jdev - 2024-09-03

When testing IPv6 and IPv4 functionality, what approaches do you take? Does something like a container-based server/test environment sufficiently work? I'm guessing that experimenting with various DNS-based settings is hard, that way?

Testing something like Happy Eyeballs gets complex, quick, I suspect.

My approach to testing IPv6 stuff is very simple. I say "Edwin, you understand this stuff, please sort it" and it gets sorted. :D

Hey, Edwin!

An answer that is truthful, accurate, and completely unhelpful. My favourite kind :)

If I am "oldnick" in a MUC, and try to rename myself to "newnick" but "newnick" is invalid, should the presence error be from="room@component/oldnick" or "room@component/newnick". I'm asking because prosody uses .../newnick, but gajim expects .../oldnick, and I don't know who is right here? Related: https://dev.gajim.org/gajim/gajim/-/issues/11930 ✏

Hi oldnick ! I'm edhelas ✏

edhelas does not look like that, I've seen the movies. FAKE NEWS.

Now my question is buried and no one will answer. Are you happy? Do you see what you have done? All for what? Laughing? How do want your country to be proud if you enjoy laughing? Tssssss

> When testing IPv6 and IPv4 functionality, what approaches do you take? Does something like a container-based server/test environment sufficiently work? I'm guessing that experimenting with various DNS-based settings is hard, that way? Guus: yes over a container network and yes with various DNS records by running a bind9 in that network like https://github.com/moparisthebest/xmpp-proxy/blob/master/integration/15-s2s-srv-record-tls/example.org.zone But it's true I can't think of a good way to test happy eyeballs this way... ↺

> If I am "oldnick" in a MUC, and try to rename myself to "newnick" but "newnick" is invalid, should the presence error be from="room@component/oldnick" or "room@component/newnick". I'm asking because prosody uses .../newnick, but gajim expects .../oldnick, and I don't know who is right here? Related: https://dev.gajim.org/gajim/gajim/-/issues/11930 Examples are not normative, but if they were, <https://xmpp.org/extensions/xep-0045.html#example-52> would mean that prosody is the culprit. Maybe the XEP should make that more clear? ↺

Logically, error stanzas are sent with the to/from flipped and a matching id.

So applying that consistently here would mean that the error should come from the newnick, as that's what the presence was sent to.

So you are saying gajim should not do that <https://dev.gajim.org/gajim/gajim/-/blob/d5b91cc6a7b5e814996bfd4e6607d1a9f0513ae7/gajim/common/modules/muc.py#L509>: ```python if properties.jid.resource != muc_data.nick: self._log.warning('Unknown error presence') self._log.warning(stanza) return ```

I wonder why that check is here. Is there a way to flood a muc with presence errors? Or maybe some disfunctional MUC component out there that send presence errors that they shouldn't?

I made slidge behave like gajim expects a while ago, but maybe both gajim and slidge are not respecting the XEP <https://git.sr.ht/~nicoco/slidge/tree/aedae1ce2622eb64d28dca364dd0c26bd185edcb/item/slidge/core/dispatcher/presence.py#L157> In any case, it would be nice if this was made explicit in the XEP I guess. (and examples updated)

nicoco__, there is no specific reason why this is there in Gajim

I just dont like to process unexpected stuff, and hope it goes well ✏

when i wrote this code i saw no reason why we would ever receive a presence error not from ourself

of course i didnt think about the abomination that is nick change in muc :D

but on that topic, its weird that prosody returns a "invalid-jid" but has no problem to set this invalid-jid as a "from" attribute?

to me that would a violation of the base rfc, server should not send stanzas with invalid jids

lovetox: I agree, generally. And if you're well-behaved, it won't. But if you are not well-behaved (you start trying to send stuff to invalid JIDs) then this seems like the least-worst way to handle it

MattJ: yes I agree it should be fine