On the Maemo5 SDK, 2 clients in the room, each accepting/offering d-bus tubes, when one client accepts the tube, it gets closed and my app tries to connect to an invalid dbus-gabble unix socket file.
The other client uses gabble 0.8.7-1ubuntu1 on ubuntu karmic.
The log is there:
You can grep for the tube 612614435 for instance.
404 when trying to download the logs :\
(In reply to comment #1)
> 404 when trying to download the logs :\
Sorry it is http://base-art.net/static/gabble.log.bz2
Here are my partial observations from this log:
- You joined email@example.com
- You created a tube T1 (1571899144) using the bytestream 1265898273-2625381383
- You created another tube T2 (1573115257) using the bytestream 1265898273-4088574537
- You created a third tube T3 (2143994341) using the bytestream 1265898273-1159475516
- Your app connected to these 3 D-Bus tubes
- "Coherence At Phil's" joins T2
- then T1
- then T3
- then created a new tube T4 1502302501 using 1265898368-3092784896
- You joined tube T4 and connect to it
- "Coherence At Phil's" created a new tube (T5) 1846207483 using 1265898368-3669707655
- You joined T5 and connect to it
- "Coherence At Phil's" created a new tube (T6) 612614435 using 1265898368-729463036
- You joined T6, connect to it and send one message to it
- You received one message on T3
- You sent one (big) message on T3
- A bit later firstname.lastname@example.org/Coherence At Phil's left the room
- T4's bytestream is removed by Gabble and so T4 is closed.
- So is T5
- and T6
- "Coherence At Phil's" rejoins the room with his tubes (T2, T5, T1, T3, T6)
I stopped at this point. I'll check the code but it seens weird to my that T4 is closed because its initiator left. That shouldn't happen.