I distinctly remember that it was once setting full JID always for all stanzas and there was some problem. Unfortunately I do not remember details. So I changed the code to not modify from attribute in if there is a correct BareJID set already. Maybe it was related to presence subscription workflow? Attached RFC excerpt suggest that presence subscription should be treated differently.

Artur Hefczyc wrote:

I distinctly remember that it was once setting full JID always for all stanzas and there was some problem. Unfortunately I do not remember details. So I changed the code to not modify from attribute in if there is a correct BareJID set already. Maybe it was related to presence subscription workflow? Attached RFC excerpt suggest that presence subscription should be treated differently.

Artur, how can I check your commit? Don't see any related info here https://projects.tigase.org/projects/tigase-server/repository/show?rev=devel

This commit probably was quite a few years ago. I do not think you can easily find it. I think the correct action would be now to work on the code which replaces form address to work correctly, that is replace all message and iq stanzas and deal properly with presences.

Ok, I think this is not fixed through #2019

Artur Hefczyc wrote:

Ok, I think this is not fixed through #2019

Actually it wasn't for all connections - only for c2s (and thus ws2s), which surfaced in https://projects.tigase.org/boards/8/topics/4917?r=5114. Corrected the logic in tigase.xmpp.impl.BindResource.preProcess() which was responsible for setting correct from and added unit test.