aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/OpenSim (follow)
Commit message (Collapse)AuthorAgeFilesLines
* Merge branch 'master' of ssh://opensimulator.org/var/git/opensimJustin Clark-Casey (justincc)2014-03-113-67/+144
|\
| * varregion: remove serialization of region terrain to floats when sending ↵Robert Adams2014-03-111-2/+6
| | | | | | | | | | | | | | | | | | | | | | | | patches. This should eliminate much memory thrashing and CPU usage while sending initial terrain. The old way of passing terrain was to convert it to an array of floats. This is really bad for large terrain (think 4096x4096 floats). This change passes a dummy float array since the real region info is used anyway and the floats are ignored. (The ignoring the terrain floats is a kludge so as to not change IClientAPI.)
| * Change terrain update sending to be triggered by frame tick ratherRobert Adams2014-03-103-66/+139
| | | | | | | | | | | | | | | | | | 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.
* | Instead of auto-creating a parcel on request if one doesn't cover a given ↵Justin Clark-Casey (justincc)2014-03-113-70/+217
|/ | | | | | | | | location, fill in gaps or extend existing parcel right after initial data load. This is in line with simpler and still existing behaviour where a default parcel is created if no parcels are in storage at all. Making this change as another step to address current problems with spurious parcels occasionally being created. Also adds regression tests for different load behaviours depending on existing stored parcel data. Relates to http://opensimulator.org/mantis/view.php?id=7035
* Don't start KeyframeMotion timers until all the regions are ready. This ↵Oren Hurvitz2014-03-111-2/+28
| | | | prevents problems in megaregions (prims that think they've crossed over to other regions).
* After an object with KeyframeMotion is copied into inventory, resume the ↵Oren Hurvitz2014-03-111-2/+15
| | | | motion (previously it remained stopped)
* Merge branch 'master' of ssh://opensimulator.org/var/git/opensimJustin Clark-Casey (justincc)2014-03-111-1/+3
|\
| * Add Varregion support to osGetRegionSize OSSL functionDev Random2014-03-071-1/+3
| | | | | | | | Signed-off-by: Michael Cerquoni <nebadon2025@gmail.com>
* | Send group notices through the same messaging module mechanism used to send ↵Justin Clark-Casey (justincc)2014-03-116-101/+195
|/ | | | | | | | group chat to avoid timeout issues when sending messages to large groups. Only implementing for XmlRpcGroups initially to test. May require MessageOnlineUsersOnly = true in [Groups] to be effective. In relation to http://opensimulator.org/mantis/view.php?id=7037
* Remove try/catch in LandManagmentModule.GetLandObject() - this is very old ↵Justin Clark-Casey (justincc)2014-03-071-70/+51
| | | | code and the caught exceptions can no longer occur.
* Lock m_landlist whilst loading parcels from storage to prevent a race ↵Justin Clark-Casey (justincc)2014-03-071-7/+18
| | | | against any parcel auto-creation in GetLandObject()
* Only auto-create a land parcel when there is none in a specified tile if ↵Justin Clark-Casey (justincc)2014-03-071-1/+13
| | | | | | | there is more than 1 existing land parcel. This is because there are still issues with bad parcels being generated in http://opensimulator.org/mantis/view.php?id=7035 Theorizing now that it's possible that something is calling GetParcel() before any parcel data has been loaded from persistence.
* Add regression test for sending group notices via xmlrpc groups connector.Justin Clark-Casey (justincc)2014-03-075-48/+253
|
* Add scene name to bad parcel add loggingJustin Clark-Casey (justincc)2014-03-061-2/+3
|
* refactor: Simplify land object by using c# get/set auto-properties where ↵Justin Clark-Casey (justincc)2014-03-061-23/+13
| | | | applicable.
* minor: slightly simplify code in LandObject.ContainsPoint()Justin Clark-Casey (justincc)2014-03-061-1/+1
|
* Remove array initialize calls in LMM - these are unnecessary as the VM ↵Justin Clark-Casey (justincc)2014-03-061-4/+0
| | | | already does this.
* minor: Increase size of parcel name field in "land show" console command ↵Justin Clark-Casey (justincc)2014-03-062-13/+18
| | | | output. Construct table using CDT rather than string formats
* Prevent adding a land object if it overlaps any existing objects that have ↵Justin Clark-Casey (justincc)2014-03-064-30/+215
| | | | | | | | not had their bitmaps adjusted. This is to prevent an immediate problem in http://opensimulator.org/mantis/view.php?id=7035 where a development code bug occasionally overlays all the existing parcels with a blank parcel owned by the estate manager and to gather more data. My guess is that this parcel is being created by the new code in LandManagementModule.GetLandObject(), probably some race between threads since this only happens occasionally. Adds regression tests for this case and for parcel subdivide.
* Add UUID and ready status (whether region has finished starting up) to "show ↵Justin Clark-Casey (justincc)2014-03-052-10/+60
| | | | regions" console output.
* Don't request a terse update after local teleport - this will be done by the ↵Justin Clark-Casey (justincc)2014-03-051-6/+0
| | | | main frame loop anyway and has the potential to race that thread.
* If an avatar is sitting, send out position updates to clients for any ↵Justin Clark-Casey (justincc)2014-03-054-5/+48
| | | | | | | | change, not just those outside the usual tolerances. 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.
* When positioning agent with PRIM_ROTATION in llSetLinkPrimitiveParams(), set ↵Justin Clark-Casey (justincc)2014-03-032-12/+38
| | | | | | | the global rotation rather than the local rotation Functionally the same as the patch in http://opensimulator.org/mantis/view.php?id=7044, thanks Aleric. This commit also extends the regression test
* Maybe slightly reduce warp3d memory leakage by disposing of decoded bitmap ↵Justin Clark-Casey (justincc)2014-02-281-8/+6
| | | | | | in Warp3DImageModule.GetTexture() However, this still appears to be leaking massively, at least for me. Possible cause is warp3D using GDI objects internally and not disposing of them.
* Set up a StreamReader and call ReadToEnd() instead of using the ↵Justin Clark-Casey (justincc)2014-02-271-41/+13
| | | | | | GetStreamString() extension method This eliminates some stream seeking that was never necessary and makes disposable of the StreamReader consistent with other code
* Remove long unused UntrustedWebRequest classJustin Clark-Casey (justincc)2014-02-271-230/+0
| | | | This purports to check web requests but doesn't appear to actually do that.
* refactor: More consistently use using construct within WebUtil to match ↵Justin Clark-Casey (justincc)2014-02-271-225/+218
| | | | other existing code there.
* Add more details to InvalidOperationException logging in ↵Justin Clark-Casey (justincc)2014-02-271-3/+4
| | | | SychronousRestFormsRequest.MakeRequest().
* Stop LSL_Api.GetLinkEntities() from creating an unused list on every ↵Justin Clark-Casey (justincc)2014-02-271-9/+6
| | | | parameter except LINK_THIS. Simplify code.
* minor: remove unused variable from recent commit 0e23374Justin Clark-Casey (justincc)2014-02-271-2/+0
|
* refactor: Make osGetPrimtiveParams() and soSetPrimitiveParams() use common ↵Justin Clark-Casey (justincc)2014-02-271-33/+12
| | | | get/set params code in LSL_Api rather than an old copy&pasted section.
* Add Util method to load OpSys env varsDev Random2014-02-263-124/+104
|
* Add some regression tests for previous commit 0e23374Justin Clark-Casey (justincc)2014-02-261-0/+115
|
* Implement PRIM_ROTATION, PRIM_ROT_LOCAL, PRIM_POSITION and PRIM_POS_LOCAL ↵Justin Clark-Casey (justincc)2014-02-261-17/+128
| | | | | | | when manipulating avatars via llSetLinkPrimitiveParams() Combination of core parts of Freaky's patch at https://github.com/ft-/opensim-patches/blob/master/opensim-llsetlinkprimitive-agent-fix.patch plus further adjustments from myself. Resolves Mantises 6121, 6421, 6573, 6657
* Improve regression test TestSetAppearance()Justin Clark-Casey (justincc)2014-02-241-42/+69
|
* Re-enable regression TestSaveBakedTextures()Justin Clark-Casey (justincc)2014-02-241-4/+2
|
* Stop mesh avatars that specify the alpha texture in their bake slots from ↵Justin Clark-Casey (justincc)2014-02-243-38/+70
| | | | | | | causing the simulator to continually request that they rebake This is because the alpha texture is not in the cache, we must continue to have the fallback of looking for these and other persisted bakes in the asset service. Relates to http://opensimulator.org/mantis/view.php?id=6927
* Make Scene.DefaultDrawDistance to be the max of the region size and theRobert Adams2014-02-231-1/+3
| | | | user setting. Seems this parameter has many side effects.
* Fix problem of hurricane speed winds. Thanks Vegaslon.Robert Adams2014-02-222-11/+6
|
* Added 2 new behaviors to pCampBotDiva Canto2014-02-211-0/+6
|
* Merge branch 'master' of ssh://opensimulator.org/var/git/opensimDiva Canto2014-02-2134-1150/+780
|\
| * If texture decode fails in Warp3D map maker, log uuid of asset that failed ↵Justin Clark-Casey (justincc)2014-02-201-1/+14
| | | | | | | | to decode along with exception
| * If GetFolderContent called by WebFetchInvDescHandler.Fetch() fails for some ↵Justin Clark-Casey (justincc)2014-02-202-2/+10
| | | | | | | | reason and returns null, log and return empty contents rather than throwing an exception that ends up terminating the simulator.
| * minor: Add some method doc to AFM,SetAppearanceAssets()Justin Clark-Casey (justincc)2014-02-201-0/+6
| |
| * In core.groups GroupsServiceRemoveConnector, if GetGroupRecord() or ↵Justin Clark-Casey (justincc)2014-02-201-4/+6
| | | | | | | | | | | | | | UpdateGroup() do not have a RESULT parameter in the result message, return null rather than fail with NullReferenceException This check was already done by other methods. Looks to resolve http://opensimulator.org/mantis/view.php?id=7012
| * For now, do not replacing missing wearables with default wearables if itme ↵Justin Clark-Casey (justincc)2014-02-201-144/+190
| | | | | | | | | | | | | | | | details cannot be found. This is causing many issues on OSGrid, possibly due to teleporting timing differences with simulators hosted in different places or HG teleports Added a bit more logging for debug purposes. See http://opensimulator.org/mantis/view.php?id=6939
| * add newRegion parameter to CrossAgentToNewRegion event and trigger the event ↵dahlia2014-02-203-4/+5
| | | | | | | | after crossing thread is invoked
| * add EventManager event OnCrossAgentToNewRegiondahlia2014-02-192-0/+25
| |
| * varregion: make scene default draw distance to be the maximum size of theRobert Adams2014-02-191-1/+8
| | | | | | | | | | | | 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.
| * Repair check for if in region position (I mean || is kinda like &&).Robert Adams2014-02-152-2/+2
| |