0ed6149463
Experimental reorder of Heartbeat loop now simulates physics and sends updates to clients prior to sleep. Existing behavior was to sleep BEFORE sending updates. We found this patch reduced latency to clients by 1-2 heartbeat periods.
2011-11-02 18:59:38 +00:00
a64be59c3b
Add missing max_listens_per_region to [LL_Functions] config section in OpenSimDefaults.ini + explanation.
...
This setting controls the maximum number of listeners in a region
2011-11-02 18:59:30 +00:00
b720454950
Remove the SyncRoot locking from Scene which was only being done around the main physics loop and ScenePresence position and velocity setting
...
This is no longer necessary with ODECharacter taints (ODEPrim was already not taking part in this). BSCharacter was already tainting.
2011-11-02 18:57:04 +00:00
dd3dc5cd91
Add taint target velocity for ODECharacters as is already done for ODECharacter position and position and velocity for ODEPrims.
...
This is to help stop surprises if the velocity is set in the middle of physics calculations, though this probably isn't a huge problem.
It's more for consistency and for the next step of removing some scene locks
2011-11-02 18:51:55 +00:00
9cd94ac6ec
tidy up OdeCharacter so that we just use OpenMetaverse.Vector3 assignment directly where possible, instead of transferring X, Y and Z components separately
...
some of this is probably a hold over from using ODE.Vector3, which is still necessary in some places.
2011-11-02 18:51:46 +00:00
2cc49d7d9a
Move position set from taint to logically better position at top of ODECharacter.ProcessTaints() though this makes no practical difference
2011-11-02 18:51:31 +00:00
69e11af475
Stop setting _position as well as m_taint_position in ODECharacter.Position
...
setting position at the same time as taint appears to undermine the whole purpose of taint
testing doesn't reveal any obvious regressions in doing this
2011-11-02 18:51:20 +00:00
aa6915f1ba
Continuation of previous checkin. Found more places where ForEachScenePresence can be changed to ForEachRootScenePresence.
2011-11-02 18:50:17 +00:00
b36ff0fd24
Added new ForEachRootScenePresence to Scene since almost every delegate passed to ForEachScenePresence checks for !IsChildAgent first. It consolidates child and root handling for coming refactors.
2011-11-02 18:48:21 +00:00
a16c9206b4
For now, comment out error message on new script engine console commands.
...
This causes false positives if a simulator has more than 1 region and the current region is 'root' since this sends the command separately to each region and each region has its own XEngine
2011-10-28 22:19:46 +01:00
b5b2541d1a
Comment out inventory folder bulk update code on InventoryAccepted message introduced in commit db91044 on Aug 22 2011
...
This should be unecessary since the folder update is already made at the time of the offer (and moved to trash if not accepted).
This code was also not taking into account the situation where an item was accepted.
Needs more fixing if this results in an aggression elsewhere.
2011-10-28 22:19:36 +01:00
1baadac59a
Don't blow our brains out if LLClientView.BulkInventoryUpdate() is wrongly passed a null node reference.
...
Addresses worst aspect of http://opensimulator.org/mantis/view.php?id=5752
2011-10-28 22:19:26 +01:00
88ef35cb23
Fix a bug I introduced yesterday in ODE physics where prim scripts would only receive the very first collision.
2011-10-28 22:19:17 +01:00
05dff4987b
Restart the event queue worker threads that I accidentally disabled earlier today in 8a0a78c.
...
Also adds these to the watchdogs with very large timeouts (should really be infinite)
2011-10-28 22:19:04 +01:00
e1e0f20c7f
In Watchdog, add ability to specific timeout for a thread.
...
This also changes the point of registration to the StartThread() call rather than the first Update()
2011-10-28 22:18:56 +01:00
70d559d1af
Add m_threads dictionary locking to Watchdog.GetThreads()
2011-10-28 22:18:48 +01:00
6c92b48143
Clear OdeCharacter CollisionEventUpdate when we subscribe or unsubscribe from collision events
2011-10-28 22:18:42 +01:00
127626edd8
Remove unused fields from CollisionEventUpdate
2011-10-28 22:18:35 +01:00
aba26c098d
Get rid of the pointless null checks on collision listeners. Add warning about synchronicity for PhysicsActor.OnCollisionUpdate event doc
2011-10-28 22:18:23 +01:00
9f19405490
When sending object collision updates, don't null out and recreate the CollisionEventUpdate() if the number of collisions falls to zero. Reuse the existing one instead.
2011-10-28 22:18:17 +01:00
871f1d0ae7
For ScenePresence collision events, instead of creating a new CollisionEventsThisFrame every time we need to send some new ones, reuse the existing one instead.
...
This assumes that the listener is using the data synchronously, which is currently the case.
2011-10-28 22:18:11 +01:00
057b78bfbe
Fix bug where collision event listeners were not removed once the listener had gone away.
...
This was causing continuous use of temporary memory even when all avatars had left the scene.
Memory does leak but it does cause more calls to the garbage collector, which would pause the scene thread for a very short while during collection.
2011-10-28 22:18:03 +01:00
d4fc07aae1
minor: rename a parameter in OdeScene.Simulate() from actor -> prim since it's an OdePrim
2011-10-28 22:17:55 +01:00
750e8ec3da
Add new LindenCaps test dll to panda run as well as "nant test"
2011-10-28 22:17:44 +01:00
15142093ad
minor: add "threads show" as synonym for "show threads" for consistency
2011-10-28 22:17:32 +01:00
f050f0fc0b
Add "threads abort <thread-id>" simulator console command that allows us to abort a watchdog managed thread.
...
This is for diagnostic purposes.
2011-10-28 22:17:20 +01:00
a3c79b399e
Make OpenSim.Framework.Servers.HttpServer rely on OpenSim.Framework instead of the other way around.
...
This is necessary so that code in HttpServer can use framework facilities such as the thread watchdog for monitoring purposes.
Doing this shuffle meant that MainServer was moved into OpenSim/Framework/Servers
Also had to make OpenSim.Framework.Console rely on OpenSim.Framework rather than the other way around since it in turn relies on HttpServer
MainConsole and some new interfaces had to be moved into OpenSim/Framework to allow this. This can be reverted if parts of OpenSim.Framework stop relying on console presence (cheifly RegionInfo)
2011-10-28 22:17:10 +01:00
424735efee
separate out future common setup code from EventQueueTests.AddForClient()
2011-10-28 22:16:49 +01:00
d5dfbc3844
Add new EventQueueTests with basic test to check that adding a client registers an http poll
2011-10-28 22:16:31 +01:00
c3127d1323
Remove earlier mistaken addition of EventQueueTests.cs
2011-10-28 22:15:03 +01:00
03c98c3b8f
Fix bugs in EventQueueGetModule.ClientClosed() and BaseHttpServer.RemovePollServerHTTPHandler() that stopped existing code in ClientClosed() from actually tearing down the poll handler
...
Actually doing the tear down appear to have no ill effects with region crossing and teleport.
2011-10-28 22:09:53 +01:00
24e02afeac
Comment out the uuid gatherer lines that I accidentally left in.
2011-10-28 22:02:54 +01:00
8d5606e09b
Rename some of the example commented out server names to mygridserver.com like all the others.
...
This is to make it easier to change all these names at once, where all grids services are running from the same location.
Also rearranges some lines in [Modules] so that StandaloneCommon.ini.example and GridCommon.ini.exmaple are consistent.
2011-10-28 22:02:43 +01:00
bd3d119a03
Make config comments about choosing assetcache modules clearer.
...
OpenSim needs an asset cache. Running without one will cause problems.
Thanks to Ai Austin for the suggestions in http://opensimulator.org/mantis/view.php?id=4850
2011-10-28 22:02:37 +01:00
cf2405385d
Get UUIDGatherer to scan notecards in the graph for asset uuids.
...
This is to support npc baked texture saving in oars and iars.
May address http://opensimulator.org/mantis/view.php?id=5743
2011-10-28 22:02:25 +01:00
a6eba09dac
redirect UserInventoryHelpers to use a different CreateNotecardAsset() so we can comment out some inconsistency
2011-10-28 22:02:18 +01:00
b2ca7c0927
correct misleading method doc on CreateNotecardAsset()
2011-10-28 22:02:09 +01:00
dd6c236253
in AssetHelpers, store the actual text passed in to the method, not the string "data"
2011-10-28 22:02:02 +01:00
b2ff680cca
encode notecard assets in proper format for tests, rather than just using whatever string is given
2011-10-28 22:01:54 +01:00
23a9a7daa7
refactor: rename some AssetHelpers.CreateAsset() methods to CreateNotecardAsset()
2011-10-28 22:01:47 +01:00
19fde57fca
Change wording on asset requests.
...
Not all 'notified missing' assets are a problem. Some are invalid references which happen to be buried in other text.
2011-10-28 21:59:53 +01:00
a556930456
Store scene identifier passed in to OdeScene for later debug messages
2011-10-28 21:58:30 +01:00
1a008b237b
Don't bother taking OdeLock during OdeScene construction, since there can be no contention until the object is constructed.
2011-10-28 21:58:21 +01:00
d6456b9ea8
Remove unnecessary lock of OdeLock in OdePrim.changeadd()
...
This taint can only ever be processed from the OdeScene.Simulate() loop, which already locks OdeLock.
2011-10-28 21:58:12 +01:00
5980d57b7a
minor: method doc to explain a lock of OdeLock
2011-10-28 21:58:01 +01:00
7c409eff53
Remove the unused CollisionLocker from ODE
...
Despite its name, this wasn't actually being used in any collision checking
2011-10-28 21:56:04 +01:00
614ea5a48a
refactor: Use SOP.SitTargetAvatar instead of calling a special GetAvatarOnSitTarget() which returned exactly the same thing
2011-10-28 21:47:08 +01:00
8905f34e18
rename OS_NPC_SIT_IMMEDIATE to OS_NPC_SIT_NOW since it's shorter and more understandable
...
This makes something like osNpcSit(npc, llGetKey(), OS_NPC_IMMEDIATE) now become
osNpcSit(npc, llGetKey(), OS_NPC_SIT_NOW);
This is why it's in development :)
2011-10-28 21:47:00 +01:00
380f2a1719
Implement osNpcStand(<npc-id>)
...
Allows you to stand an NPC that has sat.
2011-10-28 21:46:53 +01:00
1c66e08964
Implement osNpcSit(). This is still in development so don't trust it
...
Format is osNpcSit(<npc-uuid>, <target-uuid>, OS_NPC_SIT_IMMEDIATE)
e.g. osNpcSit(npc, llGetKey(), OS_NPC_SIT_IMMEDIATE);
At the moment, sit only succeeds if the part has a sit target set.
NPC immediately sits on the target even if miles away - they do not walk up to it.
This method is in development - it may change so please don't trust it yet.
Standing will follow shortly since that's kind of important once you're sitting :)
2011-10-28 21:46:36 +01:00
Dan Lake