#dev 2020-12-10
2020-12-10 UTC
[tw2113_Slack_] and [Raphael_Luckom] joined the channel
# 
[Raphael_Luckom] I remember a conversation from maybe a week ago, about how the original rss spec specified max 15 items per feed. With modern RSS and atom, is there still a max? are feeds paginated?
# [Raphael_Luckom] right--I found this RFC but it seems more complicated than I want to do if no one else does: https://tools.ietf.org/html/rfc5005#section-3
# [Raphael_Luckom] lol gotta stop doing that
# 
aaronpk i'm pretty sure the vast majority of rss feeds set some arbitrary limit and just let older things fall off, assuming readers are only interested in the latest ones. the major exception to that is podcast rss feeds which do usually contain the full list of podcast episodes, because apple treats the rss feed as the canonical list of full episodes
[tantek] joined the channel
# [Raphael_Luckom] perfect, I'll pick a number, thanks!
# [Raphael_Luckom] aaronpk++
# 
[KevinMarks] Atom had better pagination defined as part of atompub etc, I think 5005 was part of that. Standard ways to query for n to m etc and prev/next
# [KevinMarks] Messy boundary conditions if items are added while you're doing page queries
# [KevinMarks] So define the next and previous links carefully. I think granary implements a lot of this stuff
# [Raphael_Luckom] right...tbh I'd rather start out with a flat feed file and then build out pagination later
# [KevinMarks] There was a point when all Google apis were atompub. Which didn't really work for clientside js
# [KevinMarks] They had an amazingly bad bodge for that that wrapped the xml in magic json keys with $ in
# [Raphael_Luckom] AWS handles boundary conditions in a lot of APIs by having you specify the last element you saw instead of a page number. So any shenanigans at the top of the list are invisible if you're iterating
# [KevinMarks] Right, that's a better approach, especially if you can guarantee consecutive ids (hence snowflake at twitter)
# [KevinMarks] Or at least increasing ids.
maxwelljoslyn joined the channel
# [KevinMarks] You are trusting that the server is doing it right though. You do still need to dedupe by url for overlap. Trickier is catching gaps.
# [KevinMarks] You end up with TCP like "ACK up to" requests where you poll with the last one you saw.
# [Raphael_Luckom] I usually try to do the simplest thing that works. Which means one file to start out, and even after that I'd probably use multiple static pages before implementing active pagination server-side.
# [Raphael_Luckom] unless there's a really good case for it
# [Raphael_Luckom] on a different note, I'm finding URL templates pretty magical so far
sebbu and [snarfed] joined the channel
# [snarfed] heads up for anyone who uses the Goodreads API, sounds like its days are numbered. https://help.goodreads.com/s/article/Does-Goodreads-support-the-use-of-APIs
[Raphael_Luckom] joined the channel
# [Raphael_Luckom] I'm writing posts in markdown. When I put them in the feed (the `description` or `content` attribute), is it better to use the raw markdown, or parse to html and then escape the html, or something else?
# [Raphael_Luckom] s/parse/render
[tantek] joined the channel
craftyphotons__ joined the channel
jjuran and [tw2113_Slack_] joined the channel
# [tw2113_Slack_] Good to know snarfed. I wouldn't be surprised if they axed RSS soon too
# [tw2113_Slack_] i've already started considering not maintaining my GoodReads profile as heavily as my personal site and potentially a newly created account on bookhype.com
geoffo, KartikPrabhu, schmudde and jjuran joined the channel
# 
@adactioLinks Grow the IndieWeb with Webmentions | Amber Wilson https://amberwilson.co.uk/blog/grow-the-indieweb-with-webmentions/ (twitter.com/_/status/1336971799867875329)
swentel and swentie joined the channel
# 
@qubyte Grow the IndieWeb with Webmentions | Amber Wilson: https://qubyte.codes/links/1607597111642 (twitter.com/_/status/1336986118424768513)
# 
@lostfocus ⭐️ Grow the IndieWeb with Webmentions https://ift.tt/37PXhC7 (twitter.com/_/status/1336991420922597377)
[KevinMarks] joined the channel
# [KevinMarks] feeling this https://css-tricks.com/npm-ruin-dev/
KartikPrabhu, jeremy, ethanyoo, NinjaTrappeur and [Raphael_Luckom] joined the channel
# [Raphael_Luckom] Thanks [aaronpk] and [tantek]. That was basically my question--is .md "text-like" enough for inclusion in a feed. After looking at aaronpk's website feed and thinking about my rendering process, I think html in an xml cdata element is the way to go
ethanyoo, geoffo, jeremych_ and [hibs] joined the channel
# [Raphael_Luckom] I like that approach too. Professionally, I've only ever worked on teams that use big web frameworks and single-page apps. One of my favorite parts of working on my own site is not having to justify choosing not to use that stuff.
ToddontheHusting and [Murray] joined the channel
ethanyoo, [chrisaldrich] and [tantek] joined the channel
# 
[tantek] Anyone working on replying (heh) to Scott Jenson and actually demonstrating the technology (i.e. replying from your site) as part of the answer, in contrast to every other reply that is only a disconnected tweet? https://twitter.com/scottjenson/status/1337036596806356993
# 
@scottjenson #indieweb folks. there are a ton of self hosted commenting systems. But what about a "universal comment" system? Where all of my comments, on any site I use, are just links back to my server. I own/host all of my comments. I'm assuming this has been discussed? (twitter.com/_/status/1337036596806356993)
KartikPrabhu joined the channel
# 
@aaronpk @scottjenson This sounds like what Webmention enables. This tweet for example is actually a copy of my reply to you that I've posted to my website. It works directly from site to site too of course. (twitter.com/_/status/1337082365458272256)
# @aaronpk ↩️ This sounds like what Webmention enables. This tweet for example is actually a copy of my reply to you that I've posted to my website. It works directly from site to site too of course. (twitter.com/_/status/1337082365458272256)
# 
[chrisaldrich] swentel++ for the Drupal updates!
[tw2113_Slack_] joined the channel
# aaronpk not sure I understand the "but" here https://twitter.com/scottjenson/status/1337086293214253057
# @scottjenson @aaronpk Excellent thanks. What would be the 'next step'? If we could get social networks to have the tweets/comments ONLY be a link to your website what does that buy us? I appreciate the solution you have but it's about working within the current siloed approach (twitter.com/_/status/1337086293214253057)
# @aaronpk ↩️ There are at least two social networks that use webmentions natively: https://micro.blog and https://pine.blog
As for the current siloed approach, that's exactly what we're doing right now, many of us with https://brid.gy (twitter.com/_/status/1337087076089544704)
# @aaronpk ↩️ There are at least two social networks that use webmentions natively: https://micro.blog and https://pine.blog
As for the current siloed approach, that's exactly what we're doing right now, many of us with https://brid.gy (twitter.com/_/status/1337087076089544704)
# 
Loqi Pine.blog is a reader and blog hosting service that supports Webmention https://indieweb.org/Pine.blog
# aaronpk [manton] looks like micro.blog accepted a webmention of the home page and dropped it into a reply to a post on micro.blog https://micro.blog/aaronpk/10700779
# Loqi [aaronpk] @analogindex There are at least two social networks that use webmentions natively: https://micro.blog and https://pine.blog
As for the current siloed approach, that's exactly what we're doing right now, many of us with https://brid.gy
# @aaronpk ↩️ If you're talking about what existing silos can do to play better with the open web, there's no reason they can't also support webmentions natively (twitter.com/_/status/1337087303030697984)
swentel and gRegorLove joined the channel
KartikPrabhu joined the channel
# [chrisaldrich] Brid.gy has a fairly simple UI for link/no link/shortened link set of options if you use that interface.
# [chrisaldrich] (or brid.gy publish)
[KevinMarks] joined the channel
# [KevinMarks] could you rpely with just a link to your post and set the twitter preview metadata to show it?
# [tantek] [KevinMarks] that's a clever idea. Want to add that to https://indieweb.org/Twitter#Brainstorming
# 
@dracos I guess I’ll leave it alone, then ;-) https://twitter.com/_developit/status/1337101240313016320 https://pbs.twimg.com/media/Eo5aCPxXIAAG92_.png (twitter.com/_/status/1337105178055626752)
# [chrisaldrich] I was sure aaron didn't but wasn't sure about yours tantek.
# [chrisaldrich] I could have checked the source on your page as well for the necessary code to trigger bridgy too...
gxt__ joined the channel
# [chrisaldrich] I've never checked before, but I wonder if GWG's semantic linkbacks integration with bridgy silently drops the hidden metadata. I know it puts in a placeholder syndication link for some silos.
# [chrisaldrich] I have a direct syndication to Twitter myself that's labeled with my site name, but I need to fix the twitter character counter to actually use it more frequently in practice. Its been sitting dormant for too long...
# [tantek] aaronpk, do you remember (or can you look up) when you first posted a reply post? https://indieweb.org/reply#Aaron_Parecki
# [chrisaldrich] OTOH, manually syndicating to Twitter makes me do it less which has it's own benefits...
# [tantek] looks like you didn't start POSSEing notes til 2012: https://aaronparecki.com/2012/08/19/2/indieweb so it's sometime after that
rrix joined the channel
# [tantek] aaronpk, how do I navigate from https://aaronparecki.com/2012/08/19/2/indieweb to the next post chronologically? or the next note/reply chronologically?
[fluffy], rrix and [pfefferle] joined the channel
# [KevinMarks] well, trying to deploy my site to heroku and it has deprecated so I have to faff around instead.
# [KevinMarks] maybe I will just stuff it on github as it's static anyway with the smallest possible express wrapper
[schmarty] joined the channel
# [KevinMarks] well, that almost works - clearly twitter has shorter limits for external replies https://twitter.com/kevinmarks/status/1337137323453145094
# [KevinMarks] the other way may be to use large photo summary card and render text in there
# [KevinMarks] hoping for a loqi preview
# [KevinMarks] so, twitter arbitrarily truncates the summary text at different lengths https://developer.twitter.com/en/docs/twitter-for-websites/cards/overview/summary
[Emma_Humphries] joined the channel
# [KevinMarks] so proof of concept but actually more annoying than it is worth
# [KevinMarks] Though that is a way to get an editable tweet - change the card text and reload in the card test page
KartikPrabhu, schmudde, ethanyoo and [tantek] joined the channel; KartikPrabhu left the channel
# [KevinMarks] the amount of yakshaving involved was considerable as I had to recreate my static serving on heroku hack as they have deprecated the version I used originally
[manton] joined the channel
# 
[manton] [aaronpk] Oh weird, thanks for letting me know… I can see how that might happen because the permalink for that other post is actually set to “micro.blog”, so M.b tried to connect your Webmention to it. That’s clearly wrong but I’m not actually sure which part is wrong. 🙂 Maybe Micro.blog should ignore permalinks for domains other than what the user owns.
ethanyoo joined the channel
ShadowKyogre joined the channel
# ShadowKyogre Okay, in the channel
# 
petermolnar ShadowKyogre: so, what's wrong with (fed.)brid.gy?
rawtext joined the channel
# ShadowKyogre Hm. With fed.brid.gy, the webmention module for processwire returns a 200 OK status, but I can't discover the website from my mastodon account.
# ShadowKyogre I'm not sure why, but I do remember reading in the documentation that I need to interact with the fediverse at least once? Wasn't sure if that meant I had to make the first post I wrote an explicit like/etc of another toot.
# ShadowKyogre As for the brid.gy, I connected my Twitter account to my website, but it's asking for a link for a syndicated copy of the page that I posted.
# ShadowKyogre Which is a bit confusing since was also trying to use the same original post mentioned before to pub to twitter.
# petermolnar you need to send a webmention to fed.brid.gy as well; if the webmention itself went throught, it should show here: https://fed.brid.gy/responses
# petermolnar if it's there, clicking on the "time" link will give you debug logs
# ShadowKyogre ^that's the weird part. It did send a request, but that's the most feedback I get.
# ShadowKyogre will try running the same thing over curl to try getting more details
# petermolnar can you please give me the link you're trying to send the mention from?
# petermolnar the entry you're trying to get to fed.brid.gy
# ShadowKyogre yes
# ShadowKyogre https://www.aeons-library.gq/inspiration/media/
# ShadowKyogre https://www.aeons-library.gq/inspiration/media/
# ShadowKyogre https://www.aeons-library.gq/inspiration/media/
# ShadowKyogre O_o
# ShadowKyogre meant to send that one time
# ShadowKyogre Have the webmention endpoint setup, the links needed for webmentioning brid.gy instances, but can't figure out what am missing.
# petermolnar ok, from the looks, your .well-known redirects are not working: https://fed.brid.gy/ -> "How do I set it up?" -> .well-known urls
# petermolnar you need a similar redirection
# petermolnar but your webmention still should show in that list, and even if I send it by hand, it doesn't
# petermolnar so I'm checking if there's something off with your mf2
# petermolnar I can't find anything with it to be honest
# petermolnar you might want to ping [snarfed]
# 
@caleb_hearth ↩️ Thanks! I probably won’t do that but I do want to get my webmention commenting system worked on. (twitter.com/_/status/1337173164795850752)
# ShadowKyogre Gotcha. I'll go ahead and try to get a hold of him.
# petermolnar you do seem to be missing the atom feed as well
# petermolnar you do need to follow that "How do I set it up?" on fed.brid.gy :)
# ShadowKyogre Hm. I did set up the feed (it's buried up there in the header). Let me check over it one more time to see if I mistyped something potentially.
# ShadowKyogre https://www.aeons-library.gq/feed/ <- the feed that's mentioned in the head element
# petermolnar it's not as simply as normal brid.gy: webmentions, atom feed, pubsub, also pubsub on your atom feed, .well-known redirects, fed.brid.gy link, then it'll probably work.
# petermolnar *simple
# [KevinMarks] I think I have a cartoon for you
# ShadowKyogre I have the pubsub set up on the atom feed (included a link to that), but, yeah S:
# ShadowKyogre only think I can think of at this point would have to be the well-known redirects
# petermolnar but you atom feed is not linked on your site
# ShadowKyogre You mean as an HTML link?
# ShadowKyogre because the feed *is* there, but it's marked as self instead of alternate
# petermolnar <link href='https://www.aeons-library.gq/feed/' title='Feed' rel='self' type='application/atom+xml' /> -> that should be rel="alternate"
# ShadowKyogre aah, ok
# ShadowKyogre lemme fix that
# petermolnar do set up the /.well-known/host-meta and /.well-known/webfinger redirects before pinging snarfed; the last time I thought I found a bug I embarassingly forgot to have the redirect keeping the request uri.
# ShadowKyogre Yeah. What's the best way to check that those are working?
# ShadowKyogre Hm. Okay, I think I may know what's going on. It probably is the webfinger
# ShadowKyogre tweaked the url you had for my own case to see what'd pull up in browser, got a 404 page
# ShadowKyogre running curl gives this output (gimme a moment to show)
# ShadowKyogre says the document has moved to "https://www.aeons-library.gq/.well-known/webfinger?resource=acct:aeons-library.gq@aeons-library.gq" over curl
# ShadowKyogre will try sticking the rule at the bottom of the .htaccess file, thinks that ProcessWire's own rewrite rules might be overriding the initial declare
# ShadowKyogre Okay I found the rule in ProcessWire that's causing it,
# ShadowKyogre just need to figure out how to create the exception specifically for fed.brid.gy
# ShadowKyogre I think it's that
# petermolnar that's a nasty big redirect file
# petermolnar but yes, that looks suspicious
# ShadowKyogre yep @_@
# petermolnar easy test: remove the line, see what happens.
# petermolnar if fixed, that was your problem, so one needs to find a permanent fix
# petermolnar if not, the problem is not there :)
KartikPrabhu joined the channel