Chat log from the meeting on 2019-07-16
From OpenSimulator
[11:06] Sheera Khan: Firestorm split their OpenSim- and SecondLife branches today...
[11:06] Gavin.Hird @grid.xmir.org:8002: I saw that
[11:07] Misterblue Waves: what was the reason for the split?
[11:07] Sheera Khan: https://www.firestormviewer.org/second-life-opensim-are-no-longer-joined-at-the-hip/
[11:08] Ubit Umarov: yes seen that also
[11:08] Gavin.Hird @grid.xmir.org:8002: too complicated to combine the two
[11:08] Ubit Umarov: something they planed for ages
[11:08] Gavin.Hird @grid.xmir.org:8002: same as we did for KokuaOS that is now Dayturn
[11:09] Ubit Umarov: should allow cleaner code for each opensim and sl
[11:09] Gavin.Hird @grid.xmir.org:8002: I will add a new version of Dayturn later this week with the SL grids removed from the grid selector
[11:09] Ubit Umarov: lets hope they do have devs working on that opensim branch, thats the only problem
[11:10] Gavin.Hird @grid.xmir.org:8002: they might Ubit
[11:10] Gavin.Hird @grid.xmir.org:8002: repository has not changed since March
[11:10] Ubit Umarov: and they do keep adding selected sl improvements, in coordenation with us
[11:10] Gavin.Hird @grid.xmir.org:8002: but the current SL merge with eep and stuff is hellish for them
[11:11] Ubit Umarov: yeah my fear is that no one there will work on that opensim branch
[11:11] Gavin.Hird @grid.xmir.org:8002: they'll have to pull Lightshare out of that base to add eep
[11:12] Gavin.Hird @grid.xmir.org:8002: Question about Ubode, now that I have tested it more
[11:12] Gavin.Hird @grid.xmir.org:8002: if you make a very simple mesh building
[11:12] Gavin.Hird @grid.xmir.org:8002: and make a very simple correspoding collision model
[11:13] Misterblue Waves: that has always been a problem -- no one is there to give the OpenSIm version any love
[11:13] Gavin.Hird @grid.xmir.org:8002: when placed on a region uboode, one of the wall - usually one with a window cut out is as a hot knife through butter
[11:13] Gavin.Hird @grid.xmir.org:8002: you can walk right through it from the inside
[11:13] Gavin.Hird @grid.xmir.org:8002: is this a known issue?
[11:13] Misterblue Waves: Cinder used to but there was some drama
[11:13] Ubit Umarov: yeah on same code base, they had to look on each release
[11:14] Ubit Umarov: ode does single side collisions only
[11:14] Gavin.Hird @grid.xmir.org:8002: meaning?
[11:14] Ubit Umarov: all prims are double side
[11:14] Gavin.Hird @grid.xmir.org:8002: the collison model is two sided
[11:14] Gavin.Hird @grid.xmir.org:8002: so all walls but one collide perfectly
[11:14] Ubit Umarov: meaning each triangle only collides from on side
[11:14] Sheera Khan: are the normals correct?
[11:15] Gavin.Hird @grid.xmir.org:8002: yes
[11:15] Ubit Umarov: from the "outside" side
[11:15] Ubit Umarov: ie the one that has normal point outwards
[11:15] Ubit Umarov: i had some parts of cylinders doing that :)
[11:16] Gavin.Hird @grid.xmir.org:8002: the normals are pointing in the right direction both on the inside and outside mesh
[11:16] Gavin.Hird @grid.xmir.org:8002: as does the physics model
[11:16] Ubit Umarov: as you say, wrong normal :)
[11:16] Gavin.Hird @grid.xmir.org:8002: explain
[11:16] Ubit Umarov: hmm explain what gavin ?
[11:17] Gavin.Hird @grid.xmir.org:8002: what you mean by wrong normal
[11:17] Ubit Umarov: wrong normal direction
[11:17] Gavin.Hird @grid.xmir.org:8002: the normal is pointing out from the surface in the correct direction
[11:17] Ubit Umarov: you can see mesh normals in viewer
[11:18] Gavin.Hird @grid.xmir.org:8002: and they point the way they are supposed to
[11:18] Ubit Umarov: they should work
[11:18] Gavin.Hird @grid.xmir.org:8002: they don't
[11:18] Ubit Umarov: then...
[11:18] Ubit Umarov: then something wrong on that mesh :p
[11:18] Gavin.Hird @grid.xmir.org:8002: if there is a cutout like a window
[11:18] Gavin.Hird @grid.xmir.org:8002: it works everywhere else
[11:19] Ubit Umarov: are you sure it was uploaded right?
[11:19] Gavin.Hird @grid.xmir.org:8002: of course
[11:19] Bill.Blight @hg.osgrid.org: Also make sure the physics shape type is set right ... Miss those sometimes ...
[11:19] Gavin.Hird @grid.xmir.org:8002: it works correct both in bulletsim and in Havok¨
[11:19] Ubit Umarov: "of course" does not mean much on mesh uploads, viewers are creative
[11:19] Gavin.Hird @grid.xmir.org:8002: it is set right
[11:19] Ubit Umarov: bullet and havoc do double side colisions
[11:20] Ubit Umarov: at least on some meshs
[11:20] Andrew.Hellershanks @hg.osgrid.org: What program was used to create the mesh model where the normals appear to be wrong in world?
[11:20] Ubit Umarov: any prog can do that
[11:20] Gavin.Hird @grid.xmir.org:8002: bullet and havok can use a single plane for collision and it will collide from both side
[11:20] Gavin.Hird @grid.xmir.org:8002: ubode need a collision surface on both sides of a wall for example
[11:21] Gavin.Hird @grid.xmir.org:8002: the meshes I have the issue with have that
[11:21] Ubit Umarov: gavin how many times i need to tell that ode only does single side collisions??
[11:21] Gavin.Hird @grid.xmir.org:8002: so you can only make walls that work from one side?
[11:21] Ubit Umarov: yes you need 2 surfaces...
[11:21] Misterblue Waves: there is a flag in BulletSim to turn off double side collisions....
[11:21] Misterblue Waves: physics engines prefer doing single side collisions because it's a lot less computing
[11:22] Gavin.Hird @grid.xmir.org:8002: sure, but users prefers walls that work from both sides :-)
[11:22] Ubit Umarov: like if you do a wall with boxes.. all prims have outside faces
[11:22] Misterblue Waves: and one doesn't get things caught together (if a mesh passes through a little then some of the vertices are on the inside and then can't get out)
[11:22] Ubit Umarov: yes single side is lighter and has less instability issues
[11:23] Gavin.Hird @grid.xmir.org:8002: so how do you construct a room that works with ubode?¨
[11:23] Ubit Umarov: yes if a bullet gets inside a cylinder, lets say, because of tunnel, it does come out
[11:23] Ubit Umarov: you make both side walls
[11:23] Gavin.Hird @grid.xmir.org:8002: I did
[11:23] Gavin.Hird @grid.xmir.org:8002: what more?
[11:23] Ubit Umarov: with normals pointing in right directions
[11:24] Gavin.Hird @grid.xmir.org:8002: they do
[11:24] Gavin.Hird @grid.xmir.org:8002: what more?
[11:24] Bill.Blight @hg.osgrid.org: Was it a single mesh, with a hollow center or individual walls?
[11:24] Ubit Umarov: well that is a specific issue.. we can later see your model and what is its problem
[11:25] Misterblue Waves: one-sided collisions work for prims because there is always an 'inside' to the prim (a hollow cylinder has an inside and outside mesh, for instance)
[11:25] Gavin.Hird @grid.xmir.org:8002: even a single mesh with a hollow center can fail collide from one side
[11:25] Ubit Umarov: prims have faces all around
[11:25] Ubit Umarov: always
[11:25] Misterblue Waves: mesh houses are a new thing since one can have only a single mesh for a wall rather than a flattened box
[11:26] Gavin.Hird @grid.xmir.org:8002: what is a single mesh?
[11:26] Gavin.Hird @grid.xmir.org:8002: a wall is a box made of mesh faces, right?
[11:26] Ubit Umarov: i did add some code that allows avatars to collide double side.
[11:26] Ubit Umarov: but.. may not work all the time
[11:27] Ubit Umarov: ( and has instabilities on same edges )
[11:27] Gavin.Hird @grid.xmir.org:8002: that is probably what I am seeing
[11:27] Ubit Umarov: i had a model from seth i used to test that
[11:27] Gavin.Hird @grid.xmir.org:8002: particularly if the collision direction is facing south
[11:28] Ubit Umarov: but to be safe, there should be 2 surfaces
[11:28] Gavin.Hird @grid.xmir.org:8002: there are
[11:28] Gavin.Hird @grid.xmir.org:8002: I'll send you an example after
[11:28] Ubit Umarov: you can rez it here
[11:29] Andrew.Hellershanks @hg.osgrid.org: Gavin, I know some people that have made mesh houses. They find that you may have to move or rotate the house to get the physics working properly.
[11:30] Ubit Umarov: well we had a nasty surprise with .net4.8
[11:30] Ubit Umarov: as now, its new JIT fails to handle our OSD code, in Release more
[11:31] George Equus: Came by to ask a question, but can wait till current topic is done :)
[11:32] Gavin.Hird @grid.xmir.org:8002: all right
[11:32] Bill.Blight @hg.osgrid.org: Yeah that .Net issue is a pain ....
[11:32] Andrew.Hellershanks @hg.osgrid.org: Gavin, If you disappeared to test moving or rotating, did it help?
[11:32] Gavin.Hird @grid.xmir.org:8002: where to rezz?
[11:32] Ubit Umarov: the only workaround for now is to disable that new JIT
[11:32] Andrew.Hellershanks @hg.osgrid.org: Apparently you don't need to move it much.
[11:32] Ubit Umarov: err what is the mantis where i said how to disable the new JIT ?? :)
[11:33] Andrew.Hellershanks @hg.osgrid.org: Gavin, just don't drop the house on my head. I've had that done to me enough. I've even done it to myself. :)
[11:33] Bill.Blight @hg.osgrid.org: http://opensimulator.org/mantis/view.php?id=8549
[11:33] Gavin.Hird @grid.xmir.org:8002: if you walk in the door and the opposing wall is facing south, you can walk through it
[11:35] Ubit Umarov: http://opensimulator.org/mantis/view.php?id=8549#c35475 <-- the workaround for .net4.8 ( as win10 1903 installs ) is there
[11:35] Gavin.Hird @grid.xmir.org:8002: now you guys can copy it
[11:35] Gavin.Hird @grid.xmir.org:8002: it is all one mesh with a proper physics model
[11:35] Andrew.Hellershanks @hg.osgrid.org: Where is the setting to make Firestorm render the physics model?
[11:36] Gavin.Hird @grid.xmir.org:8002: Develop Render metadata probably
[11:36] Ubit Umarov: yeack hard to see the normals here :)
[11:36] Andrew.Hellershanks @hg.osgrid.org: I don't know the object so the browser doesn't show me whether physics is set on the model.
[11:37] Gavin.Hird @grid.xmir.org:8002: it is set top prim
[11:37] George Equus: I helped myself. nice bunker :) Thanks Gavin
[11:37] Gavin.Hird @grid.xmir.org:8002: you can move it now too
[11:38] Bill.Blight @hg.osgrid.org: I just rezzed one outside and all the walls are solid from inside and out
[11:39] Gavin.Hird @grid.xmir.org:8002: they should be :-)
[11:40] Ubit Umarov: so what is the problem with this ?
[11:40] Gavin.Hird @grid.xmir.org:8002: we can try and rotate it 90 deg
[11:41] Ubit Umarov: see no issues
[11:41] Ubit Umarov: inside walls and ceiling collide fine
[11:42] Ubit Umarov: hmm clastrophobic.. may i leave now ? :)
[11:42] Gavin.Hird @grid.xmir.org:8002: hehe
[11:42] Gavin.Hird @grid.xmir.org:8002: no, this is Hotel California
[11:42] Ubit Umarov: camera is also limited to inside
[11:43] Ubit Umarov: so rays also doing right collisions
[11:43] Gavin.Hird @grid.xmir.org:8002: well, then is most be something back at my place
[11:43] Gavin.Hird @grid.xmir.org:8002: thanks for the testing
[11:44] Ubit Umarov: you did select ubodemeshmerizer also right ?
[11:44] Gavin.Hird @grid.xmir.org:8002: I did
[11:44] Ubit Umarov: well don't know
[11:44] Ubit Umarov: even as i said, avatars do double side collisions on most meshes now
[11:45] Ubit Umarov: ( or get stuck badly :p )
[11:45] Ubit Umarov: well the mesh must have a opening..
[11:45] Gavin.Hird @grid.xmir.org:8002: if the ceiling height is too low and you stand up, you might get stuck in the ceiling
[11:45] Ubit Umarov: hmm i had that Seth thing somewhere
[11:46] Ubit Umarov: well seen that mantis with the workaround for .net4.8 ?
[11:47] Ubit Umarov: i hope ms does fix this :(
[11:47] Gavin.Hird @grid.xmir.org:8002: mantis #?
[11:47] Bill Blight: http://opensimulator.org/mantis/view.php?id=8549
[11:48] Andrew Hellershanks: If we are ready for a new topic, George has a question.
[11:48] Ubit Umarov: if possible just stay on .net 4.7.2,,,
[11:48] Bill Blight: It's not a specific windows 1903 issue, it is a .Net update issue, I was on 1903 through the insider program and did not have the issue, took the .Net update and now have the issue ..
[11:49] Andrew Hellershanks: Bill, ok. Good to know. My laptop is offering up 1903.
[11:49] George Equus: About water level on a region. Seems ther is only one command >set water height and that will apparently lover all water in one SIM. Is there a way to lower level on just a single region?
[11:49] Ubit Umarov: yeah im on win10 1809 but did updated .net so i could see the issue
[11:50] Bill Blight: if you happen to have KB4486153 you can try uninstalling it
[11:50] Bill Blight: but I think you will not have that if you took the full 1903 update
[11:50] Ubit Umarov: sim == region
[11:50] Andrew Hellershanks: George, did you select the region you wanted to operate on before issueing the command?
[11:50] Gavin.Hird @grid.xmir.org:8002: you can do a single region from the viewer
[11:50] George Equus: SIM 25 regions, mostly water
[11:51] Gavin.Hird @grid.xmir.org:8002: I have sims with different water levels in the regions running on the sim
[11:51] George Equus: Not tried the command yet :)
[11:51] Bill Blight: does it not work if you , Change region to that region?
[11:51] Gavin.Hird @grid.xmir.org:8002: region/estate
[11:51] Gavin.Hird @grid.xmir.org:8002: Terrain
[11:51] Gavin.Hird @grid.xmir.org:8002: water height
[11:51] Gavin.Hird @grid.xmir.org:8002: is per region
[11:51] Bill Blight: There is also an os function that does it
[11:52] Gavin.Hird @grid.xmir.org:8002: physics works fine
[11:52] George Equus: Say specifically on the server command page that this command will affect entire sim that is all regions as I understand it
[11:52] Andrew Hellershanks: I would expect that as long as you make sure to set the region you want to alter before using the command it would operate only on that one region for an instance running multiple regions.
[11:52] Gavin.Hird @grid.xmir.org:8002: agreed Andrew
[11:52] Bill Blight: not if you change region to that region, commands only affect that region
[11:52] Bill Blight: if region is root, then it is going to affect all regions in that instance
[11:52] Gavin.Hird @grid.xmir.org:8002: agreed Andrew
[11:53] Bill Blight: http://opensimulator.org/wiki/OsSetRegionWaterHeight
[11:53] Andrew Hellershanks: I think that second "agreed" was meant for Bill.
[11:53] George Equus: OK, I will test with the usual procedure then, first change to the region and see what happens on the others
[11:54] Bill Blight: if you are root in an instance affect all regions
[11:54] Ubit Umarov: better wear a breathing suite
[11:54] Bill Blight: hehe
[11:54] Ubit Umarov: :)
[11:55] George Equus: got scuba gear handy
[11:55] George Equus: thanks
[11:55] Bill Blight: hehe
[11:55] Bill Blight: flooding us
[11:56] Bill Blight: glub glub
[11:56] George Equus: you beat me to it... lol
[11:56] Andrew Hellershanks: Bah. I wish there was a search for the built-in commands. I have trouble finding some of them. I don't remember where the command is located for setting the region on which you wish to alter.
[11:56] Andrew Hellershanks: Good thing we have a lifeguard nearby.
[11:56] Ubit Umarov: nice now in summer..
[11:56] Sheera.Khan @hypergrid.org:8002: change region
[11:56] Andrew Hellershanks: The lifeguard is still above water.
[11:57] George Equus: Got the command from http://opensimulator.org/wiki/Server_Commands but don't know how relevant and up to date that page is.
[11:57] Bill Blight flüstert: yes that is right, but as it says on that page ..
[11:57] Bill Blight: change region <region name> - subsequent commands apply only to the specified region. If region name is "root" then all regions are selected
[11:58] Andrew Hellershanks: Bill, what category does that command fall under?
[11:58] George Equus: Got that Bill, thanks
[11:58] George Equus: phew..
[11:58] George Equus: General
[11:59] Andrew Hellershanks: hmm... I thought I checked General.
[11:59] Bill Blight: yes it is general
[11:59] Andrew Hellershanks: To me, it would make more sense for it to be under Regions.
[11:59] Bill Blight: set water height # - sets the height simulator wide (rather than single regions as viewer tools do).
[12:00] George Equus: Should be under Land perhaps? and with clarification that can be used on individual regions, if that is indeed the case now
[12:00] Bill Blight: well, regardless where it is if you read how the commands are supposed to work it is prety clear
[12:00] Andrew Hellershanks: ah. That indicates that selecting a region to operate on still wouldn't help if water level is changed at the console.
[12:01] Bill Blight: but that individual definition needs touched
[12:01] George Equus: "(rather than single regions as viewer tools do)." ??
[12:01] George Equus: viewer tools?
[12:02] Bill Blight: done
[12:02] Bill Blight: set water height # - sets the height simulator wide or single region if you use change region.
[12:03] Ubit Umarov: m_module.Scene.AddCommand("Regions", m_module, "set water height",
"set water height <height> [<x>] [<y>]",
"Sets the water height in meters. If <x> and <y> are specified, it will only set it on regions with a matching coordinate. " +
"Specify -1 in <x> or <y> to wildcard that coordinate.",
consoleSetWaterHeight);
[12:04] George Equus: Thanks for all info on this.
[12:04] Ubit Umarov: odd only see it changing one region
[12:05] George Equus: I should play around more with environment I guess just to get familiar
[12:06] Bill Blight: the os function might be your best bet
[12:06] Bill Blight: it does only change the region it is run on
[12:06] Bill Blight: a tad bit safer
[12:07] George Equus: always keep current backups in case I do some really bad stuff it happens lol
[12:08] Bill Blight: well water height is not gonna break much , just have to dry some things out if you screw up ... LOL
[12:08] George Equus: Is the trend of the world anyway...
[12:08] George Equus: the RW'
[12:10] Andrew Hellershanks: We have past the top of the hour. Is there any other last minute items for today?
[12:10] Andrew Hellershanks: George, did you have any other questions?
[12:10] George Equus: List long as my arm but those can wait :)
[12:11] George Equus: Kidding
[12:11] Andrew Hellershanks: ok. You can take over from Kayaker who isn't here today. He often has a lot of questions. ;)
[12:11] George Equus: lol.maybe bit more pertinent for a dev meet than mine I suspect
[12:12] Andrew Hellershanks: If they are somehow related to OpenSim they are pertinent.