XSF Discussion - 2022-01-16

https://xmpp.org/software/clients/ no feature comparison matrix?

No we hardly keep a list

moparisthebest: so a documentation PR would be rejected

moparisthebest: are you and your role mentioned on site? Many use real name yet xmpp handles is mostly what I see in rooms

Is there a more up to date this anywhere? https://xmpp.org/schemas/index.shtml

You can propose whatever, just seems there's been a lot of pushback for doing the bare minimum, something something just a standards org

Also lol it's been so long since i've seen a shtml

Ha yea good times qy

moparisthebest: matrix clients have good feature comparison

Was useful

qwestion: matrix is a company that doesn't do standards, just pushes marketing to get those investor dollars

There's the WIP https://github.com/xsf/xmpp.org/pull/986

Older render: https://linkmauve.fr/software/clients.html

Bunch of discussion in https://github.com/xsf/xmpp.org/issues/212

qwestion: this unfortunately named website has some feature comparisons https://joinjabber.org/clients/gnulinux/

Having attempted to make a feature comparison table once, I can say that it's a lot of not very rewarding work to keep them up to date and stop everyone from turning it into marketing vehicle for their particular software.

It gets outdated pretty quickly and then people get either misled or unpleasant about it.

Also hard to compare things sometimes. And comparing XEP numbers isn't really optimal

E.g. where might i find a theoretical xsd for the old omemo spec

But also, every other xep

https://github.com/xsf/xeps/ ?

https://github.com/xsf/schemas ???

> Having attempted to make a feature comparison table once, I can say that it's a lot of not very rewarding work to keep them up to date and stop everyone from turning it into marketing vehicle for their particular software. > It gets outdated pretty quickly and then people get either misled or unpleasant about it. Wiki where changes get approved?

What, create more bureaucracy? That'll just make it even more outdated.

And clients devs/reliableenthusiasts are expected to fill their own clients details ✏

The current idea is that client projects get to fill in their stuff in files that they serve, that can be collected and mangled into whatever you want.

See https://github.com/pulkomandy/xmpp-doap

> Zash wrote: > https://github.com/xsf/xeps/ ? > https://github.com/xsf/schemas ??? So, latter has e.g. no mention of omemo/axolotl, and former... i _guess_ i could parse out schemas in a hideous manner, but that seems evil

I presume its only missing from the latter because experimental?

https://github.com/xsf/schemas/issues/5 oh

>2019

Wow

```curl https://xmpp.org/extensions/xep-0384.xml | xq '.xep.section1 | map({(.["@anchor"]): .cod e}) | add | .schema' -r``` teehee

> Having attempted to make a feature comparison table once, I can say that it's a lot of not very rewarding work to keep them up to date and stop everyone from turning it into marketing vehicle for their particular software. > It gets outdated pretty quickly and then people get either misled or unpleasant about it. How Wikipedia manages to make so many sw comps? Maybe it should be done there too, as many clients don't have a page? Not notable enuf?

Zash: if i used that to dump a schema for every xep under the sun, and shoved it in a PR, would it have any hope of going anywhere?

It "works", i just compared 203 to delay.xsd from the schemas page

Any client where I can "select all" msgs or shift highlight many msgs?

Perhaps privacy friendly way to translate

qy, only one way to find out

I've dumped them, i'll keep them to myself unless anyone's interested

Not my department anyway, not the one to ask.

To have XMPP URI support on GitHub, please take your voice too: - https://github.com/github/feedback/discussions/10162

qy: why not make a PR? ✏

wurstsalat: i feel like it'll snowball and i'll end up spending way more time and effort than i intend. That's normally how nontrivial PRs go for me and it's kinda put me off them...

My intention was mainly just to see if i could autogen c++ bindings for every xep and hence have somewhat much neater code

Turns out i can autogen, but it's not exactly neat

qy, C++ client library?

Client app, but yeah i was thinking of breaking that bit off into "libXEPs" or something

qy, Which one?

weechat-xmpp

Which part of standard describes it's wrong to to take into account resolved via SRV hostname while validating xmpp-starttls certificate? I see Conversations and Gajim ignore the hostname during validation.

You don't have a trust chain between something the user entered (jid domain) and the hostname

Unless you do and then Conversations does take it into account

rion, https://xmpp.org/rfcs/rfc6120.html#security-certificates-validation-server

rion, if an attacker is able to spoof the SRV records, then clients using the target of the SRV record to validate the TLS certificate would easily authenticate against an attacker controlled service

true.

ralphm - I wrote you a direct message. Did you receive it?

ah nevermind, I will reach out to board via mail directly

rion: https://wiki.xmpp.org/web/The_Knight