jdev - 2021-10-01

Zash, what would be the practical advantage of such an approach?

Zash, flow, see https://github.com/xsf/xeps/pull/682

Link Mauve, lgtm

flow: of what approach? file transfer stuff on the same host? single thing to have A+AAAA records for http, proxy, etc and grouping of related things.

Zash, why not steer towards having everything on the same xmpp address, why the grouping?

Because it needed to be separate from the user host for those compat reasons, and what I said above.

But I've since moved everything to the main host.

Having a host matching the actual machine name and the SRV target meant it had AAAA+A already, so no need to advertise a different hostname for http, proxy connection. And I had thoughts of having those things on a machine with more availabe bandwith.

Apparently Gajim's nbxmpp doesn't support sha-1 hashes (particularly for Jingle), and <https://dev.gajim.org/gajim/python-nbxmpp/-/blob/master/nbxmpp/protocol.py#L1600> mentions it as "SHOULD NOT", but does that come from something mentioned or referenced in the XEP (which I'm failing to find)? From just reading the XEP, I thought it's a valid algorithm for use with that XEP.

XEP-0300, I mean. Forgot to mention.

i'd probably support at least receiving sha-1 for now

and sending them, together with values of newer hash algos can't hurt either (if possible by the protocol, which I think it is)

defanor, btw the source of the SHOULD NOT is https://xmpp.org/extensions/xep-0414.html#recommendations

"This specification is not meant to override recommendations or requirements laid out by other specifications. "

Thanks, at least now it's clear where that came from.

Gajim's error message was confusing too; would be nicer if it mentioned an unsupported hash algorithm, but it just says that the file was possibly damaged, and suggests to reload it. Not sure if I should open an issue, or it's enough to poke lovetox here.

no please open and issue ✏

Too late, poked already. But now created an issue too.

😂