#social 2017-07-26

2017-07-26 UTC
#
ajordan saranix: huh?
#
ajordan I'm guessing in this case Hubzilla can say A and B and silos can only say A? and they're complaining that people think being able to say A is really amazing?
#
saranix ajordan: pretty much :)
#
ajordan lol
#
ajordan I have a lot of thoughts on that but I don't want to rant so I'll shut up instead
xmpp-social, timbl and csarven joined the channel
#
jaywink > seeing just {Object} come across the wire for the first time, as the top level object, is effectively the same as {Create {Object}} <---- so much ++++ about this
#
jaywink If I got an update for an object I don't have, it would feel logical to process it as a create. All these create/update so on are artificial since it always depends on what the receiver has. It might make sense to show activities someone did to know the verbs, but for distributing content they're meaningless
#
puckipedia jaywink: well, a S2S update is basically "hey here's that object again, just fyi"
#
jaywink though, it does matter when you're doing a Delete to say explicitly Delete, so in that respect always specifying the activity would be consistent.
#
puckipedia the question is: can you Delete an Activity?
#
puckipedia or should an Activity be always undone / (inverse: create vs delete, add vs remove)
#
jaywink can you Create a Delete?`;)
#
puckipedia {"type": ["Delete", "Note"], "object": "toDelete", "content": "woops, violated copyright laws", "actor": "user", "attributedTo": "user"}
#
jaywink shouldn't the Note be wrapped in a Delete instead? {"type": "Delete", "object": {"type": "Note"}}
#
puckipedia I mean, this is a Delete and a Note at the same time :P Should you Create it or just post it as is? :P
#
puckipedia it deletes another note
#
jaywink that looks confusing to me. sounds like it should be two different "payloads", one for the {delete {note}} and one for the new note {create {note}} - at least I've thought of things as "one activity + one object"
#
puckipedia hey it's allowed by the spec :P
#
puckipedia it's a really bad idea but nothing's stopping it from existing
#
jaywink hmm which part?
#
puckipedia well, it's got two types
#
puckipedia which probably shouldn't be mixed in AP objects
#
puckipedia I also thought of a {"type": ["Create", "Note"], "object": "example.com/obj", "id": "example.com/obj"}
#
puckipedia (... is this the equivalent of an implicit Create in OStatus? D:)
#
puckipedia ehm, implicit create in AS1 Atom
#
jaywink where does AS2 say "type" can be a list?
#
puckipedia JSON-LD does
#
puckipedia https://www.w3.org/TR/json-ld/#specifying-the-type
#
puckipedia also it's mentioend in AS2 core, with an extension: "type": ["Place", "gr:Location"],
#
jaywink that would be very confusing since all the spec examples make type reference just one type. if this is to be supported it should be clarified in the spec that type could be multiple types. but how do you even process that? would be better to enforce that type should be one type?
#
jaywink " Identifies the Object or Link type. Multiple values may be specified. "
#
jaywink ok it's a part of AS2 spec itself
#
jaywink so if you have {create, delete {note, note}} how the hell do you parse that? ;)
#
puckipedia I assume the type array is a JSON-LD set so it has to be unique
#
jaywink imagines the fun parser writers are going to have :P
#
puckipedia I have a magic "first type" which is used for some unique processing
#
jaywink still, AP could enforce a single type
#
jaywink to simplify things. simple == adoption
#
puckipedia but then yo ucan't add extensions :P
#
jaywink oh hmm, well, good point :D
#
jaywink goes hide back in his corner
#
puckipedia jaywink: tbh, it might be possible to say "there may only be one activity type per object", so you can't do Create/Delete, but you can still do Create/Note if you reaally want to
#
jaywink Kind of annoying for parsing still but *shrugs*
#
cwebber2 moin moin
#
cwebber2 I think Update and Delete make sense as activities sent across the wire btw
#
cwebber2 I just think Create is a PITA
#
puckipedia cwebber2: well, an Update is for clients to parse too
#
cwebber2 yep
#
cwebber2 it conveys a lot more information about change
#
cwebber2 so!
#
cwebber2 call today in a little over an hour, eh?
#
cwebber2 puckipedia: btw I tested your site against the test suite
#
cwebber2 you may have seen from the logs but there's a language issue I have to fix :)
#
cwebber2 not on your end
#
puckipedia ah, didn't look at the logs
#
cwebber2 though! I can tell you a couple fast things you can fix already from the output I got :)
#
cwebber2 - should be returning 201 instead of 200 on posts to outbox
#
cwebber2 - add a Location header to the response too
#
puckipedia I had those two :P
#
cwebber2 really?
#
puckipedia *had*
#
cwebber2 maybe my tests were wrong. at any case, I'll have the actual thing up as soon as I add async support to https :P
#
puckipedia well, it might've disappeared as a feature
#
cwebber2 I guess that's why test suites are helpful :)
#
cwebber2 https://www.w3.org/wiki/SocialCG/2017-07-26
#
cwebber2 add topics?
#
cwebber2 and cya in an hour :)
#
puckipedia ehm it's kinda segfaulting when compiling :<
#
puckipedia because curl on debian is too new
#
cwebber2 uhoh
#
puckipedia annoyingly this is a Known Issue
#
cwebber2 wait why do you need curl to compile?
#
puckipedia libcurl
#
puckipedia and it's just dotnet that's crashing
#
puckipedia okay, I think it /should/ be fixed, and /should/ reply with 201
#
puckipedia IT SEGFAULTED
#
cwebber2 lol did I kill your server?
#
puckipedia libcrypto
#
puckipedia exactly when doing the json web key thingy
#
cwebber2 but 201 and Location now passing
#
puckipedia "The reporter for that one gave information which suggests that their system has both libcrypto.so.1.0.0 and libcrypto.so.1.0.2, and that libcurl uses the latter, we use the former."
#
cwebber2 sucks
#
cwebber2 test output btw http://sebsauvage.net/paste/?0c6b0876c02316ff#uCos1CU7quFXKMZuJqgA9mVMIGyg/YCY1Ms9QuNovuQ=
#
cwebber2 though a few of them have tests that involve addressing that will probably fail being my fault, not yours
#
cwebber2 and I don't know why the uploadMedia ones didn't run
#
cwebber2 I'll have to look at that
#
puckipedia maybe if it returns 50X it should probably just mention 'server error' instead of failed?
#
cwebber2 or it could at least say the error code returned
#
cwebber2 sorry for crashin' your stuff puckipedia !
#
puckipedia well, I'm surprised it worked as far
#
cwebber2 puckipedia: are you trying to run the MS .NET compiler on Debian?
#
cwebber2 seems tricky indeed
#
puckipedia too new debian
#
cwebber2 aha
#
cwebber2 (iirc you said you sere using newer things than Mono supported)
#
cwebber2 props for being on Debian though :)
#
puckipedia yeah, Mono will some time in the future be replaced completely by .NET Core
#
cwebber2 puckipedia: oh really?
#
puckipedia cwebber2: I ran alpine linux but that doesn't work at all because musl
#
puckipedia cwebber2: well, .NET Core is now cross-platform in itself
#
puckipedia okay, maybe Mono won't be completely gone for a while, because Xamarin
#
cwebber2 puckipedia: come to the guix side, we have reproducibility ;) (okay, and fewer packages atm)
#
cwebber2 but you can always roll back a bad upgrade so there's that :)
#
puckipedia okay I kinda just downgraded curl
#
puckipedia (it passes something from curl to openssl, or reverse, and that breaks it because libcrypto 1.0.0 is used by .NET and 1.0.2 is used by libcurl)
#
puckipedia cwebber2: back up and running :P
#
cwebber2 ran it again, did it kill your server? :)
#
puckipedia it did not
#
puckipedia it's still running
#
cwebber2 got two new passes now
#
cwebber2 Accepts Uploaded Media in submissions: OK!
#
cwebber2 Responds with status code of 201 Created or 202 Accepted as described in 6.: OK!
#
puckipedia https://puckipedia.com/images/1494446b-f074-4e42-aa42-cba86c92027b.png :P
#
cwebber2 yup :D
#
puckipedia wonder if I should put a sha256 hash in the url instead
#
cwebber2 I was lazy and used the default pubstrate icon
#
cwebber2 https://dustycloud.org/tmp/pubstrate_2016-08-17.png
#
cwebber2 old screenshot but still
#
cwebber2 fun times puckipedia
#
cwebber2 I'm gonna go shower in the meanwhile
#
cwebber2 beep beep
#
cwebber2 time for that call, eh?
#
cwebber2 where'd I put my headphones...
#
cwebber2 dialed in
#
ajordan cwebber2: heya!
#
cwebber2 ajordan: just you and me, huh? :)
#
ajordan I heard you say hi but I can't remember what key I bound to push-to-talk
#
cwebber2 lol
#
ajordan sighs
#
ajordan ironically I am exhausted because I stayed up late doing non-W3C standards stuff
#
puckipedia oh woops, one sec
#
puckipedia been doing some school stuffs
#
puckipedia apparently tudelft@puckipedia.com is a bad recovery email because it contains tudelft
#
ajordan wtf
#
ajordan apparently my push-to-talk key is right Super
#
cwebber2 lol
#
ajordan ...but I have no right Super
#
ajordan pulls out another keyboard
#
ajordan cwebber2: do we actually have any business?
#
aaronpk waves
#
ajordan waves to aaronpk
#
ajordan cwebber2: wait what?
#
cwebber2 Create
#
cwebber2 not create ;)
#
ajordan I thought the test suite was done except for the language hacking you had to do lol
#
cwebber2 :)
#
ajordan lol thanks all
#
ajordan goes back to bed for a bit :-)
#
cwebber2 :)
tantek joined the channel
#
tantek good morning #social
#
tantek did we skip CG telcon today?
#
tantek sees 0 people on the mumble
#
puckipedia tantek: we did
#
tantek cool cool. back to other things. enjoy your Wednesday!
#
aaronpk we met for 5 minutes and decided there wasn't anything we neededf to talk about :)
#
cwebber2 :)
#
xmpp-social [ajordan] I'm really happy we did honestly
#
xmpp-social [ajordan] I felt like a vegetable
timbl and ajordan joined the channel