#jeremycherfasWonderful. Thanks. I may have got some interest from @mapcyka
KevinMarks joined the channel
#ZegnatThe description value seems fine in the Like plugin, there is just no way to set it through the Micropub endpoint (I think)
#jeremycherfasThat was the general conclusion we came to before.
#ZegnatYou are probably able to post a description value in the Micropub request. I haven’t tested that. But that would be counter to Micropub and not be compatible with any other endpoint.
#ZegnatThere are definitely several people who would love to see better Micropub in Known, so if you got their attention, do feel free to rope in other people for assistance ;)
#jeremycherfasIt doesn't work with either micropub or with using the bookmark API
#jeremycherfasI've given some details of the request and the results. Now, thanks to your help, I can flesh out what is happening. I couldn't remember the details.
#jeremycherfasMaybe I'll even write a post that I can refer to.
#Zegnatjeremycherfas++ for documenting stuff for real
#Loqijeremycherfas has 5 karma in this channel (14 overall)
#dgoldpusher laravel was the key for the google search
#aaronpkOh I think I left some boilerplate stuff in the config template
#Loqiaaronpk: [eddie] left you a message 15 hours, 11 minutes ago: Created two PRs on Aperture for the scope string/array and the different channel scopes.
#aaronpkI used pusher for a project last year, it definitely made things easy, but I prefer to run that infrastructure myself
#dgoldisn't varchar(256) smaller than varchar(512) - one header byte and...
#aaronpkBecause it's utf8mb4 it's actually 4x the bytes but the indexer caps out at 767 bytes so the limit is actually 767/4
#aaronpkOh yeah 256 might not work... might need to do 190. The other fix is the limit the index to only the first 767 bytes but I don't think there's a syntax in the ORM for doing that so you have to do it manually
#aaronpkThe other fix is to update to MySQL 5.7 :-)
#dgoldhuh. i might do that - although this is 'mysql' aka MariaDB
#[gerwitz][tantek] do you have opinions on whether autolinking (or embedding) should be applied at author time, or view time?
#[gerwitz](I always lean toward author time, but links and rules get outdated)
#tantekgerwitz interesting. for such "automatic" things, I lean towards "view time" to 1. preserve precisely what was authored (rather than automatically implied), and 2. assuming that the "autolink" software may be updated for fixes or to autolink more things in the future
#tantekI have observed that Twitter however applies autolinking at publish time, and thus we are able to see a history of when Twitter started autolinking things
#[gerwitz]Meh. Autolinking is sugar, so I think I will deal with that and let the browser do it. After all, it kind of feels like something a mature browser should be doing for you, with user preferences.
#tantekI think the points you bring up are worthy of capturing for additional discussion
#sknebelfor URLs, doing it client-side means no webmention sending for them
#[gerwitz]good point. I already am in a jam around this because I have an asynchronous publishing pipeline.
#sknebelright, you mentioned having to wait for cloudfront rollout?
#[gerwitz]yeup. I’ve shortened the time enough that it works for me as a user. (In fact, I *like* that it takes more than 60 seconds from me hitting “post” to it existing on the web. It reduces the temptation to tweetstorm.)
#[gerwitz]I may go hybrid and detect at “build” time that recognizable data exists, and mark it up for handling client-side. Sending webmentions might be facilitated with a “mentions” feed that another service consumes and uses to ping
#sknebelyeah, if there is a good bound on the time until a page is available you could just delay sending webmentions
#Loqieddie has 16 karma in this channel (34 overall)
#tantekgetting the "reply flow" to be as smooth as possible has been one of the longstanding big challenges in the indieweb and frankly any "federated" systems
snarfed, [cleverdevil], [xavierroy], benwerd, [tantek], [chrisaldrich], tantek, chrisaldrich and jmac1 joined the channel
#[cleverdevil]Syndication targets aren't showing up for me in Indigenous.
#[eddie]Can you open up settings -> debug info and copy and paste that info into a DM, email or github issue?
#[eddie]That should potentially help shed some light :crossed_fingers:
#jackjamiesonsnarfed: Does bridgy do two passes when fetching posts then? i.e. first find permalinks in the h-feed, then parse each permalink individually?
#snarfed(as an optimization, it remembers if it's seen synd links in a given feed before, and skips the permalink fetches for that feed)
#snarfedjackjamieson: out of curiosity, why do you care about synd links in a reader? do you render them yourself?
#jackjamiesonI render them in the feed, but truthfully they're not a high priority. I've been including them as a matter of completion (and I guess I find it interesting to see where people are syndicating)
#jackjamiesonTo further answer [eddie]'s question about what I've struggled with (that microsub might help solve), generally I've found normalizing data to be a challenge. But much of that might be resolved for now. Probably I need to spend a bit more time with the microsub spec to understand it better
#[eddie]Ahh, that makes sense, jackjamieson. From the front-end perspective microsub definitely helps normalize data. I am NOT a great person at data manipulation, so it has made it incredibly easier to build Indigenous depending on a microsub server.
#[eddie]That said, there is only one alpha testing mode Mirosub server (called Aperture), so we definitely need more Microsub servers that handle the normalization
#[eddie]Since your project is currently handling both aspects of Microsub (parsing and viewing the feeds), your project is in a unique place where if you have resolved the majority of issues you’ve come across it might be more useful to have Yarn support the Microsub server capabilities.
#[eddie]It would give you a standard normalization to attempt to store your data as, and would allow people who use your plugin to essentially host their microsub data in Wordpress
#[eddie]Hmmm how would authentication work on those? I guess it would be like Bridgy where you log in via IndieAuth, and then into the thing such as feedly to connect the two
#snarfedoh no i don't expect individual users would ever see or auth into the feed fetching service, just the wrapper
#snarfedeg bridgy uses superfeedr, but bridgy users don't know or care
#jackjamieson[eddie] that's a good idea. I expect I'd have a fair bit of work to do to further improve my normalization, but I like the idea of being able to support other clients. I have some other goals to figure out first, but it's worth considering
#[eddie]Feedly is a personal feed storage service, though right?
#snarfedyes, but afaik also a feed fetching service for developers
#[eddie]jackjamieson: Definitely want to get your priorities taken care of first.
#[eddie]snarfed: I was imagining essentially a Microsub endpoint that allows you to get feeds from Feedly in a Microsub client rather than behind the scenes fetching. But apprently Feedly’s API only works for the $18/month Pro user
#[eddie]If people were willing to pay $18/month for Microsub, we might as well just build a pay to host service rather than build a Feedly wrapper. lol
maingo joined the channel
#jackjamiesonAt any rate, thanks for the help! I've got to run for now
#[eddie]Sounds good. Have a good one, jackjamieson!
#sknebelI have an Inoreader2microsub bridge on my list of projects, as soon as I have time for something like that and read-state is specified
#snarfed[eddie]: sure, i can see both alternatives
#snarfed1) connect to existing user accounts on existing feed reader services. requires user login, etc. the service still owns each user's feeds, read status, etc.
#snarfed2) just use services as plumbing, for fetching feeds etc. the wrapper owns user data.
#snarfed(for 2, for feedly, the microsub wrapper would pay, and that cost would be amortized across all users)
#Zegnat“microsub wrappers around existing feed fetching services” – I was even thinking the other way around, bridge Microsub to Fever API. So I could try running Aperture yet keep using my current RSS reader.
#[eddie]!tell aaronpk: Known’s syndicate-to returns: “syndicate-to”: [ “indiesyndicate::https://news.indieweb.org/en” ] is this a previous standard or just something Known made up?
#LoqiFever was a self-hosted, proprietary feed reader that is no longer developed or sold as of 2016-12-24 [1] https://indieweb.org/fever
#dgoldZegnat: pity that sooooo many of the apple-crowd seem to reject it out of hand because Shaun Inman stopped selling it
[kevinmarks] joined the channel
#ZegnatI think Shaun did the right thing stopping sales though, as he hadn’t worked on it for quite a while. Guess the alternative would have been to go open-source.
#ZegnatThe API readers could use to communicate with a Fever server got implemented by several readers and subsequently also implemented/emulated by different servers.
#KartikPrabhuyeah, like a closed-source, non-open microsub
#dgoldI found four aggregators in the space of those few minutes on github, I'm pretty sure there are others
#ZegnatThe Old Reader, I think, emulates the Google Reader API still. But most servers seem to have moved away from that one. At least the ones I have seen in the last couple of years.
#dgoldstays schtum on the Old Reader people, on the basis that I was thought if you can't say something nice, then say nothing.
#ZegnatIn as far as documentation is ever “closed-source, non-open”, yes. KartikPrabhu. But as Fever reached feature completion a while ago, I do not think the API actually got changed much if at all. So it was pretty safe to implement/emulate it.
#dgoldthe nice thing about the API is that it can spit json at you, or xml if you're really _hardcore_
snarfed joined the channel
#KartikPrabhui don't really like to be spat at especially with XML :P
#Zegnatfollowing the link in a browser I think is logical to give an unauthorized error. It probably requires an Authorization header with a token.
#ZegnatMaybe you need to reauthorise with Indigenous? I expect the token it uses needs to have certain microsub scopes, and it may not have requested those first time ’round?
#dgoldahhh. there's nothing showing because there are no entries to any of the feeds I added
#ZegnatI now kinda want to tag a microsub endpoint onto Sink that will not let you subscribe to any outside feeds but lets everyone login to read the Sink feed itself :P
#[eddie]Ahh gotcha. Okay, I don’t know anything other than Ubuntu. haha. But basically there is a file you need to run every so often. Let me track down the file, I think I added it as a service in my Ubuntu
#[eddie]Yeah, so I created a ubuntu systemctl service that starts up every time my server starts. But essentially in the /scripts directory, it runs watchtower.php using PHP
#Loqiaaronpk: [eddie] left you a message 1 hour, 54 minutes ago: Known’s syndicate-to returns: “syndicate-to”: [ “indiesyndicate::https://news.indieweb.org/en” ] is this a previous standard or just something Known made up?
#aaronpkI never claimed it would be easy to run these yet :-)
#aaronpkomg ownyourgram needs some serious refactoring
KevinMarks joined the channel
#[eddie]That’s true sknebel. I’ve thought about it before but then I didn’t necessarily encourage more people to try it out and thus broaden aaronpk’s “workload” while in the early stages 😆 It’s a no win situation
#[eddie]looks at microcasts in Indigenous and starts pondering Media Fragments
KartikPrabhu joined the channel
#[eddie]Does anyone have media fragment webmention support built in to their site? I know I’ve seen regular fragment support built in. I’m guess probably not
#Loqiloqi has 6 karma in this channel (448 overall)
#[eddie]So no one has probably headed in that direction. I guess eventually I would think a webmention that includes a media fragment would have a “jump to segment” link next to the comment
#[eddie]I guess really the most important part is when the post is embedded in the author’s site. Meaning my site should recognize the media fragment and include the media with the time stamp
#aaronpkI do want to have something like soundcloud comments on my site at some point
#aaronpkso you can comment on a specific part of an audio file
#aaronpkand then i'd show the comment at that point in the timeline somehow