| Commit message (Collapse) | Author | Files | Lines |
|
printed out periodically)
|
|
This shows the actual amount of RAM being taken up by OpenSimulator (objects + vm overhead)
|
|
|
|
whether the process is 32-bit or 64-bit
In theory, this means that a 64-bit Windows OS user can now run OpenSim.exe with ODE and use more than 2 (or 3) GB of memory.
However, this is completely untested since I don't currently own a 64-bit Windows box. Feedback appreciated.
Using OpenSim.32BitLaunch.exe should continue to work. Other platforms are unaffected.
This will currently not work with sqlite - I will add that too if this works.
|
|
|
|
This checks that the initial UseCircuitCode packet is handled correctly for a normal client login.
|
|
This is to support npc baked texture saving in oars and iars.
May address http://opensimulator.org/mantis/view.php?id=5743
|
|
bucket to prevent a viewer 3 crash.
This is the message sent to the client when the object is returned.
We were sending byte[0] in the binary bucket. This didn't kill viewer 1 but did terminate viewer 3 (don't know about viewer 2).
So sending "\0" instead.
This is to address http://opensimulator.org/mantis/view.php?id=5683
|
|
rather than letting it terminate the simulator.
Exceptions don't appear to do this with the SmartThreadPool but they do with UnsafeQueueUserWorkItem (and maybe others)
|
|
This executes the callback on the same thread that made the request. Designed for use only by regression tests that rely on a predicable event ordering.
|
|
continuing wtih the invocation.
This is to avoid bugs where the locale is not manually set on the thread and bad data values get sent to the database or over the wire.
Lots of code does this manually but as we've seen, a subtle change can hit code which has forgotton to do this.
Since en_US show be used throughout the server at present, setting it at FireAndForget seems reasonable.
Arguably, it would be better to do this where data is sent, but doing it here is much easier.
All the manual BeginInvokes() remaining in the code should probably call FireAndForget instead.
|
|
an existing one with the same id.
We compare existing and loaded asset by doing an SHA1 on both, so that a changed library asset will still update the store.
This cuts asset library load time from 10 seconds to <1 sec.
Note, a fix on the previous commit revealed a bug where a library script cannot be copied except on the first login after a cache clear.
This is unrelated to this commit and needs to be fixed at some subsequent time.
|
|
As per http://wiki.secondlife.com/wiki/Mesh/Mesh_Asset_Format, some submesh blocks may just have the flag "NoGeometry" to signal that they provide no mesh data.
If a block contains this, ignore it for meshing purposes rather than suffer a ClassCastException
This fixes physics proxy meshing, so you can now walk through mesh doorways, properly stand on the trailer of mesh trucks, etc.
To get mesh physics proxy, the UseMeshiesPhysicsMesh must be true in a [Mesh] config section in OpenSim.ini (example in OpenSimDefaults.ini).
Convex hull physics not currently supported.
|
|
security to friendship identifiers so that they can safely be deleted across worlds. Had to change Get(string) to use LIKE because the secret in the identifier is not always known -- affects only HG visitors. BOTTOM LINE SO FAR: HG friendships established and deleted safely across grids, local rights working but not (yet?) being transmitted back.
|
|
correctly handled. Friends list showing correct foreign names. TODO: GrantRights.
|
|
Often, by the time the UDPServer realizes that an entity update packet
has not been acknowledged, there is a newer update for the same entity
already queued up or there is a higher priority update that should be
sent first. This patch eliminates 1:1 packet resends for unacked entity
update packets. Insteawd, unacked update packets are decomposed into the
original entity updates and those updates are placed back into the
priority queues based on their new priority but the original update
timestamp. This will generally place them at the head of the line to be
put back on the wire as a new outgoing packet but prevents the resend
queue from filling up with multiple stale updates for the same entity.
This new approach takes advantage of the UDP nature of the Linden protocol
in that the intent of a reliable update packet is that if it goes
unacknowledge, SOMETHING has to happen to get the update to the client.
We are simply making sure that we are resending current object state
rather than stale object state.
Additionally, this patch includes a generalized callback mechanism so
that any caller can specify their own method to call when a packet
expires without being acknowledged. We use this mechanism to requeue
update packets and otherwise use the UDPServer default method of just
putting expired packets in the resend queue.
|
|
Often, by the time the UDPServer realizes that an entity update packet
has not been acknowledged, there is a newer update for the same entity
already queued up or there is a higher priority update that should be
sent first. This patch eliminates 1:1 packet resends for unacked entity
update packets. Insteawd, unacked update packets are decomposed into the
original entity updates and those updates are placed back into the
priority queues based on their new priority but the original update
timestamp. This will generally place them at the head of the line to be
put back on the wire as a new outgoing packet but prevents the resend
queue from filling up with multiple stale updates for the same entity.
This new approach takes advantage of the UDP nature of the Linden protocol
in that the intent of a reliable update packet is that if it goes
unacknowledge, SOMETHING has to happen to get the update to the client.
We are simply making sure that we are resending current object state
rather than stale object state.
Additionally, this patch includes a generalized callback mechanism so
that any caller can specify their own method to call when a packet
expires without being acknowledged. We use this mechanism to requeue
update packets and otherwise use the UDPServer default method of just
putting expired packets in the resend queue.
|
|
child agents in neighbor regions. This means you can extend
the view on a simulator beyond the default 3x3 regions.
This uses a region default draw distance and should be
replaced at some point by the avatar specified draw distance.
That will require more careful, dynamic recomputation of child
agents every time the draw distance changes.
WARNING: this is experimental and has known instabilities. specifically
all regions "within site" should be running the same default draw distance
or agents will not be closed correctly.
|
|
llInstantMessage() to stop this crashing viewer 2.4.0 (1.23.5 was fine with this).
We're putting in a string of format "<region name><x pos floor int><y pos floor int><z pos floor int> which appears to be the expected value.
This resolves http://opensimulator.org/mantis/view.php?id=5356
|
|
Made FireAndForgetWrapper a singleton class to allow us to drop
dependancy on the BclExtras35 library. BclExtras is broken in
Mono 2.8.2 and we used the library in only one function.
|
|
Made FireAndForgetWrapper a singleton class to allow us to drop
dependancy on the BclExtras35 library. BclExtras is broken in
Mono 2.8.2 and we used the library in only one function.
|
|
usage help. Moved Osp around, deleted unnecessary OspInventoryWrapperPlugin, added manipulation of SOP's xml representation in a generic ExternalRepresentationUtils function.
|
|
functions around.
|
|
|
|
|
|
|
|
|
|
in LM info. Maturity is NOT the parcel's setting, that is only for the
image and text. Parcel maturity is governed by region maturity.
|
|
OpenSim.Grid.UserServer.exe fails on startup if no previous config
probably appears to occur because mono 2.4.2.3 (and possibly later) erroneously returns a value of 0 for BufferWidth and BufferHeight in some circumstances
|
|
|
|
This will now choose the first network interface IP address, or the loopback
interface if no external interfaces are found. It will log the IP address
used as [NETWORK]: Using x.x.x.x for SYSTEMIP.
|
|
|
|
still unfinished.
|
|
amount of memory allocated to the simulator process (Moderate Threat Level).
* Cleans redundant information out of the Simulator Version. Versions now look like:
"OpenSimulator 0.6.9(dev) Unix/Mono"
* [Minor] additional log info for MySQLInventoryData
|
|
|
|
purposes. Resolves the wrap-around of the 32 bit uint.
* Teravus moved the Environment methods to the Util class
|
|
always leave a worker thread available for other tasks
|
|
delegate (which STP appears to hold on to). This removes the slow leak I was seeing when using async_call_method=SmartThreadPool and stabilizes allocated memory for an idle OpenSim instance
|
|
|
|
so initialize it to sane defaults
* Simplified the InitThreadPool() function
|
|
(like the original implementation)
* Only initialize Util's SmartThreadPool if it is actually being used
* No longer initializing Util's SmartThreadPool with a custom max stack size. From MSDN: "Avoid using this constructor overload. The default stack size used by the Thread(ThreadStart) constructor overload is the recommended stack size for threads."
|
|
Util.FireAndForget()
* Changed Util.FireAndForget() to use any of five different methods set with async_call_method in the [Startup] section of OpenSim.ini. Look at the example config for possible values
|
|
|
|
for FireAndForget. This lets us limit concurrency to make OpenSim play nice
|
|
This avoids .NET remoting and a managed->unmanaged->managed jump. Overall, a night and day performance difference
* Initialize the LLClientView prim full update queue to the number of prims in the scene for a big performance boost
* Reordered some comparisons on hot code paths for a minor speed boost
* Removed an unnecessary call to the expensive DateTime.Now function (if you *have* to get the current time as opposed to Environment.TickCount, always use DateTime.UtcNow)
* Don't fire the queue empty callback for the Resend category
* Run the outgoing packet handler thread loop for each client synchronously. It seems like more time was being spent doing the execution asynchronously, and it made deadlocks very difficult to track down
* Rewrote some expensive math in LandObject.cs
* Optimized EntityManager to only lock on operations that need locking, and use TryGetValue() where possible
* Only update the attachment database when an object is attached or detached
* Other small misc. performance improvements
|
|
work around a Mono bug with nested delegates
|
|
exceptions thrown in the async method
* Added packet stats handling to the new LLUDP implementation
* Attempting to avoid a race condition when creating a new LLUDPClient
|
|
although not everything is reimplemented yet
* Replaced logic in ThreadTracker with a call to System.Diagnostics that does the same thing
* Added Util.StringToBytes256() and Util.StringToBytes1024() to clamp output at byte[256] and byte[1024], respectively
* Fixed formatting for a MySQLAssetData error logging line
|
|
Util.UTF8 (not all references were switched since not all OpenSim libraries reference OpenSim.Framework)
* Shrinks the largest in-memory object, the LLRAW.HeightmapLookupValue struct (only used for exporting to LLRAW terrain files), to the minimum possible size. This seems to have the odd side effect of cutting the size of the two double[256,256] terrain objects in half. Possibly an alignment optimization?
|
|
|