#social 2017-10-24
2017-10-24 UTC
xmpp-social joined the channel
# 
puckipedia hehe, I added Page support into Kroeg, *and* I have a web interface to create posts/pages!
# puckipedia so now I have an AP-powered blogging system basically, with federated comments, I guess :P
timbl joined the channel
# puckipedia cwebber: eh, I can't get test.activitypub.rocks to work?
timbl, dlongley and adam joined the channel
# cwebber puckipedia: hi hi
# cwebber puckipedia: gimme a sec I'm about to push a new version
# puckipedia ah :)
# cwebber puckipedia: but then I'd like to test it against your instance
# puckipedia awesome
# puckipedia I guess it's been pushed?
# cwebber nope one sec
# cwebber I'm doing it right now
# cwebber puckipedia: https://test.activitypub.rocks/
# cwebber thre's one bug I need to fix though
# puckipedia tries client-to-server first
# cwebber the websocket connection doesn't ping you so it times out if inactivity D:
# cwebber I need to fix that by adding websocket ping
# cwebber I'm working on that bit today
# puckipedia hm. I gave it https://kroeg.azurewebsites.net/test_activitypub as actor ID for client->server and I see no messages
# cwebber puckipedia: the other sections are also now added and are a bit more prompt-y
# puckipedia wait no here it goes
# cwebber arg it's saying gnutls not available but that shouldn't be true
# puckipedia ehm, which Accept: is it sending
# cwebber sorry I just killed it because I was trying to fix the gnutls thing, shouldn't have done that
# puckipedia well,
# puckipedia ah
# puckipedia I think someone did GET on the test_activitypub actor separately from the test
# cwebber oh I think I need to upgrade the server :<
# cwebber migraine expands
# cwebber annnd of course I'm running a custom branch of guix
# cwebber because letsencrypt service hasn't made it into mainline yet
# cwebber ok! time to port some patches to guix master ;P
# cwebber sorry puckipedia
# cwebber I got the code updated and working but then not the server apparently
# cwebber I'm working on it right now
# puckipedia hehe, no problem, time to shower :P
# puckipedia just poke me whenever it's done
# cwebber puckipedia: will do
# cwebber rebased the git repo finally :P
# cwebber this is why you don't let git branches bitrot!
# cwebber jebus Loqi
# puckipedia okay, I'm back
# puckipedia cwebber: any updates?
tantek joined the channel
# cwebber puckipedia: update is that server upgrade is going worse than expected. I have a workaround if necessary where I can forward from my server but I'm trying to see if I can fix it
# puckipedia mh, ok :)
# cwebber ok, had to redo the rebase of guix
# cwebber at least this is motivating me to get letsencrypt into guix master :P
# cwebber letting branches bitrot ftl
# puckipedia I do kinda want to test the whole thing right now already :<, :P
# cwebber oh wait
# cwebber I may be able to get this up sooner without waiting for that recompile
# cwebber 1sec let's see
# cwebber puckipedia: okay I set up the liar's version
# cwebber https://test.activitypub.rocks/
# cwebber ie proxying to my computer
# cwebber see if it works
# puckipedia yep. hangs after I enter a client-to-server actor ID
# cwebber gnutls error again D:
# cwebber wtf wtf wtffffffffffff
# cwebber this shouldn't be happening aughhhhhh
# cwebber oh wait
# cwebber different gnutls error
# cwebber ERROR: Throw to key `gnutls-error' with args `(#<gnutls-error-enum The TLS connection was non-properly terminated.> read_from_session_record_port)'.
# cwebber not properly terminated?
# puckipedia huh. ID is https://kroeg.azurewebsites.net/test_activitypub
# cwebber oh right
# cwebber ughhhhh
# cwebber THAT'S why I had guile built from git in the dev environment!
# cwebber this was fixed in guile master
# cwebber ugh
# 
@uucidl @neauoire and https://www.w3.org/TR/activitypub/ (I'm not a specialist, someone simply pointed me to these when I showed them rotonde) (twitter.com/_/status/922863883404152832)
# cwebber this is the downside of deploying right up to the wire :<
# cwebber ;_;
# cwebber well that's going to take about 4 hours to compile
# puckipedia welp. I'll set up an http-only Kroeg instance lol
# cwebber so I got all the software working by last night, of course it was deploying it that was the problem :P
# cwebber well that's one "solution", heh
# 
Loqi Tantekelik made 1 edit to [[Socialwg/Github Process]] https://www.w3.org/wiki/index.php?diff=104823&oldid=103337
# puckipedia okay, I think I got an http instance working
# puckipedia cwebber: http://home.empw.nl:5000/users/test_activitypub :P
# puckipedia at least. that should work
# cwebber puckipedia: well I just trashed my environment trying to a rush patch-over to fix it. we're just going to have to wait until things compile
# puckipedia welp
# cwebber sorry :( nobody is more distressed than I am :<
# puckipedia hm, we don't have a clear extension endpoint mechanism yet?
# cwebber on the upside I guess it's better to be "stuck waiting on things to build" than not have something to present
# cwebber puckipedia: not other than the vocabulary extension mechanism
# puckipedia right. I need a few extensions for my Kroeg web client to work, mostly extra endpoints but I feel endpoints will be so packed
# cwebber we don't have a way to "declare" what extensions you're using but there's a socialcg issue open about it
# puckipedia actually, hm
# puckipedia first of all I need to fix Kroeg so it isn't quite as sensitive about ID case sensitivity
# cwebber id case sensitivity?
# cwebber why wouldn't you want case sensitive uris
# puckipedia actually, I think what I'm looking for is "if you request a specific URL and it returns an AS2 object with completely different ID, and your ID is nowhere to be found, return that object"
# puckipedia together with mapping @[user]@[domain] IDs into their proper user IDs on the server
RRSAgent joined the channel
# 
RRSAgent logging to http://www.w3.org/2017/10/24-social-irc
Zakim joined the channel
# 
ben_thatmustbeme good morning tantek
# cwebber present+
# ben_thatmustbeme present+
# cwebber not me
# ben_thatmustbeme i can scribe
# ben_thatmustbeme scribenick: ben_thatmustbeme
# ben_thatmustbeme scribe: Ben Roberts
# ben_thatmustbeme Chair: Tantek
# ben_thatmustbeme TOPIC: approval of last meeting's minutes
# ben_thatmustbeme tantek: please take a quick look
# 
tantek PROPOSED: Approve minutes from two weeks ago https://www.w3.org/wiki/Socialwg/2017-10-10-minutes
rowan joined the channel
# ben_thatmustbeme its because no one set Chair: in the minutes
# cwebber +1
# ben_thatmustbeme tantek: lets go ahead and resolve them
# ben_thatmustbeme TOPIC: telcon schedule
# ben_thatmustbeme tantek: the next item is telcon sched, but the next meetings coming up are TPAC
# Loqi Tantekelik made 1 edit to [[Socialwg/2017-10-10-minutes]] https://www.w3.org/wiki/index.php?diff=104830&oldid=104648
# ben_thatmustbeme we have a slot scheduled for the CG, monday afternoon i believe
# ben_thatmustbeme sandro, whats the date on that?)
# ben_thatmustbeme tantek: make a note of that, november 6th 3:30
# cwebber q+
# ben_thatmustbeme ... we'll see about remote attendance, we'll certainly be using the IRC
eprodrom joined the channel
# eprodrom hallo
# eprodrom present+
# ben_thatmustbeme ... not sure if thats after the time change or not
# ben_thatmustbeme cwebber: i want to point out that we said we would do it every other week but leave it open to additional weeks. I think we might want to do that for a making sure AP gets to CR
# ben_thatmustbeme tantek: we had agreed to 90 minutes today and we have that possibility of adding another week
# ben_thatmustbeme ... did you need more discussion or just confirmation
# ben_thatmustbeme cwebber: we might need to issue a new CR next week, but i think it depends on what ends up happening today and this week
# ben_thatmustbeme tantek: so you are saying you may need an additional meeting for next week either way for publishing reasons
# ben_thatmustbeme sandro: isn't it conditional?
# ben_thatmustbeme tantek: basically its 'if we need more time or we need it for publication'
# ben_thatmustbeme ... i think its a reasonable request, are there any objections?
# ben_thatmustbeme ... not hearing any obj. i'm going to approve that
# ben_thatmustbeme TOPIC: activity pub
# ben_thatmustbeme tantek: cwebber, what is left for CR
eprodrom_ joined the channel
# ben_thatmustbeme cwebber: the major thing is the test suite, and i got it written. I was deploying it this morning but ran in to servers issues. the software is good in theory, its just a matter of working through the deploy issues
# ben_thatmustbeme ... by the end of today, i should have it up and people on the IRC should be able to see it and test against it
# ben_thatmustbeme tantek: awesome progress
# ben_thatmustbeme sandro: other than puckipedia who do you expect to be submitting implementations in the near-term?
# ben_thatmustbeme cwebber: i'm going to submit one and ... and i'm going to try to push others to submit them, most notably mastodon
# ben_thatmustbeme sandro: ajordan are you expecting to have this implemented or not yet?
# ben_thatmustbeme ajordan: (via IRC) probably
# ben_thatmustbeme sandro: cwebber are you expecting these to all be complete or a feature matrix sort of thing?
# ben_thatmustbeme cwebber: its going to be a feature matrix, the test suite actually will export the implementation report and i can just plug it in to make it nice and "griddy"
# ben_thatmustbeme tantek: there is also an implementation with bridgy
# ben_thatmustbeme ... idon't know if that can be run against it
# eprodrom_ ajordan++
# eprodrom_ I have an app that should be ready to test for next week
# Loqi Tantekelik made 1 edit to [[Socialwg]] https://www.w3.org/wiki/index.php?diff=104831&oldid=104649
# eprodrom_ (It's a client app, c2s)
# ben_thatmustbeme cwebber: (lists a bunch of app) anyone else i should reach out to?
# ben_thatmustbeme tantek: i know in the past two weeks, there have been one maybe two people using fed.brid.gy to reply to mastodon
# ben_thatmustbeme ... its not a test per-se but at least its real world example
# ben_thatmustbeme ... thats one of the things we would want to bring up with you in the transition call
# ben_thatmustbeme ... you may want to consider having a section in your implementation report for real-world interop or something like that
# eprodrom ajordan: haven't pushed it
# ben_thatmustbeme ... list them or give permalinks, etc
# eprodrom I'll get it up this weekend!
# ben_thatmustbeme cwebber: i can ask that, yeah
# ben_thatmustbeme tantek: is that reasonable to you sandro?
# ben_thatmustbeme sandro: yeah, thats great
# ben_thatmustbeme cwebber: the issues that were moved to extensions have been closed, so thats moving along nicely
# ben_thatmustbeme ... everything thats left on there is editorial
# ben_thatmustbeme ... hopefully we'll have everything ready for a PR ready document by next tuesday
# ben_thatmustbeme sandro: you want to do another editorial CR next week?
# ben_thatmustbeme cwebber: i don't know the process, does it make sense?
# ben_thatmustbeme sandro: no need to, you can just wrap them in to the PR?
# ben_thatmustbeme ... the question there is when do we think we will go to PR?
# ben_thatmustbeme .... should we say nov 14th as a target deadline for imp reports
# ben_thatmustbeme cwebber: that sounds like a good goal
# ben_thatmustbeme tantek: if we are going to wait that long, there is no harm doing an editorial CR is there?
# ben_thatmustbeme ... i was under the impression we were going to try for like next week
# ben_thatmustbeme ... i would be supportive either way
# ben_thatmustbeme sandro: what is important to have the editorial changes are done in the ED at least soon
# eprodrom q+
# ben_thatmustbeme tantek: i agree, cwebber is that something that you can commit to by next week?
# ben_thatmustbeme ... incorporated in to the ED
# ben_thatmustbeme cwebber: yeah
# ben_thatmustbeme eprodrom: i can help too
# eprodrom rhiaro: great
# ben_thatmustbeme cwebber: i think thats everything i had to say
# ben_thatmustbeme sandro: we didn't quite resolve if we would do a new CR next week or just get it in to ED
# 
sandro tantek https://evancoin.com/
# ben_thatmustbeme cwebber: i think it seems like a good idea to issue a new CR
# ben_thatmustbeme sandro: it also makes the PR transition a little cleaner
# ben_thatmustbeme sandro: can you have that ready by monday so we can review?
# ben_thatmustbeme cwebber: yes
# eprodrom 30 minutes?
# ben_thatmustbeme sandro: and I guess just a quick meeting on tuesday just to do a quick vote
# ben_thatmustbeme <ben_thatmustbeme> i probably won't be able to be on next week
# ben_thatmustbeme tantek: any objections
# cwebber eprodrom, I think 30 mins should be fine
xmpp-social joined the channel
# ben_thatmustbeme tantek: i just want to record that there are no objections
# ben_thatmustbeme cwebber: 30 minutes seems reasonable to me
# ben_thatmustbeme tantek: me too
# eprodrom I can chair
# ben_thatmustbeme ... i'm also going to volunteer evan to chair
# cwebber https://test.activitypub.rocks/
# eprodrom Will be coming off of transatlantic flight but I can rock that
# ben_thatmustbeme cwebber: one more thing, this is the version of the test suite thats running, it doesn't support https but its up there and verify that version for now
# ben_thatmustbeme ... you can do the first and last one, which are just clicking buttons to say you support things
# eprodrom How do I kill people in this mush
# eprodrom j/k
# cwebber puckipedia: https://test.activitypub.rocks/ is running again with the http only interface
# ben_thatmustbeme tantek: evan: what?
# ben_thatmustbeme cwebber: this is reusing the interface i used for a MUD...
# ben_thatmustbeme tantek: i think we have answered all the questions, congratulations cwebber for getting the test suite up there
# cwebber oh and the last thing I need ot do is fix the websocket bug where it times out easily because it isn't pinging the server :)
# eprodrom I should probably note for the record that I don't want to kill people. That's just something that noobs say when they join a MUD/MUSH for the first time.
# ben_thatmustbeme tantek: i think we have some time for last call for implementations right?
# cwebber eprodrom: :)
# ben_thatmustbeme missed a little of that, boss stepped in
# cwebber total aside if you want to watch a video of the MUD where I gave the talk in the MUD itself https://www.gnu.org/software/8sync/
# puckipedia cwebber: "Unable to find OAuth endpoints..." they're definitely there
# ben_thatmustbeme tantek: cwebber do you have a stub implemntation report?
# ben_thatmustbeme cwebber: i can copy and paste the output if i just say no to everything
# ben_thatmustbeme tantek: i mean the summary
# ben_thatmustbeme cwebber: i'm doing that today
# ben_thatmustbeme tantek: so thats where it would be at that link?
# ben_thatmustbeme cwebber: yes
# ben_thatmustbeme TOPIC: websub
# ben_thatmustbeme tantek: any new issues on websub?
# ben_thatmustbeme aaronpk: yes, someone submitted a review of several ... i'll just post the link
# ben_thatmustbeme ... i have not yet started to address these
JanKusanagi joined the channel
# ben_thatmustbeme ... and then tantek opened an issue to discuss the text to refer to the text
# ben_thatmustbeme just the text links to it, its not a normative change
# ben_thatmustbeme tantek: that one (127) looks like it needs to be broken up in to sep issues
# ben_thatmustbeme aaronpk: yeah, that will be my action item
# ben_thatmustbeme tantek: other than that?
# ben_thatmustbeme aaronpk: other than taht, i don't htink i have any updates
# ben_thatmustbeme should i minute this tantek?
# ben_thatmustbeme tantek: i believe i can share that there are no negative responses to the PR yet
# ben_thatmustbeme ... it sounds like if you are able to resolve these issues we may get another in
# ben_thatmustbeme sandro: it would be good to remind people at TPAC to vote
# ben_thatmustbeme ... right now i would need to extend the response time limit
# ben_thatmustbeme tantek: can we extend it to thursday at TPAC?
# ben_thatmustbeme sandro: sure
# eprodrom ajordan: oh, har
# ben_thatmustbeme tantek: aaronpk how soon do you think you can get those issues created?
# ben_thatmustbeme aaronpk: i can probably do that either today or tomorrow
# ben_thatmustbeme tantek: sounds good
# ben_thatmustbeme ... sandro anything we should do to prepare, and editors draft, etc?
# ben_thatmustbeme sandro: sure, keeping the editors draft up to date is always good
# ben_thatmustbeme tantek: so aaronpk you have licesnse to keep updating the ED as you resolve these issues
# ben_thatmustbeme ... anything else for websub?
# ben_thatmustbeme TOPIC: JF2
# eprodrom I can scribe
# eprodrom scribenick: eprodrom
# eprodrom tantek: take us through what's happening with JF2
# eprodrom tantek: there will be a new ED
# eprodrom ben_thatmustbeme: not a lot of normative changes, editorial changes
# eprodrom ben_thatmustbeme: also examples and references
# eprodrom q+
# ben_thatmustbeme I just lost my audio responses
# eprodrom eprodrom: were there any editorial changes that could be normative?
# ben_thatmustbeme facepalm
# eprodrom ben_thatmustbeme: reviewed changes, were not normative
# 
ajordan I guess https://github.com/dissolve/jf2/commit/d766076b00d9af40521d669872fc3c4119ebf17e is technically normative but it was pretty clear so
# ben_thatmustbeme trying
# ben_thatmustbeme its not behaving
# tantek right, that red text is not in https://dissolve.github.io/jf2/
# ajordan ben_thatmustbeme: https://github.com/dissolve/jf2/pull/35
# eprodrom ben_thatmustbeme: the change ajordan is noting was not from the previous WD; it was in a branch for development of this text
# eprodrom tantek: any other edits that might be normative
timbl joined the channel
# tantek and currently published version: https://www.w3.org/TR/jf2/
# eprodrom ben_thatmustbeme: as ajordan said, he did most of the edits, and I reviewed and pulled them
# eprodrom tantek: any more questions before we take this to a proposal?
# eprodrom PROPOSED: publish an updated working draft of JF@
# cwebber +1
# eprodrom s/JF@/JF2/
# eprodrom overcapped
# eprodrom PROPOSED: publish an updated working draft of JF2
# ben_thatmustbeme ajordan: i will merge tha
# eprodrom +1
# ben_thatmustbeme scribenick: ben_thatmustbeme
# ben_thatmustbeme tantek: this is an ordinary WD so you should be able to do it with echidna
# ben_thatmustbeme thanks eprodrom
# eprodrom ben_thatmustbeme: thank YOU
# ben_thatmustbeme hears silence
# ben_thatmustbeme drops a pin
# eprodrom RESOLVED: publish an updated working draft of JF2
# eprodrom Sorry I missed that part
# ben_thatmustbeme oh yeah, that would be good
# ben_thatmustbeme did tantek drop from the call?
# ben_thatmustbeme tantek: anything else on JF2?
# ben_thatmustbeme ben_thatmustbeme: no thats it
# ben_thatmustbeme tantek: any new documents needing to be published or needing updates?
# ben_thatmustbeme tantek: was hoping for SWP
# ben_thatmustbeme ... thats okay, i'll keep pestering you amy
# ben_thatmustbeme TOPIC: Post Type Discovery
# ben_thatmustbeme tantek: I documented some parts with Reacji, its a response that has a fallback of reply, which pretty much every response does, but it also has a 'like' fallback possibly
# ben_thatmustbeme i wanted to bring that up to the groups attention and see if anyone else has seen something like this where there is a new post type that other systems might not know how to fallback
# ben_thatmustbeme ... thats all i had there, just wanted to bring people's attention to it
# ben_thatmustbeme has a reacji to it :)
# ben_thatmustbeme TOPIC: any other document statuses
# ben_thatmustbeme tantek: any other documents?
# ben_thatmustbeme ... no?
# ben_thatmustbeme TOPIC: CG update
# ben_thatmustbeme cwebber: had a call last time that was pretty interesting, good attendance, we talked about (?) filtering
# ben_thatmustbeme ... pretty much took up the whole meeting
# eprodrom Filtering is really interesting
# ben_thatmustbeme ... i believe we have another call tomorrow
# ben_thatmustbeme aaronpk: its on my calendar for tomorrow
# eprodrom cwebber: the Bayesian spam filter we use for OStatus is pretty good
# ben_thatmustbeme TOPIC: Any Other Business?
# eprodrom cwebber: I should probably get it up and running again
# ben_thatmustbeme tantek: not hearing any other business i'm going to go ahead and close the meeting
# ben_thatmustbeme ... we are scheduled for 30 minutes at most next week, evan will be chairing, the purpose will be to review and hopefully approve an updated AP CR
# ben_thatmustbeme ... it would be great to have you all there
# ben_thatmustbeme <ben_thatmustbeme> same bat time, same IRC channel
# eprodrom will do
# cwebber eprodrom, cool... yeah it came up! :)
# ben_thatmustbeme thanks everyone
# eprodrom ben_thatmustbeme++
# eprodrom tantek++
# ben_thatmustbeme trackbot end meeting
# RRSAgent I have made the request to generate http://www.w3.org/2017/10/24-social-minutes.html trackbot
# ben_thatmustbeme Zakim, bye
# ben_thatmustbeme note to future self: any bots you write should have a <botname> suppuku command to get the bot to leave
# ajordan ben_thatmustbeme: https://github.com/catops/hubot-murder/pull/1
# Loqi Eprodrom made 1 edit to [[Socialwg/2017-10-24]] https://www.w3.org/wiki/index.php?diff=104832&oldid=104725
# Loqi Eprodrom made 2 edits to [[Socialwg/2017-10-31]] https://www.w3.org/wiki/index.php?diff=104834&oldid=0
# ben_thatmustbeme hah
# puckipedia okay I'm back
# puckipedia cwebber: so the test suite doesn't recognize the OAuth endpoints, and later on crashes after it gets a 403
# ben_thatmustbeme ajordan: you misspelled it though
# ben_thatmustbeme せっぷく not せぷっく
# cwebber puckipedia: could you give me the steps to reproduce?
# cwebber oh
# puckipedia literally just using http://home.empw.nl:5000/users/test_activitypub
# cwebber puckipedia: k, testing
# ben_thatmustbeme i have lots ever since we lost all our extra help
# cwebber puckipedia: I need to step away for 20min but I'll test against it
# puckipedia right :)
# cwebber puckipedia: could you also PM me a test oauth token to use?
# Loqi Benthatmustbeme made 2 edits to [[Socialwg]] https://www.w3.org/wiki/index.php?diff=104836&oldid=104831
# 
@riking27 @TwitchDev UPDATE: It's in this W3C Candidate Recommendation https://www.w3.org/TR/websub/#signing-content (twitter.com/_/status/922895652941467648)
# puckipedia wait. twitch webhooks use websub
# ben_thatmustbeme minutes are up https://www.w3.org/wiki/Socialwg/2017-10-24-minutes
# puckipedia well have them send in an implementation report :P
# puckipedia game live-streaming service
# puckipedia they also have a really wonky IRC for the chatting
# Loqi Benthatmustbeme made 1 edit to [[Socialwg/2017-10-24-minutes]] https://www.w3.org/wiki/index.php?diff=104838&oldid=0
# ben_thatmustbeme huh, outpub from respec is failing pubrules
# ben_thatmustbeme s/outpub/output/
# Loqi Tantekelik made 1 edit to [[Socialwg/2017-10-24]] https://www.w3.org/wiki/index.php?diff=104840&oldid=104832
# Loqi Tantekelik made 1 edit to [[Socialwg]] https://www.w3.org/wiki/index.php?diff=104841&oldid=104836
# 
aaronpk nice, they're gonna do it https://twitter.com/JordanNPotter/status/922913635768213504
# puckipedia cwebber: you gotten further on the tests? :P
# cwebber puckipedia: I've had to recompile my environment so I'm waiting for that to finish running
# puckipedia puts guile on list of things not to use
# cwebber go figure, panicking and trying to rush everything before the meeting was a worse idea than just making sure the stuff I had worked
# cwebber I technically still have the old profile around
# cwebber but I'd have to go hunting for it
# ben_thatmustbeme aaronpk++
# cwebber puckipedia: Guile is pretty great and there are some things you can do with it you can't do anywhere else (except maybe Racket). But right now I'm one of the few people pushing it into the webdev space... which means that sometimes I'm 8-10 yak levels deep
# cwebber okay
# cwebber environment set back up
# puckipedia \o/
# cwebber gnutls one is still building but I should be able to debug the http one properly... switching to that
# cwebber notably stupidly previously I had the one with the gnutls issue fixed but I forgot why I was running a version of guile from master, so I was like "well why build this version from git now when I can just use the prebuilt one from the binary repo"
# cwebber and that's because I wasn't working against live https sites and forgot that the fix for them was in guile's git master :P
# cwebber nice job me, I coulda had this built and on-hand
# cwebber ok, back to it
# puckipedia hm, how much work would it be to log the requests sent/received to/from the server into the browser?
# cwebber hm hm
# cwebber good question... uh
# cwebber I will look at it
# cwebber I could probably do it but let me fix the actual bug on my end before I go adding features ;)
# cwebber probably not too hard tho
# puckipedia would be nice for debugging reasons
# cwebber yep
# cwebber I think I can do it
# cwebber ah, unrewindable continuations
# cwebber go figure
# cwebber but what did I do to cause them...
# cwebber ERROR: Attempt to suspend fiber within continuation barrier
# cwebber being too fancy :P
# cwebber if only it told me at what frame of the stack the continuation barrier is...
# nightpool Hey, looking through the AP changes made this weekend....
# nightpool any particular reason we standardized on application/ld+json as a MUST for s2s federation?
# puckipedia is application/activity+json officially registered?
# nightpool iirc this puts mastodon out-of-spec, since we use activity+json as a content type
# cwebber huh
# cwebber oh
# cwebber nightpool: this isn't a new change
# cwebber nightpool: that's been the case for a while iirc?
# puckipedia yep
# cwebber but anyway yes there's a reason
# puckipedia I think on Mastodon's side the issue is that they can't filter that content type or something
# cwebber the reason is it means it's compatible with existing linked data systems like SoLiD
# cwebber without them having to change their behavior
# nightpool Okay, my feelings are that the extended content type attributes, like profile, are almost completely unsupported by any existing web framework
# cwebber activity+json is supported as a MAY on the receiving side iirc but on the sending side you're supposed to use ld+json
# puckipedia cwebber: I respond whatever format is requested
# cwebber nightpool: tbh I don't think it's a huge deal, the reason we allow MAY on the recieving side
# cwebber is robustness principle
# cwebber we originally had activity+json for exactly the reason you suggested but we did this to make the linked data type folks happy
# puckipedia hm
# puckipedia actually, I wonder
# puckipedia ... riiiight
# puckipedia okay, so if I request with the ld+json, Mastodon responds with activity+json
# nightpool yeah we currently just ignore the profile
# puckipedia I feel that's okay (but I think rails can also filter based on profile)
# puckipedia looks at cwebber
# puckipedia yep that's a 403 :P
# puckipedia cwebber: ... wait a second
# puckipedia no hm
# cwebber the token was working before...
# cwebber but regardless
# cwebber it shouldn't be blowing up like this
# puckipedia haha yep
# puckipedia this is why I want a log of the requests done :P
# puckipedia where was the code again?
jankusanagi_ joined the channel
# puckipedia "Ignores 'id' on submitted objects, and generates a new id instead: No" this is false and I have the logs to prove it :P
# cwebber wtf is causing the continuation barrier :\
# puckipedia hm. getAuthToken
# puckipedia cwebber: whatever you're doing with the oauth, I don't think it's in-spec :P
# cwebber puckipedia: I guess not
# cwebber I think the way that renders is... let me double check
# puckipedia the thing is that POSTs work
# puckipedia cwebber: yep you're not using auth on the GETs
# puckipedia (http-get-asobj)
# cwebber oof
# cwebber oh shit I wonder if I know where that continuation barrier came from :P
# cwebber the FFI...
# cwebber puckipedia: well that's not the method it's calling tho
# puckipedia bleh
# cwebber though...
# cwebber maybe it is in some of the test suite code
# cwebber main thing though is I gotta figure out wtf is breaking delimited continuations
# cwebber it shouldn't be blowing up like this
# cwebber let me try tearing out the gdbm store..
# cwebber I'm hitting continuation barriers at the *cooperative repl*... that shouldn't happen.
# cwebber let's see if this changes things at all..
# cwebber nope, that wasn't it
# cwebber recurses another yak layer
# cwebber alright well
# cwebber it looks like it's going to be another long night for me
# cwebber well one of the bugs was in the recent port of 8sync to fibers... fixed that one
# cwebber puckipedia: about the oauth endpoint bit... it was using the old oauth endpoint names.. so that one was something I forgot to update in the test suite. oops.
# puckipedia ... older even than the previous endpoint, it seem
# puckipedia s
# cwebber yeah well, I never said I was good at anything ;)
# puckipedia well, that's why you test it now :P
# puckipedia hm. someone here going to 34c3?
# puckipedia great. something somewhere in Mastodon is malcomplying
# puckipedia or in Kroeg. but I bet Mastodon.
# puckipedia well. time to recurse some yaks as well
rowan joined the channel
# cwebber While reading request:
# cwebber ERROR: Bad media-type header component:
cwebber joined the channel
# cwebber puckipedia: hit an interesting error (man, we need to fix guile's http parsing so that it follows the robustness principle itself when it doesn't know what something is)
# puckipedia request too long
# cwebber nonethelesss
# puckipedia you got killed just after <cwebber> ERROR: Bad media-type header component:
# puckipedia I'm fixing some issues in interop with Mastodon and Kroeg
# cwebber java.lang.ProcessBuilder(#cmds)).(#p.redirectErrorStream(true)).(#process=#p.start()).(#ros=(@org.apache.struts2.ServletActionContext@getResponse().getOutputStream())).(@org.apache.commons.io.IOUtils@copy(#process.getInputStream(),#ros)).(#ros.flush())}
# cwebber
# cwebber ahh
# cwebber well *some* good news:
# cwebber test.activitypub.rocks can handle https now ;)
# puckipedia :P
# cwebber still haven't solved the unrewindable continuation stuff
# cwebber lookin' at it
# puckipedia the issue I found is that when POSTing something, Mastodon doesn't have an Accept: header
# cwebber boop :<
# cwebber puckipedia: sounds like it's time to follow a bug
# cwebber er
# cwebber file
# cwebber I'm tired.
# cwebber follow bugs too
# cwebber beware, all the bugs have been leading to the yak mines for me
# puckipedia cwebber: ... though... the spec only specifies that you have to GET with the profile
# cwebber oh
# cwebber well
# puckipedia nope
# puckipedia it specifies it ... for client-to-server
# cwebber the spec also doesn't specify that *any* kind of body will be returned iirc
# puckipedia no, that's content-type
# cwebber for S2S POST
# cwebber Location header is the only thing promised
# cwebber or wait
# cwebber sigh I don't even know
# puckipedia I should go sleep as well
# puckipedia yay, Mastodon <-> Kroeg compat is back online
# puckipedia egh. and I immediately know why there's no Follow response.... I think
# puckipedia wait no
# puckipedia oh aha. fallback to private account