aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/OpenSim/Region/ClientStack (follow)
Commit message (Collapse)AuthorAgeFilesLines
...
| * | add a Retry-After response header to the 503 sent when we are 2 busy to doUbitUmarov2015-08-191-0/+5
| | | | | | | | | | | | the request. just 503 could mean server down (?)
| * | do keepalive on mesh and texture GET. Dont use reusecontext any where.UbitUmarov2015-08-182-5/+5
| | | | | | | | | | | | | | | | | | setting of keepalive is wrong, it should follow the requested one ( or always as http1.1) only deny if needed (errors). KeepAlive may increase stress on number of avaiable file descritors.
| * | Track selected objects per clientMelanie Thielker2015-08-151-0/+6
| | |
| * | Maybe reduce incidence of ghost avatarsMelanie Thielker2015-06-121-1/+10
| |/
| * Remove the Invisible stuff and add more baked caching. Refactor selection of ↵Melanie Thielker2014-11-111-4/+1
| | | | | | | | textures to save to Bakes module.
| * Second part of invisible base avatar optionMelanie Thielker2014-11-101-1/+6
| |
| * Remove JustinCCs UDP patch - it is harmful to Avination's grid managementMelanie Thielker2014-11-101-2/+4
| |
| * scale ChildAgentThrottles with distance (internal to child server and notUbitUmarov2014-11-062-8/+19
| | | | | | | | root as was done before )
| * allow drawdistance to change between 32 and MaxDrawDistance, configurableUbitUmarov2014-11-061-2/+2
| | | | | | | | | | | | value default to 256, so should have no effect. Next steps needed: reduce client udp Throttles with distance, update childreen connections with significat movement and view range changes, Make disconnect be delayed in time, make disconnects be receiving region action not sender on region changes. Allow distance less than 256 to only connect to visible regions, even none. Make this be relative to camera and not agent position or region centers as it is now.
| * don't send updates for deleted agentsUbitUmarov2014-11-011-3/+4
| |
| * sadly revert to resend terseUpdates enqueuing them back into entityupdatesUbitUmarov2014-10-312-4/+47
| | | | | | | | | | queue. Viewers fail to handle correctly out of order updates with ugly visible effects. Make sure these packets don't include acks so they aren't lost.
| * set udp SocketOptionName.ReuseAddress to false, to not allow two regionsUbitUmarov2014-10-301-0/+2
| | | | | | | | | | | | to bind to same port, as seems to be possible at least with mono 3.2.8, same as patch just pushed to core by justin. This is not necessary on windows, possible a bug on some mono versions.
| * try to make SOG crossings full async. Simplify some borders checking....UbitUmarov2014-10-241-1/+1
| |
| * Revert "*TEST* send udp sync. Stop uncontroled and hidden use of IO threads."Melanie Thielker2014-10-132-18/+1
| | | | | | | | | | | | | | | | This reverts commit 8c41271b3312f2a02608ffc41b220f7fb018d6ad. Conflicts: OpenSim/Region/ClientStack/Linden/UDP/LLUDPServer.cs
| * revert to async send, since past experience showed sync didn't work whenUbitUmarov2014-10-131-2/+2
| | | | | | | | on mono
| * *TEST* send udp sync. Stop uncontroled and hidden use of IO threads.UbitUmarov2014-10-132-2/+19
| | | | | | | | Testing this may require several people.
| * move AnimationSet and AvatarSkeleton to OpenSimExtrasUbitUmarov2014-10-121-5/+7
| |
| * add simulator features AvatarSkeleton and AnimationSet reportUbitUmarov2014-10-111-1/+4
| |
| * changed skeleton, do parse the mesh on upload to check skeleton. Sooner orUbitUmarov2014-10-112-26/+26
| | | | | | | | later this parsing needs to validate the model cost also.
| * dont let test mesh go to meshes inventory folder that is not displayed byUbitUmarov2014-10-111-1/+3
| | | | | | | | viewers
| * bug fixUbitUmarov2014-10-111-2/+3
| |
| * check for avatar skeleton data on meshs headers on cost estimation. DontUbitUmarov2014-10-112-11/+101
| | | | | | | | | | | | | | let a model have more than one mesh with skeleton, for testing create a mesh inventory item, for this meshs. Add also option to read a avatar collider replacement. This information still needs to be saved somewhere so it can be checked on attachment, etc, without parsing the mesh asset again.
| * do global position X and Y with double precision in autopilot, soUbitUmarov2014-09-301-4/+7
| | | | | | | | corrected viewers do work
| * change it again...UbitUmarov2014-09-261-1/+1
| |
| * remove unnecessary argument refUbitUmarov2014-09-261-1/+1
| |
| * make c# more happyUbitUmarov2014-09-261-4/+1
| |
| * use central animationset permitions defineUbitUmarov2014-09-261-2/+2
| |
| * Extend upload verification to all upload pathsMelanie Thielker2014-09-251-0/+8
| |
| * fix animation asset name to "animatn", use "animset" for the newUbitUmarov2014-09-251-8/+10
| | | | | | | | animationSet
| * add also the name animset until its clear the name liru will useUbitUmarov2014-09-251-0/+7
| |
| * add animationset to upload assets, for now no costUbitUmarov2014-09-251-0/+6
| |
| * Fix a nullref in a debug message. If we have no SceneAgent, we can't output ↵Melanie Thielker2014-09-151-5/+8
| | | | | | | | data from it.
| * Reinstate closing the client if CloseAgen returns false. This was partMelanie Thielker2014-09-141-1/+2
| | | | | | | | of the last Ubit patch.
| * revert changes to m_pendingCache and remove client close if there is noUbitUmarov2014-09-141-21/+14
| | | | | | | | SP, as requested
| * do not use a blind expire of m_pendingCache. Entries must be valid whileUbitUmarov2014-09-131-15/+23
| | | | | | | | | | | | | | respective HandleUseCircuitCode, no matter how long it takes, so a viewer retry doesnt start a overlapping one. It HandleUseCircuitCode responsability to always remove entries, even on fails. Remove a client even if m_scene.CloseAgent fails.
| * If a client's SceneAgent is null, that client is not fully logged in orMelanie Thielker2014-09-131-0/+14
| | | | | | | | crashed while logging out. Don't reuse it, create a new one.
| * reduce MIN_CALLBACK_MS guard time. A value too high introduces a extraUbitUmarov2014-09-021-1/+1
| | | | | | | | | | throttle and makes more packets to be sent in bursts and not in steady state flow.
| * void time change between gettime callsUbitUmarov2014-09-021-4/+3
| |
| * *needs testing, not that good* change throttles math using floats and notUbitUmarov2014-09-024-125/+112
| | | | | | | | | | int64, etc. Limite brust bytes to the total rate client requested times a look ahead estimation time, Avoid queues starvation with updates waiting...
| * make use of it in GetMeshModule, and meshs are Assets not Task itens,UbitUmarov2014-08-291-43/+4
| | | | | | | | reduce MeshModule impact on udp rate
| * add method to get a category throttle rateUbitUmarov2014-08-292-38/+37
| |
| * reduce burstUbitUmarov2014-08-291-1/+1
| |
| * replace the tick() by a limit on the maximum number of tokens that can beUbitUmarov2014-08-292-15/+39
| | | | | | | | acumulated ( variable named BurtRate, not exactly a rate...)
| * disable first dripUbitUmarov2014-08-291-1/+1
| |
| * try to reduce insane high data rate udp bursts. This needs testing on aUbitUmarov2014-08-282-1/+11
| | | | | | | | region with a lot of contents. Should not affect much average rates.
| * remove misplaced commentUbitUmarov2014-08-281-1/+0
| |
| * try to make sense of throttle rate limitsUbitUmarov2014-08-282-4/+15
| |
| * add some functions for estimation of number of bytes that can be send in a ↵UbitUmarov2014-08-282-0/+13
| | | | | | | | category in specified time
| * send all presences terseupdates in same batchUbitUmarov2014-08-281-3/+3
| |
| * if we send wearables with ThrottleOutPacketType.HighPriority, then weUbitUmarov2014-08-282-125/+18
| | | | | | | | should send other avatarinformation with same priority on same Task category ( plus cleanup )
generated by cgit v1.1 at 2024-12-30 08:45:41 +0000