Chat log from the meeting on 2018-04-10

From OpenSimulator

Jump to: navigation, search

Short discussion about flexible sculpts and meshes
Intense discussion about the lost capability to have prims below z = 0
Short notice about the internal representation of terrain z-values, the httptests-branch and the status of SSL
There's a spreadsheet of the differences between the database tables in MySQL and SQLight

[10:39] Sheera Khan: Do you happen to know if day_night_offset is implemented in 0.9.1? It wasn't working as expected on a region - but I don't know for sure if that was due to an operator error... I could not test it myself
[10:52] tx.0h i love my spacenavigator
[10:52] Kayaker Magic: What viewer are you using with it?
[10:52] tx.0h singularity
[10:53] tx.0h the viewer with no moon
[10:54] Kayaker Magic: I've been using Singularity to help convert sculpts to mesh recently.
[10:55] tx.0h there was a patch, long time ago, which allowed flexible sculpties
[10:55] Arielle.Popstar oh?
[10:56] tx.0h funny, right?
[10:56] Arielle.Popstar didnt make it into core?
[10:56] Ubit Umarov: a viewer patch i guess
[10:56] Kayaker Magic: I saw a youtube video about it.
[10:57] Kayaker Magic: LL decided there weren't enough bones in the original design to make it visually worth it.
[10:57] Arielle.Popstar would need to be server side wouldnt it?
[10:57] tx.0h
[10:57] Kayaker Magic: Have to be both of course.
[10:57] tx.0h no, client thing
[10:57] Ubit Umarov: no flexi is viewer side
[10:58] Ubit Umarov: well server my need a few checks.. not sure
[10:58] Arielle.Popstar what about flexi mesh?
[10:58] Ubit Umarov: viewers do not have those
[10:58] Ubit Umarov: those things are heavy to render
[10:59] Ubit Umarov: remember trees moving with wind ?
[10:59] Arielle.Popstar if prims are actually mesh and prims can flex than mesh should be able to flex
[10:59] Ubit Umarov: that was removed from SL, but singu keeps the option
[10:59] Ubit Umarov: if you turn it on, viewer fs drops a lot
[11:00] Ubit Umarov: ( thats kinda a flexi )
[11:00] Arielle.Popstar sounds reasonable
[11:00] Ubit Umarov: i mean linden trees
[11:00] Ubit Umarov: well it is bad
[11:00] Arielle.Popstar but fs drop would depend on how many trees?
[11:00] tx.0h well, but iit makes things a bit more realistic
[11:00] Ubit Umarov: all games have trees and vegetation bouncing with a fake wind
[11:00] Kayaker Magic: Andrew! How goes the hunt for the Robust memory leak?
[11:01] Ubit Umarov: vital for imerssion
[11:01] Ubit Umarov: sl removed it :p
[11:01] stiofain mactomais: we have noticed since switching to 9.1.0 that u can no longer build below z=0 the sunken titanic is at -80 but when loading the oar only prims above 0 rezz
[11:01] Andrew.Hellershanks Kayaker, the hunt was delayed. I wound up working on my website and working on a 3D model.
[11:01] Arielle.Popstar i was reading article by the guy who coded prims for s/l and according to that article, they could do a lot more with prims then what s/l allows
[11:02] Arielle.Popstar which would allow for my more complicated shapes
[11:02] Ubit Umarov: all other games now do ( with meshs )
[11:03] Arielle.Popstar prob with mesh is it raises the bar so high for creation
[11:03] Andrew.Hellershanks The exception may be minecraft. :)
[11:03] Arielle.Popstar what immersion is there if everything has to be done in blender
[11:03] tx.0h yeah, minecraft doesn't need flexi
[11:03] Ubit Umarov: sl prims are only of one type, that can be generated by a 2d curve extruded along a path
[11:03] tx.0h what about a sphere?
[11:04] Ubit Umarov: with a limites set of curves and path shapes
[11:04] Ubit Umarov: a sphere is half 2d circle moved around a circle
[11:04] tx.0h right
[11:05] Ubit Umarov: a torus is a circle with offset moved on a circle
[11:05] Arielle.Popstar thing is that s/l has taken a path leading to much greater complexity in learning cirve to content creation
[11:05] Ubit Umarov: the torture thins are applied during those moves
[11:05] Arielle.Popstar so that it is no longer for the noob and amatuers
[11:05] stiofain mactomais: there was a mantis about the negative z probs but it seems to have run out of steam anyone know any more?
[11:05] Ubit Umarov: prims are the easy thing to use
[11:06] Ubit Umarov: yes there is a min Z now
[11:07] Arielle.Popstar seems it be good if Opensim started looking at adding inworld build tools and simplify ones already there
[11:08] Ubit Umarov: we cant, that be well then needs a viewer gui
[11:08] stiofain mactomais: at the time it was fixed before in 2012 there was a lot of discussion and the outcome was that it was thought a very useful addition and setr os apart from sl
[11:09] stiofain mactomais: why the change?
[11:09] Andrew.Hellershanks stiofain, have you found there is still problems with use of negative Z?
[11:09] stiofain mactomais: yes it doesnt work at all
[11:09] stiofain mactomais: since switching to 9.1.0 that u can no longer build below z=0 the sunken titanic is at -80 but when loading the oar only prims above 0 rezz
[11:09] Ubit Umarov: that mantis does have some discussion on why
[11:10] tx.0h welll...
[11:11] Ubit Umarov: by design sl height is not as rl where zero means mean water level
[11:11] Ubit Umarov: not here
[11:12] Ubit Umarov: thhink there is a option to load a oar adding a offset to all prims, did you test that ?
[11:12] stiofain mactomais: but it is a recent change was it causing probs?
[11:12] stiofain mactomais: yes
[11:12] Ubit Umarov: yes recent.. about 2015
[11:12] Andrew.Hellershanks It was causing the problem listed in the bug report.
[11:12] Ubit Umarov: :p
[11:13] stiofain mactomais: the oar is good loads on OSgrid (Dev) 90d88f2: 2016-08-28
[11:13] Ubit Umarov: ( actually on that mantis date )
[11:13] Ubit Umarov: i said. load it adding a offset
[11:13] Ubit Umarov: thing oar load does have that option
[11:13] stiofain mactomais: we did that too
[11:13] Ubit Umarov: just move all things 100m up
[11:14] stiofain mactomais: all is good but looks stupid having a shipwreck in sky
[11:14] George Equus: not an option really... pretty unrealistic in a city scenario
[11:14] Ubit Umarov: move water to 100m daaahhh :p
[11:14] tx.0h have you tried 'translate scene' after rezzing above 0?
[11:15] stiofain mactomais: it is part of 9 regions
[11:15] stiofain mactomais: it needs be negative z
[11:15] Ubit Umarov: again height here is not like RL one.. it is not relative to water level
[11:15] stiofain mactomais: was a lot of discussion at the time
[11:16] stiofain mactomais:
[11:16] Ubit Umarov: and there are issues with terrain encoding
[11:16] tx.0h i have no clue why it got changed
[11:16] Ubit Umarov: ( and terrain sets a min rez height also )
[11:17] Ubit Umarov: and sl terrain height only has 256 bits
[11:18] Ubit Umarov: so it was a pain to suport more that 256m terrain heights diferences
[11:18] stiofain mactomais: is 80m
[11:18] Ubit Umarov: and where is the zero..
[11:18] stiofain mactomais: -80
[11:18] Ubit Umarov: should terrain be from -127 to 127m only?
[11:19] stiofain mactomais: at 0
[11:19] Andrew.Hellershanks There is the terrain multiplier but you lose a bit of resolution when using it to increase the range of terrain height.
[11:19] Ubit Umarov: sl defined to be 0 to 256 ( or 512 )
[11:19] Ubit Umarov: so .. min is Zero
[11:19] tx.0h sl didn't allow 1024x1024 regions
[11:19] Ubit Umarov: it is not x,y its Z on this talk
[11:19] stiofain mactomais: or 20k prim builds
[11:19] Ubit Umarov: height
[11:20] Ubit Umarov: se we had to normalize that... so 0 is the min
[11:20] Kayaker Magic: Terrain height is floating point, you can load 16bit monochrome images to terrain and get better than 256 levels of terrain.
[11:20] tx.0h well, but opensim accept 1024x1024, they didn't asked a linden. so why not keep building below 0 ?
[11:20] Ubit Umarov: at least for now with current viewers
[11:21] stiofain mactomais: discussion got a bit metaphysical last time but ppl agreed negative z was needed and jcc implemented it
[11:21] Ubit Umarov: internaly i just gave up all that mess and made terrain height to be a float
[11:21] Ubit Umarov: but we still need to send it to viewers etc etc
[11:21] George Equus: did a z of negative value really cased any trouble for anyone?
[11:21] Ubit Umarov: at least whould force odd offset options to be added
[11:22] tx.0h kayaker, really? have you tried 16bit monochrome ? it could help me, because i use diagonal ditches
[11:22] Ubit Umarov: ( like some versions on terrain saving we had before i made them float type )
[11:23] Kayaker Magic: Yes, problem is making 16bit png files is hard, very few graphics programs support it well.
[11:23] Ubit Umarov: well we do have a 32bits raw format now :)
[11:23] Ubit Umarov: ( floats )
[11:24] Ubit Umarov: but its a long time since i touch terrain..
[11:24] Ubit Umarov: any testing on httptests branch ?
[11:24] stiofain mactomais: so will devs look at this again?
[11:25] Ubit Umarov: not for now stiofain :(
[11:25] TG.Lucan I have set one of our public grids to use httptests
[11:25] Ubit Umarov: try that oar load offset
[11:25] TG.Lucan robust and region
[11:25] stiofain mactomais: i think it is pretty important
[11:25] Ubit Umarov: save all regions to oar, then load them with proper offsets
[11:25] Ubit Umarov: it should work
[11:26] stiofain mactomais: the offset ruins the build
[11:26] Ubit Umarov: it should not
[11:26] tx.0h stiofain, can you put the titanic on a long stick and dip it deeper?
[11:26] Ubit Umarov: and on last resource
[11:26] George Equus: nothing go deeper than the standard z 0 now
[11:26] Ubit Umarov: add a prim to the the ship make it root and place it above 0 :p
[11:27] stiofain mactomais: it looks ridiculous if i move all the regions up the detroit and other regions are at diff levels
[11:27] Ubit Umarov: well that is a tradicional issue
[11:27] TG.Lucan Anything specific required from testing httptests, there is a lot to opensim?
[11:27] tx.0h well, if the pivot is above zero it should be ok, or not?
[11:27] Andrew.Hellershanks Objects outside normal region boundaries have caused a problem before. Region admins can make use of the feature that removes all items outside normal bounds. That would remove all objects with -Z values.
[11:27] Ubit Umarov: regions with different water levels nearby .. so ugly :)
[11:28] George Equus: I dug down to minus 200 and found nothing including an avatar can gop deeper than zero level anyway
[11:28] Ubit Umarov: andrew h does not want to remove them :p
[11:28] stiofain mactomais: is about 200 linksets and loads of individual items like instruments etc is 20k prims
[11:29] Andrew.Hellershanks Ubit, I'm just mentioning another issue one can have if you were using regions with prims having -Z values.
[11:30] Ubit Umarov: adn anyone testing httptests ? :p
[11:30] tx.0h what is httptests?
[11:30] TG.Lucan yes, but not yengine
[11:31] Ubit Umarov: well that is just part of it
[11:31] tx.0h what is part of what?
[11:31] TG.Lucan on linux
[11:31] Ubit Umarov: yengine is just part of the changes on httptests
[11:31] Bill.Blight htttptests what's that?? HAHAHA  :P
[11:31] Andrew Hellershanks: tx.0h: httptests is a branch of OpenSim that Ubit has been working on. Lots of changes in there.
[11:32] Bill.Blight hehe
[11:32] Ubit Umarov: it should be our master now :(
[11:32] TG.Lucan snail dev :)
[11:32] Bill.Blight yep
[11:32] Arielle Popstar: what sort of changes?
[11:32] tx.0h aha! ok. does it support https ? for hg ?
[11:32] Ubit Umarov: but we are having a few tech issues to make it master
[11:33] Ubit Umarov: no, does have some improvemente on ssl but not for full use
[11:33] tx.0h :-(
[11:33] Bill.Blight https works, but it is not happy about it
[11:33] Ubit Umarov: i did run a region fully on ssl.. but iwithou money modules and lsl secure http
[11:33] Ubit Umarov: there is a collision on those
[11:34] Ubit Umarov: and robust side it is a pain
[11:34] tx.0h ubit: sound very interesting. i hate, that everyone can sniff the protocol
[11:35] George Equus: Did we leave the negative depth discussion??
[11:35] Ubit Umarov: it is heavier of course..
[11:35] stiofain mactomais: sorry but is that the end of negative z discussion?
[11:35] Ubit Umarov: ssl is only to use where needed
[11:35] Andrew Hellershanks: George, It appears we have changed to a new topic.
[11:35] TG.Lucan it is running here, don't think you will see any difference from an avi prospective
[11:35] Bill.Blight What is there to sniff really ??
[11:35] Bill.Blight Stick it behind a https proxy
[11:36] George Equus: and why is that I wonder Andrew....
[11:36] stiofain mactomais: id like to point out i know of 2 university projects having same probs
[11:36] Ubit Umarov: a issue is that many ppl will not pay for a ssl cert
[11:37] Bill.Blight Viewers and viewer devs are picky on what types of SSL certs they will support without complaining
[11:37] Ubit Umarov: and self signed are refused by default etc etc
[11:37] Andrew Hellershanks: George, Ubit changed the topic by asking a question.
[11:38] stiofain mactomais: and that over 10 ppl worked on titanic and it was reported on in rl media and brought ppl to osg
[11:38] Andrew Hellershanks: stiofain, what range of negative values are the Uni's trying to use?
[11:38] George Equus: I question changes that do not impose any problems for anyone although it might be seen as a "bug" or "error". Just leave those please...
[11:38] Ubit Umarov: yes. even some payed certs out there are refused, bc don't have a accepted root autority
[11:38] tx.0h well, the below z=0 is a matter of politics, as i understand. you have my vote to change it to your needs. it's opensource. but i can't understand the decision making process...
[11:38] Ubit Umarov: below 0 was not political
[11:39] stiofain mactomais: ai austin goes to -100 the other is at 40 i think
[11:39] Ubit Umarov: and im not going to repeat what i already said...
[11:39] George Equus: did it cause any problems elsewhere or not Ubit?
[11:39] stiofain mactomais: i think it is a very important matter
[11:40] Ubit Umarov: it did
[11:40] tx.0h what else? it ran without this z=0 limit for a long while now.
[11:40] George Equus: what and for whom
[11:40] Bill.Blight Well "Needing" to have something at -z could be considered poor planning, raise the entire grid land mass and water 100 meters and you have an extra 100 meters to play with
[11:40] stiofain mactomais: impossible bill
[11:40] Arielle Popstar: already went over that Bill before you came
[11:41] stiofain mactomais: this was resolved in 2012 for titanic centenary
[11:41] stiofain mactomais:
[11:42] stiofain mactomais: i cant see the reason for change/regression
[11:42] George Equus: If people was aware of possibility to build down I think would be appresiated. seen a few caves... all ih sky boxes.. not terribly relistic, only possible to connect via tp
[11:43] TG.Lucan There is FATAL UNHANDLED EXCEPTION during shutdown on httptests
[11:44] tx.0h caves are a advanced capability of minecraft
[11:44] Bill.Blight I don't see that
[11:44] Arielle Popstar: voxels
[11:44] George Equus: I ask again, what problems was caused by the previous state with negative level, anyone complain of specific issues caused by this?
[11:44] Bill.Blight the Exeception
[11:45] TG.Lucan System.Threading.Monitor.Exit(object)
[11:45] Andrew Hellershanks: George:
[11:45] Ubit Umarov: null ref tg ?
[11:45] TG.Lucan yes
[11:45] Ubit Umarov: well where in code ?  :)
[11:46] TG.Lucan one sec
[11:46] stiofain mactomais: yea andrew i seen that but seemed to run out of steam
[11:46] Arielle Popstar: thats the mantis that was being discussed already Andrew
[11:46] Ubit Umarov: better mantis it with the stack dump tg
[11:46] TG.Lucan you not see this?
[11:47] Ubit Umarov: no
[11:47] Ubit Umarov: neither bill i think
[11:47] Andrew Hellershanks: Arielle, yes. George was asking whether someone had a problem with negative Z. The problems are in the bug report.
[11:48] Ubit Umarov: sorry i don't have memory to go back to a 2015 issue ( even older )
[11:48] stiofain mactomais: id also mention the belfast regions have had more visits than anywhere else on osgrid except lbsa a lot to titanic
[11:48] Arielle Popstar: only have to go back to where you brought in the regression UBit
[11:48] tx.0h before we end it up, you remember the sqlite/mysql problem? i made a spreadsheet with the differences and wrote a migration script. do you need this shite?
[11:49] Ubit Umarov: it was not a regression.. terrain should had never had z<0
[11:49] George Equus: I read that Andrew, didn't see this problem on the Titanic build, besides is 3 years old post. anyone complained recently on specific issues caused y this ability (now removed)? If not, I don't see the need to revert to current conditions, destroying a complex build in the process.
[11:49] stiofain mactomais: it was jcc instigated it after long discussion
[11:50] stiofain mactomais: melanie was also involved
[11:50] tx.0h respect the rules of the elder
[11:51] stiofain mactomais: abd all worked on OSgrid (Dev) 90d88f2: 2016-08-28
[11:52] Arielle Popstar: /
[11:52] Andrew Hellershanks: Is there two issues here? One is terrain with a -Z and the other objects with negative Z?
[11:52] Arielle Popstar: any idea when it was broken?
[11:52] George Equus: Think it worked a couple of months ago
[11:53] stiofain mactomais: i dont know of any terrain issues
[11:53] stiofain mactomais: the seabed uses a texture that still shows fine
[11:53] George Equus: loading the OAR from that region to SOAS work fine, but that is really an antique...
[11:54] George Equus: loading same OAR here now remove everything below 0 but bottom can be reached by an avatar still
[11:54] Kayaker Magic: Ubit, will httptest work with Robust v0.9.1 dev?
[11:55] Ubit Umarov: think so yes.. for now at least :)
[11:55] George Equus: including attachments (prims)
[11:55] Ubit Umarov: george deos not oar load have a offset option?
[11:56] tx.0h yes, it has, and there is also scene translation
[11:56] Ubit Umarov: so ??
[11:57] George Equus: Ubit, please don't ignore this.Consider at least to restore the ability to rez as before. It won't harm anything or anyone on the grid as far as I can tell.
[11:58] stiofain mactomais: yes please have a look into it .... i think its a reasonable request
[11:58] stiofain mactomais: OSgrid (Dev) 90d88f2: 2016-08-28 is the last version i have a record of it working on
[11:59] tx.0h 2nd try: before we end it up, you remember the sqlite/mysql problem? i made a spreadsheet with the differences and wrote a migration script. do you need this shite?
[11:59] Andrew Hellershanks: The bug report mentions that you can't enter negative values in the viewer dialog so how do you get prims with negative values other than by using the move translation handle?
[11:59] TG.Lucan Kaymaker, I am running a full grid mode using httptests
[11:59] stiofain mactomais: we built it in situ andrew
[11:59] Andrew Hellershanks: tx: It would be useful to file the information in a bug report so it doesn't get lost.
[12:00] George Equus: Unless it do cause severe problems, as far as I can see, it do not, for any user on the grid as a whole.
[12:00] tx.0h andrew: ok, i mantis both
[12:00] Bill.Blight I'm running my own flavor of httptests on my whole grid
[12:00] Andrew Hellershanks: tx: ty
[12:00] stiofain mactomais: it was a nice feature that set os apart from sl
[12:02] stiofain mactomais: here are details just for the record last working version
OSgrid (Dev) 90d88f2: 2016-08-28

we have noticed since switching to 9.1.0 that u can no longer build below z=0 the sunken titanic is at -80 but when loading the oar only prims above 0 rezz



[12:03] Andrew Hellershanks: stiofain, ok. I've never tried deliberately moving anything to a negative Z so not sure how you would build with prims below 0.
[12:04] Ubit Umarov: stofaim i asid again .. just move the all crap 100 up, including the water level
[12:04] George Equus: Andrew, this is what have been removed now,
[12:04] stiofain mactomais: was no prob andrew we built the hull above ground put in situ and then did insdside a yrss work for several ppl it was used in 30 primary schools in ireland
[12:05] George Equus: Ubit, don't dodge the issue please
[12:05] stiofain mactomais: move all crap??? nice
[12:05] Ubit Umarov: there is no z < 0
[12:05] George Equus: there was Ubit
[12:05] stiofain mactomais: there was for 6 yrs
[12:05] Ubit Umarov: that was a mistake in past that caused 2 many issues
[12:06] Ubit Umarov: b ahhh
[12:06] Andrew Hellershanks: stiofain, ok. Thought that the original build would have been done above the 0 level.
[12:06] stiofain mactomais: what issues?
[12:06] Ubit Umarov: there are things where we can go around sl viewer specs, others we can't
[12:06] George Equus: some change remove that possibility, unnecessarily it seem as no one else ever complained
[12:06] Kayaker Magic: Can you make -Z an option in the INI file that defaults to FALSE?
[12:07] Arielle Popstar: mel mentioned about making it an option back in 2015
[12:08] stiofain mactomais: so can it be looked at again?
[12:09] George Equus: " Ubit Umarov: that was a mistake in past that caused 2 many issues" Can you name a couple of those issues please, and who was affected please?
[12:09] Ubit Umarov: every thing is eventually "lookable" that case just not likely on near future
[12:09] Andrew Hellershanks: If someone absolutely needs -Z they can always modify the code. Perhaps someone should do that and see if the problems in the bug report still exist.
[12:10] stiofain mactomais: what problems?
[12:10] Arielle Popstar: what code did ubit put in or remove that broke it then?
[12:10] TG.Lucan giit will flag the changes
[12:10] Arielle Popstar: that gives an idea of where to look
[12:10] stiofain mactomais: i think it should be a priority to fix it
[12:10] George Equus: exactly, thanks Arielle
[12:10] Andrew Hellershanks: The bug report says that terrain below 0 was fixed as of commit 9928076d1a16b3b62fb6cd5e24f43b9f9c648457
[12:11] Arielle Popstar: Ubits commit?
[12:11] Andrew Hellershanks: I haven't pulled up gitk to check who made the commit.
[12:12] Arielle Popstar: think they want the one where Ubit reverted the ability to do -Z
[12:12] stiofain mactomais: the last working one was OSgrid (Dev) 90d88f2: 2016-08-28 i believe that is after the big merge with avn?
[12:12] Arielle Popstar: thats the last one you tried Stio?
[12:12] Andrew Hellershanks: hm... don't think that's the right commit.
[12:12] Ubit Umarov: it has nothing to do with the merge
[12:12] stiofain mactomais: thatys the last record i can find ari
[12:13] Ubit Umarov: stop sending sand to ppls eyeys
[12:13] George Equus: I am sure has nothing to do with merge, worked after that
[12:13] Ubit Umarov: was var regions terrain fixes
[12:13] Arielle Popstar: what commit did you do work on the -Z thing then Ubit?
[12:13] stiofain mactomais: isnt a var region
[12:13] Ubit Umarov: that included things like load and saving from the several terrain format files
[12:14] Ubit Umarov: incuding ll formats
[12:14] Ubit Umarov: do you expect me to remember things 3 yrs ago? really ??
[12:14] stiofain mactomais: it just seems a great feature to give up
[12:14] Ubit Umarov: or 2
[12:14] Arielle Popstar: wasnt 3 years ago
[12:14] Arielle Popstar: when did you remove the -Z ability?
[12:14] Ubit Umarov: and z >0 is a spec..
[12:15] Ubit Umarov: on viewers etc..
[12:15] stiofain mactomais: no ubit but do expect u to listen to reasonable concerns
[12:15] Ubit Umarov: z < 0 like a lot of opensim crap only happened to seem to work on most cases
[12:15] Ubit Umarov: then you listen !!!
[12:16] Andrew Hellershanks: I still think we are talking about two different but related issues. Lower limit on terrain height and lower limit on prims.
[12:16] stiofain mactomais: it was jcc put it in after loads of discussion
[12:16] Arielle Popstar: lot of sand throwing going on
[12:16] George Equus: jcc probably knew what he was doing here.. or would not have implemented this ability at all
[12:16] Andrew Hellershanks: stiofain, what is the main issue you are having with the Titanic? No -Z on terrain, or prims?
[12:17] stiofain mactomais: terrain isnt a prob
[12:17] George Equus: No prims allowed below zero Andrew
[12:17] stiofain mactomais: we dug down to -200
[12:17] Andrew Hellershanks: ok. Just want to make sure we are all talking about the same limitation.
[12:17] George Equus: now, that is, previously it was.
[12:18] stiofain mactomais: cant rezz below zero but can drag prims
[12:18] George Equus: I need to test some more...
[12:18] stiofain mactomais: when load oar no prims below zero rezz
[12:18] George Equus: so far I note that can dig really deep but avatar can only drop to standard zero level
[12:19] Andrew Hellershanks: Right. Dragging prims below 0 would be the only option as you can't use the Build dialog to move objects below 0.
[12:19] stiofain mactomais: true andrew
[12:19] stiofain mactomais: and is 20k prims
[12:20] stiofain mactomais: most unlinked
[12:20] George Equus: include NPC's
[12:20] stiofain mactomais: so dragging not an option
[12:21] Andrew Hellershanks: The bug report mentions a change made in BulletSim re: prims with Z less than 0. That makes me wonder if having a prim with a -Z will cause problems related to physics.
[12:22] Ubit Umarov: :)
[12:22] Andrew Hellershanks: :)
[12:22] TG.Lucan tips cap
[12:22] stiofain mactomais: so can it be looked at again?
[12:22] George Equus: Please..¨
[12:23] Andrew Hellershanks: It can but as Ubit mentioned it might not be done anytime soon unless someone here wants to take it on.
[12:23] stiofain mactomais: above my paygrade lol
[12:23] Andrew Hellershanks: First step would be find the commit that imposed the limit.
[12:23] Arielle Popstar: well if they can at least use the last good version before the code change it would be a start
[12:24] George Equus: all we ask is he undo that particular change, assume you know what code change you implemented here Ubit
[12:24] Arielle Popstar: ubit should remember something that relevant
[12:24] Andrew Hellershanks: If you know which version is good vs bad that limits the search for the commit that imposed the limit.
[12:24] Ubit Umarov: bs
[12:24] Andrew Hellershanks: Arielle, Um... no.
[12:25] Andrew Hellershanks: I barely remember some stuff I did a year ago and OS is a very large project.
[12:25] stiofain mactomais: thats what we decided ari but it really was a good feature and shud be reinstated asap
[12:25] Arielle Popstar: well was it intentional or a crept in bug to other changes?
[12:25] George Equus: Implemeted buy jcc
[12:25] Arielle Popstar: i mean the reversion
[12:25] stiofain mactomais: OSgrid (Dev) 90d88f2: 2016-08-28 is best info i can find andrew
[12:26] Andrew Hellershanks: The one I found is commit 9afe2b0
[12:26] Arielle Popstar: Stio if the particular change is known then likely some dev can revert for your particular regions at least
[12:27] Andrew Hellershanks: Mantis #7765: Add new ClampNegativeZ option. Defaults to false to restore prior functionality.
Avination code wasn't designed for deep building.
[12:27] stiofain mactomais: yea ari but should be available to all
[12:27] Arielle Popstar: there you go
[12:27] Andrew Hellershanks: That mentions a setting that can be changed.
[12:28] Arielle Popstar: it should but from past unpopular changes i suspect it will be beating head on wall
[12:28] stiofain mactomais: is it a minor fix that wud benefit all in project?
[12:28] Andrew Hellershanks: The ClampNegativeZ setting still exists in the OpenSimDefaults.ini file.
[12:28] TG.Lucan If subsequent changes rely on the non -z then you are snookered. does not sound easy to revert
[12:29] Andrew Hellershanks: I would suggest that you can override the setting in the OpenSim.ini file and see what happens.
[12:29] Bill.Blight so where do negatives stop is my question , why not -x -y , I can see the logic of not having -z
[12:30] Arielle Popstar: increasing amount of grids and regions becoming stuck on older versions due to changes that dont work for them so it is getting to be the norm rather than the exception
[12:30] stiofain mactomais: will try andrew but really think should be part of core
[12:30] Andrew Hellershanks: If the settings solves your issue then no code changes are required as it is part of core. You just don't like the default setting.
[12:31] Andrew Hellershanks: hm... actually the default is false which means it should not be clamping negative Z.
[12:31] stiofain mactomais: lol yea
[12:31] stiofain mactomais: but if opensim is to be used to build rl things then needs -z
[12:32] Ubit Umarov flüstert: NO
[12:32] stiofain mactomais: isnt really true 3d if no negative z
[12:32] Ubit Umarov: i say again inworld height is not RL height
[12:32] Ubit Umarov: default water is at 20m!
[12:33] Bill.Blight No they don't need -z because it is arbitrary to default grould level is Z in real life, well actually sea level, which is why most altitude is calculated off of sea level , not center of the earth
[12:33] Ubit Umarov: sl did designed it so terrain height is just a byte
[12:33] TG.Lucan and the moon changes sea level :)
[12:33] stiofain mactomais: which is reasonable but 2/3rd of world is in deeper water
[12:33] Andrew Hellershanks: Bill is right about asking where we stop in removing "limits".
[12:34] Ubit Umarov: thats something that does not math rl
[12:34] Ubit Umarov: but sl decision.. and viewers are like that
[12:34] Andrew Hellershanks: I've seen objects wind up at positions with X or Y values of hundreds or thousands of meters outside of region boundaries.
[12:34] Ubit Umarov: water is set at a positive value
[12:34] stiofain mactomais: was no viewer issues
[12:34] Bill.Blight so if you are at -1000 in the real world that is 1000 meters below sea level, so if you want to be "Real World" raise your water
[12:34] George Equus: def not
[12:34] George Equus: a viewer limitation
[12:35] stiofain mactomais: so we start at bottom of marianna trench and build up from there?
[12:35] TG.Lucan surely this is just numbers
[12:35] Ubit Umarov: it is a offset...
[12:35] Bill.Blight Just saying, the "Real World" argument is flawed in this case, because even the depth of the marianna trench is based off of how far below the surface of the water it is
[12:36] Ubit Umarov: and rl zero was just a convention..
[12:36] Ubit Umarov: and a very messy one, bc water does move
[12:36] Bill.Blight not how deep it is from the seabed surrounding it
[12:36] Ubit Umarov: and earth is not spheric etc etc
[12:36] stiofain mactomais: i think is reasonable request and sl conventions is a spurious reason
[12:36] Ubit Umarov: and irrelevant .. this is a simple simulation
[12:37] Ubit Umarov: damm ... sl code is made around sl conventions...
[12:37] Ubit Umarov: and assumptions
[12:38] Ubit Umarov: it may work around a lot of them.. but not all
[12:39] Ubit Umarov: not until we decide to have our own viewer and our own simulation model
[12:40] stiofain mactomais: that viewer should use -z
[12:40] Bill.Blight This is a case of a bug being a bug for so long that it is the assumption that it is the best way or the way it should be. Just because it did not cause visible issues to the few fringe cases does not mean it was the best way of doing things
[12:40] Andrew Hellershanks: We aren't going to resolve this issue today. Best to table it until next week when someone can come back with additional information.
[12:41] Bill.Blight if it SHOULD use -z , why not -x and -y
[12:41] Andrew Hellershanks: Questions are whether the setting in the OpenSimDefaults.ini file still controls clamping (or not) of -Z values.
[12:41] Bill.Blight The settings are no longer in the SceneGraph.cs Andrew
[12:41] Bill.Blight only in Scene.cs
[12:41] Bill.Blight So don't think it is working
[12:41] Ubit Umarov: that option is not on all code paths it seems
[12:42] Andrew Hellershanks: Bill, I know. It does open a can of worms. People already find a way around X/Y limits when making off region objects.
[12:43] Bill.Blight no what I am saying is why should I not be able to build at -x or -y , why can't I just walk off the negative side of my sim and just keep walking .. Same logic applies
[12:43] Andrew Hellershanks: Bill, ok. I only looked for a commit. I haven't looked at the current state of the files that were changed.
[12:44] Andrew Hellershanks: Bill, I think that is a topic for another time.
[12:44] Bill.Blight I see both sides of this argument so I am not really trying to be negative(pun intended) but from a uniformity stand point it just throws the logic side of my brain into a tail spin ..
[12:44] Andrew Hellershanks: :)
[12:45] Andrew Hellershanks: We've been going almost two hours. I think its time to call this meeting to a close.
[12:46] Andrew Hellershanks: Thank you all for being here today. See you again next week.
[12:46] George Equus: OK.. Thanks Andrew
[12:48] TG.Lucan I am not seeing any shutdown issues with httptests on windows, but that is a region with no content
[12:49] Bill.Blight I have around 150 regions running it, and don't see any shutdown issues on linux
[12:49] Andrew Hellershanks: I need to resolve a merge conflict of some kind before I can build httptests.
[12:49] Arielle Popstar: so https breaks hypergrid?>
[12:49] Bill.Blight it can
[12:49] TG.Lucan do you have mesh Bill?
[12:49] Bill.Blight I have lots of mesh
[12:49] TG.Lucan kk
[12:49] TG.Lucan I have not seen issues with HG
[12:50] Arielle Popstar: it breaks connections to non https grids?
[12:50] Bill.Blight it breaks HG because older versions trying to talk non https blows up
[12:50] Arielle Popstar: nod
[12:50] Arielle Popstar: sounds logical
[12:50] TG.Lucan I came here from httptests region
[12:50] Arielle Popstar: why did we need https again?
[12:50] Bill.Blight I run a slightly modded version of httptests and my whole grid runs it
[12:51] Bill.Blight https will encrypt connections between servers and sims to prevent or mitigate "man in the middle" attacks for sniffing data
[12:51] Arielle Popstar: we been having issues with that?
[12:51] Bill.Blight but not sure sim data is all that super secret anyway
[12:52] TG.Lucan Im aware https functionality, never seen with opensim
[12:52] Arielle Popstar: so once into core there will be a lot of pressure to come up to current
[12:52] Bill.Blight https sort of works in the httptests branch you just have to turn off the verify certs and also in the viewers because they are picky
[12:52] Ubit Umarov: httptests is not about https
[12:52] Arielle Popstar: to remain connectable
[12:52] Bill.Blight no
[12:53] Bill.Blight httptests is not just https, but it sort of works if you want to use it
[12:53] Ubit Umarov: started to be just a branch to test a few changed to httpserver
[12:53] Bill.Blight works fine without https
[12:53] Arielle Popstar: ok
[12:53] Ubit Umarov: some ssl fixes did came by logic
[12:53] Ubit Umarov: but now is more like master
[12:53] Ubit Umarov: at least for me
[12:54] Bill.Blight should be master .. More stable and secure due to advancing the .Net version
[12:54] Ubit Umarov: i only add new features on it
[12:54] Ubit Umarov: even some fixes now
[12:54] Arielle Popstar: cant merge them into present master?
[12:54] Arielle Popstar: as well as update net?
[12:55] Ubit Umarov: no bc of our jenkins tools
[12:55] Arielle Popstar: oh right
[12:55] TG.Lucan you can checkout httptests from master
[12:55] Ubit Umarov: they are on mono 2.10
[12:55] Arielle Popstar: assume noone uses mono 2.10
[12:56] Ubit Umarov: hope not :p
[12:56] Bill.Blight current mono is 5.10x series
[12:56] Ubit Umarov: and httptests may need a mono 5x
[12:56] Bill.Blight well might be 5.14 now
[12:56] TG.Lucan I run mono master for other project tests,
[12:57] Bill.Blight yeah httptests is not super secret, just check it out instead of master
[12:57] Bill.Blight but requires .Net 4.6 on windows and mono 5.x on linux

Personal tools
About This Wiki