NOTE: this code is work-in-progress, it's not live. it's so far
harmless and just wants to play. it should not bite anyone (yeah,
that's what they all say, i know...)
resending, timeouts, packet discarding. Add notification event for
packet discarding. Add priority scheduling for packet queues.
Add outgoing duplicate detection facility. Correct packet sequencing.
Make provisions for automatic server side throttle adjustments (comes
in next installment)
Add a test event handler (which would fire after a minute if a terrain packet hadn't been acked) to scene to handle the OnUnackedTerrain event, which currently just resends the terrain patch.
The idea of this packet tracking is for the region level application to be able to know if the client stack gave up on sending a packet.
* warning was:
OpenSim/Framework/EstateSettings.cs(277): warning CS0472: The result of comparing `libsecondlife.LLUUID' against null is always `false'. This operation is undocumented and it is temporary supported for compatibility reasons only
* coding style: space before flow control statements please! http://blogs.msdn.com/brada/articles/361363.aspx
Makes the estate dialog fully functional. Implements all client facing functionality. Moves estate data from estate_settings.xml, which is used to provide defaults, to the region data store. Creates one estate for each region, and places the region in it. Converts all region bans to estate bans.
HTTP method matching support. adapting OSHttpXmlRpcHandler
accordingly. dropping OSHttpXmlProcessor delegate in favour of good
old XmlRpcMethodHandler delegate (was the same signature).
* Potentially useful for diagnostics without needing to log in a client
* Packet queue statistics commented out for now pending a better way to cope with the information overload
Patch 7 of the region patches. Finish off the region parts of the estate dialog. Full user functionality. Terrain textures, heights, water, avatar counts, prim bonus, debug settings and region toggles can now be set from the dialog on a per-region basis. Estate stuff defaults to sane values where there are no defaults, to estate_settings.xml otherwise. Sun still b0rked :(
Make all of the toggles go live. Utilizes the new database table for almost everything. Remove lots of now unneeded settings from the EstateSettings class
Patch 5 in the region settings series. Adds a migration to ensure corrupted
data is purged and re-read from defaults. Some changes. Still no full
functionality
Changes the permissions module to make scripts permissive only when intended
Adds security checks to asset transfers to prevent hacked clients fron
requesting script sources.
Adds security checks to llClientView to verify all aspects of ownership
and permissions for inventory based script retrieval.
* This will work as long as those servers are actually registering any threads they use (does not include stuff plucked from the thread pool)
* command is now "show threads" rather than threads
The attached patch tries to read the SVN revision from local file
"svn_revision" before checking "../.svn/entries".
This allows simulators not running from the source tree to properly
display the SVN revision, if the installer generates the
"svn_revision" file.
UserManager.UpdateUserProfile(UserProfileData).
Adding UpdateUserProfile(UserProfileData) to IUserService interface.
Adding RemoteAdminPlugin.XmlRpcUpdateUserAccountMethod(...) to provide
a remote update capability.
* Migration should be automatic on sqlite and mysql
* Migration is not automatic on mssql, you will need to drop the invType column manually
* Migration should be fine, but as for any db change, I would recommend making sure you have backups before moving past this revision
And reverted CachedUserInfo back to revision 5262.
I don't think most of the inventory problems that people are reporting are due to these, but its easier to deal with one set of potential problems at a time, and I'm not going to get any time in the week to work on this anyway.
The Login service should now read/create new inventory on the inventory server that is set in a users profile.
Also added "Add-InventoryHost" console command to add a support for a new server to a region.
So it would be good if someone could test this. Set up the grid as normal, but then also run extra inventory server on a different computer (well actually it just has to be on a different network hostname, so one using "http://localhost:8004" and one using "http://127.0.0.1:8005" should work) then you need to manually edit the user profile database to set the new servers url in a user's "userInventoryURI" field.
Then on a region server, use the Add-InventoryHost to add the new server url (always include the full url, including http, but don't add a final /)
Login with that account and see if the inventory works.
Of course these needs to be made more user friendly.
TODO next is to make the login server read/write a users inventory from the correct server (the inventory url set in a userprofile)
On the region side, although not tested with multiple servers it should work if that inventory url was set, and the inventory servers urls have been added to the CommunicationsManager, using CommunicationsManager.AddInventoryService(string hostUrl)
* commented out [Obsolete(....)] attributes where no replacement feature
was available: if we want to attribute code that we think needs to be
reworked, we should define a new attribute and use that instead
(together with a little tool to retrieve all the attributed code then)
* commenting out unused variables
Currently module loading is done ad-hoc. I propose creating a simple
loader class that leverages Mono.Addins (and perhaps the new .NET
addins when they become available in mono). Attached is a basic
patch for review that compiles into HEAD, but doesn't yet replace
any existing ad-hoc loaders.
Fixes:
- Wearable icon and name sreset to default on copy/paste
- Cache is not updated when renaming/moving folders
- Partial refactor to make inventory less dependen on AssetBase having a "Name" field
- Add llGiveInventoryList() function
Add util functions to compress and uncompress strings.
Fixed a couple of modules so they use SceneCommunicationService rather than directly call functions on the CommsManager.
In function "ReadInExtraParamsBytes", parameter "data" is used without a boundary check.
if data.Length==0, an "out of index" exception will be thrown.
* Enables regular jpeg map images to be served directly from the region.
* EX: http://192.168.1.127:9000/index.php?method=regionImagecc4583cd269b41bfa525dd198e19a5c5
* This is actually HTTP server address + port + index.php?method=regionImage<REGIONUUID, no dashes>
* The Webmap image location gets printed on the console when the simulator starts up.
* JPEG data is cached so we only create the webjpeg once.
* This currently has various bugs which are more to do with the way its been hacked together than the feature itself (e.g. on save-oar, ghost prims will appear of the saved
contained items). These will be found and eliminated in subsequent patches.
* Not yet ready for use
* Only generates a new maptile after a refresh interval
* Maptile names have the UnixTimeSinceEpoch that they were generated and the regionUUID they're from, so you can know which ones are no longer necessary.
* Updates RegionInfo, so backup your /bin/Region/*.xml files.
* Implemented a hack so regions beyond the 10,000m range will show the map without having to click on the map before they'll start to show. The hack shows regions around the one you're in, but it won't show the one you're in.. you still need to click on the map to get that (not sure why yet). Additionally, the map still only shows pictures for regions that are hosted on the same instance (no change).
* Please feel free to comment if the periodic logging is causing you problems in some way - I'm loathe to add yet another switch to OpenSim.ini but will if it proves necessary
* This is the GC.GetTotalMemory() method, which I'm guessing does not include memory used by the VM (hence the memory usage reported in top on linux would be much higher)
* Unfortunately, there's some kludges with the Async manager and the llDetected functions that I have yet to decipher... so llDetected functions don't work with collision events at the moment....
having reached the intermediate level of .NET's XmlSudoku, i've
now figured out how to do deserialization using different
XmlSerializers (this stuff begins to grow on me, sigh).
[still not used code, work-in-progress]
* adding convenience property on OSHttpRequest.cs (from awebb)
having spent the last couple of days wrestling with .NET XmlSerializer
and trying to get it to do what is required by XMPP (RFC 3920 & 3921)
this is the preliminary result of that wrestling (you should see the
other guy!): XmppSerializer allows us to serialize Xmpp stanza (and
theoretically deserialize [or reify] them), XmppWriter helps avoiding
various gratuitous crap added in by off-the-shelf XmlSerializer.
this is currently not used anywhere but the plan is to use it for
at least an XMPPBridgeModule.
* You'll need to start-up the MessageingServer and set it up. It sets up like any of the other grid servers.
* All user presence data is kept in memory for speed, while the agent is online. That means if you shutdown the messaging server or the messaging server crashes, it forgets who's online/offline.
* Occasionally the region-cache will get stale if regions move around a lot. if it gets stale, run clear-cache on the messaging server console to clear the region cache.
* User tries to log-in but is already logged in. Userserver will send message to simulator user was in to log the user out there.
* From the UserServer, admin types 'logoff-user firstname lastname message'.
* Some regions may not get the message because they're not updated yet.
* Looks up UUIDNames for script time and colliders in a separate thread.
* Hopefully this'll allow you to look at top scripts on a region that has a lot of scripts without crashing your client thread.
llLoopSound sends out one packet to clients in view, so it doesn't work anymore
when clients enter later on, or the prim is modified in any way.
Solution: Stored sound data on prim, send full update instead.
llStartSound and llLoopSound now accept both LLUUIDs to a sound as well as object
inventory sound names. llStopSound clears prim data and sends full update.
If a request is made for an asset which is not in the cache yet,
but has already been requested by something else, queue up the
callbacks on that requester instead of swamping the asset server
with multiple requests for the same asset.
There's some oddness with the parcel counts, but if you can get past the oddness, you can return objects under an owner that you have permission to return.
Nothing huge, but the new button code for producing
a new script does well, but the script will not allow
for name change once created. It reverts back to new script.
i've added the OSHttpStatusCodes enumeration of HTTP status codes, have adapted
BaseHttpServer to use those.
then RestPlugin now has proper Failure handling returning proper HTTP status
codes. Regions/POSTHandler is work-in-progress.
* If a caller attempts to set PrimitiveBaseShape.ProfileCurve with a HollowShape or ProfileShape component which is not a valid enum, a warning is spat out and a default shape
subtituted
* This does not solve any underlying problem if we're missing some enum values (though it's not obvious what these are), but it should allow save-xml2/load-xml2 to be used
without causing invalid enum value related exceptions. The checks will also guard against badly behaved clients.
* This change alters the order of shape values in the xml, since it appears properties are serialized after fields (at least this is the case in mono). .net native
deserialization can cope with this it appears, though people manipulating xml manually may need to adapt (if there are any).
* This may be a good argument against relying on .net [de]serialization for our xml format.
provide OSHttpRequest and OSHttpResponse to our REST handler.
also, this adds proper RestPlugin.IsGod() checking against the X-OpenSim-Godkey
HTTP request header.
last, i added XML doc comments to RestPlugin.cs
the "if (request.UserAgent != null)" branch in the wrong place: as a result BaseHttpServer would not do
anything if the user agent header field was present....ARGH! BAAAAAAD.
You sure can. This change set restores pants (and the rest of the
default appearance) in grid mode. The
root issue had to do with serializing multi-faced textures to the
grid server. This also restores the lookup path through the avatar
factory module, as that seems the reasonable place to have it live.
Some clean up patches are coming later as well, plus testing on
standalone, but this should be in a good kicking around state for
grid users.
Fixes a bug in BaseRequestHandler.
If the length of the patter is equal to, or greater than, the length of
the actual request path, then an exception is thrown. System using is
added to support use of String.Empty. Exception is used to ensure most
efficient operation on (assumed to be most common) successful case.
This patch adds an additional handler to the existing BaseHttpServer.
It does not affect any of the existing behaviors except insofar as the
new handler may be selected.
It is selected first because its Agent-oriented nature means that it
should not be pre-empted. The new handler type is defined by IHttpAgentHandler
in Framework/Servers and has two interface methods: Match and Handle.
The Match function returns a boolean result based upon examination of
information presented in the User-Agent header.
The Handle function expects to get the request and response instances
associated with the flow. The handler is responsible for ALL activity
associated with the request except in the event of an unhandled exception,
in which case the HandleAgentRequest routine will generate a 500 status
message and close the stream.
There are two immediateley apparent (and VERY easy to implement)
improvements that could be made:
1. The Match call could be allowed to operate over the entire
request context., rather than just agent identity.
2. The Handler could return a boolean indication of whether or not
the request was actually handled, and if not, the remaining handler
mechanism could take a shot at it. This would eliminate issues
arising from pre-empted streams.
enhancing IStreamHandler and IStreamedHandler interfaces so that OSHttp{Request,Response} get passed in,
allowing RestHandlers to set response status code, redirections, etc.
* In theory, this should be a somewhat useless statistic since the user server will already have tried to use the inventory service to retrieve the avatar's skeleton. If this
fails, login is halted completely.
* Nonetheless I'm recording it anyway just to see whether it happens (yes, I'm too lazy to scan the logs...)
the attached patch set adds the (still work-in-progress but already
working) RestPlugin base class and the RestRegionPlugin code. we are using
XML serialization.
currently working:
GET /admin/regions/ -> returns UUID list of known regions
GET /admin/regions/<UUID> -> returns detailed region information
we'll be adding support for POST/DELETE/PUT and also for
/admin/regions/<UUID>/name etc.
* If the status cannot be retrieved, then the region startup will terminate.
* The aim of this is for earlier detection of situations where the region can send messages out but cannot accept incoming requests (often due to firewall issues)
* This is currently an extremely simplistic check which completely trusts whatever http uri is given by the region
* This contact may be problematic, though since the user service needs to be able to contact the region http uri, it doesn't seem unreasonable for the grid to have to be able to do so too at this stage
* This change will require a prebuild
* This is not functional anywhere, but is a incomplete proposed replacement class for IClientAPI. Assistance filling this out would be appreciated. Use IClientAPI.cs as a reference, but make sure not to copy it - instead rewrite and abstract.
This patch just tightens up console handling in BasOpenSimServer
and removes (or redirects) a couple of messages that were being issued
using Console.Writeline.
* Added back a sleep to the kick routine so users get a 'you have been logged off message' when they get kicked from the simulator for various reasons (like 'the simulator is going down')
* Linden client 1.19.1.4 appears to be happy using UDP for this request again, and Linden's own servers (1.21) don't appear to be supplying the CAPS version currently
* This should for now fix the bug where texture selection in the edit window both shows only already open textures, and later stops subsequent inventory requests from working.
The attached patch moves the sun module incrementally nearer where it
needs to be. Default behavior, i.e. no overriding configuration is to
match Second Life's diurnal/nocturnal rhythm.
All designated values are now sent to the client.There remain a couple
of unanswered questions about how this SHOULD be implemented though.
* Moved script errors to the debug channel.
* Typing '/2147483647 OK' results in a debug_channel message.
* Expanded the available parameters that are send-able through IClientAPI
* Fixed a bug with Terragen loader where it would do bad things on a non 256x256 sized terrain. Now loads the array correctly.
* Moved MapImageModule.cs to Modules/World/WorldMap
* Changed Location.RegionHandle to use Helpers.GetUlong instead of doing it ourselves.
* This should, in theory, let us make new servers with different protocols very easily (rather than the challenge we would have faced before).
* BREAKS LoadBalancing module for the moment.
* Commit 1/3 - Please dont attempt to update to this revision until all 3 are in.
* The resolution takes the form of not crashing the session if the folder the user is attempting to add already exists (e.g. there is one already with the same UUID).
* Printing out warnings instead
* Added new generic "Location" class to handle 2D integer locations. Going to use it to replace all RegionHandle and X,Y coordinate references throughout the entire project. You have been warned.
* Detaching from right clicking in world, detaches to your inventory.
* If you go up to a prim and attach it from in world, it appears in your inventory.
* Attachment placement is saved when you detach them.
* Choosing wear remembers your last attachment point from inventory.
* Wrote a method to update an inventory item's asset and sends the updated inventory item to the Client
* Wrote a recursive method to find the folder of a known existing inventory item.
* Removed a block on physics object position on creation. This might crash a region or two, let us know via Mantis if your region crashes because of a physics out of bounds error.
* Drop doesn't work. The menu item doesn't even come up. Don't know why :P.
* Now, if you own an item in a prim, you should be able to successfully drag it back into your inventory
* Temporarily, users which are not owners of the item cannot copy it, even if 'everyone can copy' is set
* This is pending fixes/implementation of upstream permission implementation
* Your friends can see your attachments now. People who appear in the sim after you've attached something can also see your attachments.
* You can position & rotate your attachments now. Positions do *not* save.
* You can detach attachments now the regular way.
* Attachments do not cross into other regions with you..(this isn't too far off)
* Updated ODE to not request terse updates on child prim.
* Now, emptying the trash should remove folders and the items they contain as well as items which were not in a subfolder.
* This will only work once both the region and grid servers have reached this revision.
* You may also need to clear your cache before this will work
* Refactoring to follow.
which will let you specify a connection string. Required
for Nhibernate, optional for sqlite (there is a sane default),
ignored for mysql and mssql until someone implements the
Iniatialise(string) method.
* On standalone, folders (and their items) should now be persistently deleted on trash emptying, as well as immediate child items
* An implementation for grid mode will follow.
Attached is a patch for adding the llGetSunDirection functionality. It was implemented by adding a parameter to estate settings for storing the sun position. The sun position is calculated and stored via the sun module everytime the client's sun position is updated. It was tested with several different srcipts on Linux and Windows
* This replaces the old behaviour of failing straight away, which could cause lost updates if the inventory service was slow in responding
* This is the first baby step to making all inventory requests behave this way, to reduce inventory lossage
the attached patch set is centered around RemoteAdminPlugin and focuses
mainly on making it more robust (i.e. more parameter checking and better
error reporting) but also we've re-implemented the LoadTerrain stuff that
got disabled during the terrain code reworking:
* missing PostInitialize() calls on region modules that were loaded
for regions created via RemoteAdmin's CreateRegion XmlRpc call
* re-implements RemoteAdmin's LoadTerrain XmlRpc call (probably lost
during the TerrainModule rework)
* adds lots more parameter checking and error reporting to RemoteAdmin
* adds a read-only property to RegionApplicationBase so that we can
access the CommsManager
* adds Exceptions to TerrainModule so that we get better error case
feedback (and can report more meaningful errors in turn)
* adds a CheckForTerrainUpdate() call to
TerrainModule.LoadFromFile() to make terrain changes effective
* adds TryGetCurrentScene(LLUUID) to SceneManager so that we can
retrieve Scenes not only by name but also by LLUUID
cheers,
dr scofield
* A much more significant fix is required to clean up the cache when a user moves out of a region, but really better handling of delayed inventory cache updates needs to be
written first, and possibly better affinity to cut down agent inventory requests when the move is between two regions hosted on the same server.
Linked objects won't scale together properly, only the root object scales.
This happens with scaling both up and down or inputting numbers in the edit dialog.
* Implements 'Teleport Home'
* User Server has to be updated for it to save your home in grid mode
* home position accuracy is in int because the grid comms ExpectUser method tries to convert to Uint and crashes if it gets a float. Added a convert to decimal in ExpectUser but to avoid a breaking change with old revisions, kept the save value in int for now. Eventually it needs to be a float, but lets release another incremental version before doing that.
* To make a landmark, you currently have to enable admin options in the advanced menu first. We're working on this.. however use the admin options solution in the mean time.
This patch makes the "Show in Search" checkbox on the viewer work. Additionally, I also discovered that show-in-search objects use the JointWheel flag, so this patch currently uses that flag. LibSL needs to add a flag to enum LLObject.ObjectFlags, "IncludeSearch = 32768" so we aren't using a legacy flag.
Additionally this patch also contains a small fix to BaseHTTPServer that lets the response content-type to be something other than text/html. For some reason this didn't get submitted with the DataSnapshot merge.
* This allows caps requests to be routed to regions where the agent is currently a root agent instead of the region that they logged into as it did previously.
* This fixes a wide variety of bugs related to 'can't do X once i've crossed a border'.
* The first seed cap request fails, the second one works. (this generates an error message on the console)
* Experimental.
ansgar and i have been working on an asterisk voice module that will allow
us to couple opensim with an asterisk VoIP gateway.
the patch below consists of
* AsteriskVoiceModule region module: alternative to the plain-vanilla
VoiceModule, will make XmlRpc calls out to an asterisk-opensim
frontend
* asterisk-opensim.py frontend, living in share/python/asterisk, takes
XmlRpc calls from the AsteriskVoiceModule
* account_update: to update/create a new SIP account (on
ProvisionVoiceAccountRequest)
* region_update: to update/create a new "region" conference call
(on ParcelVoiceInfo)
* a asterisk-opensim test client, living in share/python/asterisk, to
exercise astersik-opensim.py
this still does not give us voice in OpenSim, but it's another step on
this path...
* This is very early support which would only be triggered in a rather unlikely case (if the user server correctly received an inventory skeleton, but later on failed to return the whole inventory in a timely manner. Also, this only applies to the 1.19.1.4 client onwards
* Code cleanup and support for other failure cases (failure of inventory caching on region crossing, failure to actually add a folder/item, etc, should follow.
* Added license info to a few files it was missing from.
* Fleshed out the landbuy interfaces
* If you add '-helperuri http://127.0.0.1:9000/' to your list of parameters you tell the client to use when you start it up you can transfer ownership of parcels now in standalone. Structured gridmode requires a lot more work, see the documentation in the example money module. The example money module is not secure especially in standalone mode.
* A few things for testing.
* This makes a modification to the region registration with the grid server so that the region can send it a chosen password to identify itself. It will not cause any errors, if either one are not updated.
* This will reduce failures where the inventory server is simply slow
* Still need to properly inform the user if the inventory service has failed altogether
* A much larger race condition where the inventory request could be made before the region gets any data at all will also be fixed in the near future.
* This change also fixes a regression from two patches ago where items stopped appearing in standalone inventory
* This is required in order to work towards eliminating some inventory race conditions and to better deal with situations where a grid inventory server is slow or not responding.
* This patch adds support for saving a dynamically generated region to the filesystem (as a region xml file)
* Also adds some error checknig to make sure the dynamically generated region name, id or location are not already taken.
* Thanks Dr Scofield
* Several people have asked for a way to limit uploads, so I've decided to show people how to do this in the BetaGridLikeMoneyModule.
* Configure it in OpenSim.ini using the [Economy] header. See the bottom of the OpenSim.ini.example for more information.
* This also fleshes out the Economy API a bit more.
* Now properly dealing with prims which don't contain items - thanks to thomas for the patch which gave insight into this situation
* Also, an xml exception no longer crahes the client session
* Leaving in debugging lines in case there are further problems
* Not yet tested on Linux, though I'm just about to.
* Now you can drag an object from your inventory and give it to another avatar
* !!! Use at your own risk !!! Many things are unimplemented as of yet, including permissions (the person receiving your item can probably do absolutely everything with it)
* Also, items for the receiving end up in their root folder rather than the objects folder
* You will probably need to clear your client's cache before this will take effect
* The problem was that the underlying libsecondlife.AssetType was returning "primitive" rather than "object". This fix is in lieu of correcting/updating our libsecondlife.dll
* Sets the default estate parentID to 1.. which means that all regions are treated like private estates in the client as opposed to Linden Mainland. (no space server hiccup message). If that message annoys you and you want to get rid of it, set your parentEstateID to 1.
* This patch removes voice code into a region module. This required the implementation of events and other code to allow region modules to register their own caps handlers, and should allow different voice module implementations.
* CAVEAT: This does not provide complete voice support, it merely provides the hooks so that it can be plugged in.
* Introduced uint as field type
* Removed what should be superfluous Guid handling
* Introduced stub MySQLDataReader if we need to fix the Guid handling anyway
* (and Ansgar/Ansi)
* Fleshes out the voice stubs to better interact with the viewer CAPS requests - no actual voice support yet!
* In his own words
"the attached patch enhances the existing voice support by returning a proper voice account user and password and is preparing the config file so that we can specify a SIP server (not yet working). currently the SIP is hardcoded. the next step is to refactor voice support into a region module. working on that. "
* Added the new mapper to the mapper factory
* Made choice of mapper configurable
* This means, in hteory, that we can persist avatar appearance on MSSQL as well
* This update breaks inter-region communications, sorry.
* You will need to run prebuild.
Next, the good;
* This update solves the unexpected binary element when Linux simulators inform windows simulators and vice versa. So Linux Simulators and Windows simulators are 100% compatible again.
* This update introduces an Integer in the prim crossing method to tell the receiving simulator which XML method to use to load the prim that crossed the border. If the receiving prim doesn't support the method, the prim crossing fails and no prims are lost.
That being said, it's best to update all your simulators to this revision at once.
* You may also need to clear your cache before seeing any effect.
* These fixes may or may not affect inventory on the RC client.
* These fixes should make non-root folders work better, stop inventory failure on first login, allow trash to be emptied and make texture picker in object edit view work properly
* Fixes are 1) make initial root folder request wait for async inventory delivery, 2) deliver all folders in the initial login skeleton, not just the root child ones and 3) deal
with situations where we receive child folders from the inventory service before their parent is received.
Most likely doesn't really work in grid mode as the generated textures are marked as temporary and I don't think they are updated to the asset server. We have to either live with these textures being sent to the asset server, and manually clean them out from time to time or wait until there is some asset management system in place.
Also currently the texture is only generated at region startup, it is not updated after terraforming.
* This may resolve some current problems with non root child folders on standalone installations.
* A fix for the same problem in grid mode will come soon.
* This means that caps methods (editing scripts, poss map functions, etc) on non-home regions should now work with servers which are listening for http ports on a non default
(9000) port.
* If you are running a region server, this may only work properly once your grid server upgrades to this revision
* PLEASE NOTE: This shouldn't cause inter-region problems if one end of the connection hasn't upgraded to this revision. However if it does, the instability will persist until
the grid and region (and possibly all the region's neighbours) have upgraded to this revision.
* This revision also adds extra login related messages, both for success and failure conditions
* This may improve region memory usage
* This is a short-term response to a problem whereby some clients keep requesting the same texture even after we've sent it
* This treats the symptom rather than the cause.
* n can be adjusted by changing the constant at the top of UserTextureDownloadService if necessary
* We should not be using ASCII anywhere except for legacy compatibility reasons.
* A large number of UTF8 Encoders are being used in places where we should be using Util.StringToField instead. These have been tagged with // ENCODING FAULT
* This should fix Mantis#799 - Japanese Profile Text does not work.
* Should work in multi-region standalone and grid modes
* This should also solve other non-home region caps issues (map requests, RC client inventory requests, etc)
* We now pass CAPS information on to the destination region on region crossing, and set up a CAPS object when an agent becomes a master
* Current limitation is that this will only work if your http_listener_port is 9000
* This is a very early code cut (lots of bad practice, hard coding and inefficiency). However, I wanted to get this out there for feedback and my own sanity. Next few patches will clean up the mess.
* If a user logs in and they are noted as agentOnline. Set agentOnline = false and send a 'you're already logged in' message to the user asking them to wait 5 minutes. These 5 minutes are not enforced (because there's no foolproof interlock release yet without the grid operator getting a support call for every little sim crash). When the user gets the message, they can log-in immediately after it, but the user can expect weird results if they don't wait 5 minutes and log-in to the region they were in previously.
So this commit doesn't actually fix inventory in that client, it just stops the "loading" message being displayed forever next to a folder, and instead shows empty folders.
Next part will be to fill in the details of the items in the folders.
* Started OpenSim on SQLite to a mass of red 'unable to load prim' messages.
* If you are getting this message, after this update, when you start OpenSimulator, you'll get it one more time and then it'll work again.
Grass type is not persisted. Try to rez some grass several time to have different random grass types around, then immediately move a couple of them. Or restart the sim and logon again. You'll see all grass types reverted to a dull default one :-)
As per the suggestions in Mantis 687 (thanks Sakai) to resolve an exception
* This won't apply to any existing inventory tables you have, but shouldn't
cause any issues with old tables either
* Enabled new TerrainModule. (The king is dead, long live the king!)
* Use the console command: "script terrain save file.r32" / "script terrain load file.r32" to load/save terrain. Now uses the extension to determine file format.
* MANY of the old terrain features do not have a replacement function in the new module yet, this needs to be corrected, but has not been done so far. This being said, the new module is faster and more efficient and should be a good replacement.
converts Parse calls to TryParse
Replaces string.empty and lluuid.zero calls on user load to use real
values from DB (related to user profile save/load work. Finer grained patching)
appropriate names consisten with their use.
All done with all 94 handlers from handler001
through handler094. Hopefully we can move
forward without numbered handlers.
So anyway, it moves the Asset downloading (packet sending) to a module (AssetDownloadModule).
So now at last, AssetCache should be just dealing with fetching assets from the asset server and caching them.
* This might stop some client's constant requests for unfound textures, which is a candidate for the memory leak
* If a texture is not found then the "Image not found" texture will now be displayed clientside
* If it works, this should resolve mantis 676
* Non texture image requests do not receive this packet yet
* This will require a prebuild
* As such, only a request for a non cached asset, the response and failures show up now.
* I know lbsa71 only put these in not long ago, so if they are really still required, I think we should think whether we can move the default log4net level off 'Debug'
* This will show the packets waiting in each queue for each client logged into a region server
* These are displayed using 'show stats' on the region command line
* This is in pursuit of a memory leak.
* This will require a prebuild
* Resolve a bad logic bug in AssetCache.GetAsset()
* This may make some asset related things work better (possibly getting main map images will now be improved).
* This should stop the constant increase in the download requests statistics
* If you see stat numbers for download requests which are far from what you'd expect, please report
* This fixes some of the 'runaway downloads' problem but not all of it
* Also fix up logging messages so texture requests are reported as such rather than as assets
Patch provided by Openlifegrid.com
Adds locks around clientCircuits and clientCircuits_reverse Adds existance check on adding to clientCircuits for clients resending UseCircuit packet.
Adds locks around Clientmanager.m_clients for add/remove/tryGet calls
* This will make it more obvious if one accidentally starts one's region server in standalone rather than grid mode but then tries to login to a grid (as I am wont to do)
* We are now sending back a fault code (-32601) if no xmlrpc method is found rather than an ordinary message
* Converted all instances of "256" spotted to use RegionSize instead. Some approximations used for border crossings (ie 255.9f) are still using that value, but should be updated to use something based on RegionSize.
* Moving Terrain to a RegionModule, implemented ITerrainChannel and TerrainModule - nonfunctional, but will be soon.
* Made regions table name configurable (MSSQL only)
* Added a note in ini.example pointing out that the sql resources have to change if you change table names
* Removed duplicate picker method from GridData interface
[Provided by openlifegrid.com]
* Push asset update (invoked when clothing is altered) down into AgentAssetTransactions from Scene.Inventory.cs to join others
* I've tested that clothing creation and update still works, but please let me know if it suddently breaks for you.
* Add/correct comments
these two patches fixe the terrain load exception that occurs when
restarting OpenSim on PPC32 by making terrain loading endian clean.
it does NOT fix the client crash exception. need to hunt that one down.
'show assets' shows the current state of the asset cache (number of cached assets, requests, et c)
'clear-assets' forcibly re-initializes the asset cache thereby freeing all cached items.
'clear-assets' is not to be used lightly, as it probably introduces mem inconsistencies and doubling up of textures.
* Made Estate tools work for estate managers without needing to request admin status First
* Added code to make the Simulator version to be reported in the About box of the client
* Still not enough to solve the memory leak, though hopefully this is another step on the path
* All these changes are pretty temporary - this will be addressed with a more fundamental refactor in the future
* ParcelProperties are sent only when needed instead of on any movement
* Terse Updates and other temporary data packets are marked unreliable
* After a certain amount of users, the sim actually sends updates on things less
* Experimental
* Tested to 68 avatar with pCampBot (And it's surprising what actually causes the most lag.. the text chat!)
* The immediate upshot is that "show uptime" from the console will now show uptime on all server types (user, asset, grid, etc)
* DEV: This refactoring is far from complete - only just enough to makes the "show uptime" command common accross the servers. More is needed, but in this case it's somewhat like eating cabbage, which I prefer not to do all at once
* This may help with asset server mysql problems, since both the previous osgrid failures occurred when both a read and write were attempted in the same second
To use, see the appearance section in opensim.ini.example, set "persist = true", then add the correct connection string for your database.(see mysql-AvatarAppearance.sql in share folder for a example of the table mysql table structure).
This could possible be used in a very small grid, but would mean each region server would need to connect to the same mysql database.
But the work to move the code to one of the grid servers shouldn't be too much.
The ScriptDialogReply packet handler is a bit of a hack job. It is currently handled similar to ChatFromViewer, which will trigger the listen() event, however this is not exactly how LL's implementation works and will/can be fixed up later.
CONFIG SET section key value value value
CONFIG GET section key
CONFIG SAVE (it saves, but does it save correctly?:)
ScriptEngine will react correctly to any config change made while it is running.
Added option to try alternate UDP ports if the one configured is in use.
UDP packets are now bound to the actual outside IP address and hopefully won't "randomly" select IP on multihomed systems.
* You can type 'stats' at the REGION# prompt to get this information in grid or standalone mode
* Don't take these numbers as gospel yet, since for some reason textures displayed from inventory which require downloading from the server are being recorded as assets
rather than textures
* But I don't have any reason to believe they aren't broadly accurate.
* I've put these in so I can tell whether the high memory usage on regions is down to the asset/texture cache
* This will require a prebuild
* DEV: Only adds needed to be implemented since, as far as I can tell, assets cached are currently never released. For my part, seeing large cache memory numbers will
provoke me to think about doing something about this.
* DEV: Now switched to using a singleton to get the stats reporters rather than threading the object through various layers
* DEV: Will refactor the other server stats reporters to do this in one of the next commits
* The hooks are still plugged in too high at the asset server, but then next layer down is the database and this may be refactored soon.
* This change will requires a prebuild
* I intended to think about refactoring the servers to use this common base class, but temporary laziness won out.
* I'm leaving this here as a reminder/mind poke and for reasons of laziness (again)
* If people object to me leaving this kind of thing lying around, please say
* Typing 'stats' on the command line will given total number of successful logins today and yesterday
* A little bit more to come, probably
* Refactoring will follow next
* If you have Debian running, you should get a platform line that says, 'Found Debian!' when starting up your sim.
* If someone running Debian will confirm this does occur, that would be most helpful.
** If the region fails 3 times, then ChildAgentDataUpdates no longer get sent to that region
* Enabled Child_Get_Tasks in grid mode.
* When Child_Get_Tasks is enabled on neighbor regions, the neighbor region uses the client's draw distance to send out prim. This is a lot less likely to flood the client now since the ChildAgentDataUpdate contains both the throttle settings and the draw distance. This means that with this enabled, you can see prim in other regions in grid mode. Very experimental.
* On restart the latest save will be restored rather than the very first dragged in scripts
* Also add previously missed out database commits to separate prim inventory commit path (sigh)
for now. It doesn't build on new mono at the moment, and
isn't currently used. Please make sure this functions on
mono 1.2.6 prior to adding it back in for the main tree.
* Added a Logout message with a name on the console
* Added a *fixme* message to figure out why the current agent session is null
* After updating you may notice that there's a login <user> and also a logout<user>
Possibly solved startup crash.
Moved MySQL Commit() to inside each functions respective dataset access mutex because if not other threads would be waiting until function released mutex and modify the dataset before the function could commit
Melanie Thielker