(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
[12:11 PM PT] Andrew Hellershanks: Hello, everyone.
[12:11 PM PT] Ubit Umarov: hi
[12:11 PM PT] Kayaker Magic: It is finally raining in California, so the sky is dark and foreboding with or without the sun
[12:11 PM PT] Andrew Hellershanks: Nice. We went through a change in the clocks and we didn't lose anyone.
[12:12 PM PT] Kayaker Magic: WOW, 2 sec of chat lag!
[12:12 PM PT] Ubit Umarov: well abou code changes etc...
[12:13 PM PT] Ubit Umarov: i just found, well yesterday, that old ode is currently broken on linux
[12:13 PM PT] Ubit Umarov: both on master and -dotnet6 branchs
[12:13 PM PT] Ubit Umarov: og fine on win
[12:13 PM PT] Ubit Umarov: ofc
[12:14 PM PT] Kayaker Magic: ODE stands for Odious Debugging Environment. It never worked right.
[12:14 PM PT] Ubit Umarov: but that just because un,managed libs are loaded dif on win
[12:15 PM PT] Ubit Umarov: well i did broke it when i added unmanaged libs for ubode
[12:15 PM PT] Ubit Umarov: now a bit diferent from original ODE
[12:16 PM PT] Ubit Umarov: but the API, is basicly the same..
[12:16 PM PT] Ubit Umarov: same methods.. same names
[12:16 PM PT] Ubit Umarov: turns out that .net bows up
[12:16 PM PT] Ubit Umarov: blows up
[12:17 PM PT] Ubit Umarov: even with ODE and ubODE on diferent namespaces, with clear define of entry points for the dierent file,
[12:17 PM PT] Ubit Umarov: at runtime, -net ( and mono) do a total mess and call wither lib
[12:18 PM PT] Ubit Umarov: this is also because how our "modules" work
[12:18 PM PT] Ubit Umarov: they, and their dependencies, are always loaded, then just check on confing if are enabled or not
[12:19 PM PT] Ubit Umarov: so, ubode happens to load first, .net loads its libubode.so ..
[12:19 PM PT] Ubit Umarov: then ode loads, and .net/mono loads libode.so
[12:19 PM PT] Ubit Umarov: all good
[12:20 PM PT] Ubit Umarov: then using ODE, it calls-- libubode.so the first loaded
[12:20 PM PT] Ubit Umarov: BUMMM
[12:20 PM PT] Kayaker Magic: Is there a way in the INI files to tell it to never load ODE again? Can I delete the modules from the bin folder?
[12:20 PM PT] Ubit Umarov: is fact it may even do more confusing things..
[12:21 PM PT] Ubit Umarov: yeah we can remove it
[12:21 PM PT] Ubit Umarov: but i would still like to avoid it
[12:21 PM PT] Cuga.Rajal @rajal.org:9000: is there still a user base for ODE that can't switch to ubODE?
[12:21 PM PT] Ubit Umarov: some are fast crashes
[12:22 PM PT] Ubit Umarov: at
at OpenSim.Region.PhysicsModule.ODE.OdeNative:SpaceAdd
[12:22 PM PT] Ubit Umarov: 0x7f33be557cb9 - /home/ubit/opensim/bin/lib64/libubode-x86_64.so
[12:23 PM PT] Ubit Umarov: so as i said, ODE called ubode lib
[12:24 PM PT] Ubit Umarov: as is, if i delete ...ubode.dll.config, it works again.. foes not find ubode.so so only uses ode.so
[12:24 PM PT] Cuga.Rajal @rajal.org:9000: can you unload the lib that's not being used in the configs at startup time?
[12:24 PM PT] Ubit Umarov: libs cna't be unload
[12:24 PM PT] Cuga.Rajal @rajal.org:9000: make it unload
[12:24 PM PT] Cuga.Rajal @rajal.org:9000: kk
[12:24 PM PT] Ubit Umarov: and mono has no easy way to manually load things
[12:25 PM PT] Ubit Umarov: only the automatic dllimport thing, that does this
[12:25 PM PT] Ubit Umarov: dotnet also does the saem mess
[12:25 PM PT] Ubit Umarov: seems after loading it only keeps methond names not "lib/name" pairs
[12:26 PM PT] Ubit Umarov: well working on it.. somehow :P
[12:26 PM PT] Cuga.Rajal @rajal.org:9000: rename all the methods lol
[12:27 PM PT] Ubit Umarov: yeah.. on C# sources and on ode.so ones
[12:27 PM PT] Ubit Umarov: fun :p
[12:27 PM PT] Ubit Umarov: nah i may just revert to use same native dll
[12:28 PM PT] Ubit Umarov: and hammer old ode to look happy
[12:28 PM PT] Ubit Umarov: isnce he never was happy.. mb no issues :p
[12:29 PM PT] Ubit Umarov: wel idea was to keep it as it was, the reason to still keep it
[12:29 PM PT] Ubit Umarov: a few regions here have tons of things that only run on it
[12:30 PM PT] Ubit Umarov: now on dotnet6, we can consider drop it
[12:30 PM PT] Ubit Umarov: since we already had to drop X
[12:30 PM PT] Ubit Umarov: and most scripts that did use old ode and used Xengine "things"
[12:30 PM PT] Ubit Umarov: what do you think?
[12:31 PM PT] Ubit Umarov: oh no i made all sleep again
[12:31 PM PT] Andrew Hellershanks is thinking
[12:31 PM PT] Brian.Too @hg.viewtwo.net:8600: awake her, drop it
[12:31 PM PT] Cuga.Rajal @rajal.org:9000: a good scripter should be able to rescript whatever needed to remove ODE dependency
[12:31 PM PT] Cuga.Rajal @rajal.org:9000: but it is extra work
[12:32 PM PT] Ubit Umarov: well some did spent years to make things work with ode
[12:32 PM PT] Ubit Umarov: and made a ton of things
[12:32 PM PT] Andrew Hellershanks: If the dotnet6 is already a major change in and of itself and XEngine is being dropped it seems like dropping other old items at the same time makes sense.
[12:32 PM PT] Ubit Umarov: ofc.. totally gaveup, even when bullet did show up
[12:32 PM PT] Orbert.Tatham @hg.zetaworlds.com: show users
[12:33 PM PT] Orbert.Tatham @hg.zetaworlds.com: oops
[12:33 PM PT] Ubit Umarov: yeha im considering drop it on dotnet6
[12:33 PM PT] Kayaker Magic: There is a large number of people who use scripts without understanding them and cannot fix minor syntax issues that worked on X and fail on YEngine.
[12:33 PM PT] Andrew Hellershanks: Orbert, what do you want to show them? ;)
[12:33 PM PT] Orbert.Tatham @hg.zetaworlds.com: Sorry
[12:33 PM PT] Cuga.Rajal @rajal.org:9000: I agree Kayaker
[12:34 PM PT] Andrew Hellershanks: Orbert, np
[12:34 PM PT] Ubit Umarov: well on other coding, i just made some changes on dotnet6 branch
[12:34 PM PT] Andrew Hellershanks: Kayaker, that is so true.
[12:34 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I really don't know where some folks learned LSL, but some of the scripts I have seen over the years... LSL is not javascript meets python -.-
[12:34 PM PT] Ubit Umarov: in fact on ubode, doing now ray-avatar collisions on managed code only
[12:34 PM PT] Ubit Umarov: ie ported ode code into c#
[12:35 PM PT] Ubit Umarov: with simplifications
[12:35 PM PT] Ubit Umarov: bc our avatars are always vertical on physics etc
[12:35 PM PT] Ubit Umarov: so no need for a full rotating capsule
[12:35 PM PT] Vincent.Sylvester @hg.zetaworlds.com: No more walking on the ceiling heh
[12:35 PM PT] Ubit Umarov: avatar-avatar collision is now also managed only
[12:36 PM PT] Ubit Umarov: same simplifications
[12:36 PM PT] Ubit Umarov: and less managed/unmaged expensive transitions
[12:36 PM PT] Ubit Umarov: well a few ns each
[12:36 PM PT] Ubit Umarov: :)
[12:37 PM PT] Ubit Umarov: also less unmanaged code bits to maintain
[12:38 PM PT] Ubit Umarov: this bc dotnet JIT in release mode, is closer to c code
[12:38 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Easier to track bugs when they don't show up in an external lib compiled in release as well
[12:38 PM PT] Ubit Umarov: was not so much on .net framework
[12:39 PM PT] Ubit Umarov: well a bug on the unmanaged implies hard bisect etc
[12:39 PM PT] Ubit Umarov: i had to do one ofc
[12:39 PM PT] Vincent.Sylvester @hg.zetaworlds.com: dotnet still has aot?
[12:39 PM PT] Ubit Umarov: as a dif one
[12:39 PM PT] Ubit Umarov: even can make a full packaged including .dotnet
[12:40 PM PT] Ubit Umarov: :)
[12:40 PM PT] Ubit Umarov: they need AOT things for IOS
[12:40 PM PT] Vincent.Sylvester @hg.zetaworlds.com: You can do that in mono too mkbundle, though I never tested if that doesn't just blow up
[12:40 PM PT] Ubit Umarov: where code can't be changed at runtime
[12:40 PM PT] Andrew Hellershanks: AOT?
[12:40 PM PT] Ubit Umarov: yeah "evolution" of that
[12:41 PM PT] Vincent.Sylvester @hg.zetaworlds.com: On mono I only really noticed aot creating less jitter on frame times, not really a great performance increase
[12:41 PM PT] Ubit Umarov: AOY speads up load
[12:41 PM PT] Orbert.Tatham @hg.zetaworlds.com: Apple is trying to eliminate JIT from what I have heard
[12:41 PM PT] Ubit Umarov: and first use of a bit of code
[12:42 PM PT] Ubit Umarov: apple never allowed jit
[12:42 PM PT] Orbert.Tatham @hg.zetaworlds.com: Got it
[12:42 PM PT] Ubit Umarov: runtime time code changes is forbidden on good operating systems
[12:42 PM PT] Ubit Umarov: well only apple ones currently :p
[12:43 PM PT] Ubit Umarov: self modifing code is a virus playground
[12:43 PM PT] Andrew Hellershanks: Awww... but I like self-modifying code. ;)
[12:43 PM PT] Andrew Hellershanks: hehe
[12:43 PM PT] Vincent.Sylvester @hg.zetaworlds.com: jit makes sense in the way that newer ways of doing specific operations are found that are faster you don't have to recompile the whole thing, but that's not often these days
[12:43 PM PT] Cuga.Rajal @rajal.org:9000: I wanted to talk more about server on Apple M1 when time permits
[12:43 PM PT] Ubit Umarov: wel IOS forbids it
[12:43 PM PT] Ubit Umarov: so mono add AOT
[12:44 PM PT] Ubit Umarov: even framewrok has it also, just other name
[12:44 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I think it was with some uuid related code jit even makes a mess of making it a lot bigger in machine code than it needs to be
[12:44 PM PT] Ubit Umarov: i did test it
[12:44 PM PT] Ubit Umarov: don't like bc on win it stores the bin files on a global folder somewhere
[12:44 PM PT] Ubit Umarov: so a pin to clean
[12:44 PM PT] Ubit Umarov: but worked..
[12:45 PM PT] Andrew Hellershanks: Found it -> AOT is Ahead Of Time
[12:45 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Works, but not well, I think that's Microsoft motto these days
[12:45 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Well a lot of companies as well
[12:45 PM PT] Orbert.Tatham @hg.zetaworlds.com: These days?
[12:45 PM PT] Ubit Umarov: yeah versus JIT just-in-time
[12:45 PM PT] Andrew Hellershanks nods
[12:45 PM PT] Orbert.Tatham @hg.zetaworlds.com:
[12:45 PM PT] Ubit Umarov: well dotnet moved even into more complex
[12:46 PM PT] Andrew Hellershanks: Wanted to mention it as some people reading the meeting log may not know the acronym.
[12:46 PM PT] Ubit Umarov: JIT is done in stages
[12:46 PM PT] Ubit Umarov: first time a thing runs with no optimizations
[12:46 PM PT] Ubit Umarov: then keeps use stats
[12:46 PM PT] Ubit Umarov: and as it gets used, it gets more optimizations
[12:47 PM PT] Ubit Umarov: very very fun for jitter
[12:47 PM PT] Kayaker Magic: AOT: Always On Top, As Opposed To, Array Of Things
[12:47 PM PT] Ubit Umarov: i did disable that crap on current dotnet
[12:48 PM PT] Ubit Umarov: at fist time JIT will do all optimization it can
[12:48 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Yeah like unless you have a print out of what it did and found doing that what's the point might as well just apply all
[12:48 PM PT] Ubit Umarov: ppl love to waste resources on stupid runtime stats
[12:48 PM PT] Ubit Umarov: all so a application seems to load faster
[12:49 PM PT] Ubit Umarov: and they can sell see dotnet 5 loads faster...
[12:49 PM PT] Ubit Umarov: blabla
[12:49 PM PT] Vincent.Sylvester @hg.zetaworlds.com: But but it's so developers can improve their code... which they never actually then do, cause what the heck does jit know
[12:50 PM PT] Ubit Umarov: after a week its possilble they get even better code
[12:50 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Smells of github copilot
[12:50 PM PT] Andrew Hellershanks: Run time stats are fine for a developer but you don't want to burden users with it in a release.
[12:50 PM PT] Ubit Umarov: bc the stats they do, allows to spot a few more tiny optimizations
[12:50 PM PT] Vincent.Sylvester @hg.zetaworlds.com: visual studio gives you tons of suggestions too, not all work or are actually useful
[12:50 PM PT] Ubit Umarov: ( ofc stats are stats.. the thing may just break the code )
[12:51 PM PT] Ubit Umarov: and code predicability ?? forget
[12:51 PM PT] Ubit Umarov: each time we run something, it does diferent things now
[12:52 PM PT] Vincent.Sylvester @hg.zetaworlds.com: With the variables in OpenSim yeah
[12:52 PM PT] Ubit Umarov: at least uses diferent code etc
[12:52 PM PT] Ubit Umarov: no iead what they are thinking..
[12:52 PM PT] Ubit Umarov: with waht they are doing, a product is just impossible to test
[12:53 PM PT] Ubit Umarov: whatever
[12:53 PM PT] Cuga.Rajal @rajal.org:9000: I would think forward compatibility would be one of the main decision factors
[12:53 PM PT] Cuga.Rajal @rajal.org:9000: reliability of such
[12:54 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Someone up in the ranks at MS has an idea and by the time it goes to the developers it went through so many hands things just end up a murky mess
[12:54 PM PT] Cuga.Rajal @rajal.org:9000: seems reasonable to be wary of adding
[12:54 PM PT] Ubit Umarov: poor militar, space and industry
[12:54 PM PT] Ubit Umarov: will need to create own compilers again
[12:54 PM PT] Andrew Hellershanks: Less than 10 minutes before we reach the top of the hour. Before we go further on this discussion Cuga wanted to talk about something.
[12:54 PM PT] Ubit Umarov: bc this things are pure trash for them
[12:55 PM PT] Andrew Hellershanks: Cuga, go ahead
[12:55 PM PT] Ubit Umarov: well russian missiles mb using dotnet..
[12:55 PM PT] Ubit Umarov: they miss a lot :P
[12:55 PM PT] Cuga.Rajal @rajal.org:9000: well Ubit e=we had a short chat on IRC about running server on Apple M1
[12:55 PM PT] Cuga.Rajal @rajal.org:9000: with ditnet
[12:55 PM PT] Cuga.Rajal @rajal.org:9000: I told you my story, I tried it when I forst got my laptop
[12:56 PM PT] Cuga.Rajal @rajal.org:9000: screaming fast but starrtup failed wil a DLL not same architecture
[12:56 PM PT] Cuga.Rajal @rajal.org:9000: I was in touch with Gavin back then
[12:56 PM PT] Cuga.Rajal @rajal.org:9000: he said it would never work
[12:56 PM PT] Ubit Umarov: bah ofc it will work
[12:56 PM PT] Cuga.Rajal @rajal.org:9000: It does run fine in x86_64 emulation mode
[12:56 PM PT] Cuga.Rajal @rajal.org:9000: but I wanted to see if could get it running
[12:56 PM PT] Cuga.Rajal @rajal.org:9000: Ubit you mentioned it runs on another ARM arch
[12:56 PM PT] Ubit Umarov: you just need to compile the things on lib64
[12:57 PM PT] Ubit Umarov: for the m1 also
[12:57 PM PT] Cuga.Rajal @rajal.org:9000: I realize now I was trying to comile master branch, I didnt know about dotnet branch
[12:57 PM PT] Cuga.Rajal @rajal.org:9000: Should I just try again and use that>?
[12:57 PM PT] Ubit Umarov: for example make ode.dylib :)
[12:57 PM PT] Ubit Umarov: think dotnet now supports m1
[12:57 PM PT] Cuga.Rajal @rajal.org:9000: yes dotnet install was fine
[12:57 PM PT] Cuga.Rajal @rajal.org:9000: the prob was one of the DLLs on the OS chain
[12:58 PM PT] Cuga.Rajal @rajal.org:9000: you mentioerd it in our convo
[12:58 PM PT] Ubit Umarov: the main problem is that you need to compile the unmanage libs for it
[12:58 PM PT] Ubit Umarov: see what ppl did on raspberry pi
[12:58 PM PT] Cuga.Rajal @rajal.org:9000: I'm not familiar with process
[12:58 PM PT] Ubit Umarov: same problem
[12:58 PM PT] Ubit Umarov: seen get the native sources.. and compile
[12:58 PM PT] Cuga.Rajal @rajal.org:9000: there was only one file I believe though
[12:59 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Gavin has a tutorial on his forum regarding arm, you could talk to him, he uses macOS as well
[12:59 PM PT] Ubit Umarov: some we do have on opensim-libs ( like ubode and bullet )
[12:59 PM PT] Ubit Umarov: yeha gavin did that for rasp
[12:59 PM PT] Vincent.Sylvester @hg.zetaworlds.com: You may not be able to fetch or compile the versions OpenSim needs for some of the dependencies and there are also parts still broken when you use newer versions
[12:59 PM PT] Cuga.Rajal @rajal.org:9000: ok I'll check with him.. its been a while
[1:00 PM PT] Ubit Umarov: ofc maybe a bit hard, but because all the compile tools keep changing, etc
[1:00 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Famously the database connectors are so out of date any attempts at updating them is a giant can of worms
[1:00 PM PT] Ubit Umarov: wel but those are c# at least
[1:01 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Knock on wood it stays that way
[1:01 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Those getting the idea to switch to Rust or something
[1:01 PM PT] Brian.Too @hg.viewtwo.net:8600: question
[1:01 PM PT] Ubit Umarov: i did update the mysql one
[1:02 PM PT] Ubit Umarov: well possible that that utfmb3 issue again
[1:02 PM PT] Brian.Too @hg.viewtwo.net:8600: Anyone else use the mouse buttons to move their avatar?
[1:02 PM PT] Ubit Umarov: i used their lib
[1:02 PM PT] Ubit Umarov: i know.. it is slower now
[1:02 PM PT] Brian.Too @hg.viewtwo.net:8600: ok, then noted
[1:02 PM PT] Ubit Umarov: a viewers bug
[1:03 PM PT] Ubit Umarov: they do tell to move slower
[1:03 PM PT] Vincent.Sylvester @hg.zetaworlds.com: postgres is bugged on mono trunk still, not sure that'll ever get fixed, tried it before, but way too much work for me alone
[1:03 PM PT] Ubit Umarov: mouse move is a firestorm hack
[1:03 PM PT] Ubit Umarov: not fully done
[1:03 PM PT] Brian.Too @hg.viewtwo.net:8600: ok
[1:03 PM PT] Ubit Umarov: well not sure what viewer made it first..
[1:04 PM PT] Ubit Umarov: i made region honour the speed modes viewers ask
[1:04 PM PT] Brian.Too @hg.viewtwo.net:8600: speed was similar to keyboard keys prior to a code change, probably a few months ago
[1:04 PM PT] Vincent.Sylvester @hg.zetaworlds.com: You can change the constants in the code to make them both behave similarly, but of course means walking slower overall
[1:04 PM PT] Ubit Umarov: improved normal movement..
[1:04 PM PT] Ubit Umarov: that got slower :)
[1:04 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I mean we were already moonwalking forward heh
[1:05 PM PT] Ubit Umarov: when we press a key to walk, viewers sends slower move comand
[1:05 PM PT] Ubit Umarov: for 250ms, then changes to normal speed
[1:05 PM PT] Ubit Umarov: they never did had that 250ms timer to the mouse hack
[1:05 PM PT] Brian.Too @hg.viewtwo.net:8600: allow for short movements
[1:05 PM PT] Ubit Umarov: so it only sends the slow
[1:06 PM PT] Ubit Umarov: yeha its a nice hack
[1:06 PM PT] Vincent.Sylvester @hg.zetaworlds.com: public const float AgentControlNormalVel = 1.0f you can change to a lesser number and you get a more realistic speed given the walking animation we have
[1:07 PM PT] Vincent.Sylvester @hg.zetaworlds.com: You can play with those numbers and get some funky results, superman speed walking heh
[1:07 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Also some speed overrides in the config somewhere, but those are overall so apply to all speeds
[1:08 PM PT] Brian.Too @hg.viewtwo.net:8600: noted
[1:08 PM PT] Brian.Too @hg.viewtwo.net:8600: gaming thing :)
[1:08 PM PT] Ubit Umarov: well not a fix for the mouse issue
[1:08 PM PT] Ubit Umarov: just hit always run :)
[1:08 PM PT] Cuga.Rajal @rajal.org:9000: any new developments on llSetForce since summer?
[1:09 PM PT] Ubit Umarov: ( double click should work )
[1:09 PM PT] Brian.Too @hg.viewtwo.net:8600: can use double mouse button click to enter run mode
[1:09 PM PT] Ubit Umarov: llSetforce on avs will never work
[1:09 PM PT] Cuga.Rajal @rajal.org:9000: kk
[1:09 PM PT] Cuga.Rajal @rajal.org:9000: *cries*
[1:09 PM PT] Ubit Umarov: well until we kill those motors
[1:09 PM PT] Ubit Umarov: and let friction work alone
[1:10 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Might happen one day with a particularly bored Ubit perhaps :)
[1:10 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Don't hold your breath tho
[1:10 PM PT] Ubit Umarov: wel will chnage how avs move
[1:10 PM PT] Ubit Umarov: this motor make the social use of this a lot simpler
[1:11 PM PT] Cuga.Rajal @rajal.org:9000: but still a major project to undertake
[1:11 PM PT] Ubit Umarov: well ignorine the bg steps that make hard to park in front of someone
[1:12 PM PT] Ubit Umarov: i also don't like how a avatar can climb any mountain of prim
[1:12 PM PT] Ubit Umarov: or prim
[1:12 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Ubit cares very much about us being able to hug each other :)
[1:12 PM PT] Ubit Umarov: but makes life easier
[1:12 PM PT] Ubit Umarov: well i spent again hours trying to mk the hugs work
[1:12 PM PT] Andrew Hellershanks: :)
[1:12 PM PT] Ubit Umarov: again a fail
[1:13 PM PT] Ubit Umarov: for avatar avatar collsion the shape is not a capsule
[1:13 PM PT] Motoko.Karu @login.digiworldz.com:8002: lol
[1:13 PM PT] Ubit Umarov: is a capsule with a elipse shape on xy
[1:14 PM PT] Ubit Umarov: well aproximate
[1:14 PM PT] Ubit Umarov: so, it is closer to the size
[1:14 PM PT] Ubit Umarov:
[1:14 PM PT] Cuga.Rajal @rajal.org:9000: is that fixed regardless of shape?
[1:14 PM PT] Ubit Umarov: while for all other collisions is just a capsule
[1:15 PM PT] Ubit Umarov:
[1:15 PM PT] Ubit Umarov: yes
[1:15 PM PT] Cuga.Rajal @rajal.org:9000: ahh for physics
[1:15 PM PT] Ubit Umarov: and always vertical
[1:15 PM PT] Ubit Umarov: yeap
[1:15 PM PT] Ubit Umarov: physcis only sees standing avatars anyways
[1:15 PM PT] Cuga.Rajal @rajal.org:9000: flatter in front-back so they can hug?
[1:16 PM PT] Ubit Umarov: not flat
[1:16 PM PT] Cuga.Rajal @rajal.org:9000: I mean, smaller elipse axis
[1:16 PM PT] Ubit Umarov: this
[1:17 PM PT] Cuga.Rajal @rajal.org:9000: yes thats what I imagined
[1:17 PM PT] Vincent.Sylvester @hg.zetaworlds.com: A bit like a deodorant stick
[1:17 PM PT] Cuga.Rajal @rajal.org:9000: oopsahh
[1:18 PM PT] Ubit Umarov: ok a avatar is like that
[1:18 PM PT] Ubit Umarov: mostly
[1:18 PM PT] Ubit Umarov: for hugs... :)
[1:18 PM PT] Vincent.Sylvester @hg.zetaworlds.com: We are all just the insides of a kinder surprise egg heh
[1:18 PM PT] Andrew Hellershanks: hehe
[1:19 PM PT] Cuga.Rajal @rajal.org:9000: Ive noticed sometimes couples animations will bork after a few seconds when avatars are too close.. is that related to the avatar bounds?
[1:19 PM PT] Cuga.Rajal @rajal.org:9000: physics bounds
[1:19 PM PT] Cuga.Rajal @rajal.org:9000: or something unrelated
[1:19 PM PT] Ubit Umarov: the blue is the simpler on all othet collisions
[1:20 PM PT] Ubit Umarov: err no no relation with viagra :p
[1:20 PM PT] Ubit Umarov: evne if looking like a pilule and blue :)
[1:21 PM PT] Orbert.Tatham @hg.zetaworlds.com: Matrix time
[1:21 PM PT] Ubit Umarov: well capsule is a fast shape on collisions also
[1:21 PM PT] Ubit Umarov: if you see it is a sphere and a segment
[1:21 PM PT] Ubit Umarov: or even a linesegment and a radius
[1:22 PM PT] Ubit Umarov: while a sphere is a point and a radius
[1:22 PM PT] Ubit Umarov: see?
[1:22 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Basically the problem is you want to have the bounding box cover the avatar properly so you don't poke limbs and stuff through walls all the time, but still have to make it small enough to be close to each other
[1:22 PM PT] Ubit Umarov: well for collision with the prims the underside is not like that
[1:23 PM PT] Ubit Umarov: uses that to then on the lower 0.6m does like ramps
[1:23 PM PT] Ubit Umarov: so the avatar can climb prims
[1:24 PM PT] Ubit Umarov: well more conical shape
[1:24 PM PT] Cuga.Rajal @rajal.org:9000: is that the limiting factor on the climb angle?
[1:24 PM PT] Cuga.Rajal @rajal.org:9000: when avatar is climbing
[1:24 PM PT] Ubit Umarov: since it is driven by that same avatar motor.. it has a lot of power
[1:25 PM PT] Cuga.Rajal @rajal.org:9000: I see.. it just defines the collision point
[1:26 PM PT] Ubit Umarov: code changes to colison point a bit but more importna the surface normal
[1:26 PM PT] Cuga.Rajal @rajal.org:9000: are the avatar bounds collisions ignored when they are seated currently?
[1:27 PM PT] Ubit Umarov: sitting avs are phantom
[1:27 PM PT] Cuga.Rajal @rajal.org:9000: ok
[1:27 PM PT] Ubit Umarov: worse even have no shape at all on physics
[1:27 PM PT] Ubit Umarov: at SL rays can see them
[1:27 PM PT] Ubit Umarov: they get strange shapes, ppl say
[1:27 PM PT] Cuga.Rajal @rajal.org:9000: something triggers sit anims to break after a few secs, thought that might have been it.. still tracking down
[1:28 PM PT] Ubit Umarov: well like a tilted box
[1:28 PM PT] Ubit Umarov: tappared :)
[1:28 PM PT] Cuga.Rajal @rajal.org:9000: oh like walking on to of seated ppl
[1:29 PM PT] Ubit Umarov: mb something like that.. don't know
[1:29 PM PT] Ubit Umarov: but only for rays it seems
[1:29 PM PT] Cuga.Rajal @rajal.org:9000: does seem slanted, im fact you slide off if you are not perfectly centered on top of them
[1:29 PM PT] Cuga.Rajal @rajal.org:9000: "TP to avatar" puts you direct on top center
[1:30 PM PT] Ubit Umarov: some ppl did use rays to try to scan those shapes there
[1:30 PM PT] Cuga.Rajal @rajal.org:9000: on theor head
[1:30 PM PT] Ubit Umarov: here we still have none
[1:31 PM PT] Ubit Umarov: not as fast collider this tappered box also ...
[1:31 PM PT] Andrew Hellershanks: It is now half past the hour. Time to wrap up the meeting. Any final comments or questions?
[1:31 PM PT] Ubit Umarov: i wohould possible just go for a dif size capsule
[1:31 PM PT] Cuga.Rajal @rajal.org:9000: quick Q
[1:32 PM PT] Cuga.Rajal @rajal.org:9000: I still get a lot of console errors "unknown material ID" any easy SQL fix?
[1:32 PM PT] Ubit Umarov: but for now, you can't kill the driver of car with ray guns
[1:32 PM PT] Ubit Umarov: no.. you just lost a material :p
[1:32 PM PT] Andrew Hellershanks: Interesting message. I haven't seen any of those.
[1:33 PM PT] Andrew Hellershanks: s/message/question/
[1:33 PM PT] Cuga.Rajal @rajal.org:9000: I havent experienced ny problems with the warnings
[1:33 PM PT] Cuga.Rajal @rajal.org:9000: going to aassume is harmless
[1:33 PM PT] Cuga.Rajal @rajal.org:9000: but my OCD wants to fix
[1:33 PM PT] Ubit Umarov: just some prim does not look as it should
[1:34 PM PT] Kayaker Magic: LOL, yes I want to fix those missing asset errors also!
[1:34 PM PT] Cuga.Rajal @rajal.org:9000: could be my avatar clothing for all I know
[1:34 PM PT] Cuga.Rajal @rajal.org:9000: any easy way to SQL query to find the assets?
[1:34 PM PT] Ubit Umarov: no
[1:34 PM PT] Cuga.Rajal @rajal.org:9000: that need fixin?
[1:34 PM PT] Ubit Umarov: wel if missing they are not there :p
[1:35 PM PT] Ubit Umarov: you just need to fis on the prim..
[1:35 PM PT] Ubit Umarov: ie get a good copy
[1:36 PM PT] Cuga.Rajal @rajal.org:9000: if I know which prim it is, I cam do that
[1:36 PM PT] Ubit Umarov: i did mention that elipse on avatar-avatar is aproximated?
[1:36 PM PT] Ubit Umarov: i refused to full solve the equations :P
[1:37 PM PT] Andrew Hellershanks: :)
[1:37 PM PT] Ubit Umarov: but work on frontal and side collisions
[1:37 PM PT] Kayaker Magic: Cannot search for texture assets 'cause database design is crap. Texture UUIDs are hidden inside a blob for each prim.
[1:37 PM PT] Ubit Umarov: and not that bad on other angles
[1:38 PM PT] Ubit Umarov: in fact a lot of prim data should be inside a blob
[1:38 PM PT] Ubit Umarov: what we have is a total waste
[1:38 PM PT] Andrew Hellershanks: The code spitting out the messages probably doesn't have access to the object UUID or else the console messages could be made to be more helpful.
[1:38 PM PT] Ubit Umarov: you are not suposed to edit prims on mysql
[1:38 PM PT] Ubit Umarov: and materials are a extra mess
[1:39 PM PT] Kayaker Magic: Things that should be in a blob have their own column in the table, things that should be relational are in a blob. Bad design.
[1:39 PM PT] Cuga.Rajal @rajal.org:9000: 10:17:30 - [Materials]: request for unknown material ID: 546ea107-021f-a466-a3d3-7ae51eeefd38
[1:39 PM PT] Ubit Umarov: yes
[1:39 PM PT] Ubit Umarov: viewr jsut did ask
[1:39 PM PT] Ubit Umarov: and the uuid is the only thing it tells region
[1:40 PM PT] Cuga.Rajal @rajal.org:9000: I see, not useful in locating the offending object
[1:40 PM PT] Cuga.Rajal @rajal.org:9000: or prim
[1:40 PM PT] Kayaker Magic: To find the prim for that message Cuga, you have to read the blob for all prims, dig the UUIDs out of it, convert to strings, then compare.
[1:40 PM PT] Ubit Umarov: to add any form of finding it on region would be a insane waste
[1:41 PM PT] Cuga.Rajal @rajal.org:9000: ok then I won't worry about it :)
[1:41 PM PT] Ubit Umarov: in fact texture info is also stored as a bob in regions :p
[1:41 PM PT] Ubit Umarov: and blob
[1:41 PM PT] Andrew Hellershanks: If the outputting of those messages got propagated up then some code that knows the UUID of an object could report the object UUID.
[1:41 PM PT] Cuga.Rajal @rajal.org:9000: if it becomes a problem I can figure it out from process of elimination
[1:42 PM PT] Ubit Umarov: get real
[1:42 PM PT] Ubit Umarov: scan for 100000 prims, 8 faces each??
[1:42 PM PT] Andrew Hellershanks: The problem isn't unique to material IDs. Other console messages reference UUIDs that have issue where you don't know the object.
[1:43 PM PT] Ubit Umarov: ( or 48? faces per avatar or how many they are now )
[1:43 PM PT] Cuga.Rajal @rajal.org:9000: is the material ID a texture feature?
[1:43 PM PT] Ubit Umarov: regions only need to care to find prims avatars and a few othet things per uuid
[1:44 PM PT] Ubit Umarov: yes a filed per face
[1:44 PM PT] Ubit Umarov: field
[1:44 PM PT] Ubit Umarov: as i said, entire texture info is evne stored in region as a encoded blob
[1:45 PM PT] Ubit Umarov: in compact format used by viewers
[1:45 PM PT] Cuga.Rajal @rajal.org:9000: OK then worst case is a borked texture somewhere. thanks for explaining
[1:45 PM PT] Ubit Umarov: encoded/decoded on the fly on the "rare" changes
[1:45 PM PT] Kayaker Magic: Gota run, RL is calling.
[1:45 PM PT] Andrew Hellershanks: ok, Kayaker.
[1:46 PM PT] Ubit Umarov: sure a pin
[1:46 PM PT] Ubit Umarov: pain..
[1:46 PM PT] Ubit Umarov: i also have some missing matrials on my test regions
[1:46 PM PT] Ubit Umarov: well and a few more missing things.. details :)
[1:47 PM PT] Cuga.Rajal @rajal.org:9000: so common issue, nothing to worry about
[1:47 PM PT] Andrew Hellershanks: I have issues with textures in my test regions and no idea what objects are affected.
[1:47 PM PT] Ubit Umarov: those usually endup grey or white
[1:47 PM PT] Ubit Umarov: materilas are hard to see
[1:48 PM PT] Cuga.Rajal @rajal.org:9000: Oh I did have the white texture problem
[1:48 PM PT] Cuga.Rajal @rajal.org:9000: maybe thats related
[1:48 PM PT] Ubit Umarov: white dynamic textures after a tp is "Normal"
[1:49 PM PT] Ubit Umarov: a viewers issue never fixed
[1:49 PM PT] Selby.Evans @grid.kitely.com:8002: bye all
[1:49 PM PT] Andrew Hellershanks: Bye, Selby.
[1:49 PM PT] Ubit Umarov: ( like missing names, lke out lifeguard
[1:49 PM PT] Ubit Umarov: cya sleby
[1:49 PM PT] Cuga.Rajal @rajal.org:9000: So the offending proims would have some white faces
[1:49 PM PT] Motoko.Karu @login.digiworldz.com:8002: bye selby
[1:49 PM PT] Ubit Umarov: and selby.Evans
[1:49 PM PT] Cuga.Rajal @rajal.org:9000: tc Selby
[1:50 PM PT] Ubit Umarov: well i also need to RL
[1:50 PM PT] Ubit Umarov: any other fast question?
[1:50 PM PT] Cuga.Rajal @rajal.org:9000: have to reset skelly
[1:51 PM PT] Andrew Hellershanks: Ubit, I hope not. I need to be going. I've got some things I need to do. :)
[1:51 PM PT] Cuga.Rajal @rajal.org:9000: pretty sure not an OS isue
[1:51 PM PT] Ubit Umarov: should always use default viewers LOD settings
[1:51 PM PT] Ubit Umarov: changing those is BAD
[1:51 PM PT] Cuga.Rajal @rajal.org:9000: nothing unusual in my FS configs
[1:52 PM PT] Cuga.Rajal @rajal.org:9000: prob started with recent FS update but not sure if OS related, probv not
[1:52 PM PT] Ubit Umarov: well some sl builders tell ppl to change LOD settings
[1:52 PM PT] Ubit Umarov: only bc their products are crap and need that to look -
[1:52 PM PT] Cuga.Rajal @rajal.org:9000: could be badly crafted avatar parts
[1:52 PM PT] Cuga.Rajal @rajal.org:9000: wouldnt surprise me
[1:53 PM PT] Ubit Umarov: LODs should be the default, so builders can work and optimiza for all ppl to see the same
[1:53 PM PT] Cuga.Rajal @rajal.org:9000: but rob never happened till recently, sane avi for years
[1:53 PM PT] Cuga.Rajal @rajal.org:9000: prob*
[1:53 PM PT] Ubit Umarov: ok need go rl
[1:53 PM PT] Ubit Umarov: cya around..
[1:53 PM PT] Orbert.Tatham @hg.zetaworlds.com: Peace ubit
[1:53 PM PT] Andrew Hellershanks: ok, I'll close the meeting.
[1:53 PM PT] Cuga.Rajal @rajal.org:9000: kk thank you Ubit!
[1:53 PM PT] Cuga.Rajal @rajal.org:9000: tc
[1:53 PM PT] Motoko.Karu @login.digiworldz.com:8002: laters
[1:53 PM PT] Andrew Hellershanks: Thank you all for coming. See you again next week.
