Simulator User Group/Transcripts/2011.07.05
| Prev 2011.07.01 | Next 2011.07.08 |
List of Speakers
| Andrew Linden | Bronson Blackadder | Charlie Navarathna |
| Fancy Greeter | Flimsey Freenote | Flip Idlemind |
| Joshua Linden | Joshy Aulder | JS Uralia |
| Kallista Destiny | Leonel Iceghost | Liisa Runo |
| Moundsa Mayo | Pauline Darkfury | Rex Cronon |
| Smoovious Laxness | TheBlack Box |
Transcript
[12:00] Kallista Destiny: Ave Andrew
[12:01] Joshy Aulder: Heya Andrew
[12:01] Andrew Linden: Hello everyone.
[12:01] Fancy Greeter: Andrew Linden has arrived!
[12:01] TheBlack Box: hi
[12:01] Andrew Linden: Simon Linden won't be able to make it today. He's on vacation.
[12:01] Rex Cronon: hi andrew. hi everybody
[12:01] Region Performance Monitor v1.194 whispers: Configuration loaded.
[12:01] Region Performance Monitor v1.194 whispers: 29160 bytes free.
[12:02] Pauline Darkfury: Hiya Andrew :)
[12:02] Andrew Linden: I haven't done anything new since Friday's User Group meeting.
[12:02] Moundsa Mayo: No Simon - have you ever minded pulling double duty, Andrew B^D
[12:02] Andrew Linden: The latest news is that there is a crash bug on one of the RC channels for which I just found the fix.
[12:02] Smoovious Laxness: congrats
[12:02] Rex Cronon: u should have done some bbq, andrew:)
[12:03] Smoovious Laxness: log posting reminder :)
[12:03] Andrew Linden: I've got some coordinating to do by the release huddle meeting in about 1.5 hours
[12:03] Andrew Linden: where we'll decide if we're going to patch the RC channel, roll back. or let it sit very crashy.
[12:03] Kallista Destiny: is The on LeTiger? omeone was complaining that their sim rebooted every 15 min
[12:04] Fancy Greeter: Joshua Linden has arrived!
[12:04] Andrew Linden: Of course, I'm always in favor of fixing early, so we know how I'll vote.
[12:04] Pauline Darkfury: I've been hearing of multi-crashes from LeTigre & Magnum
[12:04] Moundsa Mayo: Hey, Joshua
[12:04] Kallista Destiny: If your sure of your fix, of course
[12:04] Joshua Linden: Hey folks
[12:04] Rex Cronon: hi joshua
[12:05] Kallista Destiny: Ave Josh
[12:05] Andrew Linden: The only other bit of news is that I really do plan on publishing the old transcripts from this meeting.
[12:05] Smoovious Laxness smiles
[12:05] Andrew Linden: I actually started this morning, then got sidetracked.
[12:05] Smoovious Laxness: yay
[12:05] Andrew Linden: I was able to add only one transcript, from last Friday.
[12:05] Kallista Destiny: Paper work alwast takes longer
[12:05] Smoovious Laxness puts a happy-face sticker on Andrew
[12:06] Andrew Linden: But I realized I need to fix some bugs in my converter script.
[12:06] Moundsa Mayo: We know it takes a while to excise all the IMs from Lindens about "those ()&@#&(*$$ Residents!!!"
[12:06] Smoovious Laxness: a little progress is progress nonetheless... I'm happy :)
[12:06] Kallista Destiny chuckles
[12:06] Moundsa Mayo: +1, Smoothness
[12:07] Andrew Linden: Yeah, I need to make the script smarter. Right now it doesn't handle it well when we mispell "!@#$%^&*("
[12:07] Moundsa Mayo: Oops - pardon the conflation of your names.
[12:07] Interpreter: Oops
[12:07] Interpreter: -
[12:07] Interpreter: pardon
[12:07] Interpreter: the
[12:07] Interpreter: conflation
[12:07] Interpreter: of
[12:07] Interpreter: your
[12:07] Interpreter: names
[12:07] Interpreter: .
[12:07] Smoovious Laxness: happens all the time, Moundsa
[12:07] Interpreter: happens
[12:07] Interpreter: all
[12:07] Interpreter: the
[12:07] Interpreter: time
[12:07] Interpreter: ,
[12:07] Interpreter: Moundsa
[12:07] Interpreter: .
[12:07] Kallista Destiny: That is a pain...
[12:07] Interpreter: That
[12:07] Interpreter: is
[12:07] Interpreter: a
[12:07] Interpreter: pain
[12:07] Interpreter: .
[12:07] Leonel Iceghost: sorry!
[12:07] Pauline Darkfury: argh, please KILL your tranlator spam, Leonel!
[12:07] Rex Cronon: that interepreted is spamtastic:)
[12:07] Andrew Linden: Hrm... What is up with that interpreter.
[12:07] Rex Cronon: interpreter*
[12:08] Moundsa Mayo: Yeah, and lots of dialects in that vocabulary.
[12:08] Leonel Iceghost: I'm so sorry about that
[12:08] Smoovious Laxness: we can snip that part out of the log :P
[12:08] Andrew Linden: Well, now that we're talking about it, it makes it harder to snip.
[12:08] Moundsa Mayo: Yeah, and correct typos, and maybe make me look a litle less stupid, please?
[12:08] Moundsa Mayo: OOPS!!
[12:09] Liisa Runo: Andrew. if i know a sim that been having constant peaks of net time going to 60.0ms up to 2200.0ms for 2 days, despite restartings and autoreturn, should i just Ticket it or is there something that you want to investigate first?
[12:09] Joshua Linden: Since someone will ask: the legendary SVC-6760 caps-router/499 issue fix is still in deploy-testing. Looking good, just lots of work to test.
[12:09] Moundsa Mayo: Good to hear, Joshua!
[12:10] Moundsa Mayo: Hmm. No JIRA helper active here?
[12:10] Andrew Linden: Liisa, if we had time today we might actually look into a report of a problem like that.
[12:10] Liisa Runo: ok, ill ticket it and see if the recieving linden understand that it is not fixed with remote restart
[12:10] Andrew Linden: Today is a bad day for it though. Simon is out and it looks like I'm going to be scrambling to prepare a simulator for a possible emergency fix.
[12:10] Smoovious Laxness: and, when a sim goes through a maint restart, upgrade, etc... do you guys have a way of knowing that a sim hasn't come back up? I know of another person's sim that stayed down for hours until the estate owner got on and called it in
[12:11] Pauline Darkfury: You might want to note that region on Kelly's latest performance Jira. It's not 100% on-topic for it, but could be relevant. SVC-7079
[12:11] JIRA-helper: http://jira.secondlife.com/browse/SVC-7079
[#SVC-7079] Full region simulator performance issues since 6-30 server oll
[12:11] Flip Idlemind: Oooh emergency fixes are always fun
[12:11] Pauline Darkfury: Anyone can put a "region offline" ticket in, Smoovious
[12:11] Kallista Destiny: for some very small value of fun.
[12:12] Smoovious Laxness: oh ok... next time it comes up I'll pass it along
[12:12] Andrew Linden: Yes, there is a system for bringing up downed regions.
[12:13] Flip Idlemind: The conductor
[12:13] Andrew Linden: I haven't been on that team for a few years now, so I'm not sure how they system fully works now.
[12:13] Smoovious Laxness nods
[12:14] Andrew Linden: That's all the news I have, so the table is open.
[12:14] Pauline Darkfury: So, what's your current feeling on performance with Mono 2, Andrew? Is there an actual perf issue? Is there a numbers issue? Both?
[12:15] Andrew Linden checks around to see if Kelly is here...
[12:16] Andrew Linden: I'm not well informed on the Mono2 performance issues.
[12:16] Joshy Aulder: : you just gave me an idea for my radar HUD - a search option
[12:16] Andrew Linden: All I've heard is that some cases are better (as good as legacy LSL)... either rez or derez. I can't remember which, but I think it was for rez.
[12:17] Pauline Darkfury: Some people seem a bit unhappy with it. Myself, I'm yet to be convinced there's a huge issue on full sims (not sure about HSs), although am pretty convinced that Top Scripts is giving bad data on lightly loaded sims
[12:17] Rex Cronon: btw. there is someting weird going on with the lsl compiler. it seems to give wrong line numbers
[12:17] Pauline Darkfury: it was reporting error line numbers earlier for me ok
[12:18] Joshua Linden: Rex: the pipeline changed a bit, so I'm not surprised if that happens on occasion.please JIRA that one (SCR if you can)
[12:18] Joshua Linden: With a repro case of course :)
[12:18] Joshy Aulder: Yeah, often when I compile a script, if theres an error in it my attention will usually be brought to a line either above it or below it
[12:18] Rex Cronon: i don't know what triggers i
[12:19] Rex Cronon: i can't reproduce it
[12:19] Moundsa Mayo: No repro, can't fix B^(
[12:19] Joshua Linden: (Also note that reporting line numbers correctly in compilers is tricky, so we may not be perfect, but if you can find an example we can take a look)
[12:19] Pauline Darkfury: mmm, I couldn't swear it wasn't doing an off-by-one though, but that's not unusual for compilers, as sometimes the exception isn't raised until the next line, when it gets something it doesn't expect at that point (e.g. from a missing semi)
[12:19] Rex Cronon: i look a line above, a line bellow. i see no errors. the next day i log in, and it compiles ok
[12:20] Andrew Linden: Yeah, too bad Kelly isn't here for that question.
[12:20] Rex Cronon: is like the compiler gets a borked version of my code
[12:21] Kallista Destiny: Compilers occasionally get confused.
[12:21] Rex Cronon: and it only happens at rausch
[12:21] Andrew Linden: Really?
[12:21] Joshy Aulder: Oh, one question. Regarding the fix for graphics crashers, I heard that sculpts with a surface area greater than 32 are rendered as plain spheres. What unit is that 32 in exactly?
[12:21] Rex Cronon: yes. either the center sim, or the blue one
[12:22] Pauline Darkfury: Hmm, random thought just popped into my head. If it's easy to do it, some people would be very happy if the fix for SVC-6905 got rolled into the "mesh-prep 2" regions, I think (it's currently on the BlueSteel maint-server stuff)
[12:22] JIRA-helper: http://jira.secondlife.com/browse/SVC-6905
[#SVC-6905] Immortal prims: Objects resisting take, delete, return, and autoreturn
[12:22] Andrew Linden: I dunno Rex. If you ever find a reliable repro definitely let us know.
[12:22] Rex Cronon: ok
[12:22] Andrew Linden: Figuring that one out is going to be tough with out a repro recipe.
[12:23] Andrew Linden: Joshy, my first guess for the units for that graphics crasher would be meters^2.
[12:23] Andrew Linden: But I'm not positive. What is the jira #?
[12:24] Joshy Aulder: Not sure. A friend of mine told me about it, but he didnt have any idea what the unit was
[12:24] Pauline Darkfury: Megaprim sculpts work ok, though, so it can't be larger than a certain size? Do you maybe mean the issues with sculpts being replaced by spheres due to RenderVolumeLODFactor being too low?
[12:24] Liisa Runo: i dont see how that would work without ruining lots of content.
[12:25] Andrew Linden: Pauline, that is a good idea. I'll try to track down the code changes for fixing SVC-6905 so I can see how risky it appears.
[12:25] Flimsey Freenote: excuse me, gets fire extinguisher ready.. must go put out a fire
[12:25] Andrew Linden: If it were just me of course I would do it.
[12:25] Andrew Linden: But then my sense of paranoia for breaking stuff tends to be low.
[12:25] Pauline Darkfury: Yup, you've obviously got a more significant issue on mesh-prep to deal with, so will understand if it can't be done in tomorrow's roll
[12:27] Liisa Runo: and even ruining all big sculpts will not fix the problem, can crash clients also with normal prims
[12:27] Joshy Aulder: Well I think it had something to do with the surface area of the sculpt itself, not the prim
[12:28] Andrew Linden makes a note about porting SVC-6905 so I don't forget. No promises on that one.
[12:28] Pauline Darkfury: Thanks :)
[12:28] Liisa Runo: .. i guess i better not mention the methods here in open. i guess LL know the tricks already
[12:28] Andrew Linden: Some of them, yes.
[12:30] Meeter: Timecheck : User Group is half over
[12:30] Joshy Aulder: The fix was viewer side, if that helps. Viewer 2 has the fix, but sculpt crashers will still crash 1.x viewers
[12:30] Liisa Runo: v2 has fix? challenge accepted!
[12:30] Kallista Destiny: Which Version?
[12:30] Kallista Destiny: 2.?
[12:31] Joshy Aulder: Not sure, maybe 2.5 or higher if it still crashers Firestorm
[12:31] Joshy Aulder: Sorry, higher than 2.5
[12:31] Kallista Destiny: Firestorm is 2.5.2 well the Beta is
[12:31] Flip Idlemind: The JIRA for what Joshy is talking about is SH-1696 (has permission violation on it though)
[12:31] Joshy Aulder: Oh thx Flip
[12:33] Andrew Linden: Joshy, the units were in "meters".
[12:33] Andrew Linden: Sculpts with surface areas > 32m were to be left as spheres (according to the comments).
[12:34] Liisa Runo: that will ruin oh so much content
[12:34] Joshy Aulder: So does that mean megaprim sculpts will be broken?
[12:34] Liisa Runo: yes
[12:34] Pauline Darkfury: That does sound content-breaking
[12:34] Kallista Destiny: Very content breaking
[12:34] Pauline Darkfury: and doesn't sound like it will really solve anything if the issue is complexity, as surface area is a very poor measure of complexity
[12:35] Rex Cronon: so, u can't have mega-sculpties:(
[12:35] Flip Idlemind: A recent commit to "viewer-development-shining-fixes" changes that max surface area for 32 to 384, "Fix for blocking legitimate sculpt maps with over-eager area check"
[12:35] Flip Idlemind: from 32 to 384*
[12:35] Andrew Linden: Yeah, sounds like it would break some content.
[12:35] Pauline Darkfury: I'd think a 10x10x10 sculpty could have more than 32m^2 surface, and 64x64x64 (which won't be mega soon) certainly can
[12:36] Kallista Destiny: 32 s not a very large object
[12:36] Smoovious Laxness: will the linking distance and count limits be going up? (one or the other)
[12:36] Joshy Aulder: But is this based on the prim or the sculpt itself?
[12:36] Rex Cronon: there r quite a few islands that use 768x768x256 to make their sim look biger
[12:36] Kallista Destiny: 10X10X10 is 1000 urface
[12:39] JS Uralia: is http://wiki.secondlife.com/wiki/Server/Sim/Scripting_User_Group#Agenda the agenda?
[12:39] Rex Cronon: well joshua. that will no longer be true. now u will only see a huge ball. right?
[12:39] Pauline Darkfury: Mmm
[12:39] Charlie Navarathna: I saw a HUD for mega It's a sculpty generator with hundreds of maps included, many for terrain, rocks, water, etc. I've used it to make a giant waterfall and volcano on my land. The items it generates are full perm
[12:40] Liisa Runo: anyway, as i said, can crash clients with normal prims too. the fix must be in some other place, not in limiting how sculpts get drawn
[12:40] Pauline Darkfury: I can see that one being un-fixed by TPVs, tbh, if it's the anti-fix it sounds like
[12:40] Andrew Linden: No JS, we don't have much of an agenda for this meeting. I usually give some news and we open up the table.
[12:40] Rex Cronon: u can make it a switch, that a user can turn it on/off
[12:40] Joshua Linden: Rex: I'll follow up w/ the Lindens touching that JIRA
[12:41] Rex Cronon: cool:)
[12:41] Andrew Linden: Which reminds me. Archeron was asking me via email about how to make effective "shields" agains fast moving bullets.
[12:41] Rex Cronon: u can't
[12:41] JS Uralia: my friends have nominated me to try to say something about https://jira.secondlife.com/browse/SVC-472
[12:41] Liisa Runo: shields are useless eyecandy
[12:41] JIRA-helper: [#SVC-472] Region Crossings Fail
[12:41] Smoovious Laxness: make em 64m thick?
[12:42] Joshy Aulder: Yeah thatd seem like a good idea, sorta like what I had to do on the 1.x viewers. Ctrl-Shift-Alt-9 to derender volume. Except that way the only thing getting "derendered" is the crasher
[12:42] JS Uralia: has anyone else heard that region crossings got worse a few days ago?
[12:42] Liisa Runo: shoelds are relic from the days when watermelon launcher was the most powerfull weapon
[12:42] Rex Cronon: even if u put yourself inside a solid prim, that still won't stop a some bullets:)
[12:42] Andrew Linden: That is, Archeron Gloom, was asking.
[12:42] Andrew Linden: Bullets can tunnel through thin prims.
[12:42] Andrew Linden: They'll collide with thick objects, however there are two kinds of "collision points" when objects collide:
[12:42] Smoovious Laxness: was thinking more about being behind the prim instead of inside it
[12:43] Rex Cronon: there is at least one way to hit somebody that hides behind 64m of prims:)
[12:43] Andrew Linden: (1) collisions where the two objects are near (withint about 0.1m) but not overlapping
[12:43] Andrew Linden: (2) Collisions where the objects are overlapping
[12:43] Andrew Linden: in case (2) the "penetration resolution" code kicks in
[12:44] Andrew Linden: case (1) has the most "stopping power" because the correct collision forces are computed
[12:44] JS Uralia: maybe you can use server GPUs to optimize colision detection http://portal.acm.org/citation.cfm?id=1944756
[12:44] Andrew Linden: whereas the penetration resolution code is not terribly smart and just tries to get the two objects appart
[12:44] Andrew Linden: it may push the object sideways, or even forward -- whatever it thinks would get it out the fastest.
[12:45] Andrew Linden: so a very thick shield of one prim will likely just get bullets penetrating it
[12:45] Pauline Darkfury: yup, or if you're in the middle of a large prim, it just kinda gives up pushing you out
[12:45] Moundsa Mayo: Andrew, does it separate them along a vector of least-length-to-separation, or is it velocity-dependent?
[12:45] Andrew Linden: So for best (theoretical) results you would just stack a bunch of boxes of different depths
[12:46] Rex Cronon: there is a way to make super-fast shield. although might not be that effective. u allow attached prims to use volumedetect:)
[12:46] Liisa Runo: Shields protect you only from very weak attacks. And they waste lots of sim resources. For them to be even somewhat reliable they need to have tons of sensors firing very rapidly and they rez stuff rapidly. Not worth the lag just to be protected from watermelon launcher or other old guns. Weapon makers should finally stop making these even though they look cool and flashy.
[12:46] Andrew Linden: such theat each forward shielding face is about 0.1m shifted
[12:46] Andrew Linden: that way, if a bullet penetrates the shield, it will likely have a colliding face within about 0.1 meter, so a correct collision point will be computed.
[12:47] Joshy Aulder: Attached prims using llvolumedetect? That would be a lot like when you could sit on a phantom cube and then be forced off it because it deleted itself and you stayed phantom. I heard it was pretty bouncy
[12:47] JS Uralia: are serve gpus already being used?
[12:47] JS Uralia: server* GPUs?
[12:47] Bronson Blackadder: laminate a few layers with spaces between
[12:47] Rex Cronon: that is why we need llvolumedetect to work on attachments:)
[12:47] Andrew Linden tries to parse Moundsa's question.
[12:47] Leonel Iceghost: I think you could improve physics in general, but promove Cast Ray instead of bullets for weapons
[12:48] Rex Cronon: well. castray is kind of useless if u want mortar fire:)
[12:48] Bronson Blackadder: castray as a laser rangefinder Rex
[12:48] Andrew Linden: Ah, I think Moundsa was aking about the details of the penetration resolver's intelligence.
[12:48] Rex Cronon: it wasn't bouncy, joshy:)
[12:48] Pauline Darkfury: Using the GPUs could be tricky, if the servers even have GPUs to be used. There's 4-8 full sims per server (they are 4 or 8 core systems), many more homesteads & openspaces
[12:49] Moundsa Mayo: The interpenetration algorithm must have some strategy. Is it static geometry dependent, on the particular penetration at hand, or does it react to the relative velocities??
[12:49] Leonel Iceghost: a reimplementation of the built in combat system would be very nice though
[12:49] Andrew Linden: I don't know exactly how the penetration resolver works, but it is expensive (CPU cycles) to compute the most optimal path out between two penetrating objects
[12:49] Joshy Aulder: I wasnt around back then, but I heard that a major problem with it was people falling through the ground and then bouncing back up
[12:49] Andrew Linden: so you can be sure it is not optimal
[12:49] Smoovious Laxness: sounds fun
[12:50] JS Uralia: if you just want higher resolution collision detection for fast things, and servers have GPUs that aren't being used, then maybe that's a place to see whether it's appropriate to optimize
[12:50] Rex Cronon: having a flight assist was good;)
[12:50] Moundsa Mayo: ty, Andrew.
[12:50] JS Uralia: anyway, we only got 10 minutes. Are boundary crossings failing for large linksets?
[12:51] Pauline Darkfury: How large a linkset?
[12:51] Charlie Navarathna: I read that In Second Life this problem is resolved by making all projectiles as if they were arrows; a long invisible shaft trails behind the bullet so that as the bullet teleports forward, the shaft is long enough to cover the gap between successive teleports of the bullet and collide with any object that might fit between
[12:51] Andrew Linden: I'm not aware of a reliable boundary crossing bug for large linksets.
[12:51] Charlie Navarathna: frames
[12:51] Moundsa Mayo: JS, in light of your question, would you consider an avatar with several complex attachments to be a large linkset?
[12:51] Andrew Linden: Charlie, that is a popular strategy for many "bullet games", even outside of SL.
[12:52] Andrew Linden: "long bullets" that is.
[12:52] Charlie Navarathna: nods
[12:52] JS Uralia: well https://jira.secondlife.com/browse/SVC-472 is old and needs to be recycled. I was told there would be a new comment on it but I need to get back to the person who told me that
[12:52] JIRA-helper: [#SVC-472] Region Crossings Fail
[12:52] Smoovious Laxness: where would the bullet's impact be calculated from tho, the front? middle?
[12:52] Pauline Darkfury: Too high a script count is the most reliable way to get crossing issues, vastly improved on what it used to be, but 200 scripts and a high speed crossing is a recipe for a 10-20km trip across the void, then disconnect
[12:52] JS Uralia: keeping bugs open that long for persistent issues makes it hard to use JIRA
[12:52] Andrew Linden: Ah yes. The meta jira issue for all region crossing failure modes: SVC-472.
[12:53] Moundsa Mayo: B^D
[12:53] Kallista Destiny: Smooth you detect the collision by position (objects move stepwise in SL).
[12:53] Andrew Linden: Unfortunately SVC-472 can't be closed. We'll have to leave it open until no one can complain about region boundaries.
[12:54] JS Uralia: oh, it's a tracking bug, I see
[12:54] Kallista Destiny: if the buttle intersecst with the Avatars
[12:54] Smoovious Laxness: there will always be someone who can complain
[12:54] Joshy Aulder: In other words itll be open for as long as SL is around?
[12:55] Meeter: Timecheck : User Group is almost over
[12:55] Andrew Linden: Progress can be made on SVC-472, however I wonder if it would ever be able to be completely closed.
[12:55] Andrew Linden: There are so many ways it can fail.
[12:55] JS Uralia: sure it can be
[12:55] Pauline Darkfury: Mmm, talking of region boundaries, how's the 8-sim cross-region encroachment returning? And any news on getting it switched on for mainland?
[12:55] Andrew Linden: It may be that with a large enough virtual world and complex enough hardware/network system underneath it may be that region crossings will be failing somewhere in the world at any time
[12:56] Andrew Linden: even in the "glorious future"
[12:56] Moundsa Mayo: That models RL, Andrew. NP.
[12:56] Rex Cronon: won't be that a future something similar to the matrix:)
[12:56] Andrew Linden: It would be nice to fix some of the known problems though.
[12:56] Rex Cronon: or skynet:)
[12:56] JS Uralia: well, if a sim goes down during a region crossing, that would cause a failure
[12:57] Smoovious Laxness: that's an easy fix tho, JS... just need to have 100% uptime, problem solved
[12:57] Moundsa Mayo: Sure, JS, or a backhoe operator 'recycles' some fiber optics cable ...
[12:57] JS Uralia: do linksets really get renumbered across regions?
[12:57] Pauline Darkfury: Yeah, there's always going to be some case possible when the system on the far side of the boundary could be having issues. The main thing is to cover the reasonable cases (say less than 100 scripts at a reasonable max speed, between 2 full regions that are healthy)
[12:57] Leonel Iceghost: if a server handles 4 sims, why not make them one sim, divided only at the front end as 4 sims again
[12:57] Liisa Runo: having sims full of temp rezzers marked red to the mimimap would help us avoid crashing on border. Could avoid sims that are lagged out. :P
[12:58] JS Uralia: it would seem like trying to preserve linkset numbers would help prevent llsetpos vehicle scripting bugs
[12:58] Charlie Navarathna: I think one of the limits of physics engine realism is the precision of the numbers representing the position of an object and the forces acting on that object. When the precision is too low, errors can creep into the calculations due to rounding
[12:58] Pauline Darkfury: link numbers shouldn't ever change unless someone is editing the object
[12:58] Pauline Darkfury: an AV sitting on an object will add a link number, but the prims should remain the same
[12:59] Andrew Linden: I don't believe the link numbers change on region crossing... except for the numbers of multiple sittingh agents.
[12:59] JS Uralia: Charlie: you are correct. We need to get double precision for faster formation flight
[12:59] Andrew Linden: er... link numbers for sitting agents.
[12:59] JS Uralia: that's good.
[12:59] Charlie Navarathna: good
[13:00] Meeter: Thank you for coming to the Server User Group
[13:00] Pauline Darkfury: if the prim link numbers are changing without any editing, that's a critical or showstopper issue that needs Jira'd
[13:00] Andrew Linden: Ok, sorry but I've got some emergency to work on and have to report on it in 30 min.
[13:00] JS Uralia: thanks, Andrew
[13:00] Joshy Aulder: Thanks
[13:00] Smoovious Laxness: have fun with that, Andrew o/
[13:00] Andrew Linden: So I'm going to depart on the hour today.
[13:00] Rex Cronon: tc andrew
[13:00] Pauline Darkfury: multi-AV sitting is a very complex thing, so wouldn't want to say when that is "expected behaviour" to change
[13:00] Andrew Linden: Thanks for coming everyone.
[13:00] Joshy Aulder: Baibai
[13:00] Pauline Darkfury: Thanks, Andrew
[13:00] Joshua Linden: ttfn folks!
[13:00] Pauline Darkfury: Happy coding :)
[13:01] Charlie Navarathna: byee
[13:01] Pauline Darkfury: Thanks, Joshua :)
[13:01] Moundsa Mayo: Andrew, Joshua, thanks for your time!
[13:01] Rex Cronon: tc joshua
[13:01] Rex Cronon: tc everybody, and have a nice day:)m
| Prev 2011.07.01 | Next 2011.07.08 |