#indiewebcamp 2015-04-28

2015-04-28 UTC
#
GWG snarfed: I think, after all the Bridgy requests, I owe you some Issue solving.
#
KevinMarks_ if you use the testing tool you can make facebook recrawl the page
scor joined the channel
#
tantek thanks rhiaro! that's a huge improvemet in /follow
lukebrooker joined the channel
#
aaronpk i can't find the testing tool anymore
#
tantek yeah I was about to ask for a URL :P
#
snarfed GWG: aww thanks! not at all though, definitely not quid pro quo like that
#
aaronpk this page mentions it and has a screenshot, but doesn't have a link to it https://developers.facebook.com/docs/sharing/best-practices#debug
#
aaronpk here we go https://developers.facebook.com/tools/debug/
#
tantek also https://developers.facebook.com/tools/debug/og/object/ from /OGP
#
aaronpk i clicked the button that forces them to go scrape it again, and they found the new image tag, but even that page still uses the old one
#
aaronpk oh now it works. weird caching bug
#
aaronpk although not live on facebook.com yet
KartikPrabhu joined the channel
#
GWG snarfed: Feedback makes things better
#
GWG It certainly motivates me.
tmro, wolftune and snarfed joined the channel
#
GWG Hmm...just got webactions enabled in test mode
#
GWG Who has webactions enabled?
#
tantek GWG, did you check /webactions#IndieWeb_Examples ?
#
GWG I did, yes. But, my question was more...people who have it set up, how often do they use it.
#
GWG I added the javascript protocol handler shown off at IWC UK last year. So, that is working. I just need to add a few more features for it to work properly.
#
kylewm GWG: I use webaction endpoints multiple times a day
#
kylewm voxpelli's javascript magic, I've only gotten to use a couple of times for testing
#
GWG kylewm: Any tips? I did see your indie-config page.
#
GWG I need to add links to my site.
#
kylewm don't worry about indie-config, that's my only tip
#
kylewm it's a bolt-on you can add later
#
GWG kylewm: I bolted it on already. I've moved on to improving the actions themselves.
snarfed, emmak, KartikPrabhu, wolftune, kylewm_, j12t and lukebrooker joined the channel
#
aaronpk okay well 28 mentions of this post (since I didn't POSSE it anywhere myself) is now motivating me to improve the display of non-comment mentions
snarfed, tantek and LCyrin joined the channel
#
tantek aaronpk - haha oops. ;)
#
Loqi awesome
snarfed joined the channel
#
tantek aaronpk - nice image too!
#
tantek classic pixelated view of code - very Hollywoodesque ;)
lukebrooker, loic_m, tilgovi, KevinMarks__, wolftune, KartikPrabhu, jacus and glennjones joined the channel
#
@oSoftDev Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592944223198019584)
#
@TheIndieSloth RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592944271432495104)
KevinMarks joined the channel
#
@indiegamesdevel RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592944439691190272)
#
@product_mobile RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592944606980939776)
#
@mobilephotos00 RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592944615369506816)
#
@MobileDifferent RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592944858622341120)
#
@TooStrong_Gamer RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592945171769249792)
KevinMarks__ joined the channel
#
@S78Mobile RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592945416439615488)
Jihaisse and tmro joined the channel
#
@prgmbot RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592955769106096128)
#
@mob_attack56 RT @oSoftDev: Ok--Enough punishment for one day. The new product User Controls. #indieweb #indiedev #webdevelopment https://twitter.com/oSoftDev/status/592944223198019584/photo/1 (twitter.com/_/status/592956019543707648)
glennjones, pfefferle, interactivist, jacus and csarven joined the channel
#
KevinMarks_ https://twitter.com/kevinmarks/status/592968671967883264
#
@kevinmarks @alwillis what's your URL? I only see microformats 2 markup on this profile http://pin13.net/mf2/?url=https%3A%2F%2Ftwitter.com%2Fintent%2Fuser%3Fuser_id%3D644523 (twitter.com/_/status/592968671967883264)
petermolnar and Sebastien-L joined the channel
#
csarven https://twitter.com/csarven/status/592973363800383489
#
@csarven We'd understand & embrace the impact of lower-case semantic web efforts on the upper-case #SemanticWeb #Web eg @microformats @schemaorg_dev (twitter.com/_/status/592973363800383489)
LCyrin joined the channel
#
@hmans So apparently I will be giving a nice big talk on IndieWeb and Pants in May. DIS GON B GUD (twitter.com/_/status/592986443779276800)
stream7, pfefferle_, eschnou, loic_m, AcidNerd, almereyda, scor, frzn and parzzix joined the channel
#
GWG Good morning
#
pfefferle good morning
#
GWG Hello, pfefferle.
parzzix joined the channel
#
GWG Any excitement there?
almereyda, parzzix, nloadholtes, interactivist_, sdboyer_ and tmro joined the channel
#
@MrSimonWood Anyone using Known For Education? Indieweb platform that integrates with VLEs https://withknown.com/education/ (twitter.com/_/status/593042757863788545)
#
pfefferle GWG excitement?
#
petermolnar is looking for an oembed provider wordpress plugin - aaand, surprise, one of the authors for it is pfefferle :)
#
pfefferle petermolnar :)
KevinMarks and frzn joined the channel
#
GWG Up to interpretation.
KevinMarks__ and csarven joined the channel
#
ben_thatmustbeme good morning #indiewebcamp
#
rhiaro good afternoon, ben_thatmustbeme
#
ben_thatmustbeme The logs have become too much, I don't read back anymore
#
ben_thatmustbeme other than a quick flip though
#
petermolnar indieweb is becoming too popular...
#
rhiaro clearly we need indieindieweb
#
ben_thatmustbeme haha
#
ben_thatmustbeme starts #indieindieweb
#
petermolnar #evenmoreindierweb?
#
ben_thatmustbeme or #indierweb
#
rhiaro #indiestweb
#
ben_thatmustbeme joins them all
fourtonfish joined the channel
#
ben_thatmustbeme too much indie, must leave them
#
rhiaro decides enough is enough
#
ben_thatmustbeme should probably move this discussion to indiechat at this point though
#
ben_thatmustbeme https://www.w3.org/wiki/Socialwg/Social_API/Comparison#iw:Micropub_.26_ap:Outbox_.26_ldp:Container says micropub response payload empty...
pfefferle joined the channel
#
ben_thatmustbeme oh yeah, its response if via Link:
#
ben_thatmustbeme s/Link/Location
#
Loqi ben_thatmustbeme meant to say: oh yeah, its response if via Location:
snarfed joined the channel
#
ben_thatmustbeme HI SNARFED!!!
#
snarfed whoa
#
snarfed HI BEN_THATMUSTBEME!!!!
#
petermolnar OMG CAPSLOCK LOL!
#
snarfed WHAT
#
snarfed I'VE BEEN USING SHIFT ALL THIS TIME LIKE A SUCKER
#
snarfed on an unrelated note
#
snarfed finally tuned the cache clearing on my site so webmentions show up immediately
#
snarfed thanks to tantek for the nudge
#
ben_thatmustbeme nice
#
petermolnar so if someone starts flooding you with webmentions, it'll kill your site, wouldn't it?
#
ben_thatmustbeme webmentions are up to 5 minutes delayed for me as i run it on a cron job
#
ben_thatmustbeme petermolnar: no, not really
#
ben_thatmustbeme actually its there to prevent floods
#
petermolnar not yours, snarfed's
#
ben_thatmustbeme oh
#
ben_thatmustbeme haha
#
snarfed petermolnar: why?
#
petermolnar ( by the way, is there an ' there or not? )
#
petermolnar tonnes o' webmentions -> tonnes o' cache clear
#
petermolnar tonnes o' cache generation
#
snarfed cache generation is lazy, on page requests
#
snarfed more importantly though, i think we might worry too much here about preventing DoSes, amplification attacks, etc
chalettu joined the channel
#
snarfed i expect pretty much none of us actually get attacked like that in practice
#
snarfed good to keep in mind, but low low priority
#
snarfed (btw petermolnar yes you used the ' correctly :P)
#
petermolnar (thanks)
#
ben_thatmustbeme hmm, folowing a discussion on w3c about heade fields like to, cc, bcc, bto. If we do start indiemessaging, how do we specify who we are sending to?
#
rhiaro ben_thatmustbeme: also note that possible conflation of tags and mentions just now (maybe it's fine, I don't have enough brain to devote to working that out right now)
#
ben_thatmustbeme presentation wise I wouldn't want everything to be a note with a metion of the person in there, and person-tags could work but would seem to obfuscate things
#
snarfed ben_thatmustbeme: u-in-reply-to right? if it's a new msg, maybe set it to their home page
#
snarfed display: none it if you don't want it visible
#
ben_thatmustbeme hmmm... interesting
#
rhiaro while 'in-reply-to' is semantically different to 'to', practically I think it probably doesn't matter
#
snarfed exactly
#
ben_thatmustbeme display:none is fine if its out of the content section like that
#
ben_thatmustbeme because the other end can know to strip it out
#
ben_thatmustbeme the question is do we continue to reply-to home-page or do we start threading there?
#
rhiaro I'd say start threading for sure
#
rhiaro I in-reply-to: ben_thatmustbeme with /2015/04/msg4ben, and you in-reply-to: /2015/04/msg4ben
#
rhiaro ^ where by ben_thatmustbeme I mean your homepage
#
ben_thatmustbeme the only issue that once i post my reply-to /2015/04/msg4ben I need to remember that that post is a direct message and any reply-to that would be rendered as a direct message
#
rhiaro but I probably wouldn't want direct messages to people showing up in my main/homepage feed, and filtering them out by parsing the url they're in-reply-to seems a bit heavy
#
rhiaro it depends if the direct message is meant to be public or private
#
ben_thatmustbeme where as dropping the threading you can say that all messages from a specific person on in precisely one thread
#
rhiaro in my head everything I post on my site is public at the moment, because not is too hard right now
#
ben_thatmustbeme ahh, true, with public its rather important to keep the thread
#
rhiaro possibly public but obfuscated
#
snarfed what is indieweb-messaging
#
rhiaro and I'm a fan of threading, and I can still collect all messages in-reply-to a particular person in one place, and keep threading
#
Loqi indieweb-messaging is a suggested protocol for combining microformats2, WebMention, and IndieAuth to convey private posts between two people or between groups of people https://indiewebcamp.com/indieweb-messaging
#
ben_thatmustbeme examples on there just mention homepage
#
rhiaro oh that's explicitly private anyway
#
rhiaro well, maybe they're the same just with/without indieauth
#
snarfed for threading you'd just in-reply-to the previous message(s), right?
#
rhiaro snarfed: yes, but I think the concern is then do you lose the direct-message-ness of it
#
ben_thatmustbeme my other concern is two people responding to each other too quickly
#
ben_thatmustbeme or sending 2 messages at once
#
ben_thatmustbeme so i want to reply a second time, i guess i reply-to the last message from the other party
#
rhiaro twitter had this problem with threading actually
#
ben_thatmustbeme and which one do they reply to? are they already replying to the first before they see the second
j12t joined the channel
#
rhiaro s/had/has
#
Loqi rhiaro meant to say: twitter has this problem with threading actually
#
ben_thatmustbeme does twitter thread direct messages?
#
rhiaro no there's only one thread for dms, but for public @ mentions
#
snarfed sounds like a great problem to solve after normal messaging is up and working well
#
rhiaro or anything
#
rhiaro threads are forked all over the place depending on the message people reply ot, and it's easy to miss stuff
#
ben_thatmustbeme makes me wonder if its actually better to only in-reply-to the original message
#
rhiaro tweets are only related to the tweet they're in reply to, and not the origin of the thread
#
ben_thatmustbeme thus its all in one conversation,
#
rhiaro that could work, I guess just gotta try it and see
#
snarfed tryitandsee++
#
Loqi tryitandsee has 1 karma
#
ben_thatmustbeme actually the simplest way might be i create a post that either mentions, tags, or in-reply-to rhiaro's homepage (undecided on this yet)
#
rhiaro tryitandsee++
#
Loqi tryitandsee has 2 karma
#
ben_thatmustbeme and then rhiaro replies to that post
#
ben_thatmustbeme all future messages in the thread reply to both my first post AND rhiaro's first post
#
rhiaro hmm
#
ben_thatmustbeme that way just viewing the original post on either site would get the entire thread (basically just showing up in comments)
#
rhiaro oh that makes sense
#
ben_thatmustbeme and we both end up with a full copy of the conversation
#
snarfed i had to do exactly that for twitter backfeed in bridgy
#
ben_thatmustbeme it also extends easily to 3 or more people
#
snarfed surprisingly difficult
#
rhiaro but viewing a post from the middle of the thread would only be linked to the first posts
#
rhiaro and not to the one immediately before or after it
#
snarfed https://github.com/snarfed/bridgy/issues/217
#
ben_thatmustbeme rhiaro: true, though that will always be a problem.... but you could also in-reply-to the previous post as well
#
rhiaro yeah
#
ben_thatmustbeme makes it easy for new people to join the thread... they just in-reply-to the last post they saw and the original two posts
#
rhiaro seems like a lot of webmentions being sent. Maybe that's fine
#
ben_thatmustbeme then their context pulls could take care of getting the entire thread to their site as well
#
ben_thatmustbeme rhiaro: it does, thankfully most of them go to themselves
#
ben_thatmustbeme but the polling for all those webmentions can certainly add up
#
rhiaro what would we lose with replying to the very first message, and what it's directly in reply to, rather then the first two messages?
#
rhiaro as the first message from the second person would show up as a regular in-reply-to on their site anyway
#
rhiaro brb coffee
#
ben_thatmustbeme getting lost in pronouns
#
ben_thatmustbeme person A sends a message to person B
#
ben_thatmustbeme A creates post A1 that in in-reply-to B.com
#
ben_thatmustbeme here is another issue, A wants to add to that message thread with a second message (not uncommon to do)
#
ben_thatmustbeme if A does in-reply-to B.com its seen as a new message, if it is in-reply-to A1, B doesn't get a webmention
#
ben_thatmustbeme as far as B13 not having a reference to A12, i wonder if thats not more of just a UI issue. B.com knows that B13 is part of a direct messaging thread and can show the entire thread in the context
#
ben_thatmustbeme I almost wonder if its better if the original message can have people "subscribe" to it
#
ben_thatmustbeme thus get webmentions for ALL replies to that post
#
ben_thatmustbeme thus if C.com joins the conversation by in-reply-to A1, A.com will send the webmention along to B.com
#
ben_thatmustbeme but that seems to be getting too complex again
#
rhiaro ben_thatmustbeme: so a direct messaging thread can be a h-feed
#
rhiaro I have this nagging desire for h-feeds to have unique ids, which might help here
#
rhiaro a post with an 'in-thread' property
#
rhiaro maybe not
#
rhiaro I also like tags for sorting/separating so maybe tag a post with a threads url...
tantek joined the channel
#
ben_thatmustbeme rhiaro: h-feeds can already have unique IDs, URLS
#
tantek good morning #indiewebcamp!
#
ben_thatmustbeme good morning tantek
#
rhiaro ben_thatmustbeme: I haven't seen any. Maybe I haven't been looking. Or do you just mean implicitly of the page the feed is on?
#
rhiaro does anyone have pages with multiple feeds?
#
rhiaro maybe that's not a thing
#
ben_thatmustbeme rhiaro: I used to
#
ben_thatmustbeme i took it out
almereyda joined the channel
#
ben_thatmustbeme and i realized I should have had a url for the actual feed in each of them
#
ben_thatmustbeme an h-feed can have a u-url
#
rhiaro also if a page has a h-card and a h-feed h-card wins as per discussion earliner in the week
#
ben_thatmustbeme thus have full feeds and abridged feeds
#
rhiaro a u-url on a h-feed makes sense, wasn't sure if anyone used that
#
ben_thatmustbeme rhiaro: no, whichever is listed first wins
#
tantek rhiaro: that's not quit right
#
rhiaro tantek: okay, the first one wins
#
tantek that ^^^
#
rhiaro sorry, yeah, I know
#
ben_thatmustbeme needs to update his header to not always list h-card first... only on his homepage
#
ben_thatmustbeme otherwise it wouldn't work
#
ben_thatmustbeme i like the idea of u-in-reply-to an h-feed. that would certainly work for messaging and chats... can subscribe to the h-feed by any reader that way
#
tantek ben_thatmustbeme: rather than thinking about it as "first", instead, use <body> for the "type" of page, e.g. <body class="h-card"> or <body class="h-entry"> etc.
#
ben_thatmustbeme interesting. i can make that work
#
rhiaro it's okay for h-entrys to be nested inside an h-card?
#
tantek it's the same thing that OGP makes you do - pick a page "type", and indicate in a meta, except there's no meta
#
tantek rhiaro: have you seen my home page? ;)
#
KevinMarks__ Twitter does quite a good job of following the threads back with only one in-reply-to
#
KevinMarks__ And using it as a constraint on total explosion 💥
#
ben_thatmustbeme indeed, its just a matter of a specific conversation like IMs don't follow that same threading model
#
ben_thatmustbeme they force a single thread
#
petermolnar single threaded conversation is good imho; @mentions are there to make it clear who are you replying to
#
rhiaro I find twitters anti-explosion measures annoying because I miss stuff. But maybe it would be more annoying if I did see everything though
#
ben_thatmustbeme the threading i think makes a good amount of sense for the public messaging case
#
ben_thatmustbeme reply-to a post, etc
#
KevinMarks__ What cocks it up is people using the quoted tweet thing to reply
#
KevinMarks__ Because that doesn't set in-reply
#
ben_thatmustbeme direct messaging i don't think you can afford losing some of the conversation
#
rhiaro I see QTs not threading is a bug, rather than by design
#
KevinMarks__ Eg this thread https://twitter.com/kevinmarks/status/592947443299323904 is coherent in past and future
#
@kevinmarks @sideshowbarker @LeaVerou I'm not sure even @Hixie would say microdata was simpler than microformats (twitter.com/_/status/592947443299323904)
#
KevinMarks__ Whereas this one: https://twitter.com/sideshowbarker/status/592871208716214272
#
@sideshowbarker @LeaVerou IMHO this all suggests most webdevs don’t want to mark up their content w/ structured metadata, regardless https://twitter.com/sideshowbarker/status/592870197121384450 (twitter.com/_/status/592871208716214272)
#
KevinMarks__ Has a quoted link that points to a quoted link and you have to keep clicking through
#
tantek KevinMarks: not worth having webdev arguments with people that actually have or link to their own websites on their silo profiles.
KartikPrabhu, wolftune and snarfed joined the channel
#
aaronparecki.com edited /Micropub (+88) "/* Update */ return 200 on successful update" (view diff)
#
aaronparecki.com edited /Micropub (+50) "/* Actions */ nowiki" (view diff)
snarfed, KevinMarks, elf-pavlik, KevinMarks__, Nowaker, tilgovi and KartikPrabhu joined the channel
#
ben_thatmustbeme aaronpk re:micropub remove properties... you cannot remove all values from 2 properties this way?
#
aaronpk mp-remove[]=prop1&mp-remove[]=prop2
#
aaronparecki.com edited /Micropub (+88) "/* Update */ add example of removing all values for two properties" (view diff)
#
aaronparecki.com edited /Micropub (-314) "/* Updating Objects */ update examples of updating objects" (view diff)
#
aaronpk ben_thatmustbeme: i'm wondering about your thoughts on the custom syndication content here http://indiewebcamp.com/micropub#Actions
#
ben_thatmustbeme hmm, that seems to be getting kind of messy
#
ben_thatmustbeme i'd rather assume that content is not an array
#
aaronpk maybe custom property name somehow?
#
ben_thatmustbeme content-alt[twitter.com/aaronpk] = altversion
#
aaronpk afaik nobody actually does this
#
ben_thatmustbeme alt-content
#
aaronpk ...in a micropub client anywa
#
ben_thatmustbeme something like that
#
aaronpk my native posting interface has a separate box i can enter tweet content for, and i do that all the time
#
aaronparecki.com edited /Micropub (+220) "/* Syndicating a Post */ update examples to use mp- prefix" (view diff)
cweiske joined the channel
#
cweiske Why is the pgp key link not using rel="me" but rel="pgpkey"?
#
Loqi cweiske: kylewm left you a message 6 days, 2 hours ago: I'm getting empty (thin) pings from phubb now, did something change? http://indiewebcamp.com/irc/2015-04-22/line/1429716958207
#
Loqi cweiske: kylewm left you a message 2 days ago: do either of your hubs support multi-valued "hub.url" pings? http://indiewebcamp.com/irc/2015-04-26/line/1430067568743
benwerd joined the channel
#
ben_thatmustbeme cweiske: good question... it probably should be rel=me
#
ben_thatmustbeme doesn't know who suggested it initially
#
cweiske http://indiewebcamp.com/pgp#How_to_use_PGP_with_IndieAuth
#
aaronpk i think because people were already publishing keys with rel=pgpkey
#
kylewm looks like first suggested in https://golem.ph.utexas.edu/~distler/blog/archives/000320.html
#
ben_thatmustbeme wonders if he should just do rel="me pgpkey" at least
#
cweiske kylewm, I don't know why phubb should do thin pings now
#
cweiske I didn't change anything
#
aaronpk the key isn't really a representation of yourself though
#
aaronpk whereas other profile URLs are
#
KevinMarks_ that is <picture> <source>
#
cweiske is the problem still there?
#
KevinMarks_ or srcset
#
cweiske also, about the multi-valued hub.url pings - you mean I notify your subscriber about the change of multiple urls?
#
cweiske or is that question about the publisher notifying the hub about multiple changes at once?
#
kylewm cweiske: yeah -- "content size: 0
#
kylewm "
#
aaronpk publisher notifying hub about multiple URL changes i thought
#
cweiske I support htat
#
cweiske via wildcards
#
cweiske hub.url=http://cweiske.de/tagebuch/*
#
cweiske btw, kylewm - [2015-04-25 18:24:09] phubb.NOTICE: Verification failed {"callback":"http://woodwind.dev/_notify/134","topic":"http://push-tester.cweiske.de/","mode":"subscribe","leaseSeconds":608400,"secret":"5a2fd0de15f04fd1bbe1c49d5e64dae5","reason":"verification response status not 200 but 404","job":"H:ahso2:1057"} []
#
cweiske ah. dot dev
#
aaronpk cweiske: wildcard doesn't make sense to me... does that mean all URLs under /tagebuch/ have been updated?
#
kylewm oh good catch
#
cweiske aaronpk, it does make sense
#
cweiske phubb then checks with subscriptions match that wildcard
#
cweiske and checks all that URLs for changes
#
cweiske and if there are changes, sends them out
#
kylewm cweiske: I just tried re-subscribing
#
kylewm could you check if verification succeeded on your side?
#
cweiske phubb.INFO: Subscription accepted
#
aaronpk hm that seems like it would generate more fetches from the hub to the publisher than needed
#
cweiske whenever my blog's static html gets checked out from git for update, I let phubb check all blog urls I have subscribers for
#
aaronpk also when I post an update I update URLs that don't necessarily match a wildcard
#
cweiske it's a solution for my blog
#
aaronpk https://indiewebcamp.com/how-to-push#Multiple_topic_URLs
#
kylewm I don't think the wildcard is a bad idea; honestly I kind of suspect superfeedr does that under the covers
#
cweiske so the publisher (my blog from git checkout) does not have to care about tracking the changes it sent to phubb already
#
aaronpk oh yeah i'm not saying wildcard is bad, just that it's not sufficient
#
kylewm I was trying to get it just to update a tag-specific feed, and it would still just fetch everything, as far as I could tell
#
cweiske it's sufficient for me, which is why I built it in
#
cweiske kylewm, phubb.INFO: Subscriber notified {"topic":"http://push-tester.cweiske.de/","sub_id":"2","sub_url":"http://reader.kylewm.com/_notify/134","pr_id":"120","job":"H:ahso2:1066"}
#
cweiske still content-length:0?
#
kylewm yes
#
kylewm hmm, it seems like I got a verification request from phubb with no challenge
#
kylewm PuSH verification. feed=<Feed:push-tester.cweiske.de/,http://push-tester.cweiske.de/>, mode=None, topic=None, challenge=None, lease_seconds=None
mlncn and KartikPrabhu joined the channel
#
cweiske this is beginning to get weird
#
cweiske mode=None?
#
cweiske was this just minutes ago?
benwerd joined the channel
#
cweiske 2015-04-28 20:01:12] phubb.INFO: Verifying subscription {"callback":"http://reader.kylewm.com/_notify/134","topic":"http://push-tester.cweiske.de/","mode":"subscribe","leaseSeconds":608400,"secret":"5a2fd0de15f04fd1bbe1c49d5e64dae5","job":"H:ahso2:1063"} []
#
ben_thatmustbeme i can never find this link when i want to pull it up, but according to XFN, it owuld make sense to rel=me to your pgp key
#
ben_thatmustbeme just as rel=me should point to lists of followings, followers, friends list, etc
#
aaronpk wait what really?
#
aaronpk i thought it was for profile URLs
#
cweiske kylewm, phubb accepted your subscription request and the challenge
#
KartikPrabhu rel=me for followings?
#
cweiske so the challenge had to be there
#
kylewm huh, ok maybe it's messed up on my side
#
KartikPrabhu rel=me is supposed to indicate that the linked page is the same person as the current page
#
KartikPrabhu rel=me on pgpkey does not seem to make sense
#
ben_thatmustbeme http://indiewebcamp.com/irc/2014-10-14#t1413323633233
#
kylewm cweiske: is it possible you are double-escaping the url?
#
ben_thatmustbeme found that, thats where we were discussing it
#
kylewm cweiske: I'm seeing GET /_notify/134%22,%22topic%22:%22 href="http://push-tester.cweiske.de/%22,%22mode%22:%22subscribe%22,%22leaseSeconds%22:608400,%22secret%22:%225a2fd0de15f04fd1bbe1c49d5e64dae5%22,%22job%22:%22H:ahso2:1063%22%7D" target="_blank">http://push-tester.cweiske.de/%22,%22mode%22:%22subscribe%22,%22leaseSeconds%22:608400,%22secret%22:%225a2fd0de15f04fd1bbe1c49d5e64dae5%22,%22job%22:%22H:ahso2:1063%22%7D
#
ben_thatmustbeme a pgp key is presumably a signiture representation of you
#
cweiske kylewm, I don't get why "ahso" should be in there
#
cweiske it's my server
#
cweiske it's the gearman job ID
#
cweiske should never be sent to your server
#
kylewm it looks like it's sending part of an escaped JSON blob attached to the end of the notification endpoint url
#
cweiske indeed
#
cweiske kylewm, why is it a GET?
friedcell joined the channel
#
@oSoftDev Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593116176085032960)
#
cweiske or is it the verification?
#
@TheIndieSloth RT @oSoftDev: Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593116221589061633)
#
@DataC5155 RT @oSoftDev: Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593116330594869248)
#
KevinMarks__ Facebook is having a press conference around their API changes
#
@indiegamesdevel RT @oSoftDev: Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593116386009964545)
#
KevinMarks__ http://techcrunch.com/2015/04/28/facebook-api-shut-down/
#
@mob_design RT @oSoftDev: Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593116454473469953)
#
KevinMarks__ Should we send journalists specific questions to ask?
KevinMarks joined the channel
#
cweiske kylewm, is that the notification or the verification?
#
kylewm cweiske: let's assume that JSON weirdness was me fat-fingering the copy-paste key for now
#
cweiske :)
#
kylewm I just tried subscribing again, and everything looks normal https://gist.github.com/kylewm/334975de54b4a24a570e
#
cweiske [2015-04-28 20:17:46] phubb.INFO: Subscription accepted {"job":"H:ahso2:1068"} []
#
@TooStrong_Gamer RT @oSoftDev: Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593117418269999104)
#
cweiske is the content-length:0 still there?
#
kylewm lol, no you know what happened: someone's IRC client (possibly mine) thought that log line you posted earlier was a link and clicked onit
#
aaronpk omg lol
#
kylewm cweiske: yes still content = 0
#
@mobile_way RT @oSoftDev: Last years design and this years design. Serious #Progress in #indiedev #indieweb! https://twitter.com/oSoftDev/status/593116176085032960/photo/1 (twitter.com/_/status/593118086078664704)
#
cweiske could it be that the request is just missing the content length header, but still contains the data?
#
cweiske I temporarily deactivated them some week ago
#
kylewm oh good thought, but no sadly it's reporting the actual len(request.data)
#
cweiske i didn't change phubb for a while now, so probably something changed on your side
#
cweiske updated deps?
#
kylewm OK
#
kylewm I'll look closer later
#
kylewm Flask does do something dumb with request.data
#
cweiske nevertheless, I should send the length
#
cweiske maybe request.data is empty because content-length is not set
#
cweiske so it doesn't parse it
#
cweiske maybe a lib update brought that behavior
#
kylewm oh maybe
#
cweiske I'll fix it here
fkooman joined the channel
#
fkooman aaronpk, how about supporting https://tools.ietf.org/html/draft-ietf-oauth-introspection-08 for access token verification instead of sending the Authorization header bearer token to the token_endpoint?
#
aaronpk ah i've been meaning to catch up on that
#
rhiaro aaronpk: how much work would it be to make the wiki not redirect to homepage after login? (ie redirect back to page you hit login on)
#
fkooman aaronpk, my oauth 2 server already supports that, would be nice for indieauth as well :)
#
aaronpk rhiaro: i don't know... i don't actually know when it stopped doing that
#
aaronpk fkooman: you advertise a separate endpoint for token verification?
#
fkooman aaronpk, yeah
#
rhiaro aaronpk: ah okay
#
aaronpk i guess i don't see why that can't be the same as the token endpoint, since the token endpoint is the thing generating tokens
#
aaronpk i guess it could be the same and still be compatible
#
fkooman aaronpk, yeah, I guess so, no need to create an other endpoint for that
#
aaronpk oh interesting, the introspection endpoint uses the authorization header for the client to authenticate itself
#
cweiske kylewm, please check if you got content-length now
#
fkooman aaronpk, where do you see that?
#
aaronpk https://tools.ietf.org/html/draft-ietf-oauth-introspection-08#section-2.1
#
fkooman aaronpk, the only issue is that the introspection also expects a form post, just like exchanging the code for the access_token
#
aaronpk fkooman: those requests will contain a lot of parameters, none of which are called "token", so the endpoint would still be able to determine if a request is an introspection request vs a grant request
#
fkooman aaronpk, ah, i guess that was added in later drafts
#
ben_thatmustbeme rhiaro: aaronpk: it works for me "sometimes" to redirect to the page i was on
#
fkooman aaronpk, true
#
kylewm cweiske: fascinating, yeah I did get content that time
#
werd.io edited /2015/Edinburgh (+14) "/* Interested */ Declaring interest" (view diff)
#
cweiske only because I now do send the content-length header
#
cweiske but I didn't change anything regarding the POST data itself
#
rhiaro.co.uk edited /2015/Guest_List (+303) "/* Remote Participants */" (view diff)
#
fkooman aaronpk, but it will be a bit tricky for e.g. the micropub endpoint to get credentials to verify the token... you need to manually do that or something
#
cweiske.de edited /How_to_publish_and_consume_PubSubHubbub (+518) "/* Multiple topic URLs */ Wildcards" (view diff)
#
aaronpk there is no client secret for indieauth
#
aaronpk and client id is alwauys the client URL
#
ben_thatmustbeme aaronpk: rhiaro: ahh, interesting, if I'm on https:// on the wiki before login it stays on the page... if i'm on http:// it goes back to home page on login.... losing some post values in an http -> https redirect i'm guessing
#
aaronpk also i think that spec says client authentication is optional
#
rhiaro ben_thatmustbeme: aha!
#
cweiske aaronpk, about &hub.url[]=http://aaronparecki.com/ - the array notation
#
cweiske this is specific to PHP. some programming languages are fine with simply repeating the hub.url parameter
#
fkooman aaronpk, the endpoint MUST
#
fkooman also require some form of authorization to access this endpoint,
#
cweiske it's only PHP which has trouble understanding this
#
aaronpk php and ruby support []
nt0 joined the channel
#
aaronpk i think python too
#
aaronpk superfeedr already supports hub.url[] anyway
#
cweiske but are ruby and python fine with &hub.url=&hub.url=&... ?
#
aaronpk even if they are, the fact that PHP can't understand it is kind of a dealbreaker, since that means any code written in PHP is going to have to use a custom query string parsing library instead of relying on $_POST, at which point you might as well just use JSON input
#
aaronpk fkooman: ok yeah, it says the endpoint must require some sort of authorization from the client, but it doesn't specify that it has to be client id+secret
#
fkooman aaronpk, true, but i guess at least some kind of shared secret with the server or something
#
kylewm cweiske: python is fine with hub.url=&hub.url=&...
#
aaronpk i'll have to think about that more
KevinMarks__ joined the channel
#
voxpelli hub.url=&hub.url= is an anti-pattern imho
#
aaronpk hm, are you intending on using this to allow clients to check tokens, or for your micropub endpoint to check tokens?
#
voxpelli or rather – there has been concerns about the possibility for query params to be both strings and arrays because getting an array when you expect a string can cause exceptions
#
fkooman aaronpk, for micropub endpoint
#
fkooman so only for protected resources that need to verify the access tokens that the client send it
#
aaronpk fkooman: for the micropub endpoint, i think it wouldn't be unreasonable to require people first sign up at tokens.indieauth.com to be able to use it, at which point a shared secret can be established (basically an API key)
#
fkooman aaronpk, yeah, i was thinking the same :)
#
KevinMarks_ it's not an antipattern it'a how checkboxes work in html forms
#
KevinMarks_ except that PHP has had a shitty parser forever
#
cweiske KevinMarks_, the parser is not shitty. it just makes assumptions on the input data that are not always fulfilled
#
cweiske and that assumption (no duplicate keys) is fine most of the time
#
rhiaro.co.uk created /2015/Edinburgh/Guest_List (+5691) "Make Guest List page" (view diff)
#
kylewm I vaguely remember Ruby does some clever trick for []'s?
#
KevinMarks_ passing [] in forms is the clunky workaround
#
aaronpk it's too bad php doesn't parse foo=&foo= according to spec
#
aaronpk but that's where we are
#
KevinMarks_ where's the best summary of bridgy's problems with the new fb api?
#
benwerd (Could potentially write a more standards-compliant polyfill for PHP)
#
aaronpk i'm sure there are tons
yakker joined the channel
#
aaronpk there's one on the parse_str page on php.net
#
aaronpk fkooman: i think i'll go ahead and add this to tokens.indieauth.com
#
fkooman aaronpk, i'll do the same on indiecert.net :)
#
rhiaro.co.uk edited /2015/Edinburgh/Guest_List (+11) "/* IndieWebCamp Edinburgh 2015 */" (view diff)
#
aaronpk i was waiting for the oauth-introspection draft to stablize since it's been in development for a while
#
ben_thatmustbeme aaronpk, I had figured out a hack to correct that behaviour but it wouldn't work when uploading at the same time
#
ben_thatmustbeme which sort of killed that idea
#
aaronpk ben_thatmustbeme: yeah you'd have to write a multipart parser too
#
aaronpk which sounds like the last thing i want to do ever
#
kodfabrik.se edited /link-preview (+262) "Added example of a conflicting third-party guideline for Facebook metadata" (view diff)
#
ben_thatmustbeme waits for forms to natively submit json
#
voxpelli KevinMarks_: true, no anti-pattern in the query params themselves – but even some new parser are handling it very shitty, making it possible to more or less crash them on demand by providing multiple values when they only expected a single one
#
fkooman aaronpk, a multipart parser can't be worse than a mf parser ;)
#
kylewm KevinMarks_: I think that snarfed actually wrote a wiki entry summarizing it, one sec
#
kylewm http://indiewebcamp.com/Facebook#API_Degradation
KartikPrabhu joined the channel
#
aaronpk fkooman: go for it :P http://www.w3.org/Protocols/rfc1341/7_2_Multipart.html
#
kylewm KevinMarks_: and I posted a q to StackOverflow about it http://stackoverflow.com/questions/28337331/facebook-graph-api-find-graph-object-from-post-url
#
voxpelli benwerd: better PHP query param parser: https://github.com/phacility/libphutil/blob/master/src/parser/PhutilQueryStringParser.php
#
aaronpk and good luck handling multi-megabyte uploads ;)
#
voxpelli think we actually used that one at Bloglovin to solve some weirdness in the PHP query param parsing
#
rhiaro.co.uk edited /2015/Edinburgh (+2174) (view diff)
#
aaronpk oh man that's right... i remember having to code $_POST['hub_url'] because it converts . to _
#
voxpelli wrong link, meant: https://secure.phabricator.com/diffusion/PHU/browse/master/src/parser/PhutilQueryStringParser.php$27
#
aaronpk now that register_globals is deprecated, I wonder if a future PHP version will fix the parsing
#
voxpelli aaronpk: looks like the PSR-7 format dodged doing so :/ https://github.com/php-fig/fig-standards/blob/master/proposed/http-message.md They had an oppurtunity there otherwise
eschnou joined the channel
#
tantek is this a form-encoded question?
#
tantek or are we talking hypothetical URL query string attack vectors?
#
@benwerd Discussing IndieWebCamp Edinburgh is super-weird. If you believe in owning your own data, you should go: http://indiewebcamp.com/2015/Edinburgh (twitter.com/_/status/593130931298312192)
#
Loqi [mention] Ben Werdmüller posted 'Discussing IndieWebCamp Edinburgh is super-weird. If you believe in owning your own data, you should go: http://indiewebcamp.com/2015/Edinbu...' linking to http://indiewebcamp.com/2015/Edinburgh (http://werd.io/2015/discussing-indiewebcamp-edinburgh-is-super-weird-if-you-believe-in-owning)
tilgovi and yakker joined the channel
#
rhiaro.co.uk edited /2015/Edinburgh (+134) "Photo" (view diff)
#
aaronpk tantek: talking about foo[]&foo[] vs foo&foo for sending multiple values of the same property name
almereyda joined the channel
#
aaronpk i think php is the only one that doesn't support foo&foo and requires [] for it to work
#
rhiaro benwerd: YAY I did my first repost with a webmention on your post :D
#
Loqi giggles
#
rhiaro (manual webmention)
#
tantek rhiaro++ awesome! add yourself to /repost#IndieWeb_Examples :)
#
Loqi rhiaro has 61 karma
#
rhiaro I've done reposts before, just not of anything that can receive a webmention :)
#
benwerd rhiaro++
#
Loqi rhiaro has 62 karma
#
tantek oh! that's still cool - add it as another example for yourself there - as a repost of an indieweb post!
almereyda joined the channel
#
rhiaro oh wait yeah that actually was my first repost that wasn't a test
#
aaronpk woohoo
#
rhiaro.co.uk edited /repost (+197) "/* rhiaro */" (view diff)
#
rhiaro.co.uk edited /repost (+62) "/* Amy Guy */" (view diff)
interactivist joined the channel
#
cweiske that repost display doesn't help me as a user; http://rhiaro.co.uk/2015/04/1430205825
#
cweiske I still have to click on the original link
#
aaronpk oh hm it's not really a repost unless you display the post you're reposting, right?
#
tantek hey that's a good point
#
tantek rhiaro: ^^^
#
tantek it's a called a rePOST for a reason ;)
#
cweiske currently it's a link
#
tantek similarly it's confusing when a "like" shows the whole thing you're liking
#
tantek ahem aaronpk
#
tantek ;)
#
aaronpk is anybody else confused? http://aaronparecki.com/likes
friedcell joined the channel
#
kylewm I was just enjoying looking at my /likes page yesterday (which shows the full post being liked)
#
aaronpk one of my motivations for implementing likes on my site was so that I'd have a copy of the things I'm liking
#
tantek I kinda want a LOL as a different kind of "like"
#
tantek as in - posts where I involutarily literally laughed out loud upon viewing, e.g. https://twitter.com/gaberivera/status/592521211793444864
#
@gaberivera Is tweeting about your Apple Watch part of the setup process, or are you guys doing that voluntarily? (twitter.com/_/status/592521211793444864)
#
aaronpk tehehe
#
kylewm hahaha
#
tantek class="u-😂-of"
#
kylewm like the Crossfit joke -- "how can you tell if someone has an Apple Watch" "Don't worry they'll tell you"
KartikPrabhu joined the channel
#
tantek I suppose that's still just a reply/comment, even if it's a single emoji
friedcell joined the channel
#
tantek I suppose the usecase is that I'd produce a LOLs feed that people could subscribe to - you know, to put yourself in a better mood.
#
tantek perhaps it's just a hashtag #😂
#
kylewm also -- "like posts" on facebook such as they exist, show the post being liked
#
rhiaro damnit
#
rhiaro goes to fix reposts
#
Loqi lolz
#
rhiaro I'd hoped u-repost-of was enough :p
#
aaronpk well it's enough for your webmention to work :)
j12t joined the channel
#
rhiaro spose it's about time I could consume microformats
#
aaronpk yay!
#
Loqi giggles
#
tantek rhiaro: the presentation is more important than the markup
#
rhiaro i know, i know
snarfed, KevinMarks, wolftune, scor and KevinMarks__ joined the channel
#
KartikPrabhu tantek: Twitter shows Bing translation button on your latest tweet! :P It translates from Indonesian to the same text
#
tantek KartikPrabhu: this is why: https://twitter.com/search?f=realtime&q=tantek
#
KartikPrabhu waaat!
#
KartikPrabhu I guess tantek means something in Indonesian
#
tantek as of maybe a few years ago
#
tantek that's when all the Indonesian tweets started showing up
#
tantek it's likely slang as Google Translate doesn't do anything with it
#
KartikPrabhu aah
#
aaronpk this comes up occasionally here :) http://indiewebcamp.com/irc/2012-12-31/line/1356977209000
#
KartikPrabhu :P
#
tantek lol
tantek joined the channel
#
KevinMarks_ https://twitter.com/kevinmarks/status/590212824397262848
#
@kevinmarks Hey @bing @twitter: when your language detector says it's Romanian, but translating doesn't change it, learn it's not Romanian after all. (twitter.com/_/status/590212824397262848)
#
tantek lol
#
rhiaro welp, I don't know enough about OO PHP to be able to get php-mf2 to work with my hideous mess of characters I mean blog code :/
#
rhiaro I guess I have to learn what Composer is and use it
#
KevinMarks_ I can offer you a python parser
#
KevinMarks_ I suppose unmung could be mf2py as as ervice
#
rhiaro I thought about using pin13 but figured I didn't need more http requests
#
rhiaro maybe it doesn't matter
#
aaronpk rhiaro: barnaby made the php parser easy to load without composer if you want
#
tantek rhiaro: if you figure out how to call phpmf2 without OO-PHP, let me know
#
rhiaro aaronpk: how?
#
aaronpk just include the one file
#
rhiaro I tried that
#
aaronpk what happened?
#
rhiaro but I don't know how to talk to it
#
tantek as in, I just want to included *one* file, then call *one* function
#
snarfed https://kylewm.com/services/mf2 is mf2py as a service
#
rhiaro hang on, I'll go through everything I just tried again and report back
#
rhiaro tantek: yes, that's what I want
#
tantek OO--
#
Loqi OO has -2 karma
#
aaronpk this isn't really OO
#
aaronpk it's just namespaced
#
tantek namespaces--
#
Loqi namespaces has -2 karma
#
rhiaro I dunno how php namespacing works at all
#
KevinMarks_ yes, that's the next thing feedparser guy has done, split it into multiple files
#
rhiaro I've botched it for the Markdown library I'm using, but apparently that works differently to php-mf2
#
aaronpk do you have a string with your HTML?
#
aaronpk $parsed = BarnabyWalters\Mf2\parse($html)
#
loqi.me created /Scare_Quotes (+148) "prompted by ben_thatmustbeme and dfn added by tfontaine" (view diff)
#
aaronpk oh nvm it's just Mf2
#
aaronpk $parsed = \Mf2\parse($html)
#
aaronpk oh he published it on packagist as mf2/mf2 now, usde to be barnabywalters/mf2
#
tantek is there any way to include a simple procedural function wrapper around the OOPHP syntactic vinnegar?
#
rhiaro I get Call to undefined function Mf2\fetch() following the first example on the readme
#
rhiaro I played around with various ways to include it
#
aaronpk rhiaro: what php version are you using first of all?
#
rhiaro 5.5.9
#
aaronpk ok good
#
rhiaro I also have some weird autoload function thing I borrowed when I was getting the Markdown lib to work
#
rhiaro spl_autoload_register(function($class){
#
rhiaro require preg_replace('{\\\\|_(?!.*\\\\)}', DIRECTORY_SEPARATOR, ltrim($class, '\\')).'.php';
#
rhiaro });
#
aaronpk yikes
#
rhiaro http://i3.kym-cdn.com/photos/images/facebook/000/234/765/b7e.jpg
#
aaronpk shouldn't interfere with this anyway
#
kylewm yay!
#
Loqi :D
#
kylewm what is dog science?
#
Loqi https://indiewebcamp.com/dog_science
#
kylewm.com edited /dog_science (+66) (view diff)
#
rhiaro aaronpk: I thought that function might be helping actually
#
rhiaro shrugs
#
aaronpk rhiaro: you have require 'php-mf2/Mf2/Parser.php' somewhere?
#
rhiaro oh I had use \Mf2\Parser;
#
aaronpk oh that would screw things up a bit
#
rhiaro I think I tried with require once though
#
rhiaro hmm
#
rhiaro oh yeah Call to undefined function Mf2\curl_init()
#
aaronpk if you just don't put a "use" in there, you can do it with $parsed = \Mf2\parse()
#
rhiaro in /var/www/html/lib/Mf2/Parser.php on line 66
#
rhiaro ^ that happened when I did that
#
aaronpk oh... that shouldn't happen... let me see if I have php 5.5.9 anywhere
#
rhiaro thanks!
#
aaronpk i have 5.5.23 and it definitely doesn't do that :(
#
rhiaro hm, weird
#
aaronpk oh maybe you don't have curl
#
rhiaro i use curl in my micropub endpoint
#
rhiaro It seems to be looking in the Mf2 namespace for it here though
#
rhiaro oh, actually I might have never tested my micropub endpoint locally
#
rhiaro hold that thought
#
rhiaro yeah, no curl in this docker container
#
aaronpk aha
#
rhiaro waiting for docker container to build
#
aaronpk b 20
#
aaronpk dammit sorry
torrorist joined the channel
#
tantek lol
#
parzzix New to windows, using mIrc now, but wondering what is preferred client for irc?
#
parzzix or considered the best?
#
aaronpk there are some recommendations on http://indiewebcamp.com/IRC
#
parzzix opps wrong forum...lol
#
parzzix sorry
#
rhiaro aaronpk: works now (after installing curl) :) Thanks!
#
aaronpk hooray!
#
rhiaro I guess most people parse external posts once and store contents themselves? Is anyone grabbing posts every time they need to be rendered?
#
rhiaro (and not needing to worry about if it gets edited / deleted, wrt updating own store)
#
rhiaro wait, maybe someone wrote about this on the wiki already
#
aaronpk i have a two-step process, first i fetch and store the HTML, then I parse the HTML and store that as JSON. when my site renders comments, it loads the JSON
#
tantek rhiaro: typically yes webmention sources can be cached
#
tantek and refetched only upon latter webmentions
#
rhiaro considers storing other people's posts as rdf
#
tantek cache the procssed form in whatever works best for your system!
#
rhiaro doesn't feel like writing a mf2 to rdf converter right now
#
tantek then you could cache the JSON parsed result like aaronpk
#
rhiaro I don't expect there would be massive speed advantages to caching, but maybe I'm wrong. But another reason would be to reduce polling of other people's sites? How much does that matter?
#
aaronpk you really don't want to make an http request on each view of your page
#
rhiaro mmhm
#
aaronpk caching the parsed json may be over-optmized, you can probably get away with parsing the HTML each time
#
aaronpk the reason I store the original HTML is so that if I change how I consume HTML, I don't have to re-fetch all the URLs, Ican just update my parser
#
rhiaro makes sense
#
aaronpk i'm up to 400mb of cached HTML now though :)
#
Loqi [mention] http://upon2020.com/blog/2015/04/the-patient-centric-healthcare-experience-i-want/ linked to http://indiewebcamp.com/POSSE (pingback)
#
Loqi [mention] http://upon2020.com/blog/2015/04/the-patient-centric-healthcare-experience-i-want/ linked to https://indiewebcamp.com/own_your_data (pingback)
KitB joined the channel
#
KevinMarks_ tryign out srcset, wonder if anyone cares about 3x or 4x image sizes
#
tantek which browsers are you using to test?
#
KevinMarks_ safari ios/mac, chrome android/mac
#
KevinMarks_ ff says no support on caniuse
#
KevinMarks_ my android phone is 4x
#
KevinMarks_ I was passing w values but safari can't understand them, and fails if I say 64w 2x
#
KevinMarks_ so switchd to just 2x etc
#
KevinMarks_ I don't have a 3x device
#
KevinMarks_ I suppose the way to test is to screenshare my 4x device onto a TV so I can see the pixels
#
tantek KevinMarks: you need one of those Brazil-style giant magnifying glasses
#
KevinMarks_ I do have a USB microscope
#
KevinMarks_ I can just do screenshots, they come through pixle for pixel
#
KevinMarks_ I wonder if I can set my 1920x1080 monitor to be 4x
frzn and lukebrooker joined the channel
#
rhiaro Quietly overlooking that curl warning for now, repost displayed: http://rhiaro.co.uk/2015/04/1430205825
#
aaronpk nice!!
#
rhiaro I'm not caching it yet
#
rhiaro This was just first pass
#
aaronpk it looks like setting safe_mode=Off will fix that warning
#
aaronpk http://stackoverflow.com/questions/6918623/curlopt-followlocation-cannot-be-activated
#
rhiaro yeah I have to find the php ini file on my shared hosting
#
aaronpk or somewhere in your application
#
aaronpk with ini_set
#
rhiaro oh, that's easier
#
rhiaro hmmm the time on my repost is wrong, I wonder where it got that from
KartikPrabhu joined the channel
#
rhiaro and in fact all posts I've sent from Quill in the past couple of days
#
rhiaro goes to poke her micropub endpoint with a stick
#
@SamirTalwar @NicoleRauch Hey, I'm having that issue with socrates-conference.de and IndieAuth again. Do you know if the Passport people patched it yet? (twitter.com/_/status/593189786602569729)
#
tantek random site using IndieAuth?
#
rhiaro.co.uk edited /repost (-63) "repost now display" (view diff)
#
aaronpk via openid i think
#
aaronpk "Failed to verify assertion (message: No valid providers were discovered for the asserted claimed identifier)"
#
aaronpk i have no idea what that means
#
aaronpk openid was so complicated
#
tantek blame mailing lists for making it so
tilgovi_ joined the channel
#
rhiaro quick last question before I go
#
rhiaro when you display someone's photo and name, is that from your store or do you pull that from their h-card dynamically?
#
rhiaro (question aimed at anyone listening)
#
aaronpk i display the photo and name that was in the post
#
rhiaro oh you store them each time
#
aaronpk i should actually be following /authorship better
#
Loqi it is probable
#
rhiaro so you have lots of duplicates?
#
hober tantek: http://w3cmemes.tumblr.com/post/114618598342/
#
aaronpk well i have the full HTML page of each webmention, so the author name is in each one
#
tantek hober, heh
#
aaronpk you call it duplicates, i call it built-in versioning ;)
#
tantek people do change their icons - it's interesting to have them as of the time when they posted
#
rhiaro people also change their names, and it might be insensitive to keep using the old one
#
rhiaro might be useful to have notifications of h-card changes
#
rhiaro aka h-card update posts ...
#
aaronpk people change their display name on twitter all the time as a joke, which might be more appropriate to keep displaying the name at the time of posting
#
rhiaro I would definitely want to err on displaying current name as opposed to name at time of posting
mlncn joined the channel
#
tantek interesting
#
rhiaro given the number of people I know who have (often distressing) problems with people/orgs calling them by their old names, I wouldn't want to increase the amount of old-name-occurences lurking on the web
#
tantek that's a good point!
#
rhiaro even if it's an edge case, I think it's more important than 'what quirky thing was so-and-so calling themselves on that day'
#
KartikPrabhu how would one go about updating old names on all posts/replies to posts?
#
KartikPrabhu same with updating old avatars
#
rhiaro KartikPrabhu: depends how you render it, but apart from that, I have an idea:
#
rhiaro when you update your h-card, you create an edit post
#
tantek rhiaro: you might want a /nicknames-cache
#
rhiaro anyone who is rendering your posts should subscribe to your h-card updates
#
rhiaro and when you change it, the rendering-party can update their output
#
KartikPrabhu hard on static sites I thinks
#
tantek do it lazily via homepage updates
#
rhiaro KartikPrabhu: yeah
#
rhiaro tantek: what do you mean?
#
KartikPrabhu using a /nicknames-cache and updating it occasionally when you receive update notifications or new comments might work. but again dynamic
#
KartikPrabhu as an extension of progressive enhancement it would be nice if indieweb stuff worked with static sites first and then dynamic ones
#
aaronpk static sites can easily regenerate all their pages
#
tantek rhiaro - lookup someone's display name from URL using your /nicknames-cache
#
rhiaro "do it lazily via homepage updates"?
#
KartikPrabhu aaronpk: yes. they would have to regenerate all pages everytime one nickname was changed
#
tantek update the /nicknames-cache whenever make a new post referencing that person
#
tantek via *their* homepage updates
#
KartikPrabhu must run will catch up on this topic
#
rhiaro as in check their homepage every time you need to reference that person?
#
rhiaro that works
#
tantek hence lazily - only when you need to reference them
#
rhiaro but what if they change their name and you don't reference them for a while?
#
KartikPrabhu rhiaro: check homepage whenever you reference them in a new activity. new reference or even if you get a new comment (maybe?)
#
rhiaro you're still displaying their old name and photo
#
tantek rhiaro: that problem will be solved once you integrate a reader into your site
#
rhiaro aaand you don't follow them?
#
tantek and your reader has a PuSH subscription to their home page, and updates the nickname cache accordingly
#
rhiaro anyway, I'm still interested in updates for h-card changes given that silos do that, and commenting on/liking an profile update is also possible on silos
#
tantek rhiaro: it's merely an "h-card" follow, perhaps the lightest form of following
#
rhiaro yeah. You can do a h-card follow without needing a full reader
#
rhiaro but they need to publish their changes as posts still right?
#
rhiaro or do you just diff it?
#
rhiaro in a world of everything-is-a-post...
#
tantek no need - you get an update of their homepage as a URL
#
tantek up to you what to do with that update
#
tantek including re-parsing their h-card
#
rhiaro you just get pushed the url whenever there's an update?
#
tantek that's the simple way
#
tantek it *would* be interesting to explore *actual* h-card subscriptions, and property edits
#
rhiaro sounds like an activity :p
#
tantek there was that long drawn out discussion in socialwg about people vs. profiles and editing profiles
#
tantek that of course, no one has actually implemented in the WG so it was all hand-waving
#
tantek rhiaro: nope. still a post.
#
tantek e.g. on FB
#
rhiaro okay I think we were saying the same thing just describing it differently
#
tantek when you set your profile image, or set your BG image, it makes a *post* out of it, with a permalink
#
rhiaro I *definitely* agree about the post thing
#
tantek nope - the activity thing is an unnecessary abstraction
#
rhiaro I thought you were contradicting me
#
rhiaro which confused me
#
GWG Greetings all
#
tantek only about activity :P
#
rhiaro so you're saying just a really simple post that doesn't say what property has changed... with: u-update-of:homepageurl
#
rhiaro (what's correct markup?)
#
tantek I didn't say that - I said it would be interesting to figure it out
#
rhiaro oh okay
#
tantek and the data so far is that it is "just" a post - per FB
#
rhiaro I'm trying to scope how much figuring has been done
#
rhiaro that's the point of my question
#
aaronpk what is a profile update?
#
tantek we figure things out based on concrete examples
#
Loqi It looks like we don't have a page for "profile update" yet. Would you like to create it? https://indiewebcamp.com/wiki/index.php?action=edit&title=profile+update&summary=prompted+by+aaronpk+https%3A%2F%2Findiewebcamp.com%2Firc%2F2015-04-28%2Fline%2F1430263955435
#
rhiaro and how facebook does it is also the basis for my ideas, yes
#
tantek as opposed to architectural reasoning / philosophizing
#
tantek which unfortunately is how most W3C work seems to happen
#
rhiaro which is why I'm going to implement profile update posts this week
#
tantek philosophy that is, not science
#
tantek rhiaro: want to document profile update research first?
#
tantek before desining / implementing? so you have data to reason from?
#
KevinMarks_ my concrete example is that my facepiles are half faceless as twitter discards old profile pics
#
tantek s/desining/designing
#
Loqi tantek meant to say: before designing / implementing? so you have data to reason from?
#
aaronpk KevinMarks_: lol mine too
#
tantek you can cache the old images themselves instead of old image URLs
#
tantek that's just bad caching :P
#
aaronpk yeah :(
#
Loqi aww, cheer up
#
aaronpk i keep getting stuck on how to implement that in a sane way
#
tantek what is indiearchive?
#
KevinMarks_ right, I should cache old twitter images on my static site, not the company worth billions
#
Loqi IndieArchive is a project to collaboratively grow an archival copy of pages replied to (possibly also mentioned) in indie web posts https://indiewebcamp.com/IndieArchive
#
tantek no KevinMarks_ you should cache them in free space on gihutb per /IndieArchive :P
#
rhiaro tantek: yes, will document. This conversation *is* beginngin research (which means it's a bit frustrating when you say "implement!" and I say I will and you say "reaearch!", or at least that's what' I'm getting)
#
tantek s/gihutb/github
#
Loqi tantek meant to say: no KevinMarks_ you should cache them in free space on github per /IndieArchive :P
#
KevinMarks_ or I suppose webmention.herokuapp.com should
#
KevinMarks_ or bridgy
#
tantek rhiaro: the implementation I mentioned did not require any new research - you took the leap there to profile updates of specific properties
#
tantek I say implement! and reference existing work e.g. /nicknames-cache
#
aaronpk tantek: lol in 2 years my archive of html grew from 63mb to 415mb, which isn't too bad!
#
GWG I never figured out how to do a nicknames cache
#
KevinMarks_ hm, hangon, somoen fixed things - http://www.kevinmarks.com/twitterhatespeech.html now has faces again
#
KevinMarks_ or was it ben? https://twitter.com/benward/status/561686435552854016
#
@benward @kevinmarks Though the idea that an avatar at the moment of a Tweet should be considered a part of that Tweet is also valid and interesting. (twitter.com/_/status/561686435552854016)
#
KevinMarks_ thank you middleware people for making my site better, whoever it was
#
tantek rhiaro: nonetheless you drew new insights into what can/should be cached and what should be updated
#
aaronpk tantek: one problem i've encountered with storing files on disk that match URL paths is this (and this is what's keeping me from implementing indiearchive right now)
#
aaronpk when someone has two URLs such as aaronparecki.com/tag and aaronparecki.com/tag/indieweb, I can't save them on disk because the first expects "tag" to be a file, and the other expects "tag" to be a folder
#
tantek I feel like this is the kind of thing someone would have solved
#
aaronpk i had to solve it for spiderpig but it is not an ideal solution for indiearchive
#
aaronpk i force every web page to be called index.html and be served from a folder that ends in /
#
aaronpk so on disk, i would actually have aaronparecki.com/tag/index.html and aaronparecki.com/tag/indieweb/index.html
#
aaronpk and the web server ends up sending redirects from /tag to /tag/ and such
#
aaronpk which works fine when you're intending on flattening a dynamic site into HTML as a static archive of it
parzzix joined the channel
#
tantek what is a nicknames cache?
#
Loqi A nicknames cache is a way indieweb sites store information about people to improve the user experience of the site owner referring, mention, and/or linking to those people https://indiewebcamp.com/nicknames-cache
#
tantek.com edited /nicknames-cache (+438) "/* Better Person Icon Display */ note that showing the icon as of the time of a post is also valid, add tweet citations from benward and kevinmarks" (view diff)
parzzix joined the channel