#dev 2017-05-19

2017-05-19 UTC
mblaney joined the channel
#
mblaney nice work Zegnat, dobrado is wiki-ish too :-)
[cleverdevil], arush1, [kevinmarks], [shaners], dougbeal|mb1, KartikPrabhu, arush, [colinwalker] and barpthewire joined the channel
#
@helloanselm @jeena btw wanted to manually add the link as reply but got a "504 Gateway Time-out" from your webmentions service… (twitter.com/_/status/865447981826293760)
cweiske joined the channel
#
Zegnat Does anyone here know why rel-feed didn’t make it? According to http://microformats.org/wiki/existing-rel-values is was proposed somewhere in the HTML5 spec but it doesn’t look like it made it into the final version
#
Loqi existing rel values
barpthewire and pfefferle joined the channel
#
pfefferle good morning
#
Loqi pfefferle: snarfed left you a message 14 hours, 13 minutes ago: thanks! where is the front page webmentions setting? i don't think i changed it from the default.
#
Loqi pfefferle: snarfed left you a message 14 hours, 11 minutes ago: ah, if you mean the webmention_post_id filter, i haven't added one. https://wordpress.org/plugins/webmention/#faq-header
#
Loqi pfefferle: snarfed left you a message 13 hours, 42 minutes ago: filed https://github.com/pfefferle/wordpress-webmention/issues/141
[kevinmarks] joined the channel
#
Loqi good morning
[colinwalker] joined the channel
#
mblaney Zegnat: I had no idea rel values were restricted...
sebsel joined the channel
#
mblaney I'm not sure why that needs to be the case, but the spec does say extensions are allowed via ratification on: http://microformats.org/wiki/existing-rel-values#HTML5_link_type_extensions
#
Loqi existing rel values
#
mblaney so maybe it's just a case of getting that list cleaned up.
#
mblaney got to go for now, but would be interesting to follow up on this. spec compliance seems like a worthy indieweb goal.
#
Zegnat mblaney, yes, rel keywords not documented by the HTML spec hace to be on the mf wiki, else they should be rejected by conformance checkers as per https://html.spec.whatwg.org/multipage/semantics.html#other-link-types
[jeremycherfas] joined the channel
#
voxpelli mblaney Zegnat: Or one can use a URL if it's something for oneself, like http://kodfabrik.se/rel/my-own-epic-relation
#
voxpelli like rel="http://kodfabrik.se/rel/my-own-epic-relation"
#
Zegnat voxpelli, interesting, I didn’t even stop to think that URLs would be valid tokens!
#
Zegnat Although those still wouldn’t be spec-conform though, would they?
#
voxpelli Zegnat: it at least was the recommended way to add new unregistered link-relations
#
voxpelli Webmention itself used rel="http://webmention.org/" originally
#
voxpelli Zegnat: here's one of the specs that has been made around link relations: https://tools.ietf.org/html/rfc5988#section-4.2
#
Zegnat “Conformance checkers must use the information given on the microformats wiki existing-rel-values page to establish if a value is allowed or not: […] values marked as "discontinued" or not listed in either this specification or on the aforementioned page must be rejected as invalid.” - https://html.spec.whatwg.org/multipage/semantics.html#other-lin
#
Zegnat k-types
#
Zegnat Makes it sound like the HTML spec does not allow unregistered relations
#
voxpelli I quite liked the Link Relations spec in that it tried to unify link relations across data formats – such link relations enables extension and composition of different standards in a very powerful way
[kevinmarks] joined the channel
#
[kevinmarks] The html spec allows lots of things - it won't remove unlisted release values or anything.
#
[kevinmarks] s/release/rel/
#
Zegnat No, it wouldn’t remove anything. But it does seem to define unknown rel values as nonconforming to the spec in this case.
#
Zegnat Conformance doesn’t have a real effect on HTML parsers I guess, but still a thing
#
voxpelli Well, it only seems to spec how conformance checkers should signal things
j4y_funabashi joined the channel
#
aaronpk when did HTTP 302 become "Moved Temporarily" instead of "Found"?
#
cweiske RFC 7231 still lists it as "302 Found"
#
Zegnat Maybe https://tools.ietf.org/html/rfc1945 ?
#
cweiske http 1.1 lists it as 302 found, too
[colinwalker] joined the channel
#
aaronpk oh is it an http2 thing?
#
cweiske 1945 is http 1.0
#
aaronpk just remembered that my nginx runs http2
#
aaronpk (which is where i saw "302 Moved Temporarily"
#
cweiske 10.3.3 302 Found
#
cweiske The requested resource resides temporarily under a different URI.
#
cweiske that's http 1.1, rfc 2616
sebsel joined the channel
#
cweiske the title after the code is free-text
#
cweiske only the code itself is relevant
#
cweiske so nginx could also state "302 Piggy for sale" and it would conform
#
Zegnat cweiske: 1945 is HTTP 1.0, yes, and writes about 302 Moved Temporarily: https://tools.ietf.org/html/rfc1945#page-35
#
cweiske I sometimes use that to flesh out the generic 400 status code with concrete error messages
#
cweiske "400 Invalid username"
#
aaronpk is going to start using "Piggy for sale" in random places now
#
jonnybarnes I got signed into brid.gy again with twitter
#
Zegnat That’s pretty clever, actually, cweiske. I never considered that. I should maybe do that instead of generic 403 Forbidden
#
jonnybarnes but I now have a sneeking suspicion that when I added support for JSON syntax to my client, it gets it wrong
#
jonnybarnes the HTML form has checkboxes named `mp-syndicate-to[]`
#
jonnybarnes when adding that to the json I go through a foreach loop and do `$json['properties']['mp-syndicate-to'] = [$target];`
#
jonnybarnes Shouldn’t it be $json['props']['mp-syn..'][] = $target;?
#
aaronpk yes that makes more sense
#
jonnybarnes and the end result of json_encode if there’s more than one target will be ["target1", "target2"]
#
jonnybarnes like the categories in example 5 of the spec
sebsel and [kevinmarks] joined the channel
#
[kevinmarks] The point of 302 (as opposed to 301) is that you don't cache the redirect, so renaming it clarified that.
#
aaronpk yeah it makes sense, i just can't find where it was renamed
[shaners] joined the channel
#
aaronparecki.com created /live (+108) "add embedded livestream" (view diff)
#
jonnybarnes right, if my client html form wants multiple targets, and keeping names consistent, the form name is `mp-syndicate-to[]`
#
jonnybarnes but for the generated JSON request POSTed to the endpoint, its just `mp-syndicate-to` right?
#
jonnybarnes This is an example of what my app is generating to POST to an endpoint: string(209) "{"type":["h-entry"],"properties":{"content":["Hello Fred"],"in-reply-to":["https:\/\/fredbloggs.com\/note\/abc"],"mp-syndicate-to":["https:\/\/twitter.com\/jonnybarnes","https:\/\/facebook.com\/jonnybarnes"]}}"
#
aaronpk the [] thing is a convention for sending multiple values of the same property name in urlencoded format. most backends that receive that form post will convert it to an array structure internally automatically
#
aaronpk in JSON format, there isn't a need for a workaround like that, you just send an array
#
aaronpk that looks correct. also you should be able to use micropub.rocks to check this stuff
#
jonnybarnes yeah, never quite worked out why some json used { and some uses [, specifically why is the json `"properties":{` and not "properties":[`
#
aaronpk that's just a JSON thing. {} is an object with key/values, [] is an array for containing multiple values
#
jonnybarnes cos to me it looks like properties is an array
#
jonnybarnes okay
#
jonnybarnes that akes sense
#
aaronpk really you shouldn't have to worry about it, create the data structure natively in whatever language, and use a JSON encoder to generate the JSON
#
jonnybarnes yeah, but I'm writing a PHPUnit test where I’m trying to test the JSON my client code generates is correct
#
aaronpk aha
#
seblog.nl edited /checkin (+196) "/* Sebastiaan Andeweg */ updated the way my checkins work" (view diff)
sebsel joined the channel
#
[kevinmarks] PHP doesn't really make the array/dictionary distinction
#
[kevinmarks] so your code needs to see if the keys are numbers that increase in value linearly to decide that it is an array to encode as json
#
aaronpk no json_encode handles it fine
#
aaronpk there is a difference internally, you just access arrays and hashes both using $foo[$bar] syntax
#
[kevinmarks] did they change it internally? it used to be possible to have an array $foo of ["a","b","c"] and then do $foo["bar"]="baz" and it would then be {0:'a', 1:'b', 2:'c', 'bar':'baz'}
#
aaronpk that is a theoretical situation that is only interesting if you care about the language internals. in practice you don't do that, and everything works like expected.
#
aaronpk if you're curious: https://gist.github.com/aaronpk/9f64d8483e07b0d1f499fbae9c406ebc
#
aaronpk which seems quite reasonable
[colinwalker] and sebsel joined the channel
#
[kevinmarks] yes, it gets weirder with coercion
#
aaronpk well that was a weird rabbit hole that ended nowhere. It occurred to me that technically I don't need a silo profile to include a rel=me URL on the profile page if I can check the user's website using their API.
#
aaronpk since I already have to have API keys in order to do OAuth, I could make an API request to check the user info as well
#
aaronpk that does mean more code per silo, but that's not necessarily the end of the world
#
aaronpk so I tried it with Flickr, since Flickr's API is probably the most stable of the silos
#
aaronpk first off, I had forgotten about the mega confusion between flickr profile URLs and flickr usernames
#
aaronpk my username is "aaronparecki" but my profile URL is "flickr.com/aaronpk" (someone else has the username "aaronpk")
#
aaronpk thankfully there's an API endpoint to find the flickr user ID given a flickr profile URL
#
aaronpk however, the userInfo endpoint returns the bio but doesn't return the website field that the user enters!!
#
aaronpk so much for that thought
#
[kevinmarks] sgnodemapper may have some of that https://github.com/webjay/sgNodeMapper/blob/master/sites/flickr.js
#
aaronpk i don't really want to jump through hoops for it
#
aaronpk just thought it would be nice to support silos that had decent APIs even if they didn't necessarily have a rel=me link in the HTML
#
aaronpk oh well
sebsel, KevinMarks, barpthewire, [colinwalker], j4y_funabashi and KevinMarks_ joined the channel
#
ben_thatmustbeme Social web incubator community group meeting in 2 hours https://www.w3.org/wiki/SocialCG/2017-05-19
[kevinmarks], arush, singpolyma, barpthewire, [manton] and [cleverdevil] joined the channel
#
www.svenknebel.de edited /Private-Webmention (+61) "/* Sebastiaan Andeweg */" (view diff)
tantek, [mko], miklb, KevinMarks and [manton] joined the channel
#
dougbeal.com edited /User:Dougbeal.com (+105) (view diff)
#
dougbeal Yay!
#
Loqi ?
KevinMarks_ and KevinMarks joined the channel
#
dougbeal.com edited /User:Dougbeal.com (+328) (view diff)
#
dougbeal.com edited /User:Dougbeal.com (+65) "/* ideas */" (view diff)
KevinMarks joined the channel
#
voxpelli [manton]: btw, I'm failing to connect to my Micropub endpoint in the Micro.blog app, are there any known issues with that implementation or could I help debug it?
#
voxpelli I'm getting: Authorization "code", "state" or "me" parameters were missing
#
dougbeal.com edited /User:Dougbeal.com (+19) "/* ideas */" (view diff)
tantek, KevinMarks, gRegorLove and [cleverdevil] joined the channel
#
[cleverdevil] I've been having similar issues, @voxpelli. I think manton is aware. Just other higher priority stuff to deal with at the moment. I did provide manton with an account on my Known instance to test with when he gets around to it!
KevinMarks_ joined the channel
#
voxpelli [cleverdevil]: great :+1
KartikPrabhu, KevinMarks and [manton] joined the channel
#
[manton] @voxpelli Thanks for letting me know. I thought those issues were fixed, but I'll test again and figure out what is wrong.
#
voxpelli [manton]: here's my Micropub implementation for reference: https://github.com/voxpelli/webpage-micropub-to-github And my discovery page: http://micropub-discovery.voxpelli.com/
#
Loqi [voxpelli] webpage-micropub-to-github: Self-hosteable Micropub endpoint that publishes to Jekyll by committing to GitHub
[kevinmarks] joined the channel
#
@ScalaWilliam Trackbacks vs Pingbacks vs Webmentions for WordPress @wpmudev https://premium.wpmudev.org/blog/trackbacks-pingbacks-webmentions/ (twitter.com/_/status/865675141811916800)
KevinMarks, tantek, KartikPrabhu and [chrisaldrich] joined the channel
#
tantek.com edited /2017 (+60) "/* Organizers */ welcome onboard benwerd" (view diff)
[cleverdevil] joined the channel
#
@karenbwill RT @ReadersGazette BLOG Indie Author Answers by Jim Heskett http://www.thejugglingauthor.com/indieauth/ Get help writing your book #bookbloggers 12 (twitter.com/_/status/865700269560147968)
#
@WendyandCharles ReadersGazette: BLOG Indie Author Answers by Jim Heskett http://www.thejugglingauthor.com/indieauth/ Get help writing your book #bookbloggers 12 (twitter.com/_/status/865700131521400832)
#
aaronpk wow they're using it in the slug now?
#
tantek wow indeed
#
aaronpk Might have to blacklist "indie authors" from the search now
#
tantek indeed
#
tantek.com edited /2017 (+141) "Update a bunch of organizer stuff" (view diff)