|
Daniele Ricci opened 10 years ago
|
|
What does the spec says about this use-case? It seems like, the Tigase processing might be indeed incorrect. I just looked at the code and the method: processFromUserPacket might be to simplistic indeed. The XMPP routing logic is actually more complex and it also depends on the packet type. That is, it is a bit different for presence, message and iq. Therefore, the processFromUserPacket might not work well for all possible use-cases. My suggestion to fix this error would be to reimplement the JabberIQVersion plugin and override the processFromUserPacket method instead to get a correct processing for version stanzas sent to the user himself (most likely the other connection). |
|
|
Thanks Artur, I'll begin working on JabberIqVersion as soon as I can and attach a patch to this issue. |
|
|
Isn't this enough to achive the goal? Please set the source code disclaimer to yes for me. |
|
Applied in changeset commit:tigase-server|9918d2de. |
| Type |
Bug
|
| Priority |
Normal
|
| Assignee | |
| RedmineID |
3249
|
| Spent time |
7h 30m
|
Is this intended behaviour?
I think processFromUserToServerPacket is called for the wrong case here. What do you think? I'm afraid I can break some APIs here if I touch something so I'd like your opinion before adding a patch.
Tested on Tigase 7.0.2, but both classes (JabberIqVersion and its parent class) didn't change up to now.
JabberIqVersion-self.diff