#indiewebcamp 2014-04-07

2014-04-07 UTC
#
@jtroyer @cincystorage @mikestanley @discoposse @BlueShiftBlog @halr9000 Some interesting distributed projects going on under the #indieweb banner.. (twitter.com/_/status/452959951724351488)
#
Loqi [mention] http://kylewm.com/reply/2014/04/07/1 linked to http://indiewebcamp.com/irc/2014-04-06/line/1396822084 (webmention)
#
GWG kylewm: Now I feel bad. I was just kidding.
#
kylewm lol, I know. I do think it's hilariously passive aggressive to send webmentions to IRC like I just did tho
#
GWG Would you like some pita chips?
snarfed joined the channel
#
snarfed thinking about a new project: a webmention handler for hosted blog providers like tumblr, blogger, and wordpress.com
#
snarfed debating whether it should be part of bridgy or separate
#
snarfed thoughts?
#
GWG Well, wordpress allows plugins, doesn't it?
#
GWG .com, not .org
#
snarfed afaik wp.com only allows approved plugins
#
snarfed and only a few. and they might not even describe them as plugins, just features
#
kylewm snarfed, something different than webmention.io and webmention.herokuapp.com?
#
snarfed kylewm: yeah. this would accept webmentions and use the site's API to post them as comments
#
snarfed not sure whether it makes sense as a bridgy feature, or should be separate
#
snarfed (also not sure if i should do it since it's not an itch for me personally, but meh)
#
kylewm it does seem bridgy-adjacent at least :)
#
snarfed yeah
#
kylewm and it would mean tumblr users could get comments from twitter and facebook...
#
snarfed yup
#
kylewm (i think?)
#
kylewm cool!
#
kylewm fwiw, I think my head is in the same place as yours ... i'm kinda at a temporary stopping point working on my site, want to bring the message to the masses now!
#
snarfed great!
#
snarfed i'd be happy to collaborate if you want
#
snarfed the other candidate for next project was to polish http://freedom.io/ , maybe integrate it into bridgy, and advertise it more widely
#
snarfed (it migrates silo posts to a blog platform as individual posts, w/formatting, comments, likes, etc)
#
kylewm oh! is that an older project of yours? hadnt seen it before
#
snarfed yeah, older
#
snarfed got the functionality done but got stuck at the design phase
#
snarfed lack of interest or skill :P
#
kylewm hehe, yeah sometimes it's no fun once the technical challenge is over, i can identify with that
#
snarfed any idea what you want to tackle next?
#
kylewm I was thinking about that a lot today, I think I'm interested in the gen 2 stuff... how do we make it easy enough for computer savvy people to set this up
#
kylewm I dont' know if that is via wordpress or jekyll or something like that
#
snarfed sure!
#
snarfed similar
#
snarfed i'd guess more tumblr, blogger, wp.com than jekyll, but who knows
#
kylewm ah cool, yeah good point
#
kylewm I guess i got a little discouraged thinking about it, you don't think it's jumping the gun to try and make it easy for lots more people to get involved?
#
@CaptainKurtis RT @mattl: @CaptainKurtis @GNUsocial @indiewebcamp I think it does. We use @creativecommons attribution license for posts. Plus federation … (twitter.com/_/status/452973869653966848)
#
kylewm like, e.g., without a Reader or much of a mechanism for dealing with spam ironed out?
#
snarfed eh, opinions differ, but i think it's fine to start building tools early at least
#
kylewm seems like there's a relatively small network right now that makes fast iteration possible
#
kylewm things will start to firm up quick with more people, i'd guess
#
snarfed yup, true, but that will take a while
#
kylewm haha yes
#
snarfed i doubt a new service like this would make a big change fast
#
snarfed (and it'd be a great problem if it did)
#
kylewm ah yeah, that's totally true too... cool, i'm glad to hear your opinion at least, was curious about that philosophical question earlier
#
snarfed hrm. disappointing. wp.com's api lets you create comments with any author, but blogger's requires it to be the authed user, and tumblr doesn't have comments at all, just disqus or indieweb-style replies that have to be posted as full posts on another tumblr blog first
#
snarfed https://developer.wordpress.com/docs/api/1/post/sites/%24site/posts/%24post_ID/replies/new/
#
snarfed https://developers.google.com/blogger/docs/2.0/developers_guide_protocol#CreatingComments
#
kylewm ack, that's too bad
#
kylewm i'm guessing the approval process for wp.com (e.g., to get semantic-linkbacks approved) is very onerous?
#
kylewm yeah looks like a very small handful of approved ones
#
snarfed good question. not sure but i expect so. i doubt you can even apply.
#
snarfed step 1: get a job at automattic :P
#
kylewm sounds about right
#
kylewm I ran a b2 blog before it was forked, that should count for something!
#
snarfed oh man, seriously
#
snarfed should still be doable for all three though, even if the formatting might not be ideal
#
kylewm i guess that is one thing i think is interesting about jekyll -- pretty much no limit on what you can do as long as its static
#
snarfed probably via disqus for tumblr
#
snarfed true! …which is already a pretty big constraint. and not it's probably ready for gen 2 yet. but regardless, jekyll is nice
#
kylewm oh right on, I didn't realize you meant they literally had disqus comments
#
GWG I dropped Disqus a while back on a site in favor of Jetpack comments, which is just a superset of normal Wordpress comments
#
snarfed GWG: yup
#
snarfed hmm. disqus *might* allow synthetic author. will take more investigation. http://disqus.com/api/docs/posts/create/
#
GWG Now I have Webmentions
#
GWG But the issue is that status updates, such as like, don't work with Wordpress comments.
#
GWG Wondering if I should look into the ability of the comment system to support it
#
GWG Jetpack, their extension, has a like system
#
snarfed GWG: when you say they don't "work with WordPress comments," you mean your status updates on your WP blog don't accept comments? or other people's don't show up as comments on your posts like they should?
#
GWG A like isn't a comment
#
GWG A +1 isn't a comment
#
GWG And, I may turn on Jetpack Likes
#
GWG People are too lazy sometimes to say anything
#
GWG They just want to be acknowledged
#
GWG Twitter 'favorites'/retweets, Facebook likes, Google +1s, etc
#
kylewm yeah I know that was an issue dangillmor had, that his comment feed was flooded with likes/retweets, not collected nicely like in e.g., p3k
#
kylewm speaking of good problems to have
#
GWG kylewm: There must be a way to condense t
#
GWG And put that into a plugin that can be added to a theme.
#
kylewm I just faved your post from earlier on twitter, have you tried that yet with backfeeding?
#
kylewm Ithink it will just show up as "kyle_wm liked this --via twitter.com"
#
snarfed GWG: sure! it's all just code. pfefferle and acegiak happily accept contributions to the semantic-linkbacks plugin
#
kylewm my boss once said "it's all just lists of numbers"
#
kylewm unintentionally profound
#
GWG I may disable Jetpack comments though
#
GWG That requires a link to Jetpack, which means overloading is harder
#
GWG Well, a a horrible coder, I may suggest things, but my implementation is sloppy and embarrassng
#
kylewm snarfed, to more concisely answer your question from earlier, I have no idea what I want to work on next :) Going to try to come to HWC this week and maybe come up with something
#
snarfed kylewm: sounds good!
#
GWG snarfed: I seem to remember someone writing code that linked into the Genesis framework to create a bunch of headshot profiles for likes and such
#
snarfed yeah, those were willnorris's contributions to semantic-linkbacks
#
snarfed (i think to pfefferle's fork)
#
GWG I'm wondering if I looked at it, if I could add to the documentation to make it a few lines to add to any theme
#
snarfed give it a shot!
#
GWG I want to contribute
#
GWG And, to be honest, documentation is sparse on those plugins
#
snarfed definitely
#
GWG I can always push commits to a readme, right?
#
snarfed sure!
#
GWG I may just start with a post
#
kylewm this? https://willnorris.com/2014/03/display-likes-in-a-facepile
#
GWG That's the one
#
GWG He noted it needs to be de-genesized
#
GWG Things like this should be in the webmentions plugin
#
kylewm cool, bet that would make more than a couple people happy
#
GWG I have to learn about the Wordpress comment system a bit more
#
GWG If you want to go crazy on it, someone could write a comment system based on indieauth
#
kylewm is going to make dinner, there will be no photos ;)
#
snarfed i dunno. if you believe "pics or it didn't happen," you might starve
#
GWG I just wonder if a proposal is an annoyance or a way of trying to help
#
snarfed GWG: proposals are good! working code is even better, but they both count
#
GWG snarfed: I do my research
#
GWG Anything I think of would be thought out, at least
#
ben_thatmustbeme trying to catch up on the logs is like a part time job
#
ben_thatmustbeme jeez
#
GWG Hi, ben_thatmustbeme
#
ben_thatmustbeme hey
#
ben_thatmustbeme cool, library science huh?
#
ben_thatmustbeme could have used to pick your brain at my last job GWG
#
GWG ben_thatmustbeme: Never used it
#
GWG When they spelled my name wrong on the degree, it was probably a bad sign
#
ben_thatmustbeme ouch
#
ben_thatmustbeme my brother in law's name is Brian... they spelled half of the Brian's his year as Brain
#
ben_thatmustbeme a lot of people were mad about that one
#
GWG They probably didn't permaplaque it though
#
GWG ben_thatmustbeme: What was your job that you could have used an LIS person?
#
ben_thatmustbeme I was actually the only tech guy at a small company that did digitizing of art and historic works
#
ben_thatmustbeme so logs of cataloging things came in to it
#
ben_thatmustbeme honestly, I learned enough of it that I was trying to teach the owner of the company the stuff, which he really should have learned on his own
#
GWG ben_thatmustbeme: When I was in school, they renamed cataloguing to Knowledge Management.
#
ben_thatmustbeme heh, that works
#
GWG I originally studied Archives and Records because I wanted to go into digitizing of historical records.
#
GWG Didn't work out
#
GWG I'm in a corporate setting, doing training.
#
GWG ben_thatmustbeme: Is that Niagara Falls on your site?
#
ben_thatmustbeme yeah, it is
#
ben_thatmustbeme I took that panorama when i was there last year
#
ben_thatmustbeme paid for by selling my twitter handler
#
ben_thatmustbeme i have a lot of work to do on my site graphically
#
ben_thatmustbeme I want to get the code behind it working first
#
ben_thatmustbeme I don't like the look but it works for now
lukebrooker joined the channel
#
GWG ben_thatmustbeme: I did a bunch to make it look nicer, may do more later.
#
ben_thatmustbeme I'm working on implementing webmention now
#
ben_thatmustbeme I have quite a bit to do
catsup joined the channel
#
GWG ben_thatmustbeme: I used Wordpress. You are building at a lower level.
#
GWG Yours is harder
#
GWG But as I said earlier, don't build a beautiful house and forget to put furniture in it
#
ben_thatmustbeme my problem is i build a wonderfully furnished shack
#
ben_thatmustbeme or rather, wonderfully furnished dilapidated house
#
GWG That is the poin t
#
GWG Balance
#
ben_thatmustbeme Hmm, anyone have any idea how to handle sites if they have more than one rel=webmention ?
KevinMarks joined the channel
#
GWG Depends
#
ben_thatmustbeme presumably I should be linking to a single post, so I just grab the first
#
ben_thatmustbeme or it would be a link to the entire site
#
ben_thatmustbeme so really there should only ever be one
#
@robmyers RT @mattl: @CaptainKurtis @GNUsocial @indiewebcamp I think it does. We use @creativecommons attribution license for posts. Plus federation … (twitter.com/_/status/452997961979396097)
grantmacken and bnvk joined the channel
#
@lisaradon indieweb (twitter.com/_/status/453002531849789440)
kbs joined the channel
#
ben_thatmustbeme hey kbs
#
kbs hi ben_thatmustbeme - how goes your sunday hacking?
#
ben_thatmustbeme not bad
#
ben_thatmustbeme haven't had a ton of time for it
#
ben_thatmustbeme but my webmention is getting there
#
kbs aha, nice :-)
#
ben_thatmustbeme I think i am pretty close to being able to send them
#
ben_thatmustbeme which is odd considering i was intending on making it receive them first
friedcell joined the channel
#
ben_thatmustbeme kbs, what are you writing your crawler for?
#
kbs webmention is a pretty nifty idea - so obviously simple like all great ideas
#
kbs ben_thatmustbeme: well I have some hopes of using it to layer on a private messaging layer
#
ben_thatmustbeme i was working to refine the message queue thing, have some work to do on it
#
ben_thatmustbeme oh, that would be cool
#
kbs roughly speaking
#
kbs I plug in someones profile url - and it downloads their public key locally to me. I then push a webmention pointing to an encrypted message to them - something like that.
#
ben_thatmustbeme I was wondering if we could use it to piece together some contacts list automatically
#
ben_thatmustbeme so lets say, for all my facebook friends (or some list i specify) go to all those sites, grab their data, find their sites, parse those, then perhaps do the same for G+
#
ben_thatmustbeme and Twitter
#
ben_thatmustbeme and associate them together wherever possible (two accounts both use a rel=me to the same site for example
#
ben_thatmustbeme even if it isn't indieweb
#
kbs *nod* yea, indeed that sounds like a pretty neat idea. We should ask KevinMarks for pointers to existing projects as well
#
kbs given that he's Mr. social web graph
#
ben_thatmustbeme i really dislike having my contacts all in Gmail. Don't get me wrong, it can be helpful, but then you always hit issues merging contacts on the phone between accounts, or if they have multiple accounts, it gets confused.
#
ben_thatmustbeme plus it won't have a space for a FB profile, or some other site profile
#
kbs yea, absolutely - I see what you mean
#
kbs as you say, rather quite by accident this rel=me crawling turned out to be a rather cute way to find/merge identities
#
kbs and it does seem like it might come in handy in other contexts
#
ben_thatmustbeme well I figure it could all work in together, if I can share my personal info to those that log in via indie web, whould be cool to be able to work that in
#
kbs aah, nice :)
#
kbs [fwiw, my current code is duct tape and chewing gum and it's in java - I believe you're implementing your site in php, right?]
#
ben_thatmustbeme yes
#
ben_thatmustbeme i have it up in github
#
ben_thatmustbeme i like the layout, I use it at work, you don't need to know the entire system layout to be able to really start hacking on it. I want to make it a super easy platform for people to just get a blog up and start hacking on it
#
ben_thatmustbeme well, easy for hackers to set up
#
kbs nice - fun fun :)
#
kbs Were you thinking about running your contact crawling offline, and importing the data into different places
#
kbs or more something that would just hang off openblog in some fashion?
#
ben_thatmustbeme not sure, i suppose it could work either way
#
ben_thatmustbeme i would love to have it plug in to the blog, so it could occasionally (monthly?) re-run to check for new info
#
kbs gotcha
#
ben_thatmustbeme and also have a callback URL for others to send me updates on the fly
#
ben_thatmustbeme just pushed the basic scheme for sending webmentions to github
#
ben_thatmustbeme completely untested
#
kbs cool :)
#
ben_thatmustbeme using php-mf2
#
kbs I like your contact-list crawling idea - have a feeling KevinMarks might have some pointers about a php-friendly library to use for rel=me crawling at least. [my php skills are non-existent]
#
ben_thatmustbeme php-mf2 is just that. makes it super easy
#
ben_thatmustbeme Its a project for down the way
#
kbs ah, cool.
#
ben_thatmustbeme I wonder how much you can browse with the API. Would be able to get all the account info for people who don't set everything public
#
ben_thatmustbeme would be the same issue with every site
#
kbs do you mean like a site-specific API (as opposed to scraping html and parsing for mf2 and other data?)
#
ben_thatmustbeme yeah, there is a trade off
snarfed joined the channel
#
ben_thatmustbeme scraping html would be great, more generic, but you are really limited to only things people put public
#
ben_thatmustbeme you have to be logged in to get their phone number and such
#
kbs *nod* - I see what you mean. You're thinking that once someone logs into your site with their [say] g+ account, you can get more info at that point about them
#
ben_thatmustbeme indeed
#
ben_thatmustbeme if I haven't already parsed their info from my own accounts
#
kbs yep - makes sense
#
ben_thatmustbeme so i start with my account and get all my G+, FB, & twitter contacts, and from all the links they have there try to parse those sites too
#
ben_thatmustbeme get all the info i can
#
ben_thatmustbeme then tie in info i enter manually
#
ben_thatmustbeme or perhaps dump from Gmail, etc
#
ben_thatmustbeme anyway, I'm going to head to bed
#
ben_thatmustbeme goodnight all
#
kbs it's a nifty idea :) and good night
#
Loqi don't let the bed bugs bite
#
Loqi gute nacht!
#
ben_thatmustbeme talk to you tomorrow
KevinMarks joined the channel
#
kbs idly wonders whether using data:some/type as the source in webmentions is a terrible idea
cweiske joined the channel
#
kbs probably is, I think :)
snarfed, friedcell, paulcp, lukebrooker, KartikPrabhu, eschnou, basal and tahnok joined the channel
#
jden indiewebians: would love a review of weblogin https://github.com/jden/web-login
KartikPrabhu1, lukebrooker, friedcell, KartikPrabhu and Jihaisse joined the channel
#
Jihaisse Hello
cweiske, LauraJ, jsilvestre and glennjones joined the channel
#
@bech @infonauten IndieWeb FTW. ;-) @altjoen (twitter.com/_/status/453068289074528256)
grantmacken, pfenwick, LauraJ, eschnou, krendil, pasevin, tobiastom, bnvk, cweiske, icco, adactio, friedcell1, barnabywalters, CheckDavid, basal, squeakytoy, scor, melvster and ttepasse joined the channel
#
tommorris.org edited /monoculture (+80) "/* Examples */ user-agent string is the other way WebKit monoculture exists" (view diff)
pfefferle and chloeweil joined the channel
#
Loqi chloeweil: bnvk left you a message on 4/1 at 7:07am: I met your friend Cirilia here in Iceland yesterday :)
BjornW and pfefferle joined the channel
#
bnvk heh
brianloveswords, eternicode, Sebastien-L, gRegor`, Jihaisse, jedahan, snarfed, pfefferle and ben_thatmustbeme joined the channel
#
ben_thatmustbeme hello all
#
@justnathan @laurasolomon ask if anyone is using #indieweb standards. Seems like a good open standard for libs to get involved in. (twitter.com/_/status/453189626690568193)
iangreenleaf joined the channel
#
gRegor` Morning, ben_thatmustbeme
#
ben_thatmustbeme Good morning
#
barnabywalters good afternoon
#
ben_thatmustbeme what is everyone up to today?
npdoty joined the channel
#
aaronpk getting impossible to actually catch up on these logs
#
ben_thatmustbeme ignore 90% of what I say
#
ben_thatmustbeme I did see you working more on the webmention queue stuff
#
aaronpk yep! it's implemented on my site now!
#
ben_thatmustbeme cool
#
aaronpk (the polling version, not the callback version yet)
#
ben_thatmustbeme I don't know if you saw my stab at updating the standard as I was planning to implement it
#
aaronpk on the wiki?
#
ben_thatmustbeme i have a link
#
ben_thatmustbeme i just forked it in github
#
ben_thatmustbeme https://github.com/dissolve/webmention
#
aaronpk oh!
#
aaronpk didn't notice
#
ben_thatmustbeme really I just changed around to try to say we use the same return codes for async as for sync
#
ben_thatmustbeme just encapsulate them differently
#
aaronpk ah cool updated from the wiki notes
#
ben_thatmustbeme I don't know what to do about "pending moderation"
#
aaronpk the conclusion we came to the other day was "nothing"
#
aaronpk or rather, "nothing yet"
benprew joined the channel
#
ben_thatmustbeme we'd still want something to let them know rejected vs failed parsing i'd imagine
#
ben_thatmustbeme I don't know, I'll have to look back at in when I have a few minutes
kbs joined the channel
#
ben_thatmustbeme good morning kbs
#
kbs hi ben_thatmustbeme
#
kbs gRegor`: you were quite right about _external_ rel=me s to the bare-domain being perfectly fine - there's some logs back at http://indiewebcamp.com/irc/2014-04-05/line/1396763505 from tantek, and more info in those logs if you're still curious...
#
gRegor` Interesting
#
gRegor` And perhaps more confusing. Haha
#
gRegor` But easy enough for me to add.
#
kbs :-) I shall avoid blathering about rel=me stuff - have spammed the logs enough I think.
#
barnabywalters kbs: you should distill what you’ve learned about rel=me onto http://indiewebcamp.com/rel-me
#
kbs I should probably have a large disclaimer section if I do write anything :)
#
barnabywalters kbs: it’s a wiki, don’t worry about it :)
#
Loqi yea
#
barnabywalters the important thing is for it to be on there rather than only in logs
#
kbs ah, I see what you mean.
#
barnabywalters which, despite being excellently marked up, are still quite difficult to search
#
barnabywalters also, the act of trying to express knowledge concisely can really help solidify and clarify it in your mind
#
kbs generally has difficulty displacing the solid lumps already ideated in his head :)
#
kbs I'll write something up atleast per my understanding - as you say, at least helpful for me if no-one else
#
barnabywalters kbs: that would be great!
#
ben_thatmustbeme kbs, not as bad as me writing my regex to turn my sentences in to pig latin
#
ben_thatmustbeme that was really getting spammy
#
kbs I think though, much more entertainin :)
#
ben_thatmustbeme actually, how can you search the irc logs?
#
ben_thatmustbeme i know http://indiewebcamp.com/irc/<<date>> but can you search by keyword?
#
barnabywalters ben_thatmustbeme: search google for site:http://indiewebcamp.com/irc/logs keyword
#
barnabywalters ah, yeah, site:http://indiewebcamp.com/irc keyword actually
#
ben_thatmustbeme oh yeah, duh, that works
#
ben_thatmustbeme Well, that helped me find the old regex. That was probably a bad idea honestly.
#
ben_thatmustbeme s/\b([^aeiou ]*)(\w+)\b/\2\1ay/gi
#
Loqi ben_thatmustbeme meant to say: ellWay, atthay elpedhay emay indfay ethay olday egexray. atThay asway obablypray aay adbay ideaay onestlyhay.
#
ben_thatmustbeme I'm really anxious to get webmention up and working on my site. Its slow going.
#
kbs haha. :)
#
barnabywalters ben_thatmustbeme: what are you having problems with?
#
ben_thatmustbeme oh, nothing yet, i think I'm just taking my time to do it right. I still have to figure out how to make sure a link to my own site is valid safely, and then ensuring the remote location points correctly
#
ben_thatmustbeme I do wonder how I will display posts that just link to my home page and not a specific permalink
#
barnabywalters ben_thatmustbeme: a few of us have generic mentions feeds which list all mentions to any page, e.g. https://aaronparecki.com/mentions or http://waterpigs.co.uk/mentions
#
ben_thatmustbeme I think I have the sending side almost done though, just need to parse out any links and actually send them
#
barnabywalters which are essentially analogous to the twitter interactions view, which shows all interactions with you
#
aaronpk those /mentions pages will be how the mention app sends you notifications :)
#
barnabywalters aaronpk: yay!
#
Loqi :D
#
barnabywalters it’s also how taproot knows to send client-side notifications
#
aaronpk but I didn't have enough time to get to that this weekend, and implemented my webmention status URLs instead :P
#
barnabywalters aaronpk: oh cool — did you end up caching or storing the status?
#
aaronpk caching for 24 hours
#
barnabywalters that seems sane
#
aaronpk example from bret https://aaronparecki.com/webmention/rf2_14yA
#
barnabywalters nice
#
aaronpk yes, since it will include everything even the mentions that are totally bogus like the youtube -> nonexistent wiki page one I got yesterday
#
aaronpk althought it occurs to me that I should store aggregate counts of things like that, so I know in general how many valid/invalid mentions I receive
#
barnabywalters that page is a wonderful example of how debuggin HTML+HTTP APIs can be a much better experience than, for example, JSON ones
#
aaronpk yes!
#
barnabywalters it’s a pity we can’t link to it ;)
#
aaronpk i can make that one permanent
#
aaronpk ish
#
barnabywalters it would be cool to have at least one permanent one for demonstration purposes
#
aaronpk how bout a screenshot for the wiki
#
barnabywalters that’ll do
#
aaronpk ok I removed the expiration from this one too
#
aaronpk course now it says "This page will be visible for the next -0 hours."
#
ben_thatmustbeme hah
#
barnabywalters ha ha
paulcp joined the channel
#
barnabywalters -0 is my favourite number
#
ben_thatmustbeme I'm thinking I'm going to store the queue item as part of the comment
#
ben_thatmustbeme so they persist
#
ben_thatmustbeme any that don't exist were rejected
#
aaronpk ben_thatmustbeme: what about ones that don't result in a comment? (e.g. invalid mentions)
#
ben_thatmustbeme altough you lose the invalid vs moderator rejected
#
ben_thatmustbeme I might persist moderator rejected ones
#
ben_thatmustbeme unless spam
#
aaronpk what if I send you source=http://youtube.com&target=http://ben.thatmustbe.me/
#
ben_thatmustbeme i would poll youtube find it doesn't have a link back and it would be rejected as invalid and thus deleted
#
aaronpk how can I check the status of that
snarfed joined the channel
#
ben_thatmustbeme when it creates it, it creates and entry, as a comment, thats set to not display, with no data, when I try to fill in the data, it sees invalid, and deletes the database entry
#
aaronpk but the whole point is I need a URL to check to see what happened with the mention
#
ben_thatmustbeme yeah, it would return a URL
#
aaronpk right, so where are you storing the result of the webmention attempt so that I can retrieve it at that URL?
#
ben_thatmustbeme you send me source=.... I return a URL for you to poll from. queue?id=12345
squeakytoy joined the channel
#
ben_thatmustbeme in the DB, this happens to just be in the same table as comments
#
kylewm.com edited /User:Kylewm.com (+283) "/* Todo */" (view diff)
gRegor` joined the channel
#
aaronpk ok, so you'll persist even the junk ones in the DB?
#
ben_thatmustbeme no
#
aaronpk then how can I check the status of it?
#
snarfed kylewm: saw you got yet another FB 500 from bridgy publish
#
snarfed sorry :( seems so consistent, i must be doing something wrong
#
ben_thatmustbeme after it gets deleted?
#
snarfed i'll investigate eventually
#
aaronpk ben_thatmustbeme: let me walk through this
#
aaronpk I send you a webmention (manually using curl, since i'm developing my site)
#
ben_thatmustbeme okay, i return a URL
#
aaronpk the source URL I send from turns out doesn't actually have a link to your site because of a bug on my site
#
ben_thatmustbeme okay
#
ben_thatmustbeme at first, the record is created. If you check that status, it will say pending
#
aaronpk I open your URL a few seconds later because I copy+pasted it from my terminal
#
aaronpk by the time I open it, you've already processed and rejected the mention
#
aaronpk I need to see the result of you having processed and rejected it with the "no link exists" error
#
aaronpk does that make sense?
#
ben_thatmustbeme mmm, see thats what i was just looking at, It would return a generic "failed" with no reason why
#
ben_thatmustbeme so technically any call you make to guess a random ID, would return the same "failed"
#
waterpigs.co.uk edited /webmention (+451) "/* Implementation Notes */ clarified how to verify target, source link to target" (view diff)
#
aaronpk ah yeah, that defeats much of the purpose
#
ben_thatmustbeme which is why i'm rethinking it actually
#
waterpigs.co.uk edited /webmention (-2) "/* verifying source links to target */" (view diff)
#
ben_thatmustbeme I think i might just store permanantly all of them
#
aaronparecki.com uploaded /Special:Log/upload "uploaded "[[File:aaronparecki-webmention-status-example.png]]""
#
ben_thatmustbeme its really like 3 URLs, and a status
#
ben_thatmustbeme and maybe a few IDs
#
aaronparecki.com edited /webmention (+259) "/* Asynchronous status polling */" (view diff)
#
ben_thatmustbeme 3 varchars, and a few INTs
#
ben_thatmustbeme not a huge database burden
#
barnabywalters ben_thatmustbeme: I added some clarifications to the webmention implementation notes about the things you were asking about earlier: http://indiewebcamp.com/webmention#Implementation_Notes
#
aaronpk cool
#
barnabywalters hopefully that makes things a little clearer
#
gRegor` Nice webmention status page, aaronpk
#
aaronpk thanks!
#
ben_thatmustbeme thanks barnabywalters
#
aaronpk html+http apis ftw!
#
ben_thatmustbeme okay, headed to lunch
#
gRegor` Now that I have the MySQL datetime fixed up properly, I should have my mentions page up soon.
#
snarfed.org edited /Events (+426) "update" (view diff)
#
gRegor` I went to the trouble to store webmention published/updated dates in UTC, but then found out the mysql UTC time was way off, haha
#
aaronpk yay more database fun
#
Loqi woot
#
gRegor` I still like it. :)
#
snarfed reads logs
#
gRegor` I was wondering why the twitter avatars weren't showing up for me on your status, aaronpk, then remembered I'm using the disconnect.me extension
#
aaronpk heh
#
aaronpk I still need to keep local copies of everyone's avatars
#
gRegor` I was just thinking about that yesterday
#
aaronpk I plan to do something like http://aaronparecki.com/avatar/waterpigs.co.uk.jpg
#
gRegor` Will you keep them indefinitely, or periodically refresh them?
#
aaronpk I'll probably replace them with updated versions as people update theirs
paulcp joined the channel
#
kylewm.com edited /User:Kylewm.com (+519) "trying to work out some ideas of what to work on next" (view diff)
#
barnabywalters aaronpk: what are your thoughts on extending that to content photos in reply contexts?
#
barnabywalters AFAIK we both archive HTML of post pages, but nothing else
#
aaronpk oh yeah, probably will do that once I handle photos in reply contexts better
#
barnabywalters which from an archival POV doesn’t do much for photo posts
#
aaronpk I basically want as much of my site to stand alone as possible, loading very little from external sites
#
aaronpk a good test of this is when I'm working on my site on a plane with no internet access
#
snarfed kylewm: love this idea on http://indiewebcamp.com/User:Kylewm.com : "extract webmention sending as a separate service that could watch an h-feed and send mentions for [links in] new entries"
#
kylewm snarfed: no hurry/worries re: Facebook 500 errors. I tried it again, was just curious whether it was a transient error from them
#
aaronpk totally
#
kylewm oh, thanks!
#
snarfed (i'm guessing you meant to include that "links in" there)
#
kylewm it's very much in the Bridgy mold
#
kylewm yes
#
snarfed agreed!
#
kylewm.com edited /User:Kylewm.com (+9) "/* other ideas/goals */" (view diff)
silencematters joined the channel
#
silencematters Hello all
#
snarfed hi silencematters
#
silencematters I am helping to Organize the IndieWebCamp here in NYC for later this month http://indiewebcamp.com/2014/NYC
#
silencematters We are looking for companies to help sponsor the event — to pay for food, and some of the space
KevinMarks joined the channel
#
aaronparecki.com edited /webmention.io (+125) "add sites using webmention.io" (view diff)
#
aaronparecki.com edited /User:Aaronparecki.com (-19) "/* projects */ fix link" (view diff)
#
kbs aaronpk: quick question about indieauth.com - are rel=me links from silo providers to pages "underneath" the baredomain (eg: mysite.com/about) 'supposed' to work? [I know it currently looks for an exact match for the identity url
#
kbs so I can login as mysite.com/about but not mysite.com]
#
kbs (I have rel=me links to the silo from the bare-domain)
#
aaronpk hm I suppose if your /about linked to / with rel=me then it could work
#
aaronpk in theory, that is not implemented
#
kbs so, tantek alleges that rel=me from /about to / is always implied :)
#
aaronpk oh?
#
kbs let me look up the log...
#
aaronpk I was not aware
#
aaronpk I think that is probably true for indie sites, but not for multiple-author sites or things like github.com
#
kbs that's true - but this is in the context of auth
#
aaronpk eesh...i'm not sure. i probably shouldn't try to give a half-thought-out answer
#
ben_thatmustbeme i thought the idea it to always link rel=me to the /, that is your identifier and the only way you are supposed to log in. Using no subdirectory was the easy way to say you own the domain
#
ben_thatmustbeme well always for those sites that identify you
#
kbs http://indiewebcamp.com/irc/2014-04-05/line/1396763868
#
ben_thatmustbeme i saw that before when I was reading through the logs this weekend
#
aaronpk I don't see it mentinoed here http://microformats.org/wiki/rel-me
#
kbs I think I best understand rel=me as "rel=is-controlled-by", in the context of authorization
cweiske and chloeweil joined the channel
#
kbs and the essential idea (from what I gathered from that conversation) is that, for authorization, every subpage is implicitly is-controlled-by the bare-domain (eg: github.com/user is rel=is-controlled-by github.com) etc.
#
aaronpk oh interesting
#
aaronpk "is-controlled-by" makes that make much more sense
#
ben_thatmustbeme that seems like a good way to look at it
#
aaronpk then yes obviously /about is controlled by /
#
kbs cool :) I'll continue to work on my little essay here then, to muddy the waters even more
#
aaronpk oboy
#
ben_thatmustbeme haha
#
kbs blames barnabywalters
caseorganic joined the channel
#
barnabywalters happily accepts blame for new wiki content :)
fungoat joined the channel
#
ben_thatmustbeme aaronpk, the more i think about it, you might have the right idea for only showing webmention queue items for so long
#
ben_thatmustbeme I'm thinking 7 days after moderator approval
#
aaronpk ah yeah if your'e doing moderator approval that's a good reference point
#
KevinMarks That also helps explain how we get ambiguity of two people can add rel me links to a page
#
kbs thinks aaronpk 's current indieauth implemention 'does-the-right-thing' in pretty much all situations
#
kbs don't touch it :)
#
aaronpk heh, it's close anyway
#
ben_thatmustbeme people both having rel=me links to a page could really screw up the idea I had for creating the contact lists
#
aaronpk could use better error reporting in many cases, but it mostly does the right thing :)
#
ben_thatmustbeme s/both/two
#
Loqi ben_thatmustbeme meant to say: people two having rel=me links to a page could really screw up the idea I had for creating the contact lists
scottros joined the channel
#
ben_thatmustbeme i give up
#
kbs I did discover some entertaining login urls (at least entertaining for me, that is :) it handles self-loops quite nicely
#
kbs http://indiewebcamp.com/wiki/index.php?title=Special:ListUsers&limit=500 and look for plus.google.com ...
#
aaronpk oh my
#
kbs :-)
#
kbs self-linked profile, etc...
#
kbs I have a feeling that logins to indiewebcamp.com implicitly expect baredomains
#
kbs but I kinda would love to have the system also work for myfamily.com/tommy etc
#
kbs (which indieauth does of course, quite nicely.)
#
ben_thatmustbeme kbs, i had the idea of that but just using subdomains
#
ben_thatmustbeme ben.thatmustbe.me instead of just thatmustbe.me
grantmacken joined the channel
#
kbs ben_thatmustbeme: *nod* yea - that would work nicely
#
@erinjo @jtroyer I'm working on group convo ideas for #indieweb with @teamidno. If you have suggestions, I'd love to hear them (erin @ eidyia . com) (twitter.com/_/status/453235515622313984)
tantek joined the channel
#
Loqi tantek: bupkes left you a message 1 day, 8 hours ago: thanks :)
#
tantek :)
#
tantek dietrich - how does Wednesday night look for hosting Homebrew Website Club @MozPDX?
#
aaronpk can host at esripdx (last week at our current office!)
#
tantek if it's any trouble - we can ask aaronpk to try to host at esripdx
#
tantek ah, he out-typed me :)
#
bret i have class on wed till 6:30 :(
#
tantek SF indiewebfolks (snarfed, hanni, hober, iboxifoo, jjuran, KevinMarks, mlinksva) - I have confirmed we have MozSF for 2014-04-09 Homebrew Website Club!
#
bret aaronpk did I hear you are moving?
#
bret esri that is
#
snarfed tantek: glad to hear it! fingers crossed, i might bring new people this week
#
tantek.com edited /events/2014-04-09-homebrew-website-club (+47) "note local time (yay Chicago!) and RSVP" (view diff)
#
KevinMarks Excellent, but I am going to miss it as Rosie labs at the airport on Wednesday
#
tantek.com edited /events/2014-04-09-homebrew-website-club (+329) "need indieweb event and POSSE copies, move my RSVP to actual RSVP section :)" (view diff)
#
tantek snarfed - great! Please RSVP for yourself and the friends you're bringing :) http://indiewebcamp.com/events/2014-04-09-homebrew-website-club#RSVP
#
ben_thatmustbeme hmm, not to parse out the correct data from a webmention
#
ben_thatmustbeme I can't wait for IWC Boston
#
dietrich tantek: it looks good. i even have the sekrit password for the room mics this time.
#
tantek dietrich aweseom!!
#
tantek s/aweseom/awesome
#
Loqi tantek meant to say: dietrich awesome!!
#
dietrich i will also even file the right paperwork to have the doors open for everyone.
#
tantek sweet!
#
tantek dietrich, I recommend we try to setup the Vidyo around 18:00
#
tantek would be great to let folks in that show up anytime after 18:00 - that's how we try to run it here @MozSF
#
tantek gives regulars a chance to catch-up and get settled in before the actual meeting start
#
dietrich tantek: yep, sounds good
#
tantek (I reserve the room in Zimbra from 18:00-20:00 just to give us some buffer)
#
kylewm.com edited /User:Kylewm.com (+54) "/* personal site */" (view diff)
KartikPrabhu and gavinc joined the channel
#
kbsriram.com edited /rel-me (+2821) "notes on rel=me and relmeauth" (view diff)
#
@CaptainKurtis Can @indiewebcamp sites operate effectively on @torproject hidden services? If so, is anyone doing it? #indiewebcamp (twitter.com/_/status/453242753766289408)
#
ben_thatmustbeme anyone have a link handy to a page that has actual webmention content. I'm looking to set up parsing of a page for the exact h-entry with in-reply-to in it
#
gRegor` ben_thatmustbeme: Like http://aaronparecki.com/notes/2014/04/06/2/siri ?
_6a68, scottros, fmarier and paulcp joined the channel
#
ben.thatmustbe.me edited /2014/Cambridge/Guest_List (+285) "/* Creators */" (view diff)
Kopfstein and KevinMarks joined the channel
#
ben_thatmustbeme gRegor`, that has a reply that from FB, but its not a reply itself
#
ben_thatmustbeme specifically i'm looking for a URL like someone would send me if they mentioned me
#
aaronpk ben_thatmustbeme: what do you mean "actual webmention content"?
#
ben_thatmustbeme so i'm searching for in-reply-to in the html right now
#
ben_thatmustbeme I'm setting up the piece of code that actually fetches the content from a remote site
#
ben_thatmustbeme or should i just be looking for any link back to my site?
#
aaronpk you probably should read this http://indiewebcamp.com/comments-presentation#How_to_display
#
aaronpk and if you're in PHP, then you can just use my implementation of it https://github.com/indieweb/php-comments
#
ben_thatmustbeme I see
#
aaronpk here you go: http://aaron.pk/r4VT1
#
ben_thatmustbeme thanks, aaronpk, thats just the link i needed
#
ben_thatmustbeme I skipped over that page assuming it was a discussion of UI
#
aaronpk ah interesting
#
ben_thatmustbeme given the name "comments presentation"
#
aaronpk yes I think that algorithm may deserve its own page given the number of times I keep linking to that header
#
gRegor` ben_thatmustbeme: To verify webmentions, I search for the target URL anywhere in the HTML.
#
gRegor` To determine whether it is a "reply" or a "mention", I check the in-reply-to
#
gRegor` If no in-reply-to, it's a mention.
#
gRegor` (I currently don't parse likes, rsvps, etc. But soon!)
#
aaronpk wow the Moves API has gotten much more complete since I last looked
#
ben.thatmustbe.me edited /2014/Cambridge/Guest_List (+0) "/* Venue Capacity: 20 */" (view diff)
#
aaronpk considering using it as an input to create new posts on my site from that data
#
aaronpk https://dev.moves-app.com/docs/api_storyline
#
ben_thatmustbeme i think of it as parsing, presentation being display on my site, thats probably why it gets confusing
#
ben_thatmustbeme aaronpk, I was already using your php_mf2, whats one more of your projects pulled in as a submodule :P
#
aaronpk php-mf2 is mostly barnaby actually :)
#
aaronpk I think I made one contribution to it
KevinMarks_ joined the channel
#
ben_thatmustbeme I was originally trying to find a page that I could run php-mf2 against so I could see exactly how it outputs the in-reply-to data, but the examples on here should be enough for me to work with.
#
ben_thatmustbeme didn't realize IWC cambridge had a limit of 20 people. So I signed up now before there was nothing left
#
aaronpk great!
#
kylewm.com edited /events/2014-04-09-homebrew-website-club (+33) "/* RSVP */" (view diff)
#
aaronparecki.com edited /events/2014-04-09-homebrew-website-club (+80) "/* RSVP */" (view diff)
eschnou, npdoty, tantek, yaf and chloeweil joined the channel
#
snarfed.org edited /events/2014-04-09-homebrew-website-club (+36) "/* RSVP */" (view diff)
KartikPrabhu, jacus_ and wagle_ joined the channel
#
hanni tantek: graet!
#
hanni also great
dietrich_, jtzl, acegiak_ and scottros joined the channel
#
ben_thatmustbeme so aaronpk, you just have to pull out the first hentry and basically thats it, just pass it to your magic code and it takes care of the full parse?
#
aaronpk yep, given an h-entry the magic code will return something sane
#
ben_thatmustbeme hmm, it doesn't seem to check for the presence of the URL that i give it, if its not of type=reply
BjornW joined the channel
#
ben_thatmustbeme which is fine, i can do that with my own code. And I'll assume the first h-entry in the page is what i'm looking for, since it should be a permalink
#
aaronpk that sounds right, yes
LauraJ joined the channel
#
ben_thatmustbeme okay, what does it return if it is a reply but the URL doesn't match?
shaneknysh joined the channel
#
aaronpk I believe it will only be marked as a reply if the URL is present
#
aaronpk otherwise it shows up as a "mention"
#
aaronpk yeah you'll get type=mention, but if there is an explicit in-reply-to property that contains your URL it will be type=reply
#
shaneknysh greetings I just finished listening to in Beta #90 and I can't wait to get started on my own part of the indie web
#
aaronpk shaneknysh: welcome!
#
ben_thatmustbeme so I should do my own checking to make sure the URL is present first
#
aaronpk ben_thatmustbeme: yes, this code doesn't help with that case at all
#
ben_thatmustbeme okay, thank you, that helps a lot
#
aaronpk this assumes you want to display something for a given h-entry, but it may not even be used to format a comment
basal joined the channel
#
tantek aaronpk, ben_thatmustbeme - for any given "is there an example of markup of xyz?" go to the /xyz page
#
tantek and look for a section named "Examples" or "IndieWeb Examples" for permalinks to real world examples of xyz
#
tantek which you can then use to check your parsing/consuming code against
#
ben_thatmustbeme thanks tantek
papa-away joined the channel
#
tantek ben_thatmustbeme: no problem! and if you don't find an answer by following that technique, then ask someone if there are any real world examples of xyz (and mention the specific page)
#
ben_thatmustbeme every reply i was finding was a twitter post reply just displayed on a page, and none that were actually marked up with in-reply-to a specific URL
#
ben_thatmustbeme but I found some now
krendil joined the channel
#
tantek ben_thatmustbeme: the /reply page is suboptimal in that regard, doing some minor edits to improve it
#
tantek also there's a longstanding task to split /comment into /reply and /comments
#
tantek and since you're (relatively) new here, it would be great to get your fresh perspective on that
#
ben_thatmustbeme certainly
#
tantek short version:
#
tantek a *reply* is what you call the thing you host at your own site, at its own permalink, that is in-reply-to a post (or article etc.) somewhere else
#
tantek a *comment* is what you call the presentation of a reply underneath / adjacent to the original, on the permalink of the original, perhaps copied there automatically via a webmention from the reply (which was then parsed for h-entry microformats2 etc.)
#
tantek that's my current working distinction, and I'm curious what you think about it
#
tantek.com edited /rsvp (+36) "consistent use of "IndieWeb Examples" as a heading" (view diff)
grantmacken joined the channel
#
ben_thatmustbeme actually i just jumped back in to here after thinking about it, basically thats what I got to. A reply is a post that is just referencing offsite, vs a comment is something added on to a post (or a reply)
#
ben_thatmustbeme I can see where it gets confusing though, they switch depending on which side you view it from
#
tantek ben_thatmustbeme: ok that's great to hear that that's where you got to.
#
ben_thatmustbeme a reply can become a comment on someone else's site, and a comment may have been someone else's reply
#
tantek precisely!
#
tantek and yes, the confusion is certainly easily possible, and hopefully it's something we can reduce by improving the documentation
#
ben_thatmustbeme agreed
#
tantek I've been iterating on the distinction for a while without making it fully explicit across all the wiki pages, for exactly that reason
#
ben_thatmustbeme There are so many bits and pieces I'm having to learn, its slow going.
#
tantek one of our goals is to make each individual bit and piece fairly easy to learn, (and support/implement) so that everyone can make incremental progress along whatever direction/focus is their preference (scratching their own itches)
#
ben_thatmustbeme Yes, Thats somewhat why I'm trying to just power through a basic site setup. I want to get a framework that I find easy to hack on whatever piece I want. But there is a bit of groundwork I need to get first
#
tantek.com edited /comment (+455) "use subhead IndieWeb Examples for consistency, subsections for author, iterate on reply vs. comment distinction" (view diff)
tantek joined the channel
#
tantek wow that page still needs a lot of work
friedcell and scottros joined the channel
#
aaronparecki.com edited /2012/Planning (+81) (view diff)
barnabywalters, KartikPrabhu and kbs joined the channel
#
bear if you all maintain servers and that server has openssl - check your versions… http://heartbleed.com
#
kbs yikes - CVE-2014-0160 [for any ssl....]
#
kbs ^^
_6a68 and tantek joined the channel
#
barnabywalters bear: thanks for the tip-off, updating server and local machine now
#
bear yea, it's a nasty one
#
bear it looks like ubuntu just pushed their fix
#
bear (i'm not following debian or redhat...)
#
barnabywalters hrm, homebrew seems to have the new version but I’m not seeing it when updating debian
#
barnabywalters although apache, openssh client and server have both been upgraded
#
kbs need to replace keys too, ugh
#
kbs and don't think enabling PFS necessarily prevents decryption of previously captured material
#
barnabywalters doesn’t even know how to replace relevant keys
#
kbs cf https://twitter.com/ivanristic/status/453280081897467905
#
@ivanristic @MacLemon It depends on what’s in the memory block. Leaked ticket key would compromise all sessions it signed. (twitter.com/_/status/453280081897467905)
#
gRegor` Whoa. Holy crap on that OpenSSL bug
snarfed joined the channel
#
kbs thinks the general strategy ought to be make-it-secure-without-ssl first
#
kbs then add ssl
ttepasse joined the channel
#
gRegor` How do you mean?
#
kbs my personal 0.02 - ssl is not really a panecea for security
#
kbs and sometimes tends to obscure other more real issues
#
kbs eg: if someone wants to actually communicate securely, tehre's really not much else other than encrypting end-to-end -- which doesn't need ssl (though it'd add another layer)
#
kbs also, any given web site is often easier to get into via any server code (say, a wordpress plugin) than worrying about ssl, etc
#
bear ssl is just one piece/layer of the security cake
#
kbs yea
scottros, ttepasse and bnvk joined the channel
#
pdurbin lots of sites went ssl after firesheep
#
kbs yea. The login/password/session-cookie mechanisms are pretty hacky
#
kbs I'm rather sad that something like SRP etc didn't really take hold early enough in the mosaic/netscape heydays. [https://en.wikipedia.org/wiki/Secure_Remote_Password_protocol]
#
kbs As bear succinctly points out :) it's just one piece of the system
#
kbs and it probably behooves us to not rely on it exclusively, if at all possible...
tilgovi, pfenwick and paulcp joined the channel
#
pdurbin and raid is not a backup :)
#
kbs :-)
benprew and Jefff joined the channel
#
gRegor` kbs: Sure, TLS is just a piece, but it's a pretty important piece for basic things like user authentication. It shouldn't be a "later" item in that regard.
pauloppenheim and lukebrooker joined the channel
#
kbs is a cynical old fart, apologies for inflicting my neuroses on the young :)
#
gRegor` I thought I was the one being more cynical about PGP the other day. :)
#
gRegor` Not about PGP itself, but about key/fingerprint verification. :)
#
kbs eg: user-authentication coming from (say) just utilizing https on the server is not something personally I'm a big fan of... :)
#
kbs oh - I figure that [like all interesting things] PGP puts the level of paranoia directly in your hands
#
barnabywalters heads-up: php-mf2 v0.2.6 just released with fix for multiple-property nested microformats, and a new JSON-mode to get around a PHP quirk
#
barnabywalters cc aaronpk — you should upgrade and enable JSON mode on pin13
#
gRegor` So, are you a fan of sending usernames and passwords in the clear, then? :)
#
barnabywalters and thanks for reporting that nested multiple property bug :)
#
kbs gRegor`: ah, you presume that that's the only way to authenticate yourself :) [cf https://en.wikipedia.org/wiki/Secure_Remote_Password_protocol et al]
#
gRegor` Who implements it / how easy is it? Easier than PGP, I hope, if we want adoption. :)
#
kbs nobody - just making a rather pedantic point I'm afraid... I have some faint hopes that with the emergence of apps rather than browser-based interfaces to the web, things have a chance of picking up
#
gRegor` Does look interesting
#
gRegor` Though I give up on trying to follow the protocol section, haha. :)
#
kbs it's pretty old
#
kbs but I'm very hopeful of the sorts of things Moxie is doing [eg: textsecure and other products]
#
kbs (ie, more sophisticated protocols + more usable interfaces than something geeks in the 70s cooked up)
#
kbs http://www.thoughtcrime.org/software.html - (textsecure especially) I think is really nicely done
#
aaronpk barnabywalters: I'm about to install the new php-mf2
#
aaronpk but first, why is items an array but rels not? http://pin13.net/mf2/?url=pin13.net
#
kylewm aaronpk: mf2py does the same thing http://kylewm.com/api/mf2?url=http://pin13.net
#
aaronpk oh right cause rels is normally a hash
#
barnabywalters aaronpk: yep, https://github.com/indieweb/php-mf2/issues/29
#
aaronpk so wait what is the problem barnabywalters fixed?
#
barnabywalters it’s a long-standing bug which tommorris pointed out and I fixed, but it turned out my fix broke it
#
aaronpk oh the other way around
#
barnabywalters in a different place
#
aaronpk but... the current version is correctly returning an object for rels right now
#
barnabywalters aaronpk: it seems that that is/was not always the case
#
barnabywalters this fix ensures that it will always be serialsed correctly
scottros joined the channel
#
aaronpk don't you just want the JSON_FORCE_OBJECT flag on json_encode?
#
aaronpk "Outputs an object rather than an array when a non-associative array is used. Especially useful when the recipient of the output is expecting an object and the array is empty."
#
barnabywalters aaronpk: sure, if you want empty “items” keys to be objects :)
#
aaronpk wait whaaa
#
aaronpk how can you have empty keys?
#
barnabywalters ah, sorry
#
barnabywalters if there are not microformats on the page, then the value under the “items” key will be an empty array
#
aaronpk that's correct
#
barnabywalters huh, actually it is
#
aaronpk oh.. the force object flag turns that into an object?
#
barnabywalters er… yes
#
aaronpk let me check
#
barnabywalters I believe that is the case
#
aaronpk yeah it does... seems like that could be fixed that way tho
#
aaronpk hm weird
#
aaronpk stupid php
#
barnabywalters indeed
#
barnabywalters at least json_encode($value) makes way more sense than json.dumps(value)
#
aaronpk do you have tests for the empty items / rels keys?
#
barnabywalters aaronpk: yep
#
aaronpk ok so wait, I'm still unclear what the current issue is. it seems like it's behaving correctly
#
aaronpk like, http://pin13.net/mf2/?url=pin13.net is correctly returning an empty array for items and an empty object for rels
#
barnabywalters aaronpk: on version 0.2.5?
#
barnabywalters or is it an earlier version?
#
aaronpk i think it's on 0.2.4
#
barnabywalters as I said, I did fix this bug earlier by just hard-coding a change to a new stdClass if the rels array was empty
#
barnabywalters but that cause issues when looping through it
#
barnabywalters hence the creation of JSON-mode
#
barnabywalters aaronpk: ah yep, 0.2.5 is where the bad fix was made
#
barnabywalters aaronpk: or, fixed rather
#
barnabywalters yep. < 0.2.5 bad, 0.2.5 fixed but with no built-in way to correctly serialise JSON, 0.2.6 has JSON mode
#
aaronpk ok I'm super confused
#
aaronpk so in 0.2.6 now, the default behaviour will return an empty array for both items and rels?
#
barnabywalters aaronpk: yes
#
aaronpk but that breaks the json serialization
#
barnabywalters aaronpk: yes
#
aaronpk ok then I am not actually confused
#
aaronpk did you add the json mode flag to the function version too? the example only shows it on the constructor
#
barnabywalters aaronpk: no, out of wanting to keep the function interface as clean and simple as possible
#
aaronpk ok! pin13.net is updated
#
barnabywalters most people are using php-mf2 for parsing rather than JSON serialisation, and it’s specialised enough to warrant only exposing it via the object interface
#
barnabywalters aaronpk: yay! thanks
#
Loqi giggles
#
barnabywalters aaronpk: I made a quick reference to this issue https://github.com/indieweb/php-mf2/issues/29#issuecomment-39797490
#
barnabywalters goodnight all
#
Loqi goodnight!