Fixes Mantis #3979
Applied with changes. Changed ThreatLevel to High since all discovery
functions are a high threat. Overriding that is the responsibility
of the grid owner.
Date: Wed, 5 Aug 2009 12:45:56 +0200
Subject: [PATCH] A few minor tweaks to the MRM API's in order to make it possible for MRM's to run in a separate AppDomain without poluting the primary appdomain of OpenSim
Specifically:
Added an explicit method for getting the "globals" of the MRM, removing the need to have the MRM script code loaded into the primary domain, in order to set up proxies
Added a [Serializable] attribute to TouchEventArgs, again in order to remove the need to have MRM script code loaded into the primary domain.
---------
Applied with whitespace changes
* MRM Scripts should do appropriate cleanup within this event, to allow for clean shutdowns and script updates. This means unbinding from events you are listening to, and releasing any resources.
|Date: Wed, 5 Aug 2009 09:51:52 -0700
|Subject: [PATCH] Closed two major memory leaks for scripted objects
|
|Two major memory leaks for the scripted objects were fixed
|- One leak had to do with remoting acrossing app domains. When a script and
| its controlling agent communicate across an application boundary, it calls
| functions on a stub proxy object that then invokes the remote method on
| the object in the other app domain. These stub objects (two for each script)
| were setup to have infinate lifetimes and were never being garbage collected.
|- The second leak was the result of adding a scene object part instance method
| to a scene event and never removing it. This cause the event's delegate list
| to maintain a link to that object which is then never freed as the scene event
| object is never destroyed.
Patch applied, please direct feedback to me. Possible issue: Longtime idle
scripts like vendors may fail.
[1] Added a new OnAttach event to Scene/EventManager
[2] Hooked up existing attach event handler in XEngine.
[3] Modified SceneGraph and Scene.Inventory to trigger
attach events at the appropriate places. I was forced
to distribut the changes across two files because of
the way attach processing is distributed across the
two files.
[4] In the case of RezSingleAttachmentFromInventory it is
necessary to handle event scheduling in a special way.
There is no synchronous path available, so the fact
that this object is attached, and who it is attached to,
is cached when the ScriptInstance is created. When
the script is started, the attached handler is driven
after on_rez (but before changed, this should be reviewed).
Signed-off-by: dr scofield (aka dirk husemann) <drscofield@xyzzyxyzzy.net>
* These have been mostly replaced by the PriorityQueue sender in LLClientView (it was still using threads and poll processing!)
* Thread Reduction!
* Might have been sending more image packets then necessary previously, texture times 2. We'll see if this reduces the texture packet load.
* If this doesn't melt the Http Server, this will significantly reduce the number of threads in use on regions with many users.
* Adds AddPollServiceHTTPHandler, and RemovePollServiceHTTPHandler to BaseHttpServer
* Generic enough to be used for many long poll services, not only the EventQueue.
* Made various changes to BaseHttpServer to accommodate the new interfaces.
* This version has been significantly re-architected and may fail in unusual and insidious ways.
* Please pay attention to any errors you get and post a Mantis if you can reproduce an issue with the HTTPServer. I'm including the pdb and having the http server compiled in debug for a few weeks so that when an error occurs, it'll print the pertenant data.
* Once again, this is the full C# WebServer, not the lite branch that is included in LibOMV (we need SSL!)
* Test for prim obstructions between the avatar and camera. If there are obstructions, inform the client to move the camera closer. This makes it so that walls and objects don't obstruct your view while you're moving around. Try walking inside a hollowed tori. You'll see how much easier it is now because your camera automatically moves closer so you can still see.
* Created a way to know if the user's camera is alt + cammed or just following the avatar.
* Changes IClientAPI interface by adding SendCameraConstraint(Vector4 CameraConstraint)
This involved adding a new OnUpdatePrimSingleRotationPosition event to IClientAPI so that we can get the changed position from the client.
Btw adding new events to IClientAPI is really tedious where you have to copy the change across to at least 5 or 6 other files.
[Note this doesn't fix the bug where any rotation changes to the root prim (but not the whole linkset) cause rotation errors on the child prims.]
* Remove the draconic locking around adding an avatar to the Scene
* Handle an extreme error case when border crossing fails and user uses map to teleport to a different region on the same instance causing control commands to go to a child agent.
* Make the Set Appearance method use the proper 'remove from physics scene' method.
* It *may* help border crossings.
* It *may* help the 'on avatar rez' lag, that people have been seeing the past week.
* It may also cause physics to crash more often on failed teleports (though.. I think I got the cases covered).
This fix re-introduces a small tilt into the capsule to prevent
avatar falling through terrain. Re-introduction of the tilt means
that some direction-dependent behavior when walking over prims, but
I have tried to minimize this.
Additionally this commit allows the capsule to wobble slightly when
being pushed around the terrain. This should make walking over prims
easier, as the capsule can wobble and glide diagonally over the prim's
edge, instead of rigidly being stopped vertically against the prim's
face.
this commit finally adds the VivoxVoiceModule: it supports positional
as well as conference call type voice (currently only per region
server), region and parcel voice, speaker indication (LL client
family), direct avtar-to-avatar voice chat. NOTE: you need to obtain
an customer admin account from Vivox to be able to use this module ---
DON'T ask me about how to about an admin account, i've NO clue, we
just wrote this code.
The vector parameter to llSetPos() specifies an absolute position for
an unlinked prim or the root prim of a linkset; however, when the
function is used by a child prim, the vector specifies a relative
offset from the root prim. The changed introduced in llSetPos() treats
this value as an absolute position in all cases, which has the unintended
effect of mangling the position of child prims.
This add a configuration option to the MRM module called "hidden".
if MRM is marked as enabled, the module will additionally check for
the "Hidden" flag, before registering for client side scriping events.
When MRM is running hidden, it will not respond to client side
scripting events, giving serverside scripting modules, like MRMLoader
and MRMAddin the ability to leverage the MRM engine. This way, even
a possible clientside exploit will not be possible, while still
allowing the MRM engine to run.
The admin_close_region method removes a region from the simulator without deleting it.
The region can then be recreated by calling admin_create_region with the same UUID.
There is also a change to admin_create_region to facilitate this.The reason I want
to have this functionality is to make it possible to detach regions when they are
idle and recreate them on demand through a web interface. It's probably doable
using the existing methods by saving and loading oars, but it also doesn't seem
like that should be necessary.
Set av_capsule_tilted to false in opensim.ini. Default is true, so there is
no change in avatar behavior (and no breaking of existing content which
relies on the tilted capsule).
This commit straightens up the avatar capsule so it behaves consistently
(e.g. same collision behavior against prims regardless of which direction
the avatar is coming from; ability to fit through narrow doorways).
Please note this introduces other side effects which have not been fixed.
In particular:
* The avatar frequently falls through the terrain if it is not flat, though
the avatar behaves pretty well on flat terrain. This requires investigation
of the ode terrain collider.
* The apparent foot position of the avatar with respect to the ground
is changed. This requires investigation of the avatar height/capsule height.
Please consider this as work in progress.
The patch included updates the root and child prims' AttchedAvatar
with the right UUID. It also cleans the AttachedAvatar properties
for the root and child prims on Drop and Detach
balancer plugin work again. Create a new method, GetClientEP, to retrieve
only the EndPoint for script usage. Marked the purpose of the method
in IClientAPI.cs with a warning. Also restored the corresponding SetClientInfo
functionality.
to be constantly pushed and popped uselessly, if you are not debugging packets. This showed some really big difference in a mock test, let's see how it behaves here.
Attached is a patch that changes the oar file saving of creation date/time to an integer
instead of a string. I did this after justincc emailed me saying there is a problem
with internationalisation doing it the old way and I said I'd fix it. Its been
tested with MySQL and I've made the changes for MSSQL but that hasn't been well tested.
The new IRCd module causes an error when multiple instances of OpenSim
are run on the same machine; since the port number (6666) is hardcoded,
the second and subsequent instances crash upon startup because the port
is already in use. Attached is a patch which adds a Port specifier to the
[IRCd] section of the config file, which defaults to 6666 if not present.
By extracting this code, it should make it easier to experiment with different ways of managing the update process. [Next step to make this module based, could be to create a SceneViewerFactoryModule]
I also fixed the issue where the code also loops through the m_forcesList and copies each force to the ScenePresence's movementVector. Which resulted in only the last force in the list actually be acted on. As each copy overrode the last one. So now it only copies the last force in the list.
After noticing on several occasions that the thread counts
we see when running OpenSIm were bordering on the astronomical
I decided to seriously investigate.
After much poking I discovered that the problem is actually very
simple. The XEngine secition of the example ini says that the
timeout for an iden thread is in seconds, and an example value
of 60 is specified. In fact, this is actually resulting in a 60mS
idle timeout, which is not normally enough for a smart thread
to survive. I have added a multiplier to the XEngine constructor
so that the number now matches the published behavior.
bit of GetClientInfo that is actually used seems to be userEP as part of the
OSSL osGetAgentIP() script function. Now commented are the parts where
we serialize and copy out the *entire* packet queue of the client
(locking the packet handler in the process).
- uses Environment.TickCount for all timestamps (instead of more
costly Util.UnixTimeSinceEpoch()
- takes care of Environment.TickCount overflow (which will happens
after 24.8 days of system uptime)
- avoids instantiating List copies for each check
- gets rid of one lock() invocation
- moves calculation of loop invariant variable out of the loop itself
option for LLUDPServer. On windows .NET the default socket receive
buffer size is 8192 bytes, on recent linux systems it's about
111K. both value can be a bit small for an OpenSim instance serving
many clients. The socket receive buffer size can be configured via
an OpenSim.ini config option
- adds a general catch clause to LLUDPServer.OnReceivedData() to
prevent it submerging when an unexpected Exception occurs.
This change moves texture send processing out of the main
packet processing loop and moves it to a timer based
processing cycle.
Texture packets are sent to the client consistently over
time. The timer is discontinued whenever there are no
textures to transmit.
The behavior of the texture sending mechanism is controlled
by three variables in the LLCLient section of the config
file:
[1] TextureRequestRate (mS) determines how many times per second
texture send processing will occur. The default is 100mS.
[2] TextureSendLimit determines how many different textures
will be considered on each cycle. Textures are selected
by priority. The old mechanism specified a value of 10 for
this parameter and this is the default
[3] TextureDataLimit determines how many packets will be sent for
each of the selected textures. The old mechanism specified a
value of 5, so this is the default.
So the net effect is that TextureSendLimit*TextureDataLimit
packets will be sent every TextureRequestRate mS.
Once we have gotten a reasonable feeling for how these parameters
affect overall processing, it would be nice to autonmically manage
these values using information about the current status of the
region and network.
Note that this also resolves the pathologcal problem that
previously existed which was that a seated avatar generated very
few in-bound packets (theoretically) and would therefore be the
least able to retrieve the images being displayed by a
projector script.
out which client connection has closed. So, in multi-region sims, things
can get messy fast.
This introduces a second parameters, which is a Scene object ref. Minor
adjustments to custom modules may be required due to this change.
- Set OwnerID = GroupID for deeded objects.
- Close a security loophole that would have allowed a user with deed rights in a group to deed ANY object to that group, even if it's not owned by them and/or not set to that group
- Set LastOwnerID correctly. Handle objects vs. prims correctly.
(a) Implements the ICommandableModule interface to clean up the user interface
(b) Uses a specification for a 'copse' (collected group of trees) which permits via an xml file: Tree type; Tree Line (high and low), Seed point; Radius of Copse; Number of required trees; Initial size of seeded tree; maximum size of seeded tree; growth rate; freeze growth state
(c) Multiple Copse may be defined for a region
(d) Growth on individual copse may be frozen or restarted, or all growth disabled/enabled
(e) Copse definitions are persistant, they are reloaded from the trees present on a region restart
(f) All trees in a copse may be removed and the copse definition deleted in one command
* This is necessary to avoid a dependency problem where LLStandaloneLoginModule references IInventoryService before the LocalInventoryServiceConnector is registered
* Correct some copypasta log messages in LocalInventoryServiceConnector
Changing the names of these methods because they were being picked up by
nunit as tests even though they were marked private. Naming them Check*
after the original Test*.
Several improvements in the connectors themselves.
Several improvements in configurations.
Needed to add a hack in IUserService and UserManagerBase, to be removed when that service is refactored.
Eat collision errors --- NOTE: this fix might be naive, it seems to
have helped us getting to 81 avatars (whereas we'd crash with 20
before), but it sure would benefit from some check-over by a person
skilled in the art of ODE physics.
This change addresses two issues:
[1] It adds a flag field to the blendface call which allows the
caller to indicate whether or not the generated asset is
temporary, and whether or not the asset being replaced should
be explicitly retired fromt the memory cache. The decimal
values correspond to:
0 - Permanent asset, do not expire old asset
1 - Permanent asset, expire old asset
2 - Temporary asset, do not expire old asset
3 - Temporary asset, expire old asset
'3' corresponds to the default behavior seen today, and is
the continued behavior of the non-blendface calls.
[2] The dynamic texture routines are highly-asynchronous and can
be scheduled simultaneously on a multi-core machine. The nature
of the texture management interfaece is such that updates may
be lost, and the nature of asynchornous operation means that
they may be processed out of order. A lock has been added to
ensure that updates are at least atomic. No attempt has been
made to enforce ordering. The lock applies to the SceneObjectPart
being updated and is held for the lifetime of the TextureEntry
used to carry texture updates (the one instance carries all
faces supported by the prim).
Users of these services should remember that the dynamic texture
call is asynchronous and control will be returned *before* the
texture update has actually occurred. As a result, a isubsequent
GetTexture call may not return the expected asset id. A script
must wait for the corresponding TEXTURE_CHANGED event before
retrieving any texture information.
SL's LSL supports lone idents:
integer x;
x;
as well as lone idents in for-loop assignments:
for (x; x < 10; x++) { ... }
while those are errors in C# (MONO at least). This patch skips lone
idents in such places.
Fixes Mantis #3042.
For loops with no assignment are no longer syntax errors. For example,
this is now valid:
for ( ; i < 10; i++) { ... }
Corresponding changes to lsl.{lexer,parser} in r99 in opensim-libs.
Fixes Mantis #2501. Fixes Mantis #2884.
It seems overloading the method TestllAngleBetween confused nunit a bit
and caused it to try running the private TestllAngleBetween method. The
method's access modified prevented that from happening, and nunit
complained.
* For some reason, if a null was recieved (indicating a missing asset), the code had stopped passing that on to the waiting lock, resulting in a perpetual freeze
* This change passes the null on correctly
* Many thanks to thomax for being insistent in presenting his analysis of the problem :)
* Issue was that region server was silently dropping an XmlException caused by trying to deserialize the blank asset service response
* So make asset service return http status NOT FOUND rather than OK in accordance with REST
* and interpret this correctly in the async response so that a null object is sent back
* This means that this fix won't be active until both region simulator and server reach this revision
* If an oar save fails to get responses to all asset requests to the asset service then timeout after 60 seconds
* Timeout executes abort, since missing assets in an OAR seems bad
* This means that oar saves won't permanently hang and instead can be retried if something goes wrong with the asset service
* This is not a solution to mantis 3714. Hopefully a fix will be along shortly since I can now consistently reproduce that problem
This patch adds oar file date and time (UTC) meta data to an oar file
when it is created. It also adds a unique ID, though this id does not
in anyway identify the machine that the oar file was created on.
When an oar file with this meta data is loaded this extra information
is saved with the region settings and available via LSL through:
- osLoadedCreationDate()
- osLoadedCreationTime()
- osLoadedCreationID()
If there is no meta data these fields will be blank. Subsequent oar
file loads will erase the information for the previous oar file
load. Persistence has only been implemented for MySQL, the other
backends need updating.
Overall this allows us to much more easily identify the specific version of
software that clients are using. Its very straightforward to edit the oar file
to change the ID string to be something more human friendly.
Included in the patch is a new file OpenSim/Data/MySQL/Resources/030_RegionStore.sql
required for the MySQL DB migration.
btw I had a chat with justincc about this a few weeks ago since he
wrote the oar file import/export and he sounded happy to accept
something that included date/time information but didn't want anything
that would silently leak private information like machine names.
- fixes wild swings in memory usage related to usage of GetDrawStringSize()
We've been seeing wild swings in memory usage and a large chunk of
memory leak. From analysing this it's pretty clear that the mono
garbage collector is rather buggy! When exercised heavily it looks
like it frees more than its meant to resulting in crashes.
GetDrawStringSize() measures the size in pixels of text. To do this
memory for an image is allocated and used to call the GDI text
measure functions. Although no reference to the temporary memory
for the measuring is kept, it takes quite a while for the mono
garbage collector to clean up - so if lots calls to
GetDrawStringSize() are made at once there can be a spike in memory
usage. If the garbage collector is not fast enough then the GDI
layer runs out of memory. It also looks like the garbage collector
is not always reclaiming all of the memory.
I've attached an OpenSim patch which works around the garbage collector
issues. Instead of dynamically allocating memory for measuring
text sizes, it serialises (on a per region basis) access to a single
block of memory. The effect of this is to be nicer to the garbage
collector as it has a lot less work to do, at the cost of some
theoretical loss in performance (nothing noticeable with our tests
which hit it pretty hard).
OpenSim still does leak memory slowly, but it is a lot more stable
with this patch. I suspect that either the garbage collector misses
bits of freed memory or the GDI/cairo layer leaks a bit each time a
texture is created. Thats going to be a lot harder to hunt down, but
for reference if someone has OpenSim running on Windows it would be
interesting to see if it has the same problem as it would tell us if
its a mono/GDI problem or an OpenSim problem.
Modify dynamic texture handling so that an explicitly targetted
face is not scheduled for immediate expiration. The requirement
for precaching explicitly requires these assets to persist. They
do however remain temporary.
This approach leaves the legacy mode of operation (ALL_SIDES)
unchanged in this respect.
See the files: bin/config-include/GridCommon.ini.example and bin/config-include/StandaloneCommon.ini.example to configure and enable this caching method.
See the files: bin/config-include/GridCommon.ini.example and bin/config-include/StandaloneCommon.ini.example to configure and enable this caching method.
This appears to be due to the fact that no asset cache has been
configured, possibly as a result of the configuration changes that
have been made recently. I've attached a patch to display a message
to that effect rather than throwing an error.
The IRCStackModule used Nini.Config.ConfigCollection.Contains() to
determine whether the "IRCd" section was present in the config. This
ConfigCollection, however, stores an ArrayList of IConfig objects, not
strings, so calling Contains("IRCd") always returns false since "IRCd"
is a string, not an IConfig object.
* This includes methods such as PlaySound which take a Position as an argument, allowing you to trigger sounds arbitrarily across the scene without needing a parent object in the position.
The attached patch implements llPassTouches. It has been added
to the export/import XML along with the flag for AllowedInventoryDrop.
The MySQL backend has been updated as well, though I haven't
done one of those before so could do with a check. I added
the migration mysql file as well.
The other data backends need updating as well.
* unfortunately, while the client requires uuids and we want to be able to have arbitrary string ids, these cannot be kept in sync
* I think the problems last time were due to a serialization change
* So the major inteface version has been bumped to take care of any lingering issues here.
* This means that region servers beyond this revision can only connect to similarly uptodate grid services, and vice versa
The attached patch implements llPassTouches. It has been added
to the export/import XML along with the flag for AllowedInventoryDrop.
The MySQL backend has been updated as well, though I haven't
done one of those before so could do with a check. I added
the migration mysql file as well.
The other data backends need updating as well.
Diva Canto