This is based on heavily mikemig's original patch in http://opensimulator.org/mantis/view.php?id=7149
but instead of exiting after the first IM delivery failure to presence information retrieved from the presence service
it will retry the lookup until the result matches the previous lookup.
This is to deal with the case where the agent is sent an IM whilst they are teleporting.
Offline notices can still be controlled with the [Messaging] ForwardOfflineGroupMessages setting.
Looks to address more of http://opensimulator.org/mantis/view.php?id=7037
Only for Flotsam now for testing, but if approach works should be possible with core offline notices as well.
There is a problem here with infinite recursion, but this patch loses the 'hunting' behaviour where the code will attempt multiple lookups if the avatar is teleporting rapidly around different simulators.
This patch only does a single lookup before giving up.
This reverts commit cecb446e0e.
It appears that at least Singularity 1.8.5 (but probably others) rely on attachment FromItemIDs being different to display more than one.
This commit resolves this by generating random IDs instead of always using UUID.Zero for NPCs.
Resolves http://opensimulator.org/mantis/view.php?id=7110
This happened because these notifications (the UDP packets) can only be sent to Root Agents. But the notifications were done in OnClientLogin(), at which point the client is still a Child Agent. Since a FireAndForget is used, it became a race condition as to whether the packets would be sent before or after the client became a Root Agent.
To fix this, we now only send the notifications once the client becomes a Root Agent for the first time.
This can affect which region to use. E.g., returning users may be allowed to enter any region, whereas users from other grids will have to enter a gateway region. Previously per-user decisions were only made later, but by then it's too late to change which region the user enters.
This is to prevent an immediate problem in http://opensimulator.org/mantis/view.php?id=7035 where a development code bug occasionally overlays all the existing parcels with a blank parcel owned by the estate manager and to gather more data.
My guess is that this parcel is being created by the new code in LandManagementModule.GetLandObject(), probably some race between threads since this only happens occasionally.
Adds regression tests for this case and for parcel subdivide.
This is because the alpha texture is not in the cache, we must continue to have the fallback of looking for these and other persisted bakes in the asset service.
Relates to http://opensimulator.org/mantis/view.php?id=6927
This is causing many issues on OSGrid, possibly due to teleporting timing differences with simulators hosted in different places or HG teleports
Added a bit more logging for debug purposes.
See http://opensimulator.org/mantis/view.php?id=6939
