On 2018/03/22, Matthew Wild wrote:
> We're discussing the protocol, but there is nothing stopping clients
> having their own overrides (i.e. local autojoin rooms). This could be
> as simple as, when you join a room for the first time "Do you want to
> join this room on all devices?" -> if the user answers positively,
> then a bookmark with 'autojoin' is set. If not, a bookmark may still
> be set, but the autojoin flag is only remembered locally.

I agree with the local overrides, mostly what I am (should be) doing at
the moment.

> And while we're here, I think "autojoin" is a silly concept. A client
> should just remember what rooms it has open, and keep them open. If I
> close my client and re-open it (or it crashed, or my computer crashed,
> etc.), I'd expect to still be in the same rooms, unless I explicitly
> asked it to leave them.

Indeed.

> > If my connection dropped and came back a moment later, I would want my
> > client to rejoin MUCs I was in. I use bookmarks mostly as a way to
> > remember MUC JIDs, not to know which state my clients should be in.
> 
> That's fine, and I see that as a completely valid use of the protocol.
> Just don't set autojoin on any of your bookmarks, and use clients that
> remember which rooms they are joined to.

One of your comments in another branch of this thread suggests this
would not be possible anymore, also similar to what SamWhited said
above:

> My argument is slightly different, I'm arguing that removing autojoin
> should cause clients that joined automatically to leave automatically.

It's not exactly said like this, but I would read this as "if autojoin
is false, leave the room and/or don't join it.".

You suggest that this would only be the case when first joined
automatically though, which is different from what SamWhited said above,
and probably fixes the issue I have.

-- 
Maxime “pep” Buquet

Attachment: signature.asc
Description: PGP signature

_______________________________________________
Standards mailing list
Info: https://mail.jabber.org/mailman/listinfo/standards
Unsubscribe: standards-unsubscr...@xmpp.org
_______________________________________________

Reply via email to