#dev 2022-11-15

2022-11-15 UTC
ddd joined the channel
#
angelo i have a button on my site to create a git repository. i'd like to create a token for access to it. i'm trying to reuse my indieauth server's token machinery. i'm currently simulating a standard auth flow but using an /internal/ client_id ("code-app" instead of "codesite.com") and skipping the code/code_challenge stuff. i'm currently skipping the consent screen but i'd like to eventually
#
angelo include it in this repo flow.
#
angelo does this pattern have a name?
#
aaronpk what part are you simulating if you're not doing the code exchange
#
angelo the storing it in the database and rendering it on the screen with expiration and revocation as freebies
#
aaronpk i wouldn't call that simulating. there's nothing wrong with just generating tokens
#
aaronpk some sites call this a "personal access token" and have a way to create them from the web UI when logged in
#
gRegor Yeah, I've done personal access tokens on Github
#
angelo k i basically replaced my "generate PAT" with this.. now i get to prefill in the scope and "client" details from the UX where i started (my /code/projects page). i'm storing the generated token in a htpasswd file for nginx->git-http-backend..
#
angelo has anyone heard of or tried: https://github.com/nginxinc/NGINX-Demos/tree/master/oauth2-token-introspection-oss
#
angelo or, better question, is there any way to get NGINX to use your indieauth server to validate tokens?
#
aaronpk that looks reasonable
#
aaronpk another good reason we adopted the oauth introspection draft, so stuff like this works with indieauth servers https://indieauth.spec.indieweb.org/#access-token-verification-request
#
barnaby drafted a first implementation of a function to remove false positive mf objects, initial tests seem to be working fine https://github.com/barnabywalters/php-mf-cleaner/blob/main/src/functions.php#L585 https://github.com/barnabywalters/php-mf-cleaner/blob/main/tests/CleanerTest.php#L631
#
barnaby I also made a bunch of changes to mf-cleaner including renaming master to main, so if anyone has issues installing it, that could be why.
slyduda and [tw2113_Slack_] joined the channel
#
[jacky] alright so I've outlined what I'd do for the rust parser here https://gitlab.com/maxburon/microformats-parser/-/issues/15
#
[jacky] it'll be a custom extension but opt-in and not mainline to the parser
#
[jacky] (i.e.: it's something you'd have to run _after_ parsing a document)
#
barnaby looks reasonable to me!
#
[jacky] this unlocked a memory that I wanted to do custom processing on the side and this'll allow it
Seirdy joined the channel
#
[jacky] like I can try implementing some of the experimental ideas on https://github.com/microformats/microformats2-parsing/issues and maybe add them as https://rust.microformats.io so more people can try it out
#
[jacky] this does remind me if there's other things we could experiment with (I see `<picture>` tag) to help with providing more parser implementations (or even a chance to test it out)
#
[jacky] I know I'd like to swap to using the picture tag more often
#
[jacky] like this one too https://github.com/microformats/microformats2-parsing/issues/47
#
Loqi [dshanske] #47 Implied h-feed
#
barnaby huh I’d not seen that issue before. there’s no precedent for implying entire mf2 objects (only properties, and those only under *very* specific conditions to avoid false positives) so I’m strongly against implied h-feed
#
barnaby implying a feed from a list of h-entries is a job for consumers, not the parsing spec
#
barnaby will leave a comment there
#
[jacky] I agree! And I can see that being something I can add as a 'helper method' for this
#
[jacky] Another good one (that I actually didn't consider and would like to avoid for my own wants) was raised in https://github.com/microformats/microformats2-parsing/issues/50
#
barnaby [jacky]: imo xray is the current best example of how and when to imply a feed, and worth reviewing if you’re planning on writing similar code e.g. https://github.com/aaronpk/XRay/blob/main/lib/XRay/Formats/Mf2Feed.php#L6
#
[jacky] will def look into this
#
[jacky] added it to the ticket for reference
geoffo joined the channel
#
[jacky] ugh yeah this was a good idea
#
[jacky] there's other things I'd want to sneak in here like wrapping photo URLs with a custom method (helpful when wanting to render things with a proxy, for example, or extract only content of a particular language)
#
GWG barnaby: https://microformats.org/wiki/h-feed#Implied_h-feed
#
GWG barnaby: The spec editor added something about it.
#
[jacky] oof that page reminded me of mf1 and I've been aggressively avoiding implementing that lol
#
barnaby GWG: it’s an entirely reasonable thing for consuming code to do, but doesn’t belong in the main mf2 parsing spec
#
[jacky] I understand needing to keep the spec lean but things like this _could_ live there
#
GWG barnaby: It isn't in the parsing spec, but it is helpful it is in the mf wiki
#
[jacky] yeah tbh
#
[jacky] any algos like that I'd like to see in the wiki
#
[jacky] esp if people found them useful
#
[jacky] saves time in discovery by manual labor 🙂
#
barnaby oh absolutely! to be clear, in my comment on https://github.com/microformats/microformats2-parsing/issues/47 I’m strongly against this behaviour being included in the mf2 parsing spec itself, which is what that repo is for
#
Loqi [dshanske] #47 Implied h-feed
#
[jacky] hmm bet
#
barnaby and as usual, the indieweb wiki is the better source for consuming algorithms e.g. https://indieweb.org/feed#How_To_Consume in this case
#
barnaby not sure how much sense it makes to duplicate these algorithm definitions between the two wikis
#
[jacky] has been slightly frustrated with this in the past and was curious about others' thoughts
#
[jacky] I think it'd make more sense for it to live on the mf wiki and the IW wiki link to it
#
barnaby information being spread over the two wikis, you mean?
#
barnaby ah okay
#
@typelogdev Can someone explain activity pub to me? Is it an official web spec? What does it give that webmentions don't (still leaning about these)? #indieweb #activitypub (twitter.com/_/status/1592339297637920768)
#
[jacky] one day, I'll promote do a slow wikiathon
#
barnaby I have no strong opinions about which wiki is the best place for it, I mostly want to avoid having multiple inconsistent copies of the same thing, as we’ve seen that confuse and frustrate people
#
barnaby gotta go for now, definitely interested in discussing this further and working to improve it though
#
[jacky] o/
#
GWG barnaby: I put it there because we lack a better place to discuss it. We need a 'consuming algorithm issue' repo maybe
#
[snarfed] eg https://github.com/indieweb/authorship
#
Loqi [indieweb] authorship: Authorship is an algorithm for discovering and determining the author of a post.
gRegor joined the channel
#
[jacky] hmmm
#
[jacky] attempting to deploy Sele on fly.io has aggresively forced me to test the app more (mainly found a bug in a library I use for config resolution) but hopefully I can get a demo before HWC this week!
#
[jacky] _hopefully_ lol
#
[jacky] first mistake of software dev is creating a deadline
#
@nileshtrivedi ↩️ RSS doesn't support comments by readers. WordPress had support for pingback/Webmention but users have shown preference for microblogging rather than long-form blogging. Plus WP is no replacement for YouTube/Instagram etc. ActivityPub has other flaws: (twitter.com/_/status/1592352521942892544)
geoffo and gRegor joined the channel
#
gRegor indiebookclub updates if anyone wants to try them! https://indiebookclub.biz/updates#post-2022-11-14 Doing some cleanup and testing before I add it to IndieNews.
#
Loqi Re-try publishing: If you experience an error publishing to your site with Micropub — or if you added Micropub support after you started using the app — you can now re-try publishing individual posts to your site. Browse to the post on your indie...
#
[schmarty] gRegor++ !!
#
Loqi gRegor has 18 karma in this channel over the last year (73 in all channels)
bterry joined the channel
#
gRegor [snarfed], I've got the BF redirects set up and I sent fed.bridgy a wm for a post, but I don't find myself searching the @-@. No masto account, though, so I'm guessing cache? I'm searching on https://mastodon.social/search
#
gRegor wm was successful
#
[snarfed] gRegor your redirects are dropping .well-known/ from the path
#
gRegor doh
#
gRegor ok, think that's right now
#
[snarfed] search finds you now!
#
gRegor Nice, saw the follow in the BF logs
#
gRegor Did the post show up?
#
[snarfed] no, looks like you sent it before I followed you
#
[snarfed] following doesn't backfill
#
gRegor oh right
#
[snarfed] ActivityPub 🤷
#
gRegor Not gonna get any mastodon love for my ~brilliant~ post XD
#
[snarfed] you can send the wm again if you want!
#
gRegor done
#
[snarfed] yup it's visible now
#
gRegor awesome, thanks for helping debug!
#
[snarfed] welcome!
#
gRegor and BF is using the photo from my h-card? still can't see it in search, but no biggie
#
gRegor The logs are cool. Like an AP crash course, haha
#
[snarfed] yup it has your h-card photo
#
[snarfed] oh man yeah the logs are bad UX, but they've been awesome ROI for technical users
#
[snarfed] small amount of effort, and they get used a ton for debugging Bridgy and now BF issues
#
[aciccarello] gRegor, I followed you on my toot.cafe account. I'm curious to see how BF works from the mastodon side
#
[tantek] jacky++ tailwind--
#
Loqi jacky has 30 karma in this channel over the last year (76 in all channels)
#
Loqi tailwind has -1 karma over the last year
AramZS, mro, sandy and Fokzo joined the channel
#
Fokzo Heya ! Looking for a cheap VPS to host my stuff. Mostly to prevent loss of service when my internet goes out <.>
#
Fokzo And as an offline backup
#
Fokzo So, yeah, recommendation welcome o/
#
@nezirzahirovic A Ruby gem for sending Webmention notifications. (from 28/05/2020) #ruby #rubyonrails #programming #gem #webmention https://www.rubyonrails.ba/single/a-ruby-gem-for-sending-webmention-notifications (twitter.com/_/status/1592437304487907334)
[snarfed], mro, barnaby, Fokzo and bterry1 joined the channel
#
tommorris follow up from yesterday: I emailed gijs, he removed the 'h-full' class from his html tag, now works. post on masto: https://indieweb.social/@gijs/109344119010613589
#
tommorris Fokzo: I mean, the usual - Hetzner, DigitalOcean, Linode, lotsa reviews online
#
tommorris I've mostly used Linode in the past, and have tried DO. both are fine.
#
Fokzo tommorris: True ! I'll research the topic a bit more. I was curious about what you guys used/recommended for small-scale webhosting. Thanks for the starting point :3
#
tommorris I switched to netlify and static site but at some point I may switch back to VPS for the various non-static bits.
#
tommorris (mostly because work is back-end webdev and the thought of doing more of it while not being paid isn't fun.)
[kimberlyhirsh] and [Murray] joined the channel
#
[Murray] I use a combination of Netlify for the front end, and Krystal Hosting for the back end/database/CMS, works well and fairly inexpensive
Feoh joined the channel
#
[jacky] yeah /admintax is always a thing and the level of the burden it has on the person varies on the person's technical capabilities
#
[jacky] like I can't count the number of times how people got confused about handling DNS and messed up transfers (when they wanted to move off one provider to another for a different underlying offering)
#
Fokzo Depends on your tolerance and how fun you find it. Handling sysadmin stuff for my own hosting needs is more fun than hassle for me x)
#
[jacky] that said, the last time I applied for a cell phone plan, it was all online and just was a form (one page) with Mint - domain registration _def_ requires a bit more work and if you do it wrong (forgetting WHOIS privacy which is conventionally extra for one), it can def be a privacy risk
#
Fokzo Wouldn't do it for someone else, heh
#
[jacky] yeah Fokzo, I think I was expressing that bit in "the person varies on the person's technical capabilities"
#
[jacky] like I _love_ futzing with my infra, gives me a chance to learn
#
[jacky] but if someone needed something and wanted to be online, I'm not assuming they're like me, I'd point them to micro.blog/tumblr/wordpress.org quickly
#
Fokzo Yeah, definetly. Unless you want to become their tech person ahah
#
Fokzo Hmm. Still figuring out IRC though.
Seirdy joined the channel
#
[jacky] If I can figure out https://jacky.wtf/2022/11/xt55, then I could have a demo for Sele by tomorrow tbh
#
Loqi Trying to figure out why this configuration for a Fly.io config file (source: https://git.jacky.wtf/indieweb/sele/src/branch/feature/implement-setup-flow/fly.toml) isn't allowing it to resolve when I try to connect to its URL at https://proud-pine-61...
#
[jacky] I decided to make Sele more stateful (because I wanted to provide things like WebAuthn + passkey support for authentication) but I do have an idea on how to make it such that you could _move_ providers (verrry long term)
#
[jacky] *move between providers (like from sele.jalcine.dev [not yet] to sele.onyour.domain or even your.domain/sele [preferably])
#
barnaby [snarfed]: does granary have a way of listing tweets you’ve liked? I’d like to backfeed my likes and granary.io is pretty much my only possibility for using the twitter API these days 😅
#
barnaby would be happy to send a PR if you think it’d be straightforward to add, and have an idea what the best URL scheme would be. /twitter/@me/@likes/@app maybe?
mro and gxt joined the channel
#
[snarfed] barnaby hmm! the library definitely supports it, you just do `twitter.get_activities(group_id=SELF, fetch_likes=True)`. you could grab a token from the web site and then use that in a few lines of python
#
[snarfed] (you could also plug that token into any other Twitter API tool)
#
barnaby [snarfed]: oh great, I’ll do that then. didn’t realise the library supported additional features!
#
barnaby how would using the token with other API tools work? are the granary app ID and secret public?
#
[snarfed] oh good point, no, they're not
#
[snarfed] did you have a Twitter API app that got disabled?
#
barnaby yup, had a really old one for my personal site which seems to have disappeared completely, and my application for a new one was rejected
#
barnaby not a big deal thanks to bridgy and granary though ;)
#
[snarfed] wow, disappointing
#
[snarfed] especially since they're supposedly more open now, or at least before the acquisition
#
[snarfed] when was the new one rejected?
#
barnaby yep. you have to write several short essays with enforced minimun character count about how you want to use the data, and my fairly honest description of POSSE and “personal archiving” of data from twitter was apparently flagged as “likely violating our terms of service”
#
barnaby a week ago or so? definitely post-elon, although I don’t imagine that made much of a difference
#
barnaby I could probably just make up some usage description and get approved, but I don’t particularly feel like fighting for it
#
[snarfed] wow surprising
#
[snarfed] yeah i'm familiar, I did a number of them for API v2. frustrating, sorry to hear it
#
[snarfed] the granary URL paths are based on OpenSocial, I'd be curious to see if they have anything specific for likes, but I doubt it
#
[snarfed] the other option would be to add a query param, eg ?likes=true, instead of a path
#
tommorris the one person I knew who worked on Twitter's API got uncermoniously purged the other day, without any prior notice (despite, y'know, both good manners and employment law requiring notice and consultation) so I'd assume stuff like oauth token approval is probably fairly low priority right now
#
[snarfed] small silver lining, the chaos has clearly deprioritized migrating to twitter API v2, at least for me
#
barnaby tommorris: yeah I was surprised that I got a response so quickly, given the chaos
#
barnaby also, hi tommorris! nice to see you again
#
tommorris waves
geoffo joined the channel
#
barnaby [snarfed]: what should scrape_headers be when I’m trying to use fetch_likes=True? when I pass an empty dict I still get a NotImplementedError: fetch_likes requires scrape_headers
#
barnaby hmm and when I try a simple tw.get_activities(group_id='@friends') from the docs I get a “ValueError: Only unicode objects are escapable. Got None” from somewhere deep in the oauth library
#
barnaby I’m passing my access token key and secret from the fields on https://granary.io/ to the constructor, is that the correct usage?
#
[snarfed] ah right, granary twitter has to scrape likes since they aren't available in API v1, ugh
#
barnaby just to avoid misunderstandings: I’m specifically looking for tweets which I’ve liked, not lists of people who’ve liked my tweets. skimming through the granary twitter code it looked like there were code paths for fetching both
#
[snarfed] it's an HTTP headers dict with X-CSRF-Token, Authorization: Bearer ..., and Cookie with auth_token and ct0
#
[snarfed] yup. the fetch_likes kwarg turns on both, I don't think there's currently a way to ask for one but not the other
#
barnaby ah, okay. is there an example of how to construct the headers dict somewhere in granary?
#
[snarfed] passing your accerss token and secret to the Twitter constructor is right, but to develop and test, you'd also need granary's app id and secret
#
[snarfed] not really. it's {"X-CSRF-Token": "...", "Authorization": "Bearer ...", "Cookie": "auth_token=...; ct0=..."}
#
[aciccarello] barnby, I may have missed this but if you are backfilling could you use a twitter data export?
#
[snarfed] but I'm not sure what that bearer token is, since Twitter API v1 is OAuth 1.1, which needs both access token and secret
#
barnaby [aciccarello]: do twitter exports include all the tweets you’ve liked? that’s specifically what I’m looking for
#
[snarfed] oh and re the ValueError, feel free to post the whole stack trace and I can look!
#
barnaby it’s been years since I looked at a twitter export so I’m not up to date on what’s included in them
#
[snarfed] oh! if this is one time, then definitely look at the download first!
#
[snarfed] sorry, I assumed you wanted an ongoing service/tool
#
barnaby yep, one-time. if I can get it out of the archive that’d be ideal, I just assumed that it would include likes for some reason
#
@typelogdev Anyone know of a webmentions test tool that will tell you why it failed if it did? Trying to post to /webmention ur… https://twitter.com/i/web/status/1592529735946760193 (twitter.com/_/status/1592529735946760193)
#
barnaby I will request an archive and see if it gives me something I can work with. even just a list of URLs would be sufficient
#
barnaby thanks!
#
[aciccarello] I assume it does. I just exported this week for historical purposes. Let me see if I can find that piece. There are a million json files.
#
barnaby [snarfed]: here’s the stack trace: https://gist.github.com/barnabywalters/57c3095b89506cd074c68d387f20ecba
#
[snarfed] I looked at one of my twitter exports from 2019 and it has my likes
#
[aciccarello] Yep, there is a likes.js file
#
barnaby yay! I’ll do that then. thanks a lot, apologies for the unnecessary support work 😅
#
Loqi :D
#
barnaby heh if I import all ~6000 twitter likes into my site, it’ll be the largest amount of one type of content by far
#
[snarfed] np! thanks for the thought!
#
barnaby will be interesting to see how much space it takes up, especially as I’m planning on storing local archives of all the media. maybe I’ll skip videos
#
@schnarfed ↩️ https://webmention.rocks/ ? (twitter.com/_/status/1592556941192364033)
#
barnaby [snarfed]++ hopefully you can take it as a compliment that my first thought is to use granary to get useful data out of twitter ;)
#
Loqi [snarfed] has 38 karma in this channel over the last year (68 in all channels)
#
[snarfed] definitely!
#
@schnarfed ↩️ AP and webmention are indeed both W3C recs: https://www.w3.org/TR/activitypub/, https://www.w3.org/TR/webmention/. AP is aimed more at federating multi-user servers with high traffic push JSON, webmention more at individual personal web sites, HTML w/microformats, more balanced btw push and pull. (twitter.com/_/status/1592558358376382465)
#
@schnarfed ↩️ Yes, webmention needs your HTML to be fetchable by servers, without running JS. (twitter.com/_/status/1592558555126984704)
paulrobertlloyd joined the channel
#
paulrobertlloyd @aaronpk Hello. I’m building my own IndieAuth endpoints, and looking to insure some level of interoperability with IndieAuth.com, which I know is somewhat outdated. I’ve noticed that a POST to /auth at IndieAuth.com returns a `scope` value as well as `me`. Is this a valid response? The spec suggests only `me` should be returned. Should implementations ignore anything besides `me`?
#
[Murray] @barnaby I'd be interested to know how that goes. I've seen a few of the tools for converting Twitter archives have struggled with extracting likes, e.g. https://github.com/timhutton/twitter-archive-parser
#
Loqi [timhutton] twitter-archive-parser: Python code to parse a Twitter archive and output in various ways
#
[snarfed] ^ not easy to compare ActivityPub and webmention in a single tweet, but I tried
#
barnaby [Murray]: I’ll probably end up just grabbing the tweet IDs from the archive and then using granary.io to fetch them and convert them into html+mf2 which I can process further
#
barnaby [snarfed]: looks like a good tweetsized summary to me
#
[Murray] ah nice suggestion, wouldn't have thought about that (_makes notes_)
#
barnaby my processing chain for re-importing my old twitter reply contexts (and going forward for however long it lasts) is converting the tweet URL into a granary URL, fetching the html+mf2 equivalent, passing that to XRay and then post-processing the almost-jf2 which XRay produces slightly before storing it
#
barnaby so I’ll do something similar for importing my old likes
#
aaronpk paulrobertlloyd: it'd be better to compare against https://indielogin.com, I really need to spend some time to shut down indieauth.com at this point. it still has a lot of old behavior
#
[jacky] interesting tool for 11ty users https://github.com/11ty/eleventy-activity-feed
#
Loqi [11ty] eleventy-activity-feed: Create one centralized RSS feed for all of the content you create across the web (aggregates from Twitter, RSS, Atom, Mastodon, YouTube)
paulrobertlloyd joined the channel
#
paulrobertlloyd @aaronpk Oh, I’ll check that out. I previously built a token endpoint and got that working alongside IndieAuth.com/auth, but swapping out that for my own auth server, finding a whole set of assumption that are likely misplaced!
slyduda, gRegor, [Murray], [kimberlyhirsh], [chrisaldrich] and [campegg] joined the channel
#
[jamietanna] Jacky still having issues with Fly?
paulrobertlloyd joined the channel
#
paulrobertlloyd @aaronpk IndieLogin only does authentication, not authorisation, right? I’m getting a bit lost with where a token endpoint would know the desired scope, assuming the auth and token endpoints are completely separate.
#
aaronpk oh, yeah indielogin.com might not be what you're looking for then
#
aaronpk we're kind of moving towards auth and token endpoints not being separate things in general though
#
aaronpk you can still do it if you want, but that's not a pattern i plan on supporting in anything i'm building anymore
#
paulrobertlloyd Ah! Right, because with it all under my control, I know that a auth code is a JWT… but that might not always be the case.
#
paulrobertlloyd So it sounds like I should go all in on my own server that does all the authentication journeys. Certainly makes things much easier. I’ll stop worrying about having things work with older/discrete services.
#
aaronpk yeah, when both endpoints are coupled then you can just do things however you want internally
#
sknebel thats also where the spec design overall moved to
#
sknebel we dont really spec the interplay between the two anymore I think?
#
barnaby paulrobertlloyd: are you building indieauth endpoints for use for authorizing client apps on your personal site, or are you implementing something with relmeauth support for logging other people into your site?
#
paulrobertlloyd Yup, all the sections explicitly about token endpoints has been removed.
#
barnaby if the former, then I doubt there’s much reason to imitate how indieauth.com used to work, and you’re better off just following the existing spec IMO
#
[jacky] jamietanna: nope! the founder actually replied to a tweet that sparked a solution
#
barnaby [jacky]: nice, when in doubt go straight to the top ;)
#
[jacky] lol right!
#
paulrobertlloyd Think I’m going to avoid using RelMeAuth as few services seem to work with it anymore. Might look at PassportJs strategies… need to work on the actual authentication step, so far have basic password entry.
#
[jamietanna] Awesome 😁
#
barnaby for a single user instance, password based authentication is quite sufficient and probably the quickest and easiest to set up. I bundle an auth backend with Taproot/indieauth which does exactly that, for getting set up quickly
#
paulrobertlloyd Right now my password works by storing a hash of it in an environment env, but possibly move storing that value in a database, not sure yet.
#
paulrobertlloyd Here’s my WIP indieauth endpoint FWIW: https://github.com/getindiekit/indiekit/tree/499-endpoint-authorization
#
aaronpk indielogin.com takes a different approach to relmeauth and relies on the silo APIs instead of scraping the silo's HTML
#
aaronpk so it's been super reliable for twitter and github
#
paulrobertlloyd Maybe less so the former ;-)
#
barnaby yep the taproot single user password auth backend does something similar, you just store a hashed password wherever you want an pass it to the constructor
#
paulrobertlloyd Do you have a link to your Taproot implementation?
#
barnaby https://github.com/Taproot/indieauth
#
paulrobertlloyd Ta
#
barnaby should be fairly self-explanatory but I can point you to specific parts if you’re interested
#
Loqi [Taproot] indieauth: A PSR-7-compatible PHP IndieAuth Server Implementation
#
barnaby so far it’s only interested in solving the “turn your site into an indieauth identity provider” problem though, not the “let arbitrary people log into your site using relmeauth etc” problem
#
[jacky] kind of wants relmeauth to die out
#
[jacky] only b/c it introduces more moving complexity in stuff (mainly chasing silos and hoping they don't break things, which doesn't happen often tho)
#
barnaby I’ll be happy when the indieauth/indielogin/relmeauth confusion finally starts to go away
#
sknebel Maybe we should put a large explanation of the difference on indieauth.com, and only a tiny link to it's remaining features
#
@nachtfunke Idk why I haven’t found this article earlier https://amberwilson.co.uk/blog/grow-the-indieweb-with-webmentions/ (twitter.com/_/status/1592577651948109825)
#
@CodingPotions ↩️ En la IndieWeb también mencionan Bridgy que precisamente lo que hace es eso, publicar tu contenido también en redes, pero no lo he probado y no sé si tiene soporte para lo que digo de tweets más largos en hilos, que eso me molaría mucho. https://brid.gy/ (twitter.com/_/status/1592582657111908352)
a and mro joined the channel
#
gRegor [snarfed], I know it's not really syndication, but is there a URL for posts we send to Bridgy Fed that would make sense to put in the u-syndication?
#
[snarfed] gRegor hmm, good question! I'm not sure
#
gRegor Not sure if a bug, but the following/followers URLs here are missing the /user path https://fed.brid.gy/gregorlove.com
#
Loqi gRegor Morrill
#
gRegor Loqi
#
[snarfed] ugh thanks, will fix
#
[snarfed] oh wait that's the AS2 actor
#
[snarfed] yeah that's intentional. those aren't UI links
#
[snarfed] btw interestingly Mastodon instances do have local URLs for BF posts, eg https://indieweb.social/@snarfed.org@snarfed.org/109327939703730432
#
[snarfed] probably only works if you're logged in, and specific to indieweb.social. looks like it redirects to the post on my site if you're logged out
#
Loqi [Ryan Barrett] Xfinity WiFi Hotspots is a great product. I can walk into pretty much any coffee shop or bar or park, open my laptop, and I’m immediately online. No begging baristas for a password, no sketchy FREE PUBLIC INTRNET 4 REALZ access points, no fiddly te...
#
gRegor Yeah, got the redirect
mro joined the channel
#
[snarfed] BF uses https://fed.brid.gy/r/[your full URL], eg https://fed.brid.gy/r/https://snarfed.org/2022-11-09_47872 , as the AS2 object's id and url, so you'd probably use that
#
gRegor Trying a follow post now [aciccarello]
#
jeremycherfas Throwing myself on the good people here again with a Python problem.
#
[snarfed] it redirects to your post though, so might not be useful to show visually
#
jeremycherfas I have a string in the form `[12.45924, 41.88564],[12.45935, 41.88573]` The digits are, of course, part of the one string. I would like to convert each set of digits into a float. Ideally I would end up with a list of lists, i.e. `[[12.45924, 41.88564],[12.45935, 41.88573]]` in which the numbers are floats.
#
[snarfed] jeremycherfas parse as JSON!
#
jeremycherfas I've tried everything I can think of and nothing seems to work. Very hard to search for solutions too. The numbers are lat long, if that makes any difference.
#
[aciccarello] gRegor, I unfollowed/refollowed @gregorlove.com@gregorlove.com from my toot.cafe mastodon account. Are you looking for something in particular?
#
jeremycherfas But there are no keys. Are you suggesting I add Keys?
#
gRegor Don't think so, just posted a follow post which I think BF will show me as a follower on toot.cafe
#
gRegor Thought I don't see anything in the BF logs yet. The webmention sent there was good though.
#
barnaby jeremycherfas: it should parse to a list of lists of floats, no keys necessary
#
sknebel jeremycherfas:
#
sknebel >>> json.loads("[" + "[12.45924, 41.88564],[12.45935, 41.88573]" + "]")
#
sknebel [[12.45924, 41.88564], [12.45935, 41.88573]]
#
[snarfed] jeremycherfas no, JSON objects don't always have to be mappings
#
[snarfed] right
#
jeremycherfas Smacks head!
#
jeremycherfas Thanks sknebel++
#
Loqi sknebel has 16 karma in this channel over the last year (36 in all channels)
#
jeremycherfas Snarfed++ too
#
Loqi Snarfed has 39 karma in this channel over the last year (69 in all channels)
#
jeremycherfas And, while I am about it, barnaby++
#
Loqi barnaby has 32 karma in this channel over the last year (49 in all channels)
#
barnaby jeremycherfas: and if your data didn’t happen to be parseable as JSON, provided the strings are all the same format you could do something naive like splitting on ',', stripping spaces and brackets and parsing what’s left as floats
#
[aciccarello] gRegor, I don't see a new follow on my toot.cafe account 😕
#
barnaby and then use that until it breaks
#
barnaby heh in that case you could even eval() it as raw python, although that’s much more dangerous than parsing as JSON of course
#
gRegor Showing up in the logs now, though not the following number there, so maybe takes a bit of time https://fed.brid.gy/user/gregorlove.com
#
[snarfed] gRegor when did you send the follow wm to BF? I'm not seenig it
#
gRegor oops, misread the logs
#
gRegor those are the re-follow you did I think
#
gRegor There we go!
#
[aciccarello] Hey! It showed up!
#
gRegor showing up in followers now
#
[aciccarello] (I don't have a long list to look through 😉)
#
[aciccarello] snarfed++
#
Loqi snarfed has 40 karma in this channel over the last year (70 in all channels)
#
[snarfed] 🎉
mro and [tonz] joined the channel
#
barnaby after eight years, there’s new version of php-mf-cleaner! https://github.com/barnabywalters/php-mf-cleaner
#
Loqi [barnabywalters] php-mf-cleaner: Takes microformats2 array structure and makes it more sane. Counterpart to indieweb/php-mf2
#
[schmarty] barnaby++
#
Loqi barnaby has 33 karma in this channel over the last year (50 in all channels)
[pfefferle], vilhalmer, mro, sp1ff, gRegorLove_ and geoffo joined the channel
#
[schmarty] oof talk about admintax. popular "we'll run your mastodon instance for you" host masto.host is no longer taking new signups. and the last two days of updates seem like they have been grueling. https://mastodon.social/@mastohost/109344646859979730
#
Loqi [Masto.host] I will be starting the upgrade to Mastodon v4.0.1 for everyone on Masto.host.Usually I would test a release for 24 hours before installing it but as this one includes a security fix, I will bypass the wait and start the installation.You should expect...
#
aaronpk oh gosh
[Zeina] joined the channel
#
Zegnat 4.0.2 does notmention any security fixes. So if 4.0.1 runs well for masto.host there should not really be any harm in not instantly pulling in the path release?
#
Zegnat Thought them not taking new sign-ups does not surprise me too much. Been seeing a bunch of people talking about cost of running mastodon these last couple days
#
[timothy_chambe] Dave Winer is now on Mastodon, FYI:
#
[timothy_chambe] https://twitter.com/davewiner/status/1592644579903541249
#
@davewiner I'm getting a new follower every 5 seconds on http://mastodon.social. at times one every second or more. this is pretty amazing. i'm so hyped up now. what a day. (twitter.com/_/status/1592644579903541249)
#
[tantek] [timothy_chambe] curious, how is that relevant to #indieweb-dev?
#
[tantek] while also implying a discouragement to others
#
[tantek] or something else I'm missing?
#
Zegnat [tantek]: interested to hear your thoughts, are pages like /now that have paths-by-convention another example of the /well-known antipattern? Even though the idea of them is to be complete content pages by themselves with no fixed structure?
barnaby joined the channel
#
[tantek] excellent question and I think that helps get at the heart of *why* well-known is an anti-pattern, which is consuming-code dependence
#
[tantek] e.g. pages like /now I would expect to be translated for people’s preferred / first language
#
[tantek] like if you have a German site, perhaps you'd have a /jetzt page instead of a /now page
#
[tantek] Zegnat, do we have a term for such paths-by-convention / pages like /now (and /contact or /about or /pay etc.) ?
#
[tantek] more of an emergent pattern than a prescribed pattern, that's another contrast with the /well-known antipattern
#
[tantek] what are common pages
#
Loqi It looks like we don't have a page for "common pages" yet. Would you like to create it? (Or just say "common pages is ____", a sentence describing the term)
#
[tantek] what are pages
#
Loqi pages is a place on your website where you list and link to individual (typically top level) page URLs on your own site, sometimes part of top level navigation or a list of links on a homepage https://indieweb.org/pages
#
[tantek] common pages are /page
#
Loqi ok