Simulator User Group/Transcripts/2011.08.12

From Second Life Wiki
Jump to: navigation, search


Prev 2011.08.09 Next 2011.08.16

List of Speakers

Acheron Gloom Andrew Linden ANSI Soderstrom
Ardy Lay DogWomble Dollinger Draconis Neurocam
DrFran Babcock Elite Runner Fancy Greeter
Flip Idlemind Gooden Uggla IEvangeline
Kallista Destiny Kaluura Boa Latif Khalifa
Liisa Runo Lomgren Smalls Motor Loon
Pauline Darkfury Questar Utu Ramona Criss
Rex Cronon Rodney Monitor Sahkolihaa Contepomi
Simon Linden Thomas Shikami Vincent Nacon
xstorm Radek


[16:00] Vincent Nacon: but yeah, anyone here have Mesh topic?

[16:00] Acheron Gloom: Why not physics?

[16:00] Acheron Gloom: Physics are cool D:

[16:00] Simon Linden: hey, don't knock the Meeroos until you've tried them ... they're tasty

[16:00] Sahkolihaa Contepomi: And hey Andrew.

[16:00] Motor Loon: ey Andrew

[16:00] Andrew Linden: Hello everyone.

[16:00] Meeter: Welcome to the Server User Group

[16:00] Latif Khalifa: hello Andrew :)

[16:00] xstorm Radek: simon did you see the post on google+ that hamlet did ?

[16:00] Latif Khalifa: hey Simon :)

[16:00] Vincent Nacon: you mean eat them alive, Simon?

[16:00] xstorm Radek: hi Andrew

[16:00] DrFran Babcock: hallo, Lom

[16:00] Simon Linden: Alive? Yuck, I don't like chewing fur

[16:01] Questar Utu: Late? It should only be midnight there shouldn't it? :p

[16:01] DrFran Babcock: watch the meeros, they eat micros

[16:01] xstorm Radek: lol

[16:01] Sahkolihaa Contepomi: That's still late, hon. xP

[16:01] xstorm Radek: thats why im a bug on the table

[16:01] Sahkolihaa Contepomi: Or early. Since it's technically Saturday now.

[16:01] Vincent Nacon: alrighty, let's try saving physic topic after half of the meeting time and let's see what news they got to share

[16:02] xstorm Radek: im just watching to day lol

[16:02] xstorm Radek: no bugging

[16:02] Rex Cronon: greetings everybody

[16:02] Vincent Nacon: physic is too physical, that's why

[16:02] Vincent Nacon: :P

[16:02] xstorm Radek: hi Rex

[16:02] Rex Cronon: hi xstorm

[16:03] Andrew Linden: I don't have much news. I found a few more things to fix in mesh accounting so I've been working on that.

[16:03] Andrew Linden: I didn't have time to look into the stats times not adding up.

[16:03] Andrew Linden: But it was on my little list of things to do. Just didn't get to it.

[16:04] xstorm Radek: little ? what the little

[16:04] Fancy Greeter: Andrew Linden has arrived!

[16:04] xstorm Radek: lol

[16:04] Rex Cronon: linden crashing?

[16:04] Andrew Linden: is that Xstorm's micro avatar standing on the table?

[16:05] Flip Idlemind: Hey cookie lovers. Viewer 2 kept crashing on log in T_T

[16:05] Sahkolihaa Contepomi: Yup.

[16:05] xstorm Radek: yes

[16:05] xstorm Radek: i hope

[16:05] Sahkolihaa Contepomi: That thing needs v3/viewer 3 adding to it now.

[16:05] DrFran Babcock: Hey, Lom is here, he's smaller

[16:05] Motor Loon: Mabye I should change places with Xstorm... these chairs are really kinda tiny

[16:06] xstorm Radek: yep

[16:06] Flip Idlemind: Phoenix

[16:06] DrFran Babcock: it's a amirco-off

[16:06] Lomgren Smalls: Hehe. There.

[16:06] xstorm Radek: lol

[16:06] xstorm Radek: omg ansi

[16:06] Andrew Linden: Oh yeah. that "invisi prim" hack that Phoenix used to use for a "bridge attachment"... that bit us this week on a side project...

[16:06] Vincent Nacon: excuse Ansi, send her a TP

[16:06] Kallista Destiny: off?

[16:07] Vincent Nacon: sent*

[16:07] Andrew Linden: those shapes were causing problems in some test simulator code

[16:07] ANSI Soderstrom: hi

[16:07] Rex Cronon: warning: don't step on the table. u might squish somebody:)

[16:07] Andrew Linden: I'll probably have to resurrect that old attempt to censor those shapes and actually make it work correctly.

[16:07] Kallista Destiny: Oh the extreamly tortured prim that couldnlt be rendered and so was ignored?

[16:07] xstorm Radek: are they ?

[16:07] Pauline Darkfury: does the sim care about the shape when it's attached, or just the ones that end up rezzed in world?

[16:07] Andrew Linden: Yes Kalista.

[16:08] Acheron Gloom: If you turn on wireframe you see them as a circular wire.

[16:08] Andrew Linden: the old fix was accidentally too restrictive so we have to roll back

[16:08] xstorm Radek: andrew why are you glowing

[16:08] Simon Linden: The sim cares about shapes for the physics world

[16:08] Andrew Linden: it damaged some renderable tortured prims

[16:08] ANSI Soderstrom shouts: fucking laaaaaaag

[16:08] Kallista Destiny: PG region ANIS

[16:08] Andrew Linden: but I'll probably have to try fixing that again soon

[16:08] DrFran Babcock: yikes I almost win the scripts contest

[16:08] DrFran Babcock: time to divest

[16:08] ANSI Soderstrom: ban me kallista

[16:09] ANSI Soderstrom: iam sure childrens would use more bad words for this lag

[16:09] Pauline Darkfury: seems ok here, just a little triangle-heavy with everyone here

[16:09] Kallista Destiny: I have 14 FPS, Perhaps she should be cursing Deutchetelecom?

[16:10] Rex Cronon: if u look away from table u might get less lag

[16:10] Acheron Gloom: If I can handle the area with a radeon x1600, anyone can ;p

[16:10] xstorm Radek: hhhmmm viewer 2 has another bug damn

[16:10] Kallista Destiny: 2 or 3?

[16:10] Simon Linden: fwiw the viewer 3 beta just went out today

[16:10] Sahkolihaa Contepomi: Yeah, you need to add it to the ball. :p

[16:10] DrFran Babcock: oooh

[16:10] Motor Loon: downloaded, but not tested yet

[16:11] Kallista Destiny: V3 is really just V 2.8 right?

[16:11] Sahkolihaa Contepomi: It's mesh.

[16:11] Vincent Nacon: no

[16:11] DrFran Babcock: mesh

[16:11] Vincent Nacon: V2.8.1

[16:11] Vincent Nacon: :)

[16:11] Acheron Gloom: its v3.0.5

[16:11] Acheron Gloom: or v3.0.2

[16:11] Acheron Gloom: I forget

[16:11] Andrew Linden: Yes, 3.0 is just 2.8 with a merge or two and a higher number.

[16:11] Kallista Destiny: Thank you

[16:11] Vincent Nacon: v2.8.1

[16:11] Vincent Nacon: yup

[16:11] Motor Loon: "fancy wrapping"

[16:12] Motor Loon: (kinda like andrew today)

[16:12] Elite Runner: vIEWER 3.0?!!!!!!!!

[16:12] Vincent Nacon: it seem ANSI found a bug with float

[16:12] Kallista Destiny: He If Sun can rebadge Sun OS as Solaris LL can rebadge V2.8 as V 3.0

[16:12] Pauline Darkfury: See above, it's just the latest 2.x

[16:12] Kallista Destiny: what ever floats you boat.

[16:12] Simon Linden: I believe it's ready for the grid switch and server to fully support mesh ... the number is ultimately meaningless, there are a lot to choose from

[16:13] Liisa Runo: SVC-7183

[16:13] Lomgren Smalls: Or Windows 7... heh.

[16:13] JIRA-helper:

[#SVC-7183] mathematics with float bigger than 0.8 will fail very badly !!!

[16:13] DrFran Babcock: aaaah

[16:13] Andrew Linden: Right. The mesh rollout will happen soon, so the new number is for the viewer that corresponds to it, I think.

[16:13] xstorm Radek: viewer 3? will it have localport

[16:14] Kallista Destiny: IBM 4300 series is realy just a rebadged and boxed 370

[16:14] Sahkolihaa Contepomi: SNowstorm is already v3 so just grab that and you'll see - or the beta.

[16:14] Elite Runner: Viewer 3?

[16:14] Questar Utu: It's just a version, not an entirely new thing.

[16:14] Kallista Destiny: I was just thinking that the Firestorm dvlopers would be sending hate male if that had a new code base to port

[16:15] Simon Linden: SVC-7183 is interesting, but it really is a fact-o-life, or binary math

[16:15] Sahkolihaa Contepomi: Well when OpenGL 3 comes to the viewer which is being worked on...that's more for them to do. :p

[16:15] Flip Idlemind: Do I understand this correctly, open-source versions of the new viewer can't upload mesh, because they don't have convexdecomposition (Not *entirely* sure what that is) And yeah this isn't a viewer meeting, but we're talking about it

[16:15] Simon Linden: The first comment in the jira is good

[16:15] Pauline Darkfury: Yeah, I've not looked at the details on it, but initial gut feeling is: yeah, mixing floats & ints does stuff like that

[16:15] xstorm Radek: lol

[16:15] Simon Linden: Flip - I have no idea, sorry

[16:16] Andrew Linden: I looked at SVC-7183 very quickly...

[16:17] Andrew Linden: I think that rounding bug was introduced by some work that was done on the mono script engine

[16:17] ANSI Soderstrom: 7183 makes that in fact TRUE : if(x==0.9 && x>0.9)

[16:17] Kallista Destiny: 0,999999999999999999 + 0,00000000000000000001 != 1.0

[16:17] Andrew Linden: there are two common ways to round 0.5...

[16:18] ANSI Soderstrom: but this works only with a float > 0.8 < 1.0

[16:18] Andrew Linden: you can always round up ... which is what C++ does by default, and which is what LSL used to do

[16:18] ANSI Soderstrom: LSL dont line 0.900000000

[16:18] ANSI Soderstrom: *like

[16:18] Andrew Linden: because it was implemented in C++

[16:18] Pauline Darkfury: I do know that the following fails with glow (I think) via llSLPPF: float f = 1.0; while (f > 0.0) { f -=0.1; llSLPPF(LINK_THIS,[PRIM_GLOW,f]) }

[16:18] Andrew Linden: the other option is to always round to the even number (known as "banker's rounding" I think)

[16:19] Pauline Darkfury: you end up with -0.000001 or something silly after the 10 subtractions, and get an out of bounds error

[16:19] Simon Linden: hmm, I need a sound to play for someone mentioning viewer 3...

[16:19] Andrew Linden: which has the benefit of not causing a systematic rounding error when rounding lots of stuff... some of it rounds up and some rounds down

[16:19] xstorm Radek: i think may be in time they need to stop rounding up when a sim restarts do to it makes drift errors even if its so small

[16:20] Pauline Darkfury: the sim shouldn't be rounding anything on restart (but scripts could be doing stuff that rounds)

[16:20] Rex Cronon: it might be better to be consistent. like always round either up or down, but not both

[16:21] Andrew Linden: MONO uses banker's rounding within its math, and some rounding code was moved over to work within MONO implicitly

[16:21] Pauline Darkfury: Mathematically, you're supposed to round both up and down, with 50/50 distribution, to make the errors smaller

[16:21] Ramona Criss: Hello ALL :)

[16:21] xstorm Radek: the GOV grid has drift of items

[16:21] Rex Cronon: hi

[16:21] Kallista Destiny: I remember builds getting sloppy after 20 reboots (little gaps and overlaps

[16:21] Andrew Linden: I think Kelly was going to make things use the old style "always round up" again.

[16:21] Andrew Linden: But I'm not sure of the status of that.

[16:21] xstorm Radek: yes

[16:22] Andrew Linden: I didn't have time to fully grok the example script in SVC-7183, but I'm guessing it has something to do with that rounding change.

[16:22] Pauline Darkfury: But won't that actually make errors grow larger faster (always rounding up)?

[16:22] Andrew Linden: Kallista, how long ago was that "linked object drift" bug? years ago, right?

[16:22] Andrew Linden: I did some of the fixes for that.

[16:23] Liisa Runo: 2009 i think

[16:23] Kallista Destiny: Yeah, 1.19 time fram

[16:23] Andrew Linden: Pauline, yeah the "always round up" system causes systematic error in some things, however we would switch back to that for consistency with legacy LSL scripts.

[16:24] Pauline Darkfury: Well, if we've been on the more mathematically sound method for a while, I say stick with it, tbh

[16:24] Pauline Darkfury: This isn't a recent change? (or is it?)

[16:24] Andrew Linden: I don't think it has been very long. The bug was introduced last week or before.

[16:25] Pauline Darkfury: oh, really, ok

[16:25] Kallista Destiny: That was right around the first occurqnces of Ghost Avatars and the hair/boots up the rump

[16:25] Andrew Linden: Hrm... 7183 was created two months ago.

[16:25] Pauline Darkfury: Still doesn't feel right to put it back tho

[16:25] Vincent Nacon: O-o

[16:26] Liisa Runo: 7183 shows to be created today

[16:26] Liisa Runo: and says it gets bugged only in current bluesteel

[16:26] Motor Loon: good ol' days

[16:26] Andrew Linden: On second thought, maybe the bug is not related... I'll have to study it later when I have more time to focus.

[16:26] Andrew Linden: And when it isn't at the end of Friday.

[16:27] xstorm Radek: lol

[16:27] xstorm Radek: Andrew what no over time

[16:27] Kallista Destiny: I remember slowly sliding donw a hill, standing still (really early Havok release)

[16:27] ANSI Soderstrom: it says i only tested it on bluesteel @ liisa

[16:27] Vincent Nacon: family with demands

[16:28] xstorm Radek: i remember sinking in the ground lol in 2003

[16:28] Lomgren Smalls: Heh. With these distorted animations, if I go up too high in a sim, you can see my body parts "vibrating"

[16:28] Sahkolihaa Contepomi: I think that's a common problem in general.

[16:28] Vincent Nacon: shake it like it's hot

[16:28] Lomgren Smalls: It's a lot more noticeable on micros

[16:28] Lomgren Smalls: Just like a lot of other things

[16:28] Motor Loon: mmm... bootyshake

[16:28] xstorm Radek: true

[16:29] Lomgren Smalls: I think it can be seen on me at around 700-800 meters.

[16:29] Andrew Linden: Oh I see... they are adding 0.1 several times, then checking to see if the result is eventually equal to 0.9

[16:29] xstorm Radek: us micro sink to not seeing ol

[16:29] Andrew Linden: Yeah, they are seeing floating point error in "macroscopic" numbers

[16:29] Kallista Destiny: We had a group where I worked who persisted in breaking the network in there space. I remember Friday at about 1640 they broke their netwokt just as i was logging off to go home (supposed to leave at 1600)

[16:29] Kallista Destiny: euqal no greater that perhaps yes

[16:30] Pauline Darkfury: The other thing that's worth looking at, quite possibly, is to see if IEEE-754 mandates or suggests anything for rounding. Being strongly consistent with other common implementations of float32 has to be worth something

[16:30] Meeter: Timecheck : User Group is half over

[16:30] Vincent Nacon: physic time!

[16:30] xstorm Radek: Kall was that do to network timing ?

[16:30] Pauline Darkfury: (assuming I'm remembering the standard number correctly)

[16:30] Vincent Nacon: if there isn't many of course

[16:31] Kallista Destiny: Nope, back the the days of thin net, they like to mess with the topoligy

[16:31] xstorm Radek: oh ok

[16:31] Kallista Destiny: making Ts and Ys and double terminating

[16:31] Questar Utu: Just a thought, but has there been anything regarding physics about the fact that avatars sink into the ground on land by about 0.1m, which is an easily noticeable difference between land and objects?

[16:32] Kallista Destiny: Star layouts (10-15 terminators

[16:32] Pauline Darkfury: Mmm, the joys of crawling the length of a building under desks, moving a terminator along to find the break!

[16:32] Ardy Lay: TDR!

[16:32] xstorm Radek: i remember to many packets drop will make a sim seem to other sims to be off line

[16:32] Kallista Destiny: TDR

[16:32] Kallista Destiny: Jinx

[16:32] Lomgren Smalls: It's more than that at times, depending on the lay of the land

[16:32] xstorm Radek: hi ardy

[16:32] Lomgren Smalls: I've literally been completely underground before.

[16:33] Ardy Lay: I used to drag a pulse generator and an oscilloscope around to check Ethernet and Arcnet cables.

[16:33] Andrew Linden: Questar Utu, when talking about avatars sinking into the ground, are you talking about SL everywhere? Or in the Mesh RC?

[16:33] xstorm Radek: lol i remember cable testing

[16:33] xstorm Radek: ardy ya not that old are ya

[16:34] Ardy Lay: Yeah

[16:34] Questar Utu: Everywhere. It's a still existant bug on any viewer. You stand on the sim's land, and your feet sink in by a noticeable amount, but they don't do on prims.

[16:34] Gooden Uggla: andrew - ask him to turn off his AO then test again

[16:34] Vincent Nacon: you sure it's not the shoes issue?

[16:34] xstorm Radek: damn i was thinking ya be young

[16:34] Vincent Nacon: shoes shape*

[16:34] Questar Utu: I don't have shoes on.

[16:34] Ardy Lay: I also used to pulse sweep telephone lines to find bridges and taps on them.

[16:34] xstorm Radek: ouch

[16:34] Andrew Linden: As the comments suggest: there is more than one failure mode for feet under the ground.

[16:34] Kallista Destiny: No it's related to the height of an avatar

[16:34] Pauline Darkfury: There's Falcon's tri-linear sampling thing you can turn on for better terrain collisions (if it's a private sim)

[16:35] Pauline Darkfury: there's also a common failure where the foot shape doesn't apply properly, and swapping to another and back again fixes it

[16:35] Questar Utu: What I'm getting at is that, it's as if land and objects are calculated differently for an avatar, which, isn't suprising if it is, but it's a very noticeable difference.

[16:35] Motor Loon: I tend to sink in the ground too... I just figured I was getting fat

[16:35] xstorm Radek: simon do not stay in dit to long

[16:35] Kallista Destiny: My Primary is 5' tall she sinks in to her ankles

[16:36] Kallista Destiny: Z offset 1.5M

[16:36] xstorm Radek: viewer was never tested for that long

[16:36] Kallista Destiny: yo e fla on th ground

[16:36] Kallista Destiny: to be feet flat

[16:36] Flip Idlemind: Good news for LL: My crashing on Viewer 2 (3?) was because of something I did

[16:36] Kallista Destiny: Oh wat was it Flip?

[16:36] xstorm Radek: simon fell to sleep ?

[16:37] Kallista Destiny: sll this tlak of eathernet...

[16:37] Vincent Nacon: no, he just want to wrap his arm around Andrew

[16:37] Andrew Linden: I wish the viewer would just do foot collision with the ground and adjust the whole avatar up to deal with it.

[16:37] DogWomble Dollinger: now don'

[16:37] Andrew Linden: It should be solved viewer side.

[16:37] Vincent Nacon: n/m just limping now

[16:37] DogWomble Dollinger: don't tell me all the SL servers are hooked up via thinnet :

[16:37] DogWomble Dollinger whispers: :P

[16:38] Lomgren Smalls: I've always thought that the land renderer has more "fidelity" than the collision for the land

[16:38] Gooden Uggla: no, 2 cans and a string...

[16:38] xstorm Radek: Andrew not the primtar again please nnnooo

[16:38] Vincent Nacon: muhaha!

[16:38] Pauline Darkfury: 10b5 with vampire taps ;)

[16:38] Kallista Destiny: Where I worked we pulled our last thinet segement out in 2006

[16:38] Flip Idlemind: Oh I forgot I played around with texture decoding last night. Something I definitely shouldn't be playing with, especially on no sleep

[16:38] Flip Idlemind: Then I was like "Why am I crashing when I log in !? Oh I know..."

[16:38] xstorm Radek: i see it now 2001 all over again

[16:39] Kallista Destiny: there was a printer and a single PC belongng to a Grad student on it

[16:39] IEvangeline: Hay :D

[16:39] xstorm Radek: lol

[16:39] Vincent Nacon: well anywho, Andrew, did you had luck with Falcon about the density?

[16:39] Kallista Destiny: lols

[16:40] xstorm Radek: ok simons arm is starting to twist

[16:40] DogWomble Dollinger: pauline i've repaird racecar cameras using the same technique :P

[16:40] DogWomble Dollinger: duct tape, the glue that holds the world together

[16:40] xstorm Radek: thats not right

[16:40] Sahkolihaa Contepomi: No no, gaffer tape.

[16:40] IEvangeline: Literally, I saw duct tape on the road in a photo

[16:40] Ardy Lay: SOme business networks have about the same life expectancy of a race car.

[16:40] Simon Linden: I'm doing that twisted-select move

[16:40] Andrew Linden: Vincent, no I forgot to talk to Falcon about density.

[16:40] Vincent Nacon: :(

[16:41] Ramona Criss: =))

[16:41] Vincent Nacon: write it down on a post-it note

[16:41] DogWomble Dollinger: whatr three laps?

[16:41] Vincent Nacon: next physic related I like to bring up is SVC-4562

[16:41] JIRA-helper:

[#SVC-4562] Depreciate llCollisionFilter in favor of llFilterCollisions (new functionality)

[16:41] Ardy Lay: Nah, measured in time, not distance.

[16:41] xstorm Radek: i was toldby some one on aditi falcon is the only linden testing on aditi with meshlol

[16:42] Kallista Destiny: My brother told me he saw an M-48 take with the track guard held on with Green duct tape, he said that used two rolls

[16:42] Simon Linden: What's up with 4562 then?

[16:42] Vincent Nacon: false!

[16:42] Vincent Nacon: nothing

[16:42] Vincent Nacon: just limited

[16:42] IEvangeline: Ooh gorgeous blue text

[16:43] Thomas Shikami: Type behaves a bit erratic in llSensor

[16:43] Thomas Shikami: especially the flag SCRIPTED

[16:43] Vincent Nacon: would like to see bigger filter options

[16:43] Vincent Nacon: to make physic easier to handle over lag

[16:43] Vincent Nacon: in term of scripting collision event

[16:43] Kallista Destiny: lots of things behaiv eratttically in sensor

[16:43] Vincent Nacon: I've been working on a Car derby demo system

[16:44] Vincent Nacon: and I need to cull down some events

[16:44] Simon Linden: what are you using for your car physics shapes?

[16:44] Thomas Shikami: does it really improve, if checking before calling the event, instead of using if(llDetectedType(0) & ...) and doing a shortcut return out of the event?

[16:44] Vincent Nacon: at the moment, sculpty. Will switch over to Mesh when it goes mainstream

[16:45] Rodney Monitor: Ooops. Sorry. Logged out here weeks ago.

[16:45] Pauline Darkfury: There's a texture for your ghost detector, Simon ;)

[16:45] Thomas Shikami: take that back, I just remember that collision events are cumulated, giving multiple collisions per events using the index you use with llDetected*(integer ) type of functions

[16:45] xstorm Radek: ? Collision

[16:45] Vincent Nacon: the point with filter because I need to make collision event handle all the info coming in by adding more filter by name and type

[16:45] xstorm Radek: whymess with a name

[16:45] Simon Linden: thanks

[16:45] IEvangeline: I was on the mesh server like yesterday... is it just my computer or are joints between meshes handled really poorly right now?

[16:46] Vincent Nacon: like another car, ground and walls. Not some random tiny bullets that someone trying to cheat

[16:46] Kallista Destiny: Meshes in general make poor Physics objects

[16:46] Vincent Nacon: that's your opinion

[16:46] Kallista Destiny: you need a collision shape

[16:46] Vincent Nacon: I can handle mesh's physic by making my own mesh for it

[16:47] Thomas Shikami: convex hull is appropiate for a lot of meshes anyways

[16:47] Simon Linden: Vincent ... when you switch over to mesh, also provide a really simple shape for your vehicles. That's one of the best things you can do to help the performance

[16:47] xstorm Radek: and as for Collision many time it was turn off ina simif the sim started to lag to bad

[16:47] Vincent Nacon: although I think you meant to say the mesh's upload physic generator is generally poor ;)

[16:47] Kallista Destiny: Yeah... that was what I ment

[16:48] xstorm Radek: think of sion chickens

[16:48] Vincent Nacon: yup, but right now I need more filter over what objects is allowed to trigger collision event

[16:48] DrFran Babcock: DrFran Babcock

[16:48] IEvangeline: It might just be my computer, right now I can't see anything fancy, anything above very basic windlight and shiny is beyond me

[16:48] Thomas Shikami: and sometimes, it's even better to have no collision shape at all instead of making hollow objects. Like when making a mesh like a bridge, where the posts would break the convex property of the mesh

[16:48] Sahkolihaa Contepomi: There's quite a lot of problems with enabling basic shaders.

[16:48] Sahkolihaa Contepomi: Mostly on NVIDIA 400 hardware.

[16:48] IEvangeline: I'm on a... get this... Intel G33

[16:48] xstorm Radek: if Collision events get to high it will crash a sim

[16:48] Sahkolihaa Contepomi: Well that explains everything. :p

[16:49] Vincent Nacon: llCollisionFilter is only assigned to one thing or one ID key. not enough

[16:49] Pauline Darkfury: So, before we run out of time, any progress on encroachment-return or PE-bloat-return, since Tuesday?

[16:50] Vincent Nacon: oh yeah, that too

[16:50] Motor Loon: Since we got 10 mins to go, SCR-37 any chance...?

[16:50] JIRA-helper:

[#SCR-37] llGiveMoney() method needs a money_given() event handler.

[16:50] xstorm Radek: lol

[16:50] Simon Linden: I've been out of the loop on that, Pauline

[16:50] Draconis Neurocam: you cant just loop through the colision event, and have an if else tree filter every one individually

[16:50] Draconis Neurocam: ?

[16:50] Andrew Linden: Pauline, I am convinced that encroachment can be enabled by default, and I've got one of the stakeholders also convinced.

[16:50] Simon Linden: Motor - yeah, that would be nice to have

[16:50] Rex Cronon: money_given(int amount) :)

[16:51] Motor Loon: Little more than "nice"

[16:51] Andrew Linden: I've got a few more people to round up and convince.

[16:51] Pauline Darkfury: Mmm, yeah, SCR-37 seems to be getting popularly exploited at present

[16:51] Thomas Shikami: money_given(key target, int amount)

[16:51] Vincent Nacon: Motor, there's no info that tells money is taken

[16:51] Motor Loon: read my last comment on the hira

[16:51] Motor Loon: jira

[16:51] Rex Cronon: with key even better:)

[16:51] Vincent Nacon: but how would that work?

[16:51] Pauline Darkfury: There also are some reports of people managing to raise L$0 money events (which should be doomed to failure, if the scripter has done their stuff properly)

[16:52] Thomas Shikami: that explains the cheating money out of machine stuff

[16:52] Pauline Darkfury: key dataserver_key = llGiveMoneyVerified(); is my pref, since I believe that it's problematic to add new event types

[16:53] Thomas Shikami: second that

[16:53] Motor Loon: anything is better than ... nothing

[16:53] Kallista Destiny: Remember that there can only be 32 events

[16:53] Flip Idlemind: 64?

[16:53] Thomas Shikami: and the money given event can be lost

[16:53] Pauline Darkfury: it's the sort of thing that the dataserver should be just fine for, I think

[16:54] Andrew Linden: Really? I thought it was 64 events, but am not sure.

[16:54] Thomas Shikami: so, maybe it's even better to make a llGiveMoneyTransacted()

[16:54] Thomas Shikami: and a llCheckTransaction(key id)

[16:54] Sahkolihaa Contepomi: Kelly said it's 64.

[16:54] Pauline Darkfury: (assuming it's not possible to give a real time boolean return)

[16:54] Flip Idlemind: Im pretty sure there are already at least 33 events

[16:54] Kallista Destiny: thAh OK 64 I n=nw that just over hqlf had been usd

[16:54] Pauline Darkfury: it's some sort of fairly low number that makes it a good idea to avoid new single purpose events if something like dataserver can be used, or a flag can be set to tune an existing event

[16:55] Thomas Shikami: id can be an encrypted transaction id that is being generated for transferring money

[16:55] Motor Loon: real time boolean return would sure make it easy to implement if any existing products

[16:55] Thomas Shikami: and the script could check on the transaction, if it's pending, cancelled or done

[16:55] Liisa Runo: 36 events currently

[16:55] Meeter: Timecheck : User Group is almost over

[16:55] Vincent Nacon: Simon's left arm is out of control

[16:55] Thomas Shikami: the marketplace team has done a neat job on checking, if objects have been transferred. Can we scripters have that, too?

[16:55] Vincent Nacon: put your arm down!

[16:55] Vincent Nacon: :P

[16:55] DrFran Babcock: Noooooo!

[16:55] Acheron Gloom: The limit IS 64.

[16:56] Acheron Gloom: since its stored using a 64-bit integer

[16:56] Acheron Gloom: and checked using bitwise AND

[16:56] Pauline Darkfury: Yup, llGiveMoneyVerified is the really important one, but llGiveInventoryVerified is also needed for reliable commerce

[16:56] Thomas Shikami: key id = llGiveInventoryTransacted(key target, name inventory)

[16:56] Thomas Shikami: integer result = llCheckTransaction(key id)

[16:56] Thomas Shikami: so, we could do two things in one

[16:57] Vincent Nacon: Lovecraft Linden only came across my mind

[16:57] Thomas Shikami: giving inventory in a verified way and giving money in a verified way

[16:57] Pauline Darkfury: yeah, don't care about the specific names or that detail, just as long as there's some way to determine success/fail

[16:57] Flip Idlemind: Some places hand you a "rules" notecard when you arrive, and I usually decline it (Once I already have it.) I don't want them to be able to be like "NO! YOU MUST TAKE OUR RULES!"

[16:57] Simon Linden: Yeah ... there is room to add more events, but not an endless supply

[16:57] Thomas Shikami: llCheckTransaction for both

[16:57] Vincent Nacon: but wait... viewer can quietly accept inventory

[16:58] Andrew Linden: You all have been talking to Kelly about adding these events at the Scripting User Group, right?

[16:58] Kaluura Boa: And you can mute the object so it stops bothering you...

[16:58] Acheron Gloom: I always miss the scripting user group, I've wanted to birng up the same 3 things for two weeks :(

[16:58] Pauline Darkfury: just need to know that the inventory offer hasn't been completely lost or muted

[16:58] Thomas Shikami: where can I suggest the transacted functions at?

[16:58] Pauline Darkfury: Yes, these are not new issues, about 3 or 4 years old

[16:58] Motor Loon: I'll certainly bring it up there if not Andrew

[16:58] Acheron Gloom: Oh, adn yes, the issues are old issues.

[16:58] Acheron Gloom: I think old issues tend to be forgotten some how... not sure.

[16:59] Pauline Darkfury: SCR-37 is about 2007 vintage, I think

[16:59] Andrew Linden: I don't have a good grasp of how hard it would be to add money and inventory transaction events.

[16:59] Motor Loon: yep

[16:59] Rex Cronon: we might have better chances getting watter from rocks than having a new money event:(

[16:59] Thomas Shikami: Not events

[16:59] Andrew Linden: Definitely possible, but I just don't know how hard it would be. I figure Kelly might have a better idea.

[16:59] Vincent Nacon: money more likely than inventory

[16:59] Thomas Shikami: more something like checking transaction status

[16:59] Pauline Darkfury: SCR-37 cost merchants literally millions of L$ in the Apez implosion

[16:59] xstorm Radek: inventory messing yikes

[16:59] Vincent Nacon: unless you give user an option to accept money or not

[16:59] Vincent Nacon: then it would be a problem

[16:59] Motor Loon: still does Pauline

[16:59] Pauline Darkfury: recent reports suggest that there's 100s of thousands being exploited due to it on an ongoing basis

[17:00] Motor Loon: just ONE guy costed me 150K in ONE month

[17:00] Meeter: Thank you for coming to the Server User Group

[17:00] Vincent Nacon: thank you Meeter, you can go without us

[17:02] Andrew Linden: Ok, so we need to be able to verify money and inv transactions.

[17:02] Motor Loon: holy crap its like a photo session

[17:02] Lomgren Smalls: Heh.

[17:02] Andrew Linden: I'll try to ask around about it.

[17:02] Vincent Nacon: forget the inventory for later

[17:02] Lomgren Smalls: I wonder if people are just now looking at the table...

[17:02] Motor Loon: yes pls Andrew

[17:02] Motor Loon: my man!

[17:02] Vincent Nacon: inventory would be problematic

[17:02] Andrew Linden: But I've got to go now. I've got somewhere I've got to be in an hour.

[17:02] Rex Cronon: tc andrew

[17:02] Sahkolihaa Contepomi: See you, Andrew.

[17:02] Acheron Gloom: SEe ya Andrew

[17:02] Motor Loon: tc all

[17:02] Kallista Destiny: Be well Andrew

[17:02] Andrew Linden: Thanks for coming everyone.

[17:02] Ramona Criss: *.. Bye Bye ..*`

[17:03] Rex Cronon: tc

[17:03] Draconis Neurocam: take care andrew and simon

[17:03] Vincent Nacon: yup it take him an hour to change his outfit or something

[17:03] Liisa Runo: thanks everybody

[17:03] Kallista Destiny: Thank you all for a good meeting.

[17:03] ANSI Soderstrom: tc all

[17:03] Simon Linden: Thanks everyone for coming

[17:03] Pauline Darkfury: Thanks, Andrew & Simon, take care

[17:03] Vincent Nacon: later Lindens!

[17:03] Sahkolihaa Contepomi: Yes, see you too, Simon.

[17:03] xstorm Radek: every one

[17:03] xstorm Radek: **ok i wuv u byby!!**

[17:03] Thomas Shikami: Thank you!

[17:03] Rex Cronon: tc all those leaving

[17:03] Thomas Shikami: what would it be, if you aren't talking with your residents?


Prev 2011.08.09 Next 2011.08.16