2018-05-01 UTC
[jgmac1106], eli_oat and [unoabraham] joined the channel
# 00:46 tantek now to see if Bridgy Backfeed from FB is still working, at least for RSVPs to events snarfed1, snarfed, [kevinmarks], renem_ and [snarfed] joined the channel
# 01:42 Loqi [snarfed] #814 facebook: re-apply for page events permission eli_oat2 joined the channel
eli_oat, [tantek], leg, [John], chrisaldrich, [chrisaldrich] and [jgmac1106] joined the channel
tomasparks joined the channel
tantek joined the channel
tantek, [unoabraham] and snarfed joined the channel
# 04:46 Loqi chrisaldrich has 14 karma in this channel (85 overall) tantek, KartikPrabhu, swentel and [unoabraham] joined the channel
# 08:47 sknebel might not be available to people from strange countries? [kevinmarks] joined the channel
[jeremycherfas] joined the channel
jeremycherfas joined the channel
# 10:24 Zegnat Does anyone know any good word-cloud scripts? Possibly something that goes well with SQLite? Mandrake, [unoabraham], iasai and [jgmac1106] joined the channel
# 12:16 skippy is it supported / encouraged to include in a in-reply-to post the full contents of the tweet to which I am replying? # 12:16 skippy considering POSSEing my tweets, i lose a lot of context and history if I don't grab the original for inclusion / preservation. [jgmac1106] joined the channel
# 12:20 swentel I'm planning to add that option to the drupal plugin # 12:20 swentel because now I have to click the link itself to figure out to what I've been replying # 12:21 swentel I'm not sure whether to include it into e-content though for parsers # 12:21 swentel maybe that could some other extension to microformats # 12:23 skippy i'm (tentatively) planing <div class="u-in-reply-to h-cite">Reply to <a href="{{ $.Param "in-reply-to" }}" class="u-url p-author">{{ $.Param "in-reply-to-name" }}</a>:</div><div class="e-content"><blockquote>{{original tweet}}</blockquote>My reply...</div> # 12:24 Zegnat I am pretty sure someone is doing that h-cite pattern... Maybe aaronpk ? # 12:24 Zegnat It is a good one to use. Though I am on mobile and can't really comment on that specific markup snippet you posted. [unoabraham] joined the channel
# 13:04 skippy XRay looks for "extended_tweet", and "full_text"; neither of which are in the response I got. # 13:05 skippy maybe i should scrap what i have and just start using xray. # 13:09 Zegnat If you are using PHP, using XRay as a library should be easy enough # 13:13 skippy sure. but it provides way more than I need, right now. iasai and [snarfed] joined the channel
# 13:54 Loqi Since @kanyewest’s tweets have apparently made this topic unavoidable, some thoughts on the history of the parties’ switch on civil rights.
short tweet test. [John], [unoabraham], snarfed, KartikPrabhu, eli_oat and kaushalmodi joined the channel
# 14:42 kaushalmodi snarfed++ Thanks for quickly fixing the case of retweeting mentions #bridgy # 14:42 Loqi snarfed has 61 karma in this channel (383 overall) KartikPrabhu, tantek and leg joined the channel
# 15:16 tantek aaronpk: with minimal code for the additional variants, now ~25 lines of code for my single emoji heuristic function [miklb] joined the channel
# 15:20 Loqi Since @kanyewest’s tweets have apparently made this topic unavoidable, some thoughts on the history of the parties’ switch on civil rights.
short tweet test. # 15:20 Loqi Reply to Kevin M. Kruse: Since @kanyewest’s tweets have apparently made this topic unavoidable, some thoughts on the history of the parti... # 15:22 Zegnat I would have expected the content of the tweet you are replying to to be in the content property of the h-cite. Note within the content of your own h-entry. # 15:28 skippy fair; but not something I'm likely to fix soon. # 15:30 tantek agreed with Zegnat, the h-cite should be inside the h-entry but not inside your own e-content # 15:31 skippy the h-cite is in the h-entry applied to the URL. the contents of the target URL are included in my own e-content inside a blockquote. # 15:32 tantek (and on that emoji detection code, that's the PHP version, the JS version is not as thorough, so I'm going to keep it just serverside for now until I have a need to put it in CASSIS for client and server) # 15:32 skippy i am trying to present the contents of a tweet to which I am replying. rhiaro joined the channel
# 15:32 Zegnat Would it be hard to port it to CASSIS tantek? What are you using, is the PHP version published somewhere? # 15:33 tantek skippy, then those contents should go in the reply context in the h-cite, not in your own content # 15:33 tantek otherwise it will be consumed as a quoted tweet, not as a reply # 15:34 tantek Zegnat: just finished writing / testing it, haven't published it somewhere yet # 15:34 tantek hmm looks like it turned to be pure PHP, not using any non-native functions # 15:35 tantek I wonder where I should put it, maybe in another file in the cassis project [schmarty] and [jonathanprozzi] joined the channel
# 16:00 tantek Zegnat: I'm thinking of putting such code I'm considering for CASSIS into a file like cassis-lab.php # 16:01 Zegnat Putting side projects under the lab label seems popular enough that it makes sense. snarfed joined the channel
# 16:18 tantek Zegnat I will test my is_one_emoji function in my own code (with actual content / live use-cases) before I share it # 16:18 tantek integration testing as it were, beyond just the unit testing I've done # 16:18 Zegnat Looking forward to see how you tackled it. It is an interesting topic # 16:19 tantek I did add a few comments to the various if clauses to explain which emoji cases each catches # 16:20 tantek skippy, I too am now having to redo a bunch of my code to put reply-context information *outside* of my e-content # 16:32 tantek which is now making me rethink how I markup replies in general when they're in-stream # 16:38 jeremycherfas Very leased to report that after several head-scratching hours, the fold all command in bbedit came to my rescue to find an errant closing div tag. # 16:40 tantek neat! I don't think I've known what to use that command for before, but that makes sense! # 16:40 jeremycherfas I wish I had thought of it sooner. Helped me get a bunch of links outside e-content. [Thom_M] joined the channel
# 16:52 snarfed we've had it a few times. i believe it's on the wiki now yes # 16:53 tantek I'm bashing on the claim of "helping protect against ad malware" by a domain registrar absent any new expiration policy # 16:53 tantek this is more dev-like so fine to switch channels # 17:13 tantek haven't implemented it yet, but figured I'd write down current thinking first [eddie] joined the channel
# 17:28 skippy i'm new to all of this, but it seems weird to me to place e-content around only my own contnet, and not also the blockquote to which I am actually creating contnet. # 17:29 tantek skippy that's the difference between entry and content # 17:29 tantek the purpose of the entry is to place it around both your own content and the blockquote or whatever else is the context for your own content swentel joined the channel
# 17:32 tantek perhaps that would be a good clarification to note explicitly in h-entry # 17:44 tantek assuming that's good for minimal in-stream reply context, then next I'll work on in-stream "normal" reply context, e.g. comments on github issues, replies to tweets etc. AngeloGladding and [miklb] joined the channel
# 17:45 [miklb] tantek yes, that is definitely worth documenting/clarifying. I’m not sure I’ve heard that explanation before. Makes perfect sense though. # 17:48 aaronpk I don't think I've given enough thought for how i want to show reply contexts in my reader yet swentel_ joined the channel
# 18:05 sknebel ^^^ github pages now officially does SSL for custom domains # 18:11 snarfed i'm not optimistic i'll get many of those permissions back # 18:12 tantek snarfed, hoping for at least user_events _photos _posts _link # 18:12 snarfed f8 fb api blog posts relevant to us (i haven't read yet): # 18:13 sknebel "Introducing Personal Data Deletion Callback: providing a callback URL to receive a person’s request that the info an app received from Facebook be deleted". that's an interesting addition # 18:19 sknebel the requirement to be a business for some kinds of access also sounds like it could become a problem [Thom_M] joined the channel
# 18:23 snarfed guess i'll have to try and see what happens, as usual # 18:28 tantek aaronpk - wait I'm confused - I thought you were already reading replies from people in your reader, and the problem was some had reply-contexts and others didn't? # 18:28 tantek looks for the screenshots others took of all the thumbs-ups in a row # 18:30 aaronpk the problem wasn't that some had reply contexts, it's that some home page feeds didn't even have the in-reply-to URL on the reply # 18:31 tantek I think I'll still markup the h-cite because it's not that much incremental work for a reader to use it if it is there # 18:31 tantek to optionally show more human readable text than just the URL # 18:32 aaronpk i'll probably make it work like my website's reply contexts soon, but i still want to think through the differences between using the h-cite that's there vs fetching the in-reply-to URL and using the canonical content # 18:32 tantek (since that h-cite will auto-imply name and url properties) # 18:32 tantek and you could show that while lazy fetching the in-reply-to # 18:33 aaronpk also "just use the h-cite" isn't a shure thing because of all the differences in how those end up looking in practice. e.g. posts with a name vs only a note, how long is the text content, is there a photo, etc # 18:33 tantek I mean use the h-cite properties in a design of your own # 18:34 tantek e.g if the h-cite has a name, then show that instead of just the u-in-reply-to, but still link it to the u-in-reply-to # 18:34 tantek since you're already showing the reply to URL, that kind of incremental improvement should be fairly painless and design-neutral # 18:35 skippy grrrr.... Omnibear is still having token troubles. :( # 18:35 tantek anyway, hopefully when I get this working you'll be able to see it all (with h-cite p-name) in my h-feed and see how that would look sebsel joined the channel
# 18:37 skippy oh, speaking of photos: post-type-discovery says "If the post has a "photo" property with a valid URL, # 18:37 skippy so if I want to include a photo in an article, I *cannot* use a photo property? maingo joined the channel
# 18:38 aaronpk correct, you just include the photo in the article content, not with the u-photo property # 18:38 aaronpk adding the u-photo class tells a consumer that it's a photo post # 18:42 skippy can anyone help me figure out why Omnibear keeps giving me "unknown error verifying the authorization code" ? # 18:53 skippy cleared cached. restarted browser several times. Quill login works just fine. # 19:01 Zegnat I’ll be happy to have a look in a moment, just grabbing some tea :) [jgmac1106] joined the channel
# 19:16 Zegnat is wondering how he is going to recreate skippy’s case locally pstuifzand joined the channel
# 19:36 Zegnat I added the indieauth token provider to my site, and a dummy micropub endpoint, and I cannot get the token # 19:38 Zegnat So the IndieAuth Token Endpoint is correctly not issueing a token # 19:38 Zegnat Trying to figure out what about the request isn’t OK here. Feels like some data gets lost somewhere # 19:42 Zegnat I should just learn to listen to sebsel instead of assuming everyone is doing The Right Thing # 19:43 Loqi [sebsel] #35 base64 encode the scopes? # 19:43 Zegnat At some point the spaces get turned into + signs rather than %20 # 19:44 Zegnat The IndieAuth Token Endpoint gets the code from Omnibear with + in it, and will use that to ask selfauth about validity, and selfauth has never issued a code with scopes including + signs # 19:44 aaronpk TIL spaces are a valid character for an authorization code # 19:45 aaronpk the other thing about that... why are the scopes included in the string at all? # 19:46 Zegnat And scopes are `\x21\x23-\x5B\x5D-\x7E`, so they fit within there. # 19:46 aaronpk i hope they are stored server-side as well, since otherwise an app developer could just change the list of scopes when exchanging the code for a token # 19:46 Zegnat Because we are stateless, nothing is stored server-side # 19:46 aaronpk ah so you made up your own signed token format, I see # 19:47 Zegnat I didn’t want to reimplement JWTs, and the idea is to have a single file so no pulling in libraries # 19:48 aaronpk yeah i wonder how small of a library it takes to implement JWS # 19:48 aaronpk JWT has vocab about all the key names, JWS is the enclosure/signing format # 19:49 Zegnat But this one function I wrote is pertty clean. It contains both a known message for checking (from static info, like your URL), and can optionally have extra variable data (e.g. scopes). # 19:49 Zegnat But I didn’t consider that URL encoding has 2 ways to transform spaces ... so there it falls apart. # 19:50 Zegnat If only it wasn’t handled as part of a URL, he. # 19:50 Zegnat I wonder what the URL spec says, so if Omnibear is doing anything wrong here by using + for encoding. # 19:50 Zegnat Or if maybe selfauth is already not encoding it correctly. # 19:50 Zegnat Odd that Quill works though, does it ask for only 1 scope? # 19:51 Zegnat Oh. Or it uses %20 I guess. Argh. base64 encoding patch coming up. # 19:51 sknebel the + thing is in formencoded post requests, so I think omnibear is right # 19:53 sknebel but the requests we are looking at are post, so that's the relevant one # 19:53 skippy thanks for finding that Zegnat ! I'm running out for a bit, but will be watvhing the repo. Cheers! # 19:54 Zegnat They aren’t POST, sknebel. They are in redirect URLs as GET # 19:55 aaronpk first it's in the GET redirect, then the client POSTs it # 20:00 Zegnat But on receiving a POST request, PHP does not decode the + into \x20 # 20:02 Zegnat I really want to blame someone somewhere, but honestly, I am just going to write the base64 patch and call it a day # 20:03 Zegnat aaronpk, default PHP base64 should be fine in codes, right? # 20:03 aaronpk you might have a similar problem with the = in base64 # 20:04 aaronpk there's a pretty common base64 variant that chooses alternate chars for + and = # 20:05 aaronpk replace + with / and = with _ then remove all the tailing = [colinwalker] joined the channel
# 20:06 Zegnat was checking jwt.io to see if it documents the base64 [chrisaldrich] joined the channel
# 20:09 Zegnat I’ll highlight my name in the IndieAuth spec as a reference for applying to these security centric companies! ;) # 20:12 Zegnat Looks like the lib you use is doing exactly that # 20:13 Zegnat I just have to write it myself in PHP, else we can’t put it under our public domain waiver. # 20:13 Zegnat Now that I have survived the headache part, it is time to grab another cup of tea and get this patch out. # 20:14 Loqi zegnat has 58 karma in this channel (203 overall) [Thom_M] joined the channel
# 20:35 schmarty aaronpk: i'm moving more feeds into aperture but running into issues w/ sites that don't advertise their feeds correctly. should i be able to plug in a URL directly to a feed into Aperture? # 20:35 schmarty or maybe i can make a dummy HTML page with that as a rel=alternate to get aperture to pick it up ;} # 20:36 aaronpk you can definitely plug an XML feed URL directly into aperture # 20:36 aaronpk if it's not recognizing it then it's getting tripped up on the data in the feed # 20:37 aaronpk hm xray can handle that URL so i'm not sure why aperture would fail on it # 20:47 Loqi zegnat has 59 karma in this channel (204 overall) # 20:53 Loqi [Zegnat] #40 Use URL safe base64 encoding for scopes [jgmac1106] joined the channel
# 20:56 schmarty aaronpk. hmm. updated apeture, which pulled a new version of xray. it now thinks for longer but still says no feeds found. :/ # 20:56 skippy Flickr uses base58 or something? They remove all ambiguous characters (O, 0, I, l, etc) # 20:57 aaronpk schmarty: that's really weird! is there anything unusual about the feed? # 20:59 schmarty aaronpk: i'm not sure. maybe i'll dig into the code and increase the timeout, but the feed loads in less than a few seconds in my browser. # 21:00 schmarty backend is returning {"feeds":[],"error":false,"error_description":false} # 21:02 schmarty is there a way to skip the preview step and say "this is a feed please deal with it"? # 21:02 skippy Zegnat: what's weird is that Omnibear + indieauth/token was working recently. no idea what changed where to trigger this problem. # 21:03 aaronpk schmarty: if the preview step doesn't find anything then it probably won't find anything later either # 21:03 aaronpk unless it's a timeout problem, chances are it's choking on the actual feed data somehow # 21:04 Zegnat skippy, I couldn’t say. Did you server infra change, which may have affected the decoding step? Did you start using different scopes? So many possibilities. But this change should save us all the future headaches [kevinmarks] joined the channel
# 21:06 Zegnat Yeah, I would have used a different encoding if it was meant for people. But this is meant for machines only. The only extra requirement was for it to be URL safe, which the JWS is. And JWS is a standard with decent following # 21:07 Zegnat merged to master, feel free to update, skippy! # 21:07 aaronpk yeah and using newbase60 would involve writing code to actually encode/decode the text, whereas this just is a simple string replace leveraging the exsting base64 encode/decode functions # 21:07 sknebel if even more people suggest encodings I'm changing it to base2 over an alphabet of O and 0... :P [ryan339] joined the channel
# 21:27 Zegnat [chrisaldrich], I just realised I received a webmention from you but the source page doesn’t exist. A bug, or did you remove a post? # 21:37 skippy now omnibear never performs the redirect to obtain the token. Shows "Retrieving access token..." but that's it/ # 21:41 Zegnat Huh? I am pretty sure my last commit worked for me ... # 21:44 Zegnat I don’t actually use a token endpoint with my selfauth setup, so I never really run into this # 21:44 skippy so what do tokens do? I'm confused. i guess i just blindly assumed i need them. # 21:45 aaronpk you need tokens for micropub, you don't need tokens if you're just signing in to apps # 21:46 Zegnat And my homepage is purely my identity, can’t post to it, so I do not need tokens. # 21:46 skippy i'm signing in to a micropub client app to perform micropub actions. # 21:46 Zegnat Yes. The Micropub client needs a long-lived token that it sends along with the Micropub requests to the server :) # 21:47 sknebel yes, and the actions need the token, so when it makes the action it can prove to your site that its allowed # 21:47 Zegnat For pure logins, short-lived tokens are used. So those are the only ones my website ever produces. And those never contain scopes. # 21:48 Zegnat From #-chat. Wow. That JSONFeed issue list looks like a real mess ... is anyone actually planning to push the JSONFeed spec to a v2? Or is it basically dead? # 21:51 Zegnat Slightly tempted to add the mf2 version of my h-card as a rel="alternate" type="application/json", and then complain about how feed servers (e.g. Aperture) think that is a feed! :P # 21:52 skippy WTF? The redirect with token is still not succeeding. POST to selfauth succeeds, but the token isnt beting collected? Cleared cache, restarted browser. I am confused! # 21:54 Zegnat What step is failing? Can you check the logs for PHP errors again? # 21:55 Zegnat I don’t know how to debug that as I do not have a micropub endpoint to test with # 21:55 Loqi [timbray] #32 Suggest alternate media type. # 21:56 Zegnat Any more info you can give me skippy? What micropub endpoint are you using? # 21:56 Loqi [skpy] micropub: a minimal PHP micropub endpoint, with media support tantek joined the channel
# 21:57 Loqi [timbray] #32 Suggest alternate media type. # 21:58 Zegnat skippy, please log the full $response, that should probably tell us # 22:02 Zegnat Urgh, this is what you get for working on a function you do not use yourself [snarfed] joined the channel
# 22:05 Zegnat That should be the last of that. Bit messy. Oh well. # 22:07 Zegnat Feel free to PR that, and link to this chat as the fact that I checked that piece of code, and maybe someone will merge it. I am on my way to bed iasai_ joined the channel
# 22:14 skippy great, now Omnibear is not presenting me any way to syndicate notes to Twitter. Huh. *sigh* # 22:21 Loqi [Tantek Çelik] h-entry is a simple, open format for episodic or datestamped content on the web. h-entry is often used with content intended to be syndicated, e.g. blog posts. h-entry is one of several open microformat standards suitable for embedding data in HTML.
... # 22:23 aaronpk [kevinmarks]: that won't help encoding text as newbase60, you need more code for that barpthewire and eli_oat joined the channel
# 22:50 pstuifzand skippy, it's a bug in omnibear currently, it won't show your syndication targets # 22:52 skippy "Select syndication targets (NOTE: you may need to log out and back in after the update takes effect)" # 22:54 Loqi [skpy] Sorry to revive this old issue, but I'm not seeing any syndication targets in Omnibear. Both Quill and Indigenous for Android correctly show that my MP endpoint supports syndication to Twitter, but Omnibear provides nothing. I've tried setting "Cus... # 22:54 GWG pstuifzand: How is your app going? # 22:55 pstuifzand on back burner for a few days, that latest build should be pretty stable # 22:55 GWG pstuifzand: Did you push a new one? I still have an issue with redirecting post URL change # 22:56 GWG The custom scheme worked, but got hit by the fact it didn't match the client ID # 22:57 GWG pstuifzand: But now it isn't redirecting back # 22:58 pstuifzand It's problem with the mode of the activity, it should be implemented differently # 22:58 pstuifzand I tried to implement it with single_instance for the activity, but that has it's own problems # 22:58 schmarty (jjuran.org doesn't seem to have any authorization endpoint data) # 22:59 pstuifzand I think you can close all browser windows and Wrimini "windows" and it should work. Or perhaps only from the Android Account settings in # 23:01 jjuran After submit, the page is actively in loading mode, but nothing happens # 23:02 GWG schmarty: Did you ever upgrade Screech? # 23:03 schmarty GWG: not yet. i need to hack on kylewm's Flask::IndieAuth to get it working to spec and haven't had the stomach for that yet. :} # 23:03 GWG schmarty: What Indieweb plans are you making of late? # 23:05 schmarty a couple of them are being discussed here in the chat right now, haha :} # 23:05 schmarty i have been spending most of my hobby web time on another project, crowdsourced 3d printing of sculptures: wethebuilders.com # 23:05 jjuran I didn’t RTFM. I saw a field on an indieweb asking for a domain URL, so I entered the same domain I use for indieweb.org. :-) # 23:06 aaronpk aha, well this is a good demonstration of the difference between logging in to a site and giving a site permission to post to your website # 23:06 schmarty that launched yesterday and hasn't completely melted the site yet, so i am trying to turn my sights back to indieweb projects. # 23:06 jjuran And a good demonstration of what real users are going to even if you tell them not to :-P # 23:07 schmarty jjuran: that makes sense! i'll get a more helpful error message up there. # 23:07 aaronpk yep. better error messages or even good fallback behaviour would be great # 23:09 Loqi skippy has 3 karma in this channel (8 overall) # 23:09 schmarty haha, this toy app is nowhere near ready for real users. :} # 23:10 schmarty the randomly-generated glitch.com name stands as a warning # 23:10 schmarty abandon all hope of elegant failure handling, ye you enter here # 23:10 aaronpk schmarty: lol yea it has "glitch" in the name, what can you expect :shrug: tantek joined the channel
# 23:17 GWG aaronpk: I keep meaning to add something like that # 23:17 schmarty it actually does support in-reply-to if you slap it in the URL # 23:19 schmarty i don't think the root of the app is checking whether you're logged in, yet. # 23:19 aaronpk pstuifzand: i don't remember, but feel free to post a gif reply to mine :D # 23:20 aaronpk pstuifzand: ohhh no somehow i have a bug where it un-gifs a gif # 23:27 GWG How can I better integrate my site with micro # 23:28 GWG People seem to be coming from there # 23:31 GWG I read that, but I thought I saw someone, maybe [eddie] doing more [cleverdevil] joined the channel
# 23:32 [cleverdevil] I send only certain content, and I format it specifically to show up how I want in the timeline. # 23:32 [cleverdevil] That said, I've pretty much abandoned updating it until there is a way for me to work on it and test it live. # 23:33 snarfed yeah good tight iteration feedback cycles are so important, it's almost not worth doing something if you can't get/make one # 23:34 [cleverdevil] That said, I have reported the request to [manton] and I am guessing he'll get to it someday 🙂 # 23:34 GWG I am convinced that I will lose Facebook backfeed, so I am looking for other opportunities # 23:41 snarfed GWG: or maybe just care about interactions less :P # 23:41 snarfed eg micro.blog doesn't have reposts or likes for exactly that reason # 23:42 GWG snarfed, I care about replies mostly # 23:42 GWG I like to communicate with others. [manton] joined the channel
# 23:43 [manton] [cleverdevil] A feed preview would be useful to debug broken feeds too. Micro.blog spends a lot of time parsing invalid feeds, then trying again, and again... So it's on the to-do list but (as you said) hasn't been a priority yet. # 23:43 [cleverdevil] Totally fair, you've been doing a great job filling out higher priority things in the platform. # 23:44 GWG [manton] Any suggestions for better integration with my site? # 23:44 aaronpk that's exactly why i made the preview thing in aperture that schmarty was annoyed by :-) # 23:44 aaronpk avoids it getting into the situation of fetching feeds it know it can't parse # 23:45 aaronpk GWG: do you have a micro.blog account and have you added your site's feed to it yet so that people there can subscribe to you? that would be step 1 # 23:46 aaronpk are you getting webmentions to your site when people reply to your micro.blog posts? # 23:47 [manton] @GWG You're feed looks pretty good to me, except for the "Also on Twitter/Facebook" links, which ideally would be removed from the feed you have in Micro.blog. [Thom_M] joined the channel
# 23:48 GWG [manton] How should syndication URLs work in jsonfeed? # 23:49 tantek GWG, another way to ask that question, how does jsonfeed represent the h-entry property u-syndication? # 23:50 tantek GWG, I'm kinda implying that when in doubt, use h-entry properties as they would appear in mf2json, in your jsonfeed items/entries # 23:51 tantek it's a simple consistent way to extend jsonfeed based on existing format / spec # 23:51 GWG aaronpk, what are you highlighting? # 23:51 tantek aaronpk, interesting, so you'd like to see in-stream h-entry's have explicit u-syndication to show in the UI? # 23:52 tantek GWG< presumably the bar at the bottom that shows date time then f t links # 23:52 GWG So, that is coming from my u-syndication, I assume # 23:56 [manton] In the feed I'm looking at, the end of each post has `<span class="syn-text">Also on:</span>` and then u-syndication links. Doesn't seem quite enough info to do anything smart with that part of the post. # 23:58 GWG [manton] I would be happy to exclude them from the json feed and enbed them as json metadata. You have a filter for that. eli_oat joined the channel
# 23:59 GWG [manton]: It also doesn't like my citations.
[snarfed] nope. at least temporarily. https://github.com/snarfed/bridgy/issues/814
loqi.me created /pronunciation (+119) "prompted by [chrisaldrich] and dfn added by [chrisaldrich]" (view diff)
tantek chrisaldrich++ that's a good start on /pronunciation !
[kevinmarks] All countries are strange in French
[jeremycherfas] Boom boom.
skippy anyone know how to get the expanded text of a truncated tweet? consider the tweet object described at https://skippy.rocks/micropub/t.php
@ringmaster Contemplating herd mentality today, and how it affects our choice of breakfast cereal, among other things. Also, how eating a “salad” sounds healthy, until you realize it has more junk calories in it than a Big Mac. (twitter.com/_/status/991283064750903297)
kaushalmodi snarfed++ Thanks for quickly fixing the case of retweeting mentions #bridgy
tantek.com edited /reacji (+698) "/* Brainstoming */ How to markup reacji with minimal reply-context" (view diff)
Zegnat Hmm. So if the redirect from selfauth to omnibear at the code request step uses %20, and everything after that is POST following application/x-www-form-urlencoded, all spaces should be + per https://url.spec.whatwg.org/#concept-urlencoded-byte-serializer
aaronpk i've seen this technique used before http://php.net/manual/en/function.base64-encode.php#103849
schmarty i'm wondering if i have found a feed too gnarly for Aperture to handle: https://www.sarasoueidan.com/blog/index.xml
Loqi [indieweb] newBase60py: Python implementation of http://tantek.pbworks.com/w/page/19402946/NewBase60
pstuifzand skippy, it's a bug in omnibear currently, it won't show your syndication targets
jjuran I tried https://jjuran.org and https://www.jjuran.org
@smerrill First POSSE! https://indieweb.org/POSSE #indieweb If this works, this is being posted to my site first, then to Twitter. (twitter.com/_/status/991454249170202624)
[cleverdevil] So, I have a custom feed for Micro.blog.