There is a possibilty that some V2 teleport failures are due to the viewer triggered CompleteMovement thread not seeing the change of m_originRegionID by the UpdateAgent thread.
Need an info message since this is currently important in detecting teleport issue when not at debug log level.
CompleteMovement message occurs before MakeRootAgent() one did
If this is attempted, they get a "Try moving closer. Can't sit on object because it is not in the same region as you." message instead, which is the same as current ll grid.
Sitting on ground is okay, since viewer navigates avatar to required region first before sitting.
This is giving much better results on teleports between simulators over my lan where for some reason there is a pause before the receiving simulator processes UpdateAgent()
At this point, v2 teleports between neighbour and non-neighbour regions on a single simulator and between v2 simulators and between a v1 and v2 simulator
are working okay for me in different scenarios (e.g. simple teleport, teleport back to original quickly and re-teleport, teleport back to neighbour and re-teleport. etc.)
The root cause was that v2 was only closing neighbour agents if the root connection also needed a close.
However, fixing this requires the neighbour regions also detect when they should not close due to re-teleports re-establishing the child connection.
This involves restructuring the code to introduce a scene presence state machine that can serialize the different add and remove client calls that are now possible with the late close of the
This commit appears to fix these issues and improve teleport, but still has holes on at least quick reteleporting (and possibly occasionally on ordinary teleports).
Also, has not been completely tested yet in scenarios where regions are running on different simulators
These were genuine failures caused by ScenePresence.CompleteMovement() waiting for an UpdateAgent from NPC introduction that would never come.
Instead, we do not wait if the agent is an NPC.
This prevents an issue if the user teleports back to the neighbour simulator of a source before 15 seconds have elapsed.
This more closely emulates observed linden behaviour, though the timeout there is 50 secs and applies to all the pre-teleport agents.
Currently sticks a DoNotClose flag on ScenePresence though this may be temporary as possibly it could be incorporated into the ETM state machine
In this new protocol, and as committed before, the viewer is not sent EnableSimulator/EstablishChildCommunication for the destination. Instead, it is sent TeleportFinish directly. TeleportFinish, in turn, makes the viewer send a UserCircuitCode packet followed by CompleteMovementIntoRegion packet. These 2 packets tend to occur one after the other almost immediately to the point that when CMIR arrives the client is not even connected yet and that packet is ignored (there might have been some race conditions here before); then the viewer sends CMIR again within 5-8 secs. But the delay between them may be higher in busier regions, which may lead to race conditions.
This commit improves the process so there are are no race conditions at the destination. CompleteMovement (triggered by the viewer) waits until Update has been sent from the origin. Update, in turn, waits until there is a *root* scene presence -- so making sure CompleteMovement has run MakeRoot. In other words, there are two threadlets at the destination, one from the viewer and one from the origin region, waiting for each other to do the right thing. That makes it safe to close the agent at the origin upon return of the Update call without having to wait for callback, because we are absolutely sure that the viewer knows it is in th new region.
Note also that in the V1 protocol, the destination was getting UseCircuitCode from the viewer twice -- once on EstablishAgentCommunication and then again on TeleportFinish. The second UCC was being ignored, but it shows how we were not following the expected steps...
- The existing event to scene has been split into 2: OnAgentUpdate and OnAgentCameraUpdate, to better reflect the two types of updates that the viewer sends. We can run one without the other, which is what happens when the avie is still but the user is camming around
- Added thresholds (as opposed to equality) to determine whether the update is significant or not. I thin these thresholds are ok, but we can play with them later
- Ignore updates of HeadRotation, which were problematic and aren't being used up stream
Revert "Trying to reduce CPU usage on logins and TPs: trying radical elimination of all FireAndForgets throughout CompleteMovement. There were 4."
This reverts commit 6825377380.
BestAvatarResponsiveness introduces the region rez delay in cases where the region is full of avatars with lots of attachments, which is the case in CC load tests. In that case, the inworld prims are sent only after all avatar attachments are sent. Not recommended for regions with heavy avatar traffic!
Justin, if you read this, there's a long story here. Some time ago you placed SendInitialDataToMe at the very beginning of client creation (in LLUDPServer). That is problematic, as we discovered relatively recently: on TPs, as soon as the client starts getting data from child agents, it starts requesting resources back *from the simulator where its root agent is*. We found this to be the problem behind meshes missing on HG TPs (because the viewer was requesting the meshes of the receiving sim from the departing grid). But this affects much more than meshes and HG TPs. It may also explain cloud avatars after a local TP: baked textures are only stored in the simulator, so if a child agent receives a UUID of a baked texture in the destination sim and requests that texture from the departing sim where the root agent is, it will fail to get that texture.
Bottom line: we need to delay sending the new simulator data to the viewer until we are absolutely sure that the viewer knows that its main agent is in a new sim. Hence, moving it to CompleteMovement.
Now I am trying to tune the initial rez delay that we all experience in the CC. I think that when I fixed the issue described above, I may have moved SendInitialDataToMe to much later than it should be, so now I'm moving to earlier in CompleteMovement.
This is because scripts (at least on XEngine) start unsuspended - deceptively the ResumeScripts() calls in various places in the code are actually completely redundant (and useless).
The solution chosen here is to use a copy of the SP attachments and not have the list locked whilst creating the scripts when an avatar enters the region.
This looks to address http://opensimulator.org/mantis/view.php?id=6557
This purports to fix the issue described in http://opensimulator.org/mantis/view.php?id=6653 where the camera can end up following the requested sit prim rather than the actual.
The original spot was by Vegaslon, this commit just goes about it in a slightly different way
This commit also makes m_requestedSitTargetUUID to be the actual UUID, which is consistent with m_requestedSitTargetID which was already doing this.
However, this adjustment has no practical effect since we only currently need to know that there's any requested sit UUID at all, not which one it is.
to avatar when it becomes root. This packet shows up in the viewer
logs as an error and appears to cause problems for completing the
texture rebake process for v1 viewers in some cases.
This is to reduce race conditions where neighbours may be responding erratically, thus mixing up create and close agent requests in time.
This mirrors OpenSimulator behaviour on enabling child agents where each region is contacted separately.
By checking against the grid's Avatar data, we can ignore viewer side attachments but still initiate these calls simulator-side.
Initiating simulator-side is always necessary for version 1 viewers.
This is a further commit to resolve http://opensimulator.org/mantis/view.php?id=6581
This was because the script resumption in AttachmentsModule was firing the attach event instead.
Had to reinstate the code in 285bd3a do we can resume the scripts there instead, though the bug existed before its removal.
This is to resolve http://opensimulator.org/mantis/view.php?id=6578
This is a further effort to reduce v3 viewer race conditions where this call may clash with the viewer signalling attachment wearing from its current outfit folder.
This is to simplify the code (no tricky 'wasChild' signalling required)
and to reduce the risk of a thread clash between simulator-side attaching (necessary for v1 viewers)
and the viewer-side attaching the v3 viewers perform.
This now does a check for border crossing (required to trigger the teleport) immediately in SP.MakeRootAgent().
If left any later, it looks like the physics scene changes the position and stops the cross happening.
If done any earlier, nothing happens because the cross-code currently requires a PhysicsActor to be in place, thuogh it's probably not necessary for this case.
For some reason, sending updates before this will not have this effect.
This may be something related to some viewers (e.g. LL 3.3.4) or something OpenSimulator isn't getting quite right.
This addresses http://opensimulator.org/mantis/view.php?id=6567
This creates a ScenePresence.GetWorldRotation() with the same semantics as SOP.GetWorldRotation()
SP.Rotation can't be used since it's relative to the sat upon prim if the avatar is sitting.
Fix bug that allowed only login access to regions with mis-configured telehubs. Administrators now have teleport access when there exists a mis-configured telehub in the region. Estate owners are now placed at region center in the absence of spawnpoints instead of being denied access. Grid Gods are unrestricted. All others are denied access to the region until spawnpoints are assigned to the telehub object.
property multiple times every time Update is called.
This eliminates zillions of settings which is better for BulletSim.
The should be no functionality change.
Even where checks are being made they aren't enough since they all assume that the Animator they just checked is still there in the next line, which is not necessarily the case without locking.
The memory used is small and these should be GC'd anyway when the SP is released. If this is not happening then the wider problem of old SPs being retained needs to be resolved.
This also makes the detection in SP.FindNextAvailableSitTarget() and SendSitResponse() identical.
Previously they varied slightly (SendSitResponse didn't check for an older type of invalid quaternion) but the practical effect is most probably zero.
This prevents a stack overflow where a get position on the avatar will refer to the attachment which will in turn refer back to the avatar.
This required recording of all sitting avatars on a prim which is done separately from recording the sit target avatar.
Recording HashSet is null if there are no sitting avatars in order to save memory.
If delete doesn't occur first then the update thread can outrace the IsAttachment = false necessary to save attachments and send hud artifacts to other viewers.
Justin Clark-Casey (justincc)