#dev 2017-07-21

2017-07-21 UTC
tantek, snarfed, KevinMarks, AngeloGladding, dougbeal|mb1, snarfed1 and [eddie] joined the channel
#
@NandoKstroNet Implementing Webmentions https://allinthehead.com/retro/378/implementing-webmentions (twitter.com/_/status/888209603506438144)
#
@JavaScriptKicks Implementing Webmentions by drewm https://javascriptkicks.com/r/159964?url=https%3A%2F%2Fallinthehead.com%2Fretro%2F378%2Fimplementing-webmentions #css #javascript via JavaScriptKicks (twitter.com/_/status/888226320307638272)
#
@gradar Webmention Client https://www.reddit.com/r/PHP/comments/6oapw4/webmention_client/?ref=search_posts https://twitter.com/gradar/status/888226957590093824/photo/1 (twitter.com/_/status/888226957590093824)
KartikPrabhu, Defenestrate and j12t joined the channel
#
prtksxna !tell tantek and aaronpk No one seems to have specific word for it. nirzar at wmf pointed out, Apple uses: "Entire screen", "Part of the screen", and "App window". melchoyce at wp doesn't differentiate between them. quiddity at wmf calls it a screenshot and says the tiny size would make it obvious that it was partial.
#
Loqi Ok, I'll tell them that when I see them next
#
aaronpk nice
j12t, snarfed, tantek, [allegoryanalyti, [allegory], gRegorLove and barpthewire joined the channel
#
@citrusui Very cool https://allinthehead.com/retro/378/implementing-webmentions (twitter.com/_/status/888292102798323713)
[colinwalker], cweiske and [pfefferle] joined the channel
#
@superxteam Implementing Webmentions - http://www.superxteam.com/blog/implementing-webmentions-25/ (twitter.com/_/status/888327676838203396)
[kevinmarks] and [pfefferle] joined the channel
#
@weblings_co Implementing Webmentions https://allinthehead.com/retro/378/implementing-webmentions (twitter.com/_/status/888345386322059264)
[kevinmarks] and jeremycherfas joined the channel
#
jeremycherfas Hey, Loqi, any messages?
#
Loqi jeremycherfas: Zegnat left you a message 15 hours, 2 minutes ago: looks like I got a webmention from your main site? Not sure if that was the notify script or something else.
#
@grabaperch Read about the Perch Webmention support in @drewm's blog post (which uses Webmention!) https://allinthehead.com/retro/378/implementing-webmentions https://twitter.com/grabaperch/status/888354139398443008/photo/1 (twitter.com/_/status/888354139398443008)
KartikPrabhu, [kevinmarks] and jeremycherfas joined the channel
#
jeremycherfas I can be certain now that a PHP script that works fine, although slowly, in my production site does not complete in my development site.
#
jeremycherfas Any suggestions as to where I might look for clues?
#
petermolnar jeremycherfas what are the differences between the prod/dev setup? where are they running out from, is there a connection (proxy) difference?
j12t joined the channel
#
jeremycherfas Petermolnar: Production runs on Dreamhost and Development runs locally under MAMP
#
jeremycherfas DreamHost shared.
#
Zegnat Hmm … did you get any errors locally? Did the script ever stop or hang?
#
jeremycherfas The script has hung.
#
jeremycherfas No errors in Grav logs. Just cheking MAMP
#
voxpelli Weird that is has hung – I would suggest that you run phpinfo(); to print all version and config info both locally and on your server and look for things that are different
#
voxpelli If you are eg. running different versions of PHP itself or are missing some common extension in one environemtn, then that may be it
#
voxpelli When it comes to macOS dev environments one thing that can differ is file names – as Linux file systems are mostly case sensitive while macOS by default isn't
#
Zegnat Another thing you might check is max_execution_time in MAMP’s php.ini file. Though that should not apply to CLI scripts, if I recall correctly
#
Zegnat It seems to interface with only a single file, voxpelli, so it shouldn’t run into file system problems
#
jeremycherfas Couple of errors. Need to make a gist.
#
Zegnat Please :)
#
voxpelli no includes in that file?
#
voxpelli has even had casing issues with names in MySQL :/
#
Zegnat voxpelli, well, the CMS has includes, but those seem to work in all other cases.
#
Zegnat The Webmention sender reads some sort of gigantic single data file.
#
Zegnat So I would be tempted to say it is a memory or execution time difference between the server and local. Though really most dev set-ups allow near infinite use of those.
#
jeremycherfas These all seem to be from Known, not from the Grav plugin https://gist.github.com/jeremycherfas/89fcd393a2d93768555c3f798d7df9c1
#
petermolnar Zegnat define gigantic
#
jeremycherfas What can I add to the Grav plugin to increase error recording?
#
jeremycherfas But Zewgnat, the thing is that the server executes properly, while the local is the one that hangs, and the local is dealing with a much smaller data file.
#
jeremycherfas Sorry, Zegnat
#
Zegnat It is smaller locally? Then it really shouldn’t hang :s
#
cweiske didn't you say it hang because it tried to process 500+ webmentions?
#
cweiske s/say/write/
#
voxpelli Unless it is something with network communications and some request that doesn't work correctly due to the locality of things
#
Zegnat That wouldn’t surprise me voxpelli. I seem to recall the notifier script just synchronously working through the list and doing all the HTTP requests for sending the webmention 1 by 1.
#
Zegnat goes looking for source again
#
jeremycherfas Cweiske I thought it had hung, but it was just going very slowly, possibly because it was doing something bad and DreamHost was killing and restarting, or something, but eventually it finished properly. That was the live version.
#
petermolnar jeremycherfas there is a 'plugins.webmention.sender.automatic' setting, what is that set to?
#
jeremycherfas The local version currebntly only has 9 items to process and does not get past the first one.
#
Zegnat voxpelli cweiske the source: https://github.com/Perlkonig/grav-plugin-webmention/blob/master/cli/NotifyCommand.php#L221-L310
#
jeremycherfas Currently set to false.
#
petermolnar good
#
jeremycherfas Need to get a phpino.php into the live site.
#
Zegnat It doesn’t even seem to cache webmention endpoints :/ If you link to the same thing on multiple pages within your site it will go and rediscover the endpoint every time, if I read that right
#
jeremycherfas I think that is right.
#
petermolnar jeremycherfas https://stackoverflow.com/questions/10308915/php-default-curl-timeout-value
#
petermolnar there is no TIMEOUT in the https://github.com/Perlkonig/grav-plugin-webmention/blob/master/classes/MentionClient.php
#
petermolnar which means if there is a dead link it'll wait 300s on it
#
petermolnar per url
#
jeremycherfas That would slow things down, I agree.
#
petermolnar in theory, it should respect the ini default_socket_timeout
#
Zegnat Oh MentionClient might include some endpoint caching. That’s good then.
#
petermolnar can you compare those?
#
jeremycherfas Well, one BIG difference. Live site uses PHP 5.6.29 while dev uses 7.0.15
#
jeremycherfas So maybe the very first thing would be to configure MAMP to use 5.x.x
#
petermolnar no, don't do that
#
petermolnar 7.x is stricter, it if runs on 5.6, it should be fine on 7.x
#
petermolnar well, at least everything I tried was
#
petermolnar I might be mistaken though
#
jeremycherfas OK, I’m still looking through the two phpinfo files
#
jeremycherfas HaRD TO KNOW WHAT TO LOOK FOR THOUGH. :(
eli_oat joined the channel
#
petermolnar can you share them somehow?
#
petermolnar the more eyes on it the better
#
jeremycherfas I saved as PDF. Another obvious difference is that local uses Apache 2.0 handler while live uses CGI/FastCGI
#
Zegnat Your CLI notify tool does not run through apache, so that makes no difference
#
jeremycherfas OK, thanks.
#
jeremycherfas What does expose_php mean? It is on local and off server
#
cweiske show php + version in the headers
#
cweiske that's ok as it is
#
jeremycherfas Thanks. So not really relevant.
#
jeremycherfas There seem to be some differences in the curl section
#
jeremycherfas AsynchDNS, GSS-Negotiate, IDN are off locally, on in production
tantek joined the channel
#
jeremycherfas Local does not have mysql, only mysqli and mysqlnd
#
petermolnar that is completely irrelevant
#
petermolnar grav doesn't have a db backend
#
jeremycherfas Clutching at straws
#
petermolnar do you know if any of the links it's trying to ping is pointing to a dead service?
#
petermolnar in case you're hitting that 5 mins timeout per iteration
#
jeremycherfas The very first link it is trying to ping is my local install of Known, and that is where it is hanging.
#
prtksxna seen mko?
#
prtksxna Loqi: seen mko?
#
jeremycherfas If I do the curl by hand, that results in an error, for which I raised an issue https://github.com/idno/Known/issues/1830 but have not yet had a reply.
#
Loqi [jeremycherfas] #1830 Error: source not supported
#
jeremycherfas But I do not get that error in production site. At least, I don’t think so.
#
jeremycherfas I’ll check that now.
#
@antwinkansah Implementing Webmentions https://allinthehead.com/retro/378/implementing-webmentions #webdesign #feedly (twitter.com/_/status/888389884234407936)
#
prtksxna.com edited /User:Prtksxna.com (+874) (view diff)
#
prtksxna.com edited /User:Prtksxna.com (+8) (view diff)
jeremycherfas_ joined the channel
#
jeremycherfas I’m as mystified as ever, probably more so. Going to think about maybe testing the XRay version live, or else rethinking the whole project.
#
voxpelli jeremycherfas: tried https://github.com/voxpelli/node-webmention-testpinger ?
#
Loqi [voxpelli] node-webmention-testpinger: A tool to ping your site with a variety of real world WebMentions
#
voxpelli or https://www.npmjs.com/package/node-webmention-testendpoint
#
voxpelli Dependning on if your testing sending or receiving
#
voxpelli As those are dedicated test tools you could easily have others test the exact same pings as you do
#
jeremycherfas Yes, I did test with that, thanks. Not one of them seemed to get through.
#
@JTWebDesigns #WeKnowTheWeb 50 #WP #Tips for #beginners - Nº 30: #Trackbacks vs #Pingbacks vs #Webmentions for #WordPress... https://premium.wpmudev.org/blog/trackbacks-pingbacks-webmentions/?platform=hootsuite (twitter.com/_/status/888400726178885632)
#
jeremycherfas Just to check, voxpelli, that will be OK with localhost endpoints?
#
@perchology Yours truly has been implementing webmentions [in Perch Blog 5.6] on a couple of sites. If you need help this, I'm available to hire :-) (twitter.com/_/status/888401465458425856)
#
voxpelli jeremycherfas: those test tools are meant to be used with localhost endpoints
#
jeremycherfas I thought so. And just checked again, all 14 failed to be pinged with 404
#
jeremycherfas This was the command webmention-testpinger --endpoint=http://localhost:8888/grav-admin/mentions --target=http://localhost:8888/grav-admin/blog/bird-migration-getting-to-the-point
#
prtksxna !mf2 https://aaronparecki.com/2014/12/26/8/
#
Kaja https://php.microformats.io/?url=https%3A%2F%2Faaronparecki.com%2F2014%2F12%2F26%2F8%2F
#
Loqi [Aaron Parecki] https://indiewebcat.com/notes/2014/12/23/3/businesscatual
#
voxpelli jeremycherfas: that endpoint doesn't look correct – have an online page I can look at?
#
jeremycherfas No, but this is in the head <link href="http://localhost:8888/grav-admin/mentions" rel="webmention">
#
voxpelli weird that it links to /grav-admin/
#
voxpelli that sounds like the admin page, not something where open endpoints would be exposed
#
jeremycherfas Well, that is an oddity that zegnat and I have worked on, and it seems to be correct in other cases. I think.
#
voxpelli and apparently the script doesn't find the endpoint :/
#
jeremycherfas Doesn’t seem to.
#
jeremycherfas That URL always returns a 404, but it does seem to receive webmentions.
snarfed joined the channel
#
jeremycherfas That is another strange difference between the dev and the live: in dev, curl to the advertised endpoint results in a 404 while in live it results in a 301
#
jeremycherfas That is, it returns 301 to curl, but 404 to a browser
#
jeremycherfas And I can confirm that the webmention is in fact received.
Defenestrate joined the channel
#
jeremycherfas In the end, I am beginning to believe that there is something very different about the way this plugin behaves in the development environment versus the way it behaves in the production environment.
#
jeremycherfas I’m going to stop testing and testing and testing and try to think about basics.
#
snarfed so, fetching twitter block lists is problematic. the API call is paginated, only returns 100 users per call, and is rate limited to 15 calls per user per 15m. i've already hit one user whose block list has more than 1500 users on it. :/
#
snarfed so i can't reliably get a user's full block list at once
[kevinmarks] joined the channel
#
[kevinmarks] Is it sorted by recency of block?
#
snarfed [kevinmarks]: maybe? i wouldn't know
#
voxpelli snarfed: which API resource are you using?
#
snarfed voxpelli: blocks/list
#
snarfed blocks/ids is better, returns 4k per page
#
snarfed but not very useful since it doesn't even include usernames, just numeric ids
#
voxpelli And you can't look up the id:s through some kind of batch endpoint?
#
aaronpk tweet objects include IDs tho so can't you match on that?
#
snarfed i don't see block lists accessible through any other api or endpoint, no, just those two
#
aaronpk like i think it's fine to store user IDs only because you'll always have a user ID to compare to?
#
snarfed aaronpk: yes, user ids are useful internally, but not externally, e.g. if i want to publish an mf2 blocklist that wm.io could consume
#
aaronpk aha
#
aaronpk yeah
#
voxpelli Seems like you can look up 100 user ids 900 times per 15 min with /users/lookup
#
aaronpk also don't forget that usernames can change :P
#
snarfed sure
#
snarfed voxpelli: ugh, yes, but ugh
#
voxpelli And if you cache user data by I'd then you may often not have to look it up at all
#
aaronpk so if I block user ID 1000 which currently has a username "spam_pk", and then spam_pk changes his username to spampk, I'd want to make sure it's still blocked
#
voxpelli Aren't tweets returned with user ids as well, so perhaps enough to just do user ids?
#
snarfed voxpelli: yes, see above :P
#
snarfed aaronpk: understood
#
voxpelli Ah, missed that, yeah one would have to keep the Twitter user data up to date then, or have the author link in tweets be a Bridgy user id-based proxy URL
#
snarfed aaronpk: the catch is that i don't think bridgy currently includes twitter user id anywhere in its mf2 translation
#
snarfed uid is tweet id, but i think not including user id
#
snarfed so currently i don't think we can easily follow username changes. ah well
#
snarfed on the plus side, https://granary-demo.appspot.com/ now fetches twitter block lists and converts to mf2 etc
KevinMarks joined the channel
#
snarfed so aaronpk feel free to prototype wm.io using that to filter returned wms if you want
#
aaronpk Wow nice
#
aaronpk Wait what... the first in my block list is a bitly URL?
#
aaronpk Is that taking the website field of the twitter profile?
#
snarfed yes
#
snarfed clearly needs more thinking to provide more permanent identifiers/urls also
#
aaronpk There's nothing there I can block tweets from
#
aaronpk i need the twitter profile URL
#
snarfed well there's username
#
snarfed which gives you profile URL
#
snarfed but agreed i should add full profile URL and numeric id
#
snarfed oh there's also numeric id, but that's not in back
#
snarfed er in backfed posts
#
aaronpk There's no username in the block list
#
aaronpk I'm also hoping to do this with no knowledge these are twitter accounts
#
snarfed right, that will take more talking btw us
#
snarfed in my blocklist i see numeric_id and username fields. do you not?
#
aaronpk no
#
snarfed ?! in any of the user items?
#
aaronpk http://i.imgur.com/6WM62t4.png
#
snarfed oh sorry i was looking at AS (ie json)
j12t joined the channel
#
aaronpk Oh!
#
snarfed yes agreed they're not in mf2, i can add them ther
#
aaronpk "nickname" is an mf2 h-card property right?
#
snarfed yes
#
snarfed i'll plan to include profile url though, that'll maybe be best
KevinMarks_ joined the channel
#
aaronpk yeah i think so. it should be fine to include both URLs in the "url" property too
#
@avalonhost Implementing Webmentions #webdesign https://allinthehead.com/retro/378/implementing-webmentions (twitter.com/_/status/888426617244200962)
#
Zegnat nickname is on my h-card (because of course it is), but you wouldn’t know for sure whether that meant Twitter username, right?
barpthewire, KevinMarks and KevinMarks_ joined the channel
#
snarfed Zegnat: in this case we do because they're explicitly twitter users, coming from granary or bridgy
#
sknebel wouldn't it make sense for the uid to include the twitter userID, not the nickname?
#
aaronpk i would think so yeah. the point of uid is that it's an arbitrary string that doesn't change
#
snarfed yup
AngeloGladding, jeremycherfas_, gRegorLove and tantek joined the channel
#
tantek ooh block lists! I wonder how many blocks I have on Twitter
#
Loqi tantek: prtksxna left you a message 13 hours, 14 minutes ago: and aaronpk No one seems to have specific word for it. nirzar at wmf pointed out, Apple uses: "Entire screen", "Part of the screen", and "App window". melchoyce at wp doesn't differentiate between them. quiddity at wmf calls it a screenshot and says the tiny size would make it obvious that it was partial.
#
tantek reads up on https://github.com/aaronpk/webmention.io/issues/86 wow
#
Loqi [tantek] #86 Feature: persistent webmention delete
#
tantek https://webmention.io/dashboard Recent Webmentions is all Bridgy Twitter homepage mentions ???
#
aaronpk of course
#
aaronpk i had a feeling that would happen
#
tantek thinking of showing a likes facepile etc. from a certain date forward, so I can check those at least. but even that doesn't scale eventually since a spammer can always go like a really old tweet
#
tantek considers making a request for "Recent home page webmentions" vs "Recent permalink webmentions"
#
tantek where "home page" could just be whatever you used to sign into indieauth
#
aaronpk this is again reaffirming my thoughts that home page mentions are different enough they could be handled by a totally different service
#
tantek aaronpk except they do show up unified in typical notification UIs
#
aaronpk sure but that's a different problem
#
tantek aaronpk - hah I could split the difference and delegate my homepage webmentions to webmention.heroku
#
tantek while still using wmio for permalink webmentions
#
aaronpk when i was doing the SWAT0 work, the hardest part was constructing the notification text for home page webmentions
#
aaronpk it's much easier to do it for responses to posts
#
aaronpk "tantek liked a photo you were tagged in" <-- there's a lot going on there
#
tantek indeed! did any of that end up in libraries?
#
aaronpk sadly no. some of it is in webmention.io, but kind of broke when i tried to add clustering
[mko] joined the channel
#
[mko] Hi @prtksxna -- what's up?
#
aaronpk i need to totally rethink the clustering implementation there too
#
prtksxna Hey [mko] o/ I was looking at passport-indieauth and the README said I should ping you :)
#
prtksxna [mko]: Are there any pending issues there? Can I help?
#
[mko] Oh, yeah. I think I wrote that at some point a while ago. I haven't touched it in probably over a year though.
#
[mko] No idea. I really haven't had time to hack on it in a long time. You're welcome to contribute there if you find any issues.
#
[mko] Hm. Not sure why the Git Repo has such an aggressive item in the Readme. It definitely works. I've used it on my personal site since ~2014.
#
[mko] Haven't updated it for newer versions of Passport though.
#
tantek wow the micro.blog responses to https://aaronparecki.com/2017/07/19/24/webmention-io
#
Loqi [Aaron Parecki] The https://webmention.io dashboard just got a whole lot more useful! You can now see the last few dozen webmentions you've received, and delete them if you want! This is the first step towards a "block" feature in webmention.io: https://indieweb.org... https://aaronparecki.com/2017/07/19/24/photo.png
#
tantek aaronpk did you have to crosspost, or are those replies a result of people seeing your posts automatically in micro.blog and replying to them there in their UI?
#
aaronpk yeah! I don't have to do anything manual to have my posts show up in micro.blog!
#
aaronpk it's pulling my Atom feed
#
tantek oh yeah me too!
#
tantek including my @-replies!
#
aaronpk and micro.blog sends webmentions, so i get the comments back automatically!
#
tantek that's pretty amazing. I wonder if the people replying know that is happening?
#
aaronpk a lot of people are doing that!
#
tantek that might be worth a short blog post
#
tantek "Federating IndieWeb With Micro.blog in Practice"
#
tantek or something
#
aaronpk hehe
#
tantek since that's literally what's happening
#
tantek along with screenshot of https://aaronparecki.com/2017/07/19/24/webmention-io responses etc.
#
Loqi [Aaron Parecki] The https://webmention.io dashboard just got a whole lot more useful! You can now see the last few dozen webmentions you've received, and delete them if you want! This is the first step towards a "block" feature in webmention.io: https://indieweb.org... https://aaronparecki.com/2017/07/19/24/photo.png
#
tantek Just explaining how it works, from your post on your own site, to automatically federating with PuSH+Atom to micro.blog, to people their responding "locally", to Micro.blog automatically federating those responses with Webmention back to your site
#
tantek all without users on either side having to do anything "special"
gRegorLove_ joined the channel
#
tantek !tell schmarty your "favorite" https://martymcgui.re/2017/07/20/064451/ needs explicit "u-like-of" markup for it to be recognized as a like :)
#
Loqi Ok, I'll tell them that when I see them next
#
Loqi [Marty McGuire] ★ Favorited https://aaronparecki.com/2017/07/19/24/webmention-io
#
schmarty hmm. i just changed my markup a bit, must have broken it.
#
tantek.com edited /stream.thatmustbe.us (+309) "indieweb examples, see also" (view diff)
#
schmarty ah, yep, i had dropped u-like-of, u-repost-of, and u-bookmark-of from my templates. whoopsies. thanks for the heads up.
#
tantek.com edited /webcal (+433) "indieweb examples, client support, see also" (view diff)
#
tantek anyone else have a "webcal:" link on their home page to their upcoming events? https://indieweb.org/webcal#IndieWeb_Examples
#
gRegorLove I bought a domain with Google Domains today. Super simple and super fast DNS update propagation too.
#
tantek what is Google Domains
#
Loqi It looks like we don't have a page for "Google Domains" yet. Would you like to create it?
#
tantek sounds like an experience worth capturing!
#
gRegorLove Google Domains is [[Google]]'s domain name registrar https://domains.google
#
Loqi ok
#
loqi.me created /Google_Domains (+103) "prompted by tantek and dfn added by gRegorLove" (view diff)
#
tantek what is a domain name registrar
#
Loqi A domain name registrar is an organization that will register domain names for you typically for an annual fee https://indieweb.org/domain_name_registrar
#
gRegorLove Now to decide if I want to move all my domains there or not.
#
gregorlove.com edited /Google_Domains (+6) "links" (view diff)
#
gregorlove.com edited /domain_name_registrar (+21) "/* Registrars */ Google Domains" (view diff)
#
gregorlove.com edited /Google_Domains (+383) "== IndieWeb Experience == +willnorris citation" (view diff)
[jeremycherfas] and eli_oat joined the channel
#
gregorlove.com edited /Google_Domains (+246) "/* IndieWeb Experience */ +me" (view diff)
#
gregorlove.com edited /Firefox_Focus (+2) "link dfn" (view diff)
#
schmarty tantek: i like the h2vx webcal link idea! any way to have it work for https-only sites? ;}
#
gregorlove.com edited /authN (-23) "r" (view diff)
#
schmarty http://h2vx.com/ics/martymcgui.re returns an empty calendar named "301 Moved Permanently"
#
schmarty oh, i can just slap the protocol in there. i still get an empty vcalendar back, though.
#
schmarty tantek: do i understand right that h2vx only picks up mf1 hCalendar-style events w/ vevent?
#
tantek schmarty: yeah :/ https://github.com/microformats/h2vx.com/issues/3
#
Loqi [tantek] #3 support microformats2
#
schmarty ah, thanks for the link.
#
tantek huh, just realized I should add vevent (etc.) as a stopgap to my event posts too so they get included beyond the manual sidebar of upcoming events
#
schmarty mind if i add to that ticket? i put my h-events on my homepage into an h-feed
#
tantek go for it
#
tantek though h-feed support isn't needed for that
#
tantek you can use an escaped fragment id to have h2vx convert only part of your page
#
tantek though it might as just grab all the h-event it can find
#
schmarty haha fragments is a clever idea!
#
Loqi schmarty: lol
#
gRegorLove It's been about a year, but I was tinkering with using php-mf2 for h2vx
#
schmarty this bumps up against my ignorance of mf2 parsing - i wouldn't expect an mf2 consumer to "look inside" an h-feed unless it was coded to do so.
#
tantek mf2 consumers can do whatever they want with the entire parsed tree
#
tantek so something which only cares about h-entry could just walk the whole tree and grab all the h-entry
#
tantek same for h-card, h-event etc.
[kevinmarks] joined the channel
#
[kevinmarks] Though that can lose implications of nesting
#
gRegorLove barnabywalters mf-cleaner lib adds that to php-mf2. It's handy. Don't know about the other parsers.
#
tantek KevinMarks_: only if those matter for the use-case
#
tantek otherwise, nothing lost
#
schmarty that makes some sense. i would expect diff't tools to make diff't explicit choices about walking the whole tree vs. stopping at the "top level"
#
tantek right
#
aaronpk snarfed: what do I have to do to make the photos from my mf2 appear inside the content html in the atom feed? https://granary-demo.appspot.com/url?input=html&output=atom&url=https%3A%2F%2Faaronparecki.com%2Fphotos
#
tantek e-content?
#
aaronpk not e-content, atom content
#
tantek right, make sure your photos are in the e-content in your mf2
#
tantek atom content was literally the primary use-case for e-content
#
aaronpk oh, but, that's not the recommended way for marking up photo posts...
#
tantek huh? your photo is part of your content
gRegorLove joined the channel
#
aaronpk that..hasn't been true for a while
#
tantek not AFAIK, I've been doing it for a while
#
tantek since I started posting photos I think. videos too
#
aaronpk i shut down monocle when? I remember when building it, it was a nightmare dealing with figuring out whether i should display the `photo` property, because sometimes it would appear in the content html too
#
Guest35190 photos in e-content works for me in granary: https://granary-demo.appspot.com/url?input=html&output=atom&url=http%3A%2F%2Fgregorlove.com%2Fnotes%2F
#
aaronpk and most of the sites i was parsing didn't include photo in the content
#
schmarty i had issues w/ this for photo, video, and audio posts on my site in terms of getting them to look "right" in woodwind
#
tantek larger point remains. if you want it in the atom content, put it in the e-content. that's pretty simple / straightforward
#
tantek whether a photo or anything else
#
aaronpk i'm not sure doing something exceptional for Atom compatibliity is worth changing my HTml
#
tantek like your articles include everything in your e-content for example
#
aaronpk at that point i'd rather generate the Atom feed myself
#
schmarty it's why i put my <audio> inside e-content for wehavetoask
#
aaronpk yeah but there's a distinct difference between an article containing images and a photo post
#
tantek it's not exceptional - conceptually the photo is (the biggest part of) the content
#
aaronpk i'm confused because i thought this was a done deal
#
aaronpk this feels like a debate that should have happened and been done with years ago
#
schmarty opens up his site templates and a bag of popcorn.
#
tantek if it was it would have been documented
#
aaronpk hm, there is no "how to mark up" section here https://indieweb.org/photo
#
tantek i'm confused because I never had any problems with my photo posts
#
tantek with u-photo on photo(s) inside e-content. has worked great in all consuming apps
#
prtksxna I am done with the post-type-discovery implementation https://github.com/twozeroone/post-type-discovery
#
Loqi [twozeroone] post-type-discovery
#
tantek whoa
#
aaronpk it just means that consumers have to check whether the u-photo URL appears in the content, and if it doesn't, then it should display the photo
#
tantek sounds like a specific kind of consumer
#
schmarty i don't remember why but i got the sense that nesting further properties inside of e-content was weird or bad.
#
tantek for example, Bridgy Publish has no such need
#
tantek schmarty: not sure where you got that idea
#
prtksxna tantek: I have a bunch of test cases now. Do we want to create central repo and publish on npm so that others can use as dependency?
[gregorlove] joined the channel
#
[gregorlove] This is the "photo" vs "note with a photo" post discussion: https://indieweb.org/photo#Discussion
#
aaronpk schmarty: putting additional e- p- u- properties inside e-content doesn't make them appear "nested" in the parsed result
#
tantek prtksxna: the spec repo is the "central" repo
#
schmarty i wish i had made a note of it :|
#
tantek prtksxna: when the WD goes to CR we move it to a w3c repo if I recall
#
prtksxna tantek: Ok. So the tests should live there? I got confused looking at https://github.com/microformats/tests
#
Loqi [microformats] tests: Microformats test suite
#
tantek prtksxna: no in the post-type-discovery spec repo
#
www.svenknebel.de edited /featured_image (-31) "redirect" (view diff)
#
schmarty my thinking has to do with that e-content appearing later inside of a) readers and b) reply-contexts
#
prtksxna tantek: Right, that is what I meant
#
[gregorlove] Yeah, reply-contexts inside e-content would be weird in most cases I think
#
schmarty gRegorLove: i am thinking the other way around
#
[gregorlove] Oh. Hm
#
tantek [gregorlove]: yes because the stuff in the reply-context is not your content
#
aaronpk tantek: if your photo is inside the e-content, what's the purpose of marking it up with u-photo at all?
#
prtksxna tantek: But would it be ok to publish an npm package from https://github.com/tantek/post-type-discovery
#
Loqi [tantek] post-type-discovery: W3C Post Type Discovery specification
#
tantek schmarty: also your content is not part of the thing you're replying to either so that also doesn't make sense
#
[gregorlove] I use e-content > u-photo for Bridgy Publish
#
tantek has no idea about publishing npm packages
#
aaronpk prtksxna: i would say the indieweb org is a better place for code that gets published to npm
#
tantek aaronpk - you already pointed out the difference
#
schmarty tantek: i am referring to showing the e-content of the post i am replying to
#
prtksxna tantek: We could, as long as you're ok with one extra json file
#
tantek aaronpk is right about code
#
prtksxna aaronpk: Hm, for the test suite as well?
#
tantek that probably makes sense too then
#
schmarty it would not be part of my post's content, but in my case it would appear as an h-cite inside the h-entry for my post.
#
tantek aaronpk - u-photo was introduced because we needed a way to say *this is a photo post*
#
prtksxna The way I have it right now is a bunch of jsons with a prefix that tells their type, very simple - https://github.com/twozeroone/post-type-discovery/tree/master/tests
#
tantek as opposed to hey this is a random image / photo in a blog post
#
prtksxna aaronpk tantek: Could you create a repo then?
#
aaronpk https://github.com/indieweb/post-type-discovery-node
#
Loqi [indieweb] post-type-discovery-node
#
tantek thanks aaronpk!
#
schmarty ah, ok, this has me thinking about making reasonable fallbacks
#
prtksxna aaronpk: Should it be post-type-discovery-test-suite?
#
prtksxna aaronpk: Or did you mean to transfer https://github.com/twozeroone/post-type-discovery to indieweb?
#
Loqi [twozeroone] post-type-discovery
#
schmarty so an mf2 consumer that "understands" photo posts (or audio, video or x-awesome-thing) would go looking for those mf2 properties
#
schmarty For Everything Else, there's E-Content.
#
aaronpk prtksxna: oh i misunderstood, you want to publish the test suite separately from your repo?
#
prtksxna aaronpk: My repo is already published. I was thinking that I'd move the test suite out so that its helpful to others
#
tantek schmarty - see discussion about post-type-discovery :)
#
schmarty ok, i am digging through some of my site changelogs
#
tantek attempts combining threads
#
aaronpk prtksxna: gotcha, is the test suite specific to node or useful for any language/
#
prtksxna aaronpk: https://www.npmjs.com/package/post-type-discovery
#
schmarty i moved my u-photo markup outside of e-content because woodwind was displaying my images twice
#
prtksxna aaronpk: The way I have it right now is a bunch of jsons with a prefix that tells their type, very simple - https://github.com/twozeroone/post-type-discovery/tree/master/tests
#
tantek schmarty: wasn't a problem for my posts
#
schmarty once because it was in e-content, once because it found and displayed my u-photo
#
tantek again, I didn't see that problem
#
[gregorlove] Likewise, I haven't seen that in Woodwind
#
[gregorlove] I didn't realize it did special handling for u-photo
#
aaronpk i saw that a lot in monocle until i wrote a thing that checks whether the u-photo appears in the e-content and didn't show the u-photo in that case
#
aaronpk prtksxna: https://github.com/indieweb/post-type-discovery-tests
#
Loqi [indieweb] post-type-discovery-tests
#
schmarty hmm
#
prtksxna Thanks aaronpk
#
tantek aaronpk, makes sense to have that type of cleverness constrained to a few readers rather than burdening publishers
#
aaronpk well the burden is on me right now to move my u-photo inside e-content
#
aaronpk or rather, expand my e-content to include the photo
#
aaronpk which is... difficult
#
aaronpk and is going to add a lot of whitespace junk to my content
#
tantek only if you care about using granary / atom?
#
aaronpk i mean i want people to see my photos if they follow my atom feed
#
tantek or you could file an issue on granary asking it to implement Post Type Discovery, and thus support u-photo -> AS "photo" object
#
aaronpk nobody consumes AS photo in atom :P
#
schmarty here's one of my photo posts as it appears in woodwind now
#
schmarty https://media.martymcgui.re/4d/b2/aa/19/db775238220e4beb9b83e5a7d26c324c8c033d72b3231efe570750e4.png
#
aaronpk this all started because micro.blog is missing my photos
#
tantek I can't find anyone who consumes AS in atom in practice any more
#
aaronpk micro.blog does not have special photo handling, it expects photos to be inside the atom or jsonfeed content
#
schmarty !mf2 https://martymcgui.re/2017/07/16/222914/
#
Kaja https://php.microformats.io/?url=https%3A%2F%2Fmartymcgui.re%2F2017%2F07%2F16%2F222914%2F
#
Loqi [Marty McGuire] Great time doing a set at the Unified Scene Theater tonight with Remote Possibilities! https://martymcgui.re/imageproxy/960,fit,sRFFyagSHGxzqrknwLoZ5Xcvdg05Rj7kdfdhVX6edmXk=/https://media.martymcgui.re/e3/36/73/d8/0fd74af126e77f49a3e004bb07ec50f3a4c3b825034a0da6fd552a5e.jpg
#
tantek aaronpk, so that's where the requirement is coming from
#
aaronpk well and anyone else who's using an atom feed reader
#
tantek you could file an issue on micro.blog asking them to consume h-feed / h-entry and do post type discovery
#
aaronpk that's a bit bigger of an ask, and also isn't likely to happen if we're still debating whether the u-photo should be inside the html content :P
#
aaronpk i.e. there isn't a "how to mark up" or "how to consume" on /photo
#
tantek aaronpk - u-photo doesn't have to be inside e-content - that's the point. it's publisher flexibility / freedom. but if they care about e-content to Atom content then they can decide accordingly
#
schmarty this is one of the things i really struggle with in terms of mf2 on my site
#
tantek wouldn't surprise me if there were folks who wanted their photos viewed only in the context of their site design (like nice lightboxes etc.) and not in some janky ancient feed reader
#
tantek this affords them that freedom
#
schmarty keeping track of all the consumers i want my site to "work right" with
#
tantek (and not a theoretical, lots of Flickr API discussion threads about people not wanting their photos specifically in rando API apps / feeds because they slap bad design around them)
#
tantek (found while looking for Flickr "block" contacts support in the API. turns out people wanted to block the API itself )
#
tantek schmarty: hence why we steer folks towards consuming h-feed / h-entry instead of legacy feed formats that make this all awkward
#
tantek problem is that RSS/Atom etc. are all "low-fi
#
tantek "
#
aaronpk everything is low-fi when it starts until it expands to include all post types
#
tantek so no matter how you "adapt" to them, the adaptation will be restrictive / lossy
#
aaronpk there's always going to be a "fallback" problem
#
tantek fallback is different
#
tantek hence we debated using p-summary for that
#
tantek speaking of
#
[gregorlove] Perhaps tantek and I and others who have u-photo in e-content should move our examples to https://indieweb.org/photo#Notes_with_photos_examples
#
tantek https://github.com/tantek/post-type-discovery/issues/25
#
Loqi [tantek] #25 Response Type: consider "reply" for 2nd to last for fallback use-cases
#
aaronpk so, in summary, i should not change my HTML, and that means i can't get my photos to appear in micro.blog
#
tantek that's not the summary no - because I provided a bunch of "OR ..." above
#
tantek that's one possible option you have to choose from
#
tantek [gregorlove]: nope. u-photo is enough to make it a "photo post"
#
tantek that section is specifically for "Not quite photo posts, but similar, notes"
#
tantek and there's only one example there because it predates u-photo, research from before we figured it out
#
aaronpk would it be a granary feature to add the photo to the atom content if it's not already there?
#
schmarty sounds to me like the next step is to open a granary issue to support post type discovery
#
tantek sure - that's kinda what I was implying by the u-photo -> AS photo
#
[gregorlove] tantek: Then /photo needs other updates, since there's the "photo" vs "note with photo" distinction on there.
#
tantek [gregorlove]: what other updates/distinctions?
#
tantek both variants (in or out of e-content) are valid u-photo posts
#
tantek so one list makes sense
#
[gregorlove] Well, sounds like you're saying "note with photo" isn't a thing. It's just a "photo post" regardless of u-photo location?
#
tantek if you want to add notes to your example describing markup details of what you do, that's helpful (and we've done that elsewhere)
#
[gregorlove] There's two lists currently.
#
tantek [gregorlove]: I have a handful of notes with photos, where the photo is not the first thing in the post, and my code knows to *not* mark it up as u-photo, deliberately
#
tantek there's one list of "photo" posts
#
tantek there's lots of post types with multiple lists overall like that. previous examples. prototypes. etc.
#
[gregorlove] https://indieweb.org/photo#IndieWeb_Examples and https://indieweb.org/photo#Notes_with_photos_examples
#
tantek things that are either dead/gone or "not quite" but similar enough in appearance to be worth documenting but explicitly distinguishing so as to not mislead folks
#
[gregorlove] is confused
#
tantek previous examples = examples which are dead for some reason, link stopped working, markup broke, etc.
#
[gregorlove] I'm not talking about previous examples.
#
tantek prototypes = examples which have a vaguely similar appearance (perhaps just due to content editing) but have no explicit way of distinguishing themselves as that specific type of post
#
tantek point is there are multiple such lists for research purposes
#
tantek but only one list of real live working examples
#
tantek for each post type
#
tantek examples that you can go look at, copy the markup, and things should work
#
aaronpk i'm going to continue this discussion here for now https://github.com/snarfed/granary/issues/113
#
Loqi [aaronpk] #113 Include mf2 photo in Atom content?
#
[gregorlove] My understanding was "note with photo" is a note with the u-photo inside the e-content based on the 2014 discussion on /photo
#
sebsel I don't include the u-photo in the e-content either at the moment. I do sometimes post a note with an image, without u-photo on it.
#
sebsel https://seblog.nl/2017/06/04/1/ik-las-feesten vs https://seblog.nl/2017/07/12/14/foto
#
Loqi [Sebastiaan Andeweg] Ik las 'feesten'.
#
[gregorlove] But it sounds like today, a note with u-photo inside the e-content is still considered a "photo post"
#
[gregorlove] If that's the case, I don't think the separate lists on /photo portray that clearly, today.
#
schmarty my read is that a "note with photo" is a note with <img> tag inside but without u-photo.
#
tantek [gregorlove]: the separate lists have nothing to do with that
#
tantek schmarty is right
#
[gregorlove] It's causing me confusion today, so apparently it does have something to do with it.
#
[gregorlove] Ok, so "Not quite photo posts, but similar, notes with embedded photos are another approach to photo-like posts." should probably indicate "without a u-photo"
#
tantek u-photo => photo post
#
tantek (taken in the context of post-type-discovery obv)
#
[gregorlove] (that quote is from /photo)
#
tantek [gregorlove]: why do we have to caveat when NOT to mark something up?
#
prtksxna aaronpk tantek: Things are all setup now. https://github.com/indieweb/post-type-discovery-tests has the tests that I had, and is published to npm too - https://www.npmjs.com/package/post-type-discovery-tests. And my implementation doesn't have tests any more and uses this one as a dependency instead https://github.com/twozeroone/post-type-discovery/blob/master/package.json#L33
#
tantek shouldn't "less work" be the default / obvious?
#
Loqi [indieweb] post-type-discovery-tests
#
[gregorlove] :thumbsup: I'm on board with u-photo => photo post
#
aaronpk prtksxna++
#
Loqi prtksxna has 1 karma in this channel (10 overall)
#
tantek prtksxna++ nice work!
#
Loqi prtksxna has 2 karma in this channel (11 overall)
#
tantek alright adding at least the class="u-photo" bit to the opening
#
tantek I think this post type is older than when we started explicitly adding How to Markup sections
#
prtksxna We should have post-type-discovery.js.org soon too :) https://github.com/js-org/dns.js.org/pull/1368
#
Loqi [prtksxna] #1368 Add post-type-discovery.js.org
#
[gregorlove] Perhaps "note with image" is a clearer phrase than "note with photo"
#
Loqi yea
#
sebsel agree
#
tantek.com edited /photo (+457) "/* How */ How to markup stub" (view diff)
#
tantek [gregorlove]: there you go: https://indieweb.org/photo#How_to_markup including caveat which could likely be rephrased better
#
tantek ?
#
tantek we had a similar problem with over-use of u-photo in blog posts that should have been using u-featured
#
tantek a while back, I think related to WordPress plugin -> Bridgy Publish -> Twitter
#
tantek when similarly, the request was, but I want that photo to show up in the tweet
#
tantek and "Bridgy Publish says to use u-photo to do this, so that's what I'm going to do" therefore turning their article into a photo post
#
tantek so we had to get both the WordPress Plugin to use u-featured instead, and Bridgy Publish to use u-featured as a fallback for when POSSEing to Twitter
#
tantek [gregorlove]: irony, the one "note with image" example has a dead image link
#
tantek I'm ok with dropping that whole section since it's dead anyway and has caused more confusion at this point than anything helpful
#
[gregorlove] My proposal:
#
[gregorlove] == Notes with image examples == [[Notes]] with embedded photos that are not marked up with <code>u-photo</code> are similar to photo posts.
#
tantek !tell barnabywalters image src seems to have died on your note with photo https://waterpigs.co.uk/notes/686/ found from https://indieweb.org/photo#Notes_with_photos_examples
#
Loqi Ok, I'll tell them that when I see them next
#
Loqi [Barnaby Walters] Guitar Monster: http://photos.waterpigs.co.uk/photos/custom/201212/IMG_1166-97e81d_870x550.jpg
#
tantek [gregorlove]: go for it - I don't think it makes that much difference, but if you think it helps
#
[gregorlove] Will do. It captures the definition that clears up my confusion.
#
tantek the reason I brought up the WP/article/tweet photo problem is that I think that's the larger confusion
#
tantek people jump from "I want my image to show up (insert reader use-case)" to "this must be a photo post!"
#
tantek when really, to make an image show up, perhaps you want it to show up in a link-preview etc.
#
tantek or more likely, "fix / improve the reader" is the answer
#
gregorlove.com edited /photo (+57) "/* Notes with photos examples */ clarify, note currently dead image src" (view diff)
#
aaronpk unfortunately, typically it's harder to "fix the reader" than make a small change on your one site
#
tantek aka break the spec on your site
#
tantek yeah, web browsers used be more stubborn like that until the mid 2000s
#
tantek took years of advocacy by the Web Standards Project etc.
#
tantek now web browsers actually keep up with specs, so you can get browsers fixed by contributing to specs
#
tantek "feed readers" haven't yet caught up to that kind of positive ecosystem interaction
#
tantek this is one of the reasons why I feel de-motivated to put much work into maintaining my Atom feed
#
tantek the side file ecosystem is not really a healthy "community
#
tantek "
#
aaronpk that's why i like granary doing iit for me
#
tantek aaronpk - note on my micro.blog/t that where my posts say just "Event"
#
aaronpk ah
#
tantek I couldn't be bothered to do anything for Atom for my events
#
tantek because Atom consumers are dumb and don't do anything useful with events
#
tantek so if someone cares they can read the title and click the permalink to see the event
#
tantek and I'd rather get a "feed reader" to support h-feed and h-event than try to 1) figure out how to do AS1/Atom "event" objects, 2) convince Atom consumers to add support for that
#
tantek this is kind of why Atom consumers are a dead end in the world of "social media"
#
tantek the incentives are just not there to upgrade everything to do with Atom / AS to get "social media parity"
#
tantek proxies/converters like granary are a decent stopgap / fallback but that's it
#
tantek also why all the "I'm just going to go back to RSS!" folks are horribly misguided
#
tantek It's kinda akin to, Netflix sucks, I'm going to back to buying VHS tapes because they're so cheap now!
#
tantek or Laserdiscs for the higher quality Atom analogy :P
#
prtksxna.com edited /User:Prtksxna.com (+114) (view diff)
#
tantek goes back to trying to add legacy vevent support, just to make events show up in a specific consumer that should be upgraded to just support h-event ?
#
prtksxna tantek: How many posts per type do you think we'd typically need?
#
tantek good q - depends on where we see problems maybe?
#
tantek prtksxna: would appreciate your thoughts on this in particular https://github.com/tantek/post-type-discovery/issues/25
#
Loqi [tantek] #25 Response Type: move "reply" to 2nd to last to enable p-summary fallback use-cases
#
tantek probably the biggest change in the algorithm that's being considered
#
prtksxna tantek: Oh, yeah, sorry, forgot about that. Will take a look.
#
tantek schmarty, let me know if you get stuck with h2vx
#
tantek or if you'd like to help with https://github.com/microformats/h2vx.com
#
Loqi [microformats] h2vx.com: H2VX microformats to vCard/iCalendar converter, test dev.h2vx.com for issues
#
tantek one thing we really should do is move http://microformats.org/wiki/h2vx#issues to https://github.com/microformats/h2vx.com/issues
#
Loqi H2VX is a production deployment of the X2V hCard and hCalendar conversion transforms. It converts hCard contacts and hCalendar events on web pages to .vcf and .ics respectively for use in desktop and other client software applications. Contents ...
#
tantek ok, captured that here so it's not just lost in IRC http://microformats.org/wiki/h2vx#issues
#
Loqi H2VX is a production deployment of the X2V hCard and hCalendar conversion transforms. It converts hCard contacts and hCalendar events on web pages to .vcf and .ics respectively for use in desktop and other client software applications. Contents ...
#
tantek alright I'm dogfooding dev.h2vx.com on my personal site, that's one step toward updating the production h2vx to that version (which does have a bunch of fixes and HTML5 improvements)
#
tantek.com edited /webcal (+33) "/* IndieWeb Examples */ note using dev version of h2vx" (view diff)
[kevinmarks] joined the channel
#
[kevinmarks] Hm p-summary can be inside e-content or separate, u-photo is similar
#
tantek [kevinmarks] it's interesting, from an HTML perspective (class names) it's obvious that can be true. From a "programming data structure" perspective (like if you think in JSON trees of disjoint data) it seems confusing.
#
[kevinmarks] Also, as atom (and to a lesser extent, rss) can contain html, just put the h-event in the e-content and tunnel it through
#
[kevinmarks] That is true, but microformats have always had that DRY polymorphism
#
tantek I'd rather avoid the extra wrapper (h-event in h-entry)
#
tantek [kevinmarks]: perhaps that's it - "DRY polymorphism" is confusing to a strict datastructure view of the world
#
[kevinmarks] But that was your point earlier - if we have something looking for h-event it can dig through nesting
#
[kevinmarks] So h2vx 2.0 can
#
tantek h2vx does now
#
tantek when looking for vevent or vcard because they're "just" class names
#
[kevinmarks] Right
#
tantek it "digs" for them the same way the simple CSS class selector does .vevent or .vcard
#
[kevinmarks] Because we didn't define nesting generally in mf1
#
tantek which is why I think this is more obvious to front end designer devs than programmers that think in terms of data structures all day
#
tantek we defined very little "generally" in mf1 :)
#
tantek we did define nesting of specific microformats in others e.g. vevent location vcard etc.
#
tantek gets closer to having his new event posts show up in iCalendar via dev.h2vx.com, fails, files the necessary issues of things to fix in H2VX - in particular https://github.com/microformats/h2vx.com/issues/5 and https://github.com/microformats/h2vx.com/issues/6
#
Loqi [tantek] #5 support datetime values with space instead of T separator
tbbrown joined the channel
#
ascraeus.org edited /User:Ascraeus.org (-259) "/* My Indieweb Setup */" (view diff)
#
schmarty tantek: i think i made it pretty close to a working webcal: subscribe link
#
schmarty ran into the same "datetime value with a space" that you did
#
schmarty dev.h2vx.com outputs something that looks right to me but google calendar produces no events.
#
ascraeus.org edited /Micropub/Servers (+137) "/* IndieWeb Examples */" (view diff)
#
ascraeus.org edited /Micropub/Servers (+0) "/* Daniel Goldsmith */" (view diff)
#
ascraeus.org edited /Micropub/Servers (+2) "/* Daniel Goldsmith */" (view diff)
#
ascraeus.org edited /Micropub/Servers (+27) "/* Daniel Goldsmith */" (view diff)
#
tantek schmarty: would be interested in your feedback on using p-summary as fallback, it impacts this issue: https://github.com/tantek/post-type-discovery/issues/25
#
Loqi [tantek] #25 Response Type: move "reply" to 2nd to last to enable p-summary fallback use-cases
#
tantek I'm pretty sure it's a good idea but I want to get at least *some* positive feedback before I go ahead and make all the changes, do all the documentation etc.
#
GWG Post Type Discovery is getting some attention?
tbbrown joined the channel
#
tantek GWG it's the quiet workhorse of enabling different kinds of indie web posts to actually work (federate) across systems
#
GWG tantek: I know. I'm pleased.
tantek joined the channel
#
schmarty p-summary fallback for posts is interesting. i feel like a "Liked a post <URL>" is a note - so p-name should be enough?
#
schmarty i am confused about when a post would have a p-summary but not a name.
#
[kevinmarks] Hmm. Maybe an image post with a description, but it is a bit odd
tantek, gRegorLove, [cleverdevil], snarfed, KevinMarks, davidmead, [manton], KartikPrabhu and KevinMarks_ joined the channel