#snarfed…on to the next ruby breakage. :/ bundle exec … says "Could not find rake-10.4.0 in any of the sources; Run `bundle install` to install missing gems."
#snarfedbundle install and bundle check both ran happily. all of the libs are installed in vendor/bundle/ruby/2.0.0/ , which seems odd, since i'm using ruby 2.1.6 (checked with ruby -v)
#snarfedi'm starting to feel like god is trying to tell me something
Kongaloosh, KevinMarks_, KartikPrabhu, tantek and j12t joined the channel
#GWGsnarfed: I fixed it and sent it back to you a few days ago
#Loqislack/snarfed: GWG: oh! thanks, I'll look. i don't automatically follow your fork and get notified of your commits, so you'll want to comment in the PR when you do that
#acegiak"nono, when you like/repost/bookmark a thing it should look like this, if you want to use these fields you can't because they're not the right fields for that post kind"
#GWGI was thinking of the specialized fields for things like audio, where I'd want to possibly add fields.
#GWGFor example, on a piece of music, I might want things like Genre and Release Date
#GWGben_thatmustbeme: I actually like audio metadata. Of course, I don't post that very often. I was trying to get to RSVPs, which have a field only relevant for RSVPs.
#acegiakGWG: Just do what you need for now but make your structure easily expandable
#GWGTrying to dictate too many restrictions is a problem because people want to use things they way they want to use them.
#GWGTrying to have no restrictions is harder for newcomers.
#GWGacegiak: I've been rewriting the meta storage stuff because I was working on the Micropub plugin. I want to be set up to store more microformats2 properties in the meta, even if they aren't used right now.
#GWGIt came out of the Micropub design discussions. Micropub can store any arbitrary data, even if the receiving end doesn't know how to use it.
#GWGSo, aaronpk, for example, has Teacup, which does food and drink. He uses the non-standard tags p3k-food and p3k-drink. Even if I don't know what to do with those, I want to keep them.
#GWGThat was on the Micropub side. On the Post Kind side, I want to have the functions work with all mf2_ prefixed values, and understands the established ones.
KevinMarks_, eschnou and KevinMarks__ joined the channel
#kylewm!tell KartikPrabhu no you don't (Bridgy doesn't use submodules anymore!) basically you can install deps into a virtualenv and include the virtualenv folder when you deploy to app engine. there are some tricks to setting up the path though http://stackoverflow.com/a/25564125/682648
#JarOfGreenSo, my hackday plan is adding the ability to import events on a webpage that are marked up with h-event or schema.org markup to OpenACalendar .... first step, a good HTML parsing lib for PHP, and taking the data to some solid models ....
#rhiaroGood morning everyone! Anyone remote-hacking today, and want a stream / to be streamed?
#moredhelmorning all, I'm looking at getting ghost to support microformat2 markup. does anyone know of any themes that do this already? I have looked at this already: https://indiewebcamp.com/Ghost
#tbrbaaronpk: do you have a way to sign into quill with a dev copy of a site? ie, I'm running a dev server at localhost:8080 rather than harryreeder.co.uk?
catsup joined the channel
#aaronpktbrb: i haven't tried that before, but you should be able to use a ssh tunnel to give your localhost a public address
#atomiculesanm: The code in that post is already out of date, but the approach is still the same
j12t joined the channel
#aaronparecki.comcreated /ngrok (+318) "Created page with "{{stub}} '''<dfn>Ngrok</dfn>''' is a utility for making local servers available on a public address https://ngrok.com/. Ngrok is useful for testing development websites with pu..."" (view diff)
#JarOfGreen"Ohhhhh, that's how it works" just heard from next door ....
#tbrbit worked in the sense of it getting my providers fine, but then google throws a wobbly because primary.indieauth.com isn't a registered callback handler
#rhiaroif you ask for text/html it redirects to rhiaro.co.uk/year/month/slug which is html version with comments etc
#aaronpktbrb: if your first segment is year then that's a pretty strict pattern to match, [0-9]{4}
#csarvenI'd ditch the dates altogether. Come up with a proper mechanism to generate unique paths, which includes handling scenarios where you come up with the same slug for some reason. e.g., if 'foo' exists, go with 'foo-n' or whatever. There is a trade off: how complicated do you want the URL to be and how likely are you going ot run into collisions. Work around the collision possibilities (which is far less likely than introducing date
#csarvens and other things upfront and all the time)
#rhiaroHaving two urls for everything is a problem. People are going to reply/like/etc (ie send webmentions to) the /date/ version, because that's the most obvious html version, even though I'd technically want stuff at the non-date version. But I have to accept that.
#rhiarothe ids for posts in my store are the non-date version
#anmI think date makes sense in URLs for content where date is actually relevent, but I am now thinking that for many blog articles, it is not, and really I should have a post type which does not have dates.
#rhiaroSo when I get webmentions to the date version, I have to map that to store the replies
#voxpelliI believe I will get a problem with URL:s when I move to implement update/delete in my Micropub endpoint as Jekyll requires full date in file names I believe but I only have month+year in URL
#voxpellibut that's kind of an edge case – doing Jekyll + full dynamic IndieWeb makes for some edge cases
catsup joined the channel
#anmalso there is a yerba mate going around if anyone is interested
j12t joined the channel
#csarvenInteresting that /url-design talks about yyyy/mm/dd pattern, instead of yyyy-mm-dd pattern which is actually the ISO 8601 pattern.
#aaronpkit's not a far stretch to use slashes in URLs to separate date segments. it's still in yyyy-mm-dd order which is the important part
#_franI was just cursing against EXIF dates, which are in yyyy:mm:dd format :)
#rhiarousing slashes makes yyyy and yyyy/mm collection pages obviously human-discoverable
#JarOfGreenadvantage of using slashes in URLs is then you have a natural page that lists all content from one year, or one month I guess
catsup joined the channel
#csarven- and / are merely strings. One of which in the date pattern is an ISO date pattern, the other is not. So, sure, we can chalk that also up as a personal preference.
#JarOfGreenOk, so my PHP lib to parse events from HTML with both schema.org and h-events syntax is going well. With unit tests! Title, Start and End date done. https://github.com/JMB-Technology-Limited/HTMLIsAnEvent Anyone have some example web pages with marked up events I could test it against? (And they don't mind me saving the HTML in my repo so I have stable tests)
#JarOfGreenvoxpelli: thanks, may pick those up, might try and grab lots of real world examples tho. So I've used someone else's PHP HTML parser, then I'm just selecting tags out of the result. But there is still a big variance in possible options, even in what I've already seen
#aaronpkoh man people are trying to sign in to teacup with their email address
#rhiaroterminalpixel_ just told me he's dumping the POST request and not getting anything at all
#csarven.caedited /URL_design (+50) "Study on Persistent URIs with identification of best practices and recommendations on the topic for the Member States and the European Commission" (view diff)
#rhiaroand recently teacup mysteriously stopped working for me when I didn't think I'd made any changes to my endpoint
#7YUAACSOSrhiaro: Does your server accept access_token in the form body?
#anmrhiaro: maybe, it depends on the person probably.
#anmatomicules: Is all your blog code available or just the syndication bit? I'm also working out how to do notes using jekyll collections, wondered if you did that.
#7YUAACSOSbut when I try to use this token with tokens.indieauth.com it says that error=unauthorized&error_description=The+token+provided+was+malformed
#7YUAACSOSgot everything else working, just the micropub server was rejecting me
#tantek.comedited /URL_design (+189) "move URL template to brainstorming (not a common practice), and add warning of why to avoid "literal" ISO8601 date format without path separators in a URL" (view diff)
#csarventantek I can show you million URLs using ISO8601 date pattern if that qualifies for IWC? Or is the max set to 100?
#csarventantek YOu have got to be kidding me. PROVE "not a common practice"
#csarventantek "Avoid the "literal" ISO 8601 date patterns of use /YYYY-MM-DD/ or /YYYYMMDD/ because by omitting path separators they are are less URL friendly (less hackable). " -- that's complete BS
#tantekcsarven - there are tons of URLs with long strings of numbers, database ids also - and they're also wrong
#tantekwhich as nothing to do with such 1:1 arguments
#csarvenYour "friendlier" argument is entirel ybased on preference.
#aaronpk7YUAACSOS: the code is a one-time use thing, only used to get the token
#aaronpkwonders how best to document this path of making a micropub client
ttepasse joined the channel
#7YUAACSOSwhy bother with the verify step on the code given that it will be checked when exchanging for a token?
#aaronpknow that I think about it, all the wiki pages are written from the perspective of writing micropub servers, would you agree 7YUAACSOS?
#aaronpkthe verify step on the code is between the token endpoint and the authorization endpoint, in case they are separate systems like indieauth.com and tokens.indieauth.com
#aaronpk7YUAACSOS: that page is just about login/identification, so there is no access token in that step
#csarvenThose are arguably more approriate reading than some blog post
#csarvenThe W3C notes/recs go through a lot of people befor epublishign.. whereas a blog post is someones POV
#tantekcsarven: just because I crticize something does not imply allergy. I organize things to help others be quickly productive instead of getting lost of W3C/academic esoterica (have plenty of experience with that)
#csarvenwhich you happen to "prefer" over what the W3C/RFC/IETF... prefers..
#aaronpkcsarven: the point is that in the real world, people have come to associate the "/" character with separate components of a URL, expecting to find content by removing components
#ben_thatmustbemethats not a rec, its a note from an IG which usually doesn't go through the same level of work
#tantekcsarven: lol no. "W3C notes/recs go through a lot of people" the majority who don't actually publish their own websites or practice what they talk - so thus aren't really qualified.
#csarvenT, I see that your gang is all ready to go at me :) No worries, I will go through the responses one at a time.
#tantekcsarven - focus on being productive instead of pushing academic theoreticals / edge cases.
#csarventantek THAT is only your measure of qualification. Not the rest of the world.
catsup joined the channel
#tantekcsarven: rest of the world, shipping counts.
#csarvenWould like me to show you URLs in the millions that I'm publishing which use ISO8601 or is that above the IWC bar (100 URLs?)?
#ben_thatmustbemecsarven: my main complaint is that you linked to a spec in a list of readable, real-world experience, and blatently moved it to the top to try to give it more authority
#rhiaroIf the URL design page on the wiki is specifically about blog post/article URLs, rather than all URLs (which I think it is, as there are plenty of other domains where the stuff on that page doesn't apply, eg. long identifiers are useful) then that should be made clearer on the wiki page
#voxpellicsarven: if one has the URL of "YYYY/MM/DD" and something is relative to the month, then one just does "../" – because its hierarchical – as pointed out by the URI RFC if you want an RFC :)
#csarvenben_thatmustbeme Don't even bother to wave TimBL's Cool URIs doc with that justification. Like I said earlier, everything stems from that. Read his doc several times, embrace it.. then we can talk about the rest of those blog posts
#csarvenis reading aaronpk 's message above. will get to the rest.
#rhiaroand then csarven arguing that certain types of url are fine based on RFCs, and tantek arguing that they're not based on they're not 'friendly' becomes moot
#rhiarobecause non-friendly URLs are out of scope for articles
#pokadidnt LEe Feigenbaum writ e CoolURIs. too lazy to loook
#7YUAACSOSalso, I'm now managing to 500 tokens.indieauth.com
#tantekcsarven, please consider focusing more on creating on your own site, and documenting your own results rather than taking offense at criticisms of RFCs / W3C specs etc.
#voxpelliMy Micropub endpoint now does automatic language detection on content – meaning I have multilingual support with all Micropub clients!
#csarvenben_thatmustbeme I would say that a Note or whatever still has higher importance than a blog post. Of course the content matters, so this is only a generalization/default POV.
#aaronpki find these diagrams hard to read personally, but i will try to extend the existing one to the other flow
#csarventantek Like I said, I have shipped URLs in the millions. Does that not count? how many have you shipped?
catsup joined the channel
#voxpelliaaronpk: if you want a CLI-tool that generates SVG:s for sequence diagrams instead: https://github.com/knsv/mermaid Never gotten along with ordinary UML-tools myself :/
#csarvenben_thatmustbeme Seriously? != God argument? If we are going to talk about "anyone can get things wrong", please allow me to put my default bet on TimBL over IWC. After all, one has a remarkable track record which not many even comes close to.
#pokaw3c spec-writers seem to heavily overlap with .edu paper-writers + corporate-countrycllub phd-welfare dudes who mainly seem to come up with vague r&d guidance , if there's "Trenches" involved there, theyre definitely different , IBMORACLEGOOGLE trenches than the typical wordpress admin's trenches
#moredhelvoxpelli, I'm just messing around with setting up a storage plan at the moment, but will try and actually get something more better, slightly further down the line.
#kylewmaaronpk: having a little trouble wrapping my brain around this. writing an indieauth implementatio that delegates to tumblr. where should i stick the incoming request params (me, client_id, redirect_uri, etc.) while I redirect to tumblr's auth endpoint?
#kylewmaaronpk: is a server side session (i think that means in memcache/redis?) a good way to store the oauth request token too? this is oauth1 so i need to save some info between calls
#atomiculesanm: I used to have the rest of my jekyll templates available, etc, but no more since I pulled it off Github
#aaronpkyeah that's pretty typical. or in php it just uses a file on disk by default
#KevinMarkscsarven: you see standards as legislation; we see them as documentation.
#tbrbwe'll be doing them from a different laptop than the current hackroom stream, but yeah, and I'll post hangout / youtube etc in ~ 30 minutes
#KevinMarksthanks to wordpress, about a quarter of the web has YYYY/MM/DD urls
loic_m joined the channel
#JarOfGreenaaronpk: thanks for examples, both basically working and the later one was an event that had both formats of markups which was something I was going to look out for - obvs we only want one event coming thru, which is now done!
#JeenaI took 12 beers with me from Portland but I don't want to open them because once I've dank them I will not be able to get anything like that here :-(
snarfed joined the channel
#voxpellirhiaro: how would one demo remotely? I guess I could show some of my Micropub work
#rhiarovoxpelli: you could screen share. tbrb is in charge of that though
#tbrbvoxpelli: the stream will be done on hangouts so I'll fire you over the hangouts link when we do them and you'll be able to use that to screenshare
#voxpellitbrb: cool, screensharing in hangout I've done multiple times :)
#aaronpkit's actually a sequence diagram but i guess it's uml syntax
#rhiaroI may have just become prejudiced against them at uni
#aaronpki can't get my head around the syntax this app uses so i feel like that dog
#rhiaroseemed like something we were made to do for lecturer's entertainment only
catsup joined the channel
#JeenaI liked them at the university because I was able to do them and then just to generate Java code so I didn't have to write all the boilerplate by hand :p
#aaronpkyeah i just wanted to make a few changes to this diagram, it'll look the same and i will also include the source like on indieauth-login-flow.png
catsup and KitD joined the channel
#KevinMarksOK - I liked that it made SVG instead of bitmaps
#kylewmthere are lots of different types of UML, i think class diagrams are pretty worthless, sequence diagrams are great
#snarfedtransient; they'd disappear when you send their notif
#aaronpki guess i couldn't get my head around how that would work
#KartikPrabhusnarfed: the thing is I can't just make my service install hfeed2atom. I have to install hfeed2atom, and all it dependecies as submodules or whatever
#snarfedor via pip/virtualenv, sure. but again, you'd need to do that anywhere else too, or else hfeed2atom wouldn't work, right?
#aaronpkit seemed like it would generate duplicate notifications
#KartikPrabhuusually setup tools does all that for me. now it seems I have to manually install each dependency or something
#snarfedaaronpk: i think step 4 avoids them: "Otherwise, send a single notif for the cluster, and remove the wms from the other clusters they're in (source vs target) so we don't send dupe notifs."
#aaronpkfor the set ((a -> z),(b -> z)) you would end up with three buckets, the "a" bucket: (a -> z), the "b" bucket, (b -> z), and the "z" bucket, (a -> z),(b -> z)
#snarfedKartikPrabhu: huh. i guess i assumed you were already using those. no matter, you can do the same thing with just setuptools, without them. just find the flag to install to a specific dir (e.g. local/ inside your app) instead of systemwide
#kylewmaaronpk: I have to say it's a little weird that you GET to an authorizaton_endpoint to get the code and POST to verify it, and POST to a token_endpoint to get the token and GET to verify it
#GWGkylewm: You are using the WordPress.com JSON API?
#GWGkylewm: Just thinking that if you also support the slightly different JSON Rest API, you could support any WordPress site that installed the REST API plugin
#aaronpkkylewm: the get/post on the authorization endpoint is part of OAuth 2 (it's actually the browser making the GET), and posting to the token endpoint is also oauth 2.
#aaronpki thought it made sense to use GET requests to verify codes/tokens because it's not changing state of anything
#kylewmI agree with that, but to be consistent, wouldn't the "verify auth code" bit need to be a GET?
#kylewm(but this is quibbling, I don't feel strongly)
#aaronpkthat would make that step more similar to verifying auth codes
#aaronpkmany micropub servers end up implmeenting looking for tokens in both the header and post body, so i could see having the token endpoint check both as well
#kylewmhuh, I apparently didn't implement token verification in my site's token endpiont at all!
#aaronpkyou don't need to if your micropub endpoint can verify the token on its own
#kylewmok, I think I would like to make that clearer, which bits of indieauth and token_endpoint are required for micropub, and which are just required for interoperation with other endpoints
#kylewmty aaronpk, sorry you always have to straighten this out for me. i will make a point to improve the wiki so hopefully you won't have to next time :p
#kylewmi don't know that it qualifies as multi-user or not
KartikPrabhu and parzzix joined the channel
#KartikPrabhugives up again on this app engine nonsense
#aaronpkhere's a question, is this something you'd expect me to use by signing in as aaronparecki.com? or would I be signing in as aaronpk.wordpress.com
#aaronpk(i think the answer is the wordpress subdomain)
#aaronpkyeah, the user would basically just skip past silo.pub with a redirect and see the tumblr login screen. you wouldn't need to show them anything on silo.pub for that
#KartikPrabhuwhat is a good status code to return if the atom conversion fails?
#LoqiIt looks like we don't have a page for "good status code to return if the atom conversion fails" yet. Would you like to create it? http://indiewebcamp.com/s/100S
#KevinMarks_Or lack of a h-feed/h-entry? Or would that just make an empty atom feed?
_fran and snarfed joined the channel
#7YUAACSOSAn invalid feed as in "does not exist" smells like a 400, but a problem with the feed itself.. There is a code for an upstream error I think.
#kylewmKartikPrabhu: yeah, i think in general if you try to initialize your own Response object (in any framework) you're going to have a bad time
#myfreewebhi everyone! haven't been there in a while. made some vanilla js web components https://github.com/myfreeweb/indieweb-components right now there's indie-action and fragmention-target, mostly based on existing code
#pokawhy not a 2-line shellscript to finger your friends to grab the latest .plan file. so much easier than fiddling with post-SGML dreck
jciv, mlncn and indie-visitor joined the channel
#LoqiWelcome, indie-visitor! Set your nickname by typing /nick yourname
#anomalilyHey all. I'm here to tell you guys that Aaron has a book to turn into O'Reilly and he's not allowed to come play with you guys until he finishes!
#GWGanomalily: And, where is the material you all said you were going to create at IWC earlier this month? Is there anything I can add to the Indieweb plugin?
#anomalilyI failed miserably to document the stuff at IWC, because I stupidly decided to redesign my website instead
#anomalilyI'm so excited that Teacup works with the dev version of micropub!
#rhiaroGWG: there's a Le Pain Quotidien in central park, they (in other locations I've been to at least) have free wifi, power sockets and awesome food
#rhiaroIt's my go-to when I'm stuck somewhere in London for a couple of hours with nothing to do
#GWGrhiaro: We solved the HWC issue, and were going to try a few mid-Manhattan venues. Then zachdonovan got busy.