XSF Discussion - 2017-10-10

What I tried to say was what happens when MAM is implemented to a server?

Assuming there's an accompanying server-side message archive, then users can query it.

guys, when pubsub#publish-options form definition will be fixed?

or, if everyone is lazy doing this, where can I find definitions of every fields of pubsub#publish-options form

I recall Push, Persistent Storage, what else?

zinid, make a PR? ;-)

but I'm lazy

zinid: what is broken?

Anyway, I wonder if MIX shouldn't really be PubSub2. Persistent group chat requires PubSub at the lowest level, MIX could provide that, while additionally fixing PubSub for everyone else (while keeping the old spec around)

Flow: pubsub#publish-options doesn't define pubsub#persist_items defined in XEP-0223 and 'secret' (without pubsub# ???) defined in XEP-0357

define where? in xep60? don't we have a registry for that?

yes

in xep60

I have no idea what "having a registry" means

https://xmpp.org/extensions/xep-0060.html#registrar-formtypes-publish

https://news.ycombinator.com/item?id=15437448

zinid, this one I think: https://xmpp.org/registrar/formtypes.html#http:--jabber.org-protocol-pubsubpublish-options

still not defined there ;)

right, PRs welcome 😉

also, this registry is unusable, it doesn't define possible values of pubsub#access_model

zinid, it just defines the fild name, everything else is written in the defining XEP

field name, type and description

ok

so should be fixed both XEP-0060 definition and the registry?

zinid, I found persist_items in the registry. what else is missing?

An 'secret' from the push xep

Seems like 'push#secret' would be more idiomatic, also xep357 doesn't have a registry submission for that

If i recall correctly, Kev wanted to take of xep357

*take over

Flow: persist_items should be defined in every form where it's used, just like pubsub#access_model (it's defined in several forms)

what are the other forms where persist_items is used besides node_config?

publish_options?

see XEP-0223

https://xmpp.org/extensions/xep-0223.html#store

example 4

possibly in others, dunno

I'm not sure if it's righ to use persist_items with publish-options

like access_model it appears more like a per node thing, not per item

could it be that ex1 in xep223 is wrong?

Flow, no, that was discusesd on the ML

jonasw, ahh, must have missed that, do you have a pointer?

nobody objected to the interpretation that <publish-options/> is a precondition

will do

ahh, it's used as precondition

then it probably makes sense that it doesn't appear in publish-options's registry

now I remember that there was a discussion around that

ahh and xep60 reads like preconditions must also be registered

everything should be registered, even if XEP doesn't say that

because this form definition can be used by automatic validation (I do this in ejabberd)

still looking for the discussion, hang on

jonasw, no worries, I think it was daniel who brought it up

any French people around?

Flow, https://mail.jabber.org/pipermail/standards/2017-May/032631.html

got it

Moi

yes, preconditions must be registered to define the meaning of the precondition

in a certificate, what's the most appropriate value for "state or Province Name" for Paris?

Guus: does anyone care about the value of that field?

Île-de-France? Nothing?

Keep that empty. Nobody cares.

Flow: I don't want to find out, which is why I'm asking for the most appropriate value :)

kk

But probably it is that... See other certs...

The one of OVH...

(I don't know of any other certs)

oh, paris probably has a website...

this is how the gandi.net certificate looks: CN = www.gandi.net OU = COMODO EV SSL OU = Ops O = Gandi SAS Object Identifier (2 5 4 9) = 63-65 Boulevard Massena L = Paris ST = Paris Object Identifier (2 5 4 17) = 75013 C = FR Object Identifier (2 5 4 15) = Private Organization Object Identifier (1 3 6 1 4 1 311 60 2 1 3) = FR Object Identifier (2 5 4 5) = 423 093 459 00042

so apparently it’s simply "Paris"?

Lol

thanks guys :)

Certbot keeps that empty. Really nobody cares.

Flow: whatever, I generate code from this form now: https://github.com/processone/xmpp/blob/master/specs/pubsub_publish_options.xdata

secret without prefix 'pubsub#' looks odd though

I think 'push#secret' would be more to the point

I agree

zinid, would be greate if you did an PR for the registry too

can we change the filed name?

https://github.com/xsf/registrar

zinid: Not without a namespace bump I fear

then let's don't do this

we have the same in some MUC forms, this is not a major problem

we should have errata sections in XEPs where we can note things for the next Namespace bump

re registrar: I bet that stuff is currently not updated on the website

jonasw, I once started https://wiki.xmpp.org/web/XEP_and_RFC_Remarks

yeah, nobody will look there when preparing a PR against a xep :)

jonasw [09:56]: > we should have errata sections in XEPs where we can note things for the next Namespace bump There used to be wiki pages for each xep containing that information

Who's going to be at the GSoC menter summit this weekend? Arc, Daniel, Edwin and me - anyone else?

we should meet up. I'll at least bring some stickers :)

Edwin and me are in the Sheraton

Guus: Bring stickers *to* XMPP.

?

Guus: +1 on the stickers.

> I'll at least bring some stickers :) 😉

The requests for adding stickers...

mimi89999: ah.

Please note that my two-year-old kept me up all night, and I'm very likely to miss puns, sarcasm, and any other subtilities today.

Guus: Take some sleep then...

mimi89999: customers, deadlines.

my wife is actually very jealous of me going to the GSoC meeting. Not because of the trip, but because of the four days of hotel stay without the kids :)

four nights of uninterrupted sleep... sheer bliss.

Four nights of fighting jetlag ;)

intosi: that probably sets in only after I return

2x no sleep...

I'm usually much better going west to east than the other way around. Getting up early is easy, sleeping in is something I'm terrible at.

I'll find out in a couple of days

oh, at least 4 guys from Jitsi will be there too

they're in the other hotel though

Guus: If you meet Ingo, then please say hello from me

Flow: not sure if he'll be there. by-the-by: the Jitsi guys all but finished merging all of his Smack 3->4 miration work.

\o/

it was a lot of work on Ingo's part, but the merge itself was relatively painless, from what I understand. There was one issue with certificate acepptance that they quickly resolved.

not that it matters: but it appears to be not yet merged: https://github.com/jitsi/jitsi/tree/smack4-2

that's the desktop app I think? I think they were referring to their backend stuff.

ahh, k, great