If n > 1 for RootTerseUpdatePeriod only every n terse update is actually sent to observers on same region, unless velocity is effectively zero (to stop av drift).
If n > 1 for ChildTerseUpdatePeriod only every n terse update is sent to observers in other regions, unless velocity is effectively zero.
Defaults are same as before (all packets are sent).
Tradeoff is reduction of UDP traffic vs fidelity of observed av mvmt.
Increasing n > 1 leads to jerky observed mvmt immediateley for root, though not on child, where experimentally have gone to n = 4 before jerkiness is noticeable.
Rapid polls are more expensive than triggered events (several polls vs one trigger) and may be problematic on heavily loaded simulators where many threads are vying for processor time.
A triggered event is also slightly quicker as there is no maximum 200ms wait between polls.
This is already going to be correctly set by WaitForUpdateAgent() earlier on in that method, which is always called where a callback to the originating region is required.
These govern when AgentUpdates are sent to observers on position, rotation and velocity changes to an avatar (including the avatar themselves).
Higher values reduce AgentUpdate traffic but at a certain level will degrade smoothness of avatar and perceived avatar movement.
Allows experiments in manually reducing updates under heavy load.
Activated by "debug scene set client-upd-per" console command.
In a simple test, can send as few as every 4th update before observed movement starts becoming disturbingly rubber-banded.
This governs when child agent position changes are sent to neighbouring regions.
Corresponding config parameter is ChildReprioritizationDistance in [InterestManagement] in OpenSim.ini
For test purposes.
I forgot that a null 'their version' would not be passed over the wire and ends up as an empty string instead (like older simulators).
So instead pass through the correct simulator protcol version instead (SIMULATOR/0.3) when querying from login or hg login.
Also removes a debug console write for agent limit accidentally left in for the same commit.
Relates to mantis 7276
Another manifestation of this bug is that after a Hypergrid teleport, when you click on one of the avatar's attachments the object doesn't show its name. This means that the viewer knows the attachment is there, but the simulator does not.
The problem was caused by treating Hypergrid teleports as if they're Logins (because the teleport flag ViaLogin is enabled).
This may fix: http://opensimulator.org/mantis/view.php?id=7238
This may fix: http://opensimulator.org/mantis/view.php?id=7220
This caused the client to stop responding, and even the simulators to have problems. The solution is to disallow crossing before the previous cross has completed.
For non-root prim, eye offsets now need to be made relative to root prim if either camera-at or camera-eye are set.
Probably a regression since November 2013 when all sits were made relative to root prim to match viewer expections (and fix other bugs).
Addresses http://opensimulator.org/mantis/view.php?id=7176
This was due to a silent uint overflow in ScenePresence.UpdateChildAgent() corrupting child agent positions
when the child agent was in a region with a greater x or y map co-ord than the root agent region
Probably introduced in beeec1c.
This still will not function properly with very high region map co-ords (in the millions) but other parts of the code don't handle this properly anyway.
Looks to address http://opensimulator.org/mantis/view.php?id=7163
Previously we used a heuristic of checking if the entry position is 0 to differentiate between Teleport and Cross, but that doesn't work anymore since we've started providing the precise entry position for cross, too. That's required in order to ensure that the user is allowed to enter the parcel that he's walking into.
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.
are kept on another. When users rez items from inventory or take objects from world, an HG-like asset copy takes place between the 2 servers, the world asset server and the user's asset server. This makes the simulators independent of the central asset server.
Note that this an advanced configuration and requires some security strengthening coming up.
Per http://wiki.secondlife.com/wiki/Tips_for_Creating_Heightfields_and_Details_on_Terrain_RAW_Files#Notes_for_Creating_Height_Field_Maps_for_Second_Life
terrain heights up to 508 are possible on the LL grid (and were available on previous releases of OpenSimulator).
The obvious way to allow both this and equivalent -z values, is to rewiden the internal terrain height storage from short to int.
The memory tradeoff is most noticeable on the maximum 8192x8192 var region (equiv to 1024 normal regions), where it adds 128mb to resident use (128k on a normal region)
This is still better than the double used in previous releases.
This does not affect physics or data storage since they already use float and double respectively.
This may not be the final solution if we actually want to sacrifice -z, >327 or something else.
Relates to http://opensimulator.org/mantis/view.php?id=7076
This uses an offset of 0.05 on the up vector of the sit orientation, after extensive analysis on
http://opensimulator.org/mantis/view.php?id=7096 and https://wiki.secondlife.com/wiki/Talk:LlSitTarget
This supersedes the previous adjustment in 5b73b9c4 which had been wrongly applied.
The maximum change is about 0.1 with a maximum height avatar.
This patch is all Aleric's work - I am applying manually in order to add these additional notes
objects in the world.
This was originally changed in an attempt to make vehicles work better
but the effect was not that large and it causes avatars to float
above the terrain.
When an avatar is in their home grid, their name appears as "First Last". In other grids the name appears as "First.Last @grid.example.com". However, viewers have a bug and they don't always show the new name. We use a trick (changing the Group Title briefly) in order to make the viewers show the new name. This is only done after a Hypergrid teleport.
to the physics engine.
This operation was accidentally removed while in the middle of improving
the computation of the avatar height. This is a temp fix until the real
solution is added.
Also fixes an issue where sitting avatar counts became inaccurate after any cross.
Part of the problem was due to cloning code using MemberwiseClone() but not resetting certain collection structures.
Adds regression test for this case.
In relation to http://opensimulator.org/mantis/view.php?id=7050
Root cause was that PrimLimitsModule was not properly handling the case where the parcel it was asked to check was outside the current region's bounds.
If this is the case, we can abort the check since the receiving region will perform it.
Added a regression test for this case.
Recent versions of Firestorm and Singularity have started sending the viewer name in the 'Channel' field, leaving only their version number in the 'Viewer' field. So we need to search both of these fields for the viewer name.
This resolves http://opensimulator.org/mantis/view.php?id=6952
It does not indicate that a SOG has been set to be backed up or not. It indicates that a SOG can be backed up in principle.
In practice, this was added by lbsa71 in 2007 and is set to always true, albeit as a virtual property.
This is because xmlrpcgroups currently always checks visibility for the requesting agent ID (unlike Groups v2, which can accept UUID.Zero)
But group notice IMs have a from agent which is the group rather than the sending agent.
Further addresses http://opensimulator.org/mantis/view.php?id=7037
than everytime terrain is changed.
The TerrainModule now hooks the frame event and, if terrain has changed,
sends terrain updates to the clients. This polling pattern replaces
the previous push on change pattern and will make it easier to do per
client throttling and per scene presence terrain update ordering.
This is to allow small adjustments of less than 0.05m in functions such as llSetPrimitiveLinkParams() to work
This is another fix for http://opensimulator.org/mantis/view.php?id=7044
Extends regression test for this case.
region. This is a temp fix for the use of draw distance to compute child
regions. Eventually must use the draw distance from the viewer for the
computation.
Fix odd "cannot cross into banned parcel" viewer error message when crossing
into non-existant region. Proper permission failure messages are now returned.
This commit eliminates all of the 'border' class and list code and replaces
it with testing if in the current region.
Impacts: can make a mega-region out of varregions of the same size; and
mega-region combinations must be rectangular (not square but rectangular)
the size of the input stream. This is required since the raw heightmap
format (.r32) does not contain any size information.
The estimation relies on terrain being square.
Modify archiver to use terrain merging when loading oars.
This makes displacement AND rotation properly work on terrain when loading oars.
Especially useful when loading legacy region oars into large varregions.
Locking attachments then launching script instances on a separate thread will not work, attachments will simply be unlocked and vulnerable to race conditions.
removed, when a script is removed, or when a script is reset. Also
added a stats command to track the number of json stores used by
a region. Will probably add some more commands later.
Changed UuidGatherer to use 'sbyte' to identify assets instead of 'AssetType'. This lets UuidGatherer handle Materials, which are defined in a different enum from 'AssetType'.
Adds displacment to all objects and terrain loaded from the oar.
As an example, if you have a 512x512 region and an old 256x256 oar, doing
load oar --displacement "<128,128,0>" oarFile.oar
will load the object (and terrain) into the middle of the 512x512 region.
If displacement is not specified, 'load oar' works like it always has.
If you have a 5
Justin Clark-Casey (justincc)