Remko Tronçon wrote: >> Now due to the offline message, they BOTH think the other end doesn't >> support XEP-85 and chat state notifications are never used. I already >> triggered that several times in Gajim. > > This particular problem is solved by *always* sending <active/> in our > messages. It's not 100% clear whether this is breaking the spec of > 'not sending chat state notifications when the initial message didn't > have <active/>', but it's the easiest solution for this offline > message problem, for contacts changing resources, ...
I don't think Dizzy and I thought much about the offline messaging case. What you suggest seems reasonable in that case. >> Solution: Caps, it's already in the XEP. But sadly, no client determines >> support via caps. > > Hmm, I always thought Psi did, but it seems I removed that. I think it > was because of a discussion on whether or not it was acceptable (from > the user's POV) to send composing events before the initial message. > Some found this scary, and I didn't have the time to implement > delaying sending composing events until the next message. > > Btw, what's the general opinion on sending 'composing' events before > the initial message? I know some other protocol clients like aMSN pop > up windows whenever a user starts typing, which some consider a > privacy problem. I think it's better not to send an empty composing event before you even send a message -- that would feel weird to me. But I don't know that the spec needs to forbid it (maybe it does now, though, eh?). > But I agree that caps should be preferred over the old method. +1 Peter -- Peter Saint-Andre https://stpeter.im/
