cross-parser way) https://jacky.wtf/2022/11/9QFM
#
aaronpk yeah, xray has a bunch of stuff like that too. just takes a different approach
#
[jacky] oh dang y'all said the same thing lol
#
barnaby [jacky]: great minds, etc ;)
#
Loqi If we go with the third option, I'd be slightly pressing the suggestion to parsers to allow for an option to restrict any of the `h-` values to be ones that can represent known values, namely those with defined specifications This could be an evolvin...
#
[jacky] I can easily throw this into microformats-rust
#
aaronpk plz don't put stuff like that in the actual parsers
#
aaronpk the parsers should stay true to the spec
#
barnaby yeah I don’t think this belongs in a parser library, unless it’s explicitly designed with a batteries-included approach, and puts stuff like this under a separate namespace which makes it clear that it’s an optional extra
#
[jacky] That, I can do, as well
#
barnaby seeing as XRay can consume raw MF2 data, it’d even be possible for people to use mf-cleaner to remove the unwanted objects, and then use XRay to get a normalized representation of them
#
barnaby which is probably what I’ll end up doing
#
barnaby anyone know if we’ve encountered any false-positive property classnames? does tailwind have classnames which might get picked up?
#
barnaby I’m wondering if “has no properties” is an additional heuristic I can use for detecting false-positive root classes, although, hm, I guess something with just a root classname would get implied properties
#
barnaby I’ll stick with a list of known false-positives for the moment and consider additional heuristics as and when they become necessary
#
[snarfed] we looked at this in Indie Map recently for h-about etc. finding the queries...
#
slyduda just prefixes that are microformats specific: p- is padding for example
#
slyduda but i don't think that causes any large issues
#
[snarfed] eg https://indiemap.org/docs.html#mf2_by_sites
#
aaronpk it's much less a problem to get random properties inserted than creating whole new h-* objects
#
barnaby true
#
gRegor Bootstrap 5 has some I think, padding like p-5
#
gRegor Most have a "side" like px-5 for left and right, but for all 4 sides they have p-*
#
slyduda don't think it was linked, but here are all of the built-in heights for tailwind. issue is it's very easy to make custom classes as well like h-screen-nav
#
slyduda https://tailwindcss.com/docs/height#viewport-height
#
barnaby slyduda: any idea how common that is? for the moment my approach would be to document the built in “official” tailwind classes which could be false positives, assuming that they’re going to be the most widely used
#
aaronpk numbers aren't allowed in mf2 classes so p-5 doesn't matter
#
aaronpk lol barnaby i just got an email that you merged my 8.5 year old PR
#
barnaby lol yep, sorry for the delay xD
#
barnaby hmm I could also take the opposite approach of allowing the caller to specify a list of types they *do* want, that way it’s up to them to decide and they can list all the mf2 vocabs they support. Downside is that it’s much more likely to mess with the data structures than removing a short list of known false-positives
#
gRegor ah right, keep forgetting numbers are skipped. nice.
#
slyduda i've seen h-screen-[variations] and h-full-[variations] a built in a lot of many tailwind sites, but that's it
#
barnaby hmm, matching prefixes would be possible, but add to the complexity of the calling API
#
[snarfed] gRegor the Atom part of BF is only for Mastodon interop, which I don't totally understand and is years old, so it may have changed or dropped entirely since then
#
gRegor Ah, so sending a wm to BF for a note (with the link included) should be enough to syndicate/federate the note?
#
gRegor I might try it without the Atom feed and see :)
#
[snarfed] yes!
#
[snarfed] please do and let me know!
#
[snarfed] BF only uses Atom for OStatus, not for AP at all
#
[snarfed] I should look at which services still support OStatus and maybe drop it entirely
#
Loqi definitely
[jamietanna] joined the channel
#
[jamietanna] I've been adding the BF syndication links to my posts since yesterday and they're appearing nicely for followers ☺
#
IWDiscordRelay <c​apjamesg#4492> BF?
#
[campegg] From context, I’m guessing Bridgy Fed
#
gRegor what is BF
#
Loqi It looks like we don't have a page for "BF" yet. Would you like to create it? (Or just say "BF is ____", a sentence describing the term)
#
gRegor BF is /Bridgy_Fed
#
Loqi ok
#
@voxpelli ↩️ In my view the IndieWeb content is microformatted and delivered as HTML over HTTP and WebSub with WebMentions used for pings. Then there are the MicroPub and MicroSub parts as well, but no need to couple publishing tool with publishing place and/or reading tool. (twitter.com/_/status/1592298888891072512)
#
[snarfed] ^ "no need to couple," spicy!
#
[snarfed] oh sorry I misunderstood it
gRegorLove_ joined the channel