Simulator User Group/Transcripts/2013.08.20

From Second Life Wiki
Jump to navigation Jump to search


Prev 2013.08.13 Next 2013.08.27

List of Speakers

Adamburp Adamczyk Andrew Linden ANSI Soderstrom
Baker Linden darkenedshadow Duckie Dickins
Hexapoda Kallista Destiny Kelly Linden
Kitto Flora Lucia Nightfire MartinRJ Fayray
Minuet Voir Mona Eberhardt Nalates Urriah
Petr Topaz Qie Niangao Rex Cronon
Simon Linden Whirly Fizzle Yuzuru Jewell


[12:00] ANSI Soderstrom: hi other employees

[12:00] Kallista Destiny puts her antibling sun glasses on

[12:00] Simon Linden: I can get started with the server news ... the main event being that the Server-Side Appearance code was rolled onto the main release channel this morning

[12:00] Rex Cronon: hello andrew

[12:01] Simon Linden: So that's now on about 80% of regions

[12:01] Rex Cronon: hello kelly

[12:01] Simon Linden: with tomorrow's updates it will be on all regions

[12:01] Kallista Destiny: Ave OZ

[12:01] Mona Eberhardt: Hi Oz!

[12:01] Hexapoda: It has work great for me today.

[12:01] Kallista Destiny: I presume that everything is going well on the ovens?

[12:02] Simon Linden: I haven't heard any smoke alarms going off

[12:02] Hexapoda: Hello Kito.

[12:02] Kitto Flora: Hello

[12:02] Simon Linden tries to think of a joke about being half-baked

[12:02] Rex Cronon: if u r late,can't stay, u r crashing... than u can follow group meetting live here:

[12:02] Mona Eberhardt: I haven't had much of a chance to assess SSA today, as RL has been ridiculously busy, but it seems to work well.

[12:02] ANSI Soderstrom: i dont like SSA, look below my tummy

[12:03] Simon Linden: Andrew, Kelly or Oz ... any news or things to bring up?

[12:03] Andrew Linden: I don't have any news today, except that I'm trying to track down a bug in an up and coming simulator release that makes parcel ban lines not show up.

[12:03] Hexapoda: Ansi, you look okay for me.

[12:04] Rex Cronon: is it just me or is the floor transparent?

[12:04] Kallista Destiny: Just you rex

[12:04] Kelly Linden: I do not have any news today.

[12:04] Simon Linden: oh, the RC updates tomorrow .. those will be the same 2 maintenance releases that are out now

[12:04] Simon Linden: ok, so the virtual table is open for questions or topics ....

[12:04] Hexapoda: Ctr-shift-r twice

[12:04] Mona Eberhardt: Hello Baker!

[12:05] Rex Cronon: thanks hexapoda:)

[12:05] Conover's Flight-Helper 6.3.3 (WEAR ME!): Flight-helper is ready and operational.

[12:05] darkenedshadow: Oculus viewer update please Simon if you can

[12:06] Simon Linden: I don't have anything new about the Occulus other than work continues. I know recently they were working on how to have the SL UI appear ... having menus hanging out in your vision is an interesting design

[12:06] Mona Eberhardt: Speaking of Oculus, I think Oz would like to look at the Myo armband, which seems to be fixing to give Leap Motion a run for its money.

[12:07] Simon Linden: but you're not in a "window" anymore

[12:07] darkenedshadow: Yay!!!

[12:07] Mona Eberhardt: And it could possibly work well with the Rift.

[12:07] Hexapoda: Just learn the Rifters to use hot keys.

[12:07] Simon Linden: Have you tried one of those, Mona? I played with the Leapmotion and it was kinda fun

[12:08] Mona Eberhardt: Haven't had the opportunity, but I'm not sure how easy it'd be to keep your arms stretched out for long periods of time (which isn't required by the Myo).

[12:09] Simon Linden: yes, that was an issue with the Leapmotion ... I think it'll take some work and experimentation to figure out how it might work best. Either one would be interesting along with the Rift

[12:10] Mona Eberhardt: Here's the website for Myo:

[12:10] Kitto Flora would just settle for getting the old 3-D display back.

[12:11] Mona Eberhardt: The Rift needs a system like that, or the Leap Motion.

[12:11] Mona Eberhardt: Although I must say not all of us need the 1st-person view of the Rift.

[12:11] Andrew Linden: which 3D display are you talking about Kitto? A stereoscopic view?

[12:12] Kitto Flora: Dale Glass's Stereo version of the client ... made in....

[12:12] MartinRJ Fayray: Yes, some third party viewer's had a stereoscopic view (kirsten's viewer for example)

[12:12] MartinRJ Fayray: I never got to try it, sadly.

[12:13] Kitto Flora: Humm - cant see a date but it was based on - so several years back

[12:13] Whirly Fizzle: CtelAltStidio viewer has been experimenting with stereoscopic

[12:13] Kitto Flora: Stereoscopic Patch: Support for 3D with anaglyph or shutter glasses.

[12:13] Minuet Voir: lol had the link here too

[12:14] Whirly Fizzle: PLys it has an awesome purple skin :D

[12:14] Kitto Flora: I used the anaglyph - worked as well as that does

[12:14] Minuet Voir: yes! lol I use that skin

[12:14] Whirly Fizzle: Me too haha

[12:14] Kitto Flora: Vid cars supporting glasses are more common now, maybe the glasses are a little cheaper too

[12:15] ANSI BagPipe (wear): Hello, Avatar!

[12:15] ANSI Soderstrom: Hello ANSI BagPipe (wear) with the UUID of 9e807942-b0b6-b455-a873-c4b955efa697

[12:15] Andrew Linden: ah yes, that CtelAltStidio trick used the quick-n-dirty nVidia stereoscopic wrapper

[12:15] Kitto Flora:

[12:15] Andrew Linden: I think I remember hearing about that, but never tried it out.

[12:15] Kitto Flora: Dunno how well a 1.18 client would work with latest SL

[12:15] Whirly Fizzle: I was on Dales viewer the other day actually to check something else (particle blocking) & no textures wouyld rez for me at all. Seemed to be pre windlight too.

[12:16] Simon Linden: well, you're certainly not gonna bake right :P

[12:16] Andrew Linden: Does nVidia still offer that stereoscopic feature on its video cards?

[12:16] Kitto Flora: Andrew - yes on the one I have

[12:16] Nalates Urriah: Andrew, yes. They even update it with each driver if you have installed it.

[12:17] Nalates Urriah: it = 3D Vision

[12:17] Kitto Flora: Well 'that'... ? Doe they have more than 1 system?

[12:18] Kitto Flora: I have a "3D Vision Photo Viewer", shows 3d photos with options of Ana and Glasses... The Ana works

[12:18] NightDreamer100: bell! 8O

[12:18] Andrew Linden: Well, perhaps Oculus will deliver us into the promise of 3D immersive virtual reality.

[12:18] Kitto Flora: Yeah but you loose the KB?

[12:19] Nalates Urriah: nvidia anaglyph:

[12:19] Rex Cronon: it will be so much more fun when somebody gets griefed by male body parts:)

[12:20] Rex Cronon: u will really have it in your face:)

[12:20] Andrew Linden: what happens to the viewer UI in the stereoscopic view?

[12:20] darkenedshadow: .... yeah not looking forward to that one :D

[12:20] Simon Linden: In the Rift it's projected on a surface around you ... so you look up to see the menus

[12:21] Simon Linden: and they float there in mid-air

[12:21] darkenedshadow: Ah... sounds much like the UI for Minecrift :D

[12:21] Nalates Urriah: Andrew, in anaglyph and shutter the UI is stuck on the surface of the screen, which is like a window you look through. The world is like outside the window.

[12:22] Andrew Linden: ok, so the stereoscopic vision puts the UI at some fixed distance from the eyes.... a meter or less probably

[12:22] Mona Eberhardt: Guys, when you can, I have a question regarding limiting the server RAM usage of a script - there's an LSL function for that, but I'd like to know how one can determine how much they can limit a script's RAM usage.

[12:23] Kitto Flora: 3D wireless shutter glasses, $129 from NewEgg. Of course your vid card and monitor must be 'compliant'

[12:23] Simon Linden: I think they're experimenting with the shape of that surface too ... if it's flat, the text can look funny as it's slanting away from you

[12:23] Andrew Linden: Mona, there is an upper limit of memory for each script, I think. Kelly knows the details I believe.

[12:24] Andrew Linden: You want to know what the absolute limit is? Or you want to be able to specify a lower limit?

[12:24] Mona Eberhardt: Both, Andrew.

[12:25] Andrew Linden: Kelly, perhaps you could enlighten us?

[12:25] Andrew Linden: My memory says that the old LSL scripts all had a fixed size no matter how simple or complext, but the MONO scripts take up whatever size they need... up to some maximum.

[12:26] Kelly Linden: Currently mono scripts can set their memory limit as low as the memory being used at the time the call is made - if that limit is hit the script will crash.

[12:26] Kelly Linden: Scripts not compiled to mono are 16k always

[12:26] Kelly Linden: Mono scripts start at 64k and can be set lower to that but not higher.

[12:27] MartinRJ Fayray: Is the 64k limit going to be changed in a near-ish future?

[12:27] Kitto Flora: If Mono script exceeds the 64 k limit you usually get a run-time 'stack-heap collision' error message and the script stops.

[12:27] MartinRJ Fayray: It's no fun to see the script-crash message everytime I try to create something :/

[12:28] Nalates Urriah:

[12:28] Kelly Linden: As far as a theoretical 'absolute' limit on the internal systems for scripts I'm not aware of one that is meaningful. Probably something around 4gigs. However at this time we don't have any concrete plans to raise the 64k limit.

[12:28] Kitto Flora: Divde your script into 2 parts that comm as necessary with Link Messages

[12:29] MartinRJ Fayray: I spend more time trying to split up my scripts and reorganise them, than I actually spend on writing code.

[12:29] Mona Eberhardt: So, can I set a 16K limit for a mono script?

[12:29] Kelly Linden: you can

[12:29] Kitto Flora: Martin: Welcome to the club :)

[12:29] Mona Eberhardt: Could that be set even lower? Maybe 4K? Or 2K?

[12:29] Andrew Linden:

[12:30] Kelly Linden: I think 6k was the practical limit I found, and even then the script had to be really simple

[12:30] Meeter: Timecheck : User Group is half over

[12:30] Mona Eberhardt: Thanks.

[12:30] Qie Niangao: In my experience, it's pretty difficult to get much below 4K.

[12:30] Kelly Linden: the size of the bytecode itself is included in that size, and some other script overhead.

[12:30] Mona Eberhardt: By overhead, you mean...?

[12:31] Rex Cronon: vars? constants?

[12:31] Kelly Linden: data surrounding every scripts required for that script to run which is not accessible by the script directly

[12:31] MartinRJ Fayray: 'Commissioners' (for scripting-jobs) in Second Life tend to have even less knowlegde about how scripts work than customers in a real-world software-company- environment.

[12:31] MartinRJ Fayray: It's quite common that they want to 'add' features later...

[12:31] MartinRJ Fayray: And then your scripts break, and you have angry commissioners and alot of time to spend to re-organise your LSL

[12:32] MartinRJ Fayray: I bet almost every scripter here had similar issues

[12:32] Rex Cronon: u need a contract tha says once the requirments have been written down any changes will cost extra and take longer:)

[12:33] Simon Linden: I'd bet any programmer has had the same thing too :)

[12:34] MartinRJ Fayray: Well the sl viewer for example plays in another league - 64k compared to a 2 gigabyte memory limit

[12:34] Qie Niangao: Tis the nature of the beast, although LSL is more constrained than most environments, so harder for "the commissioner" to understand why so expensive to extend.

[12:34] Rex Cronon: i agree that it is a pain indeed having to break down your scripts when your script crashes due to not enough memory:(

[12:35] Lucia Nightfire: Is anyone working on a deterrence to the asset pending dl/ul blocking vector thats going around in certain forms recently?

[12:35] Rex Cronon: btw. how is the fix for the exploit that allows passwords to be stolen with media on a prim and/or streams:)

[12:36] Andrew Linden: Someone was asking me earlier today if it was possible to detect attached lights by an LSL script on a regular (non-attached) object.

[12:36] Simon Linden: I'm not sure which attack you mean, Lucy

[12:36] Rex Cronon: simon. if u go to swt now u will see what lucy is talking about:)

[12:37] Andrew Linden: swt = ?

[12:37] Whirly Fizzle: Its still broken there?

[12:37] Rex Cronon: sandbox weapons testing

[12:38] Lucia Nightfire: its a vector that blocks assets from entering pedning ul/dl state

[12:38] Rex Cronon: i think somebody is selling something that does this to the sim. u can't rez anything, and attachments don't show. but u can built and drop-in scripts

[12:38] Duckie Dickins: that should really be renamed to sandbox grief testing

[12:38] MartinRJ Fayray: Hehe

[12:38] Rex Cronon: aka LagLand:)

[12:39] Andrew Linden: hrm. I hadn't heard of that exploit yet.

[12:39] Simon Linden: There have been some crashes using the age-old techniques of rezzing too much stuff or scripts, so that probably needs some more attention

[12:40] Rex Cronon: here is a link of the exploit regarding media on a prim:

[12:40] Kitto Flora: Andrew : Light sources - afaik not readable by an external script. Either set Midnight and wave a white-painted sphere around to see where light comes from, or if your client has the feature - Tools/ ShowLightRadiusForSelection, and click on stuff.

[12:40] Qie Niangao: Yeah, that media / password thing seems very scary.

[12:41] Andrew Linden: Ok, so maybe the llSensor() should be able to look for lights.

[12:41] Adamburp Adamczyk: how that posisble? password via media?

[12:41] Qie Niangao: Adam: not public knowledge. it's an exploit, so... buffer overrun in QT? in WebKit? who knows?

[12:42] Kitto Flora: Andrew: Are you suggesting an extention to llSensor() for finding light sources?

[12:42] Adamburp Adamczyk: yeah hard to beleive, i mean a lot of people won;t use media cos of the ip explouit, bu t never ehard that one

[12:42] Rex Cronon: its new adamburp

[12:42] Andrew Linden: Yeah. I'm sure there are other ways that llSensor() should be enhanced.

[12:43] Andrew Linden: As I recall Kitto you were asking for a filter for objects by owner.

[12:43] Qie Niangao: Lights: frankly, I'd rather be able to *set* all the properties by script, finally, more than be able to *sense* them.

[12:43] Kitto Flora: Andrew - maybe but fixing region border crossing is 100 to 1000 times more impartant. Theres 2 work-arounds for pesky lights. Theres no work around and frequent crashes for region crossing.

[12:44] Mona Eberhardt: What's this exploit about? XSS in media in SL?

[12:44] Lucia Nightfire: I would like to see a sensor pos input as well as a rot

[12:44] Lucia Nightfire: not make teh function bound to the object it is in and the object's pos

[12:44] Rex Cronon: the only info about that mona seems to be the blog post

[12:44] Nalates Urriah: Is there any nwes about being able to script materials changes?

[12:44] Lucia Nightfire: also, wth is up with the lag

[12:45] Lucia Nightfire: and rot

[12:45] Kitto Flora: Andrew = llSensor that would use regex in string fields would be useful, would cut down on sim load. BUT since it was no on horizon part of recent track-follower re-design designed out the need for regex.

[12:46] Mona Eberhardt: *sigh* I'd thought such exploits should have been killed long ago... Rather than allowing scamware/malware makers and black hats to capitalise on them.

[12:46] Kelly Linden: regex support of any kind and 'reducing load' rarely go in the same sentence. ;)

[12:46] Adamburp Adamczyk: I read that thread, adn still find it hard to beleive.

[12:47] Adamburp Adamczyk: but like owt else, have to wait and see

[12:47] Qie Niangao: Adam, if it didn't come from Cristiano, I'd doubt it too.

[12:47] Kitto Flora: Kelly: regex in script engine would surely be more efficient than the present regex written in LSL?

[12:47] Kelly Linden: it is harder to time slice appropriately

[12:47] Kitto Flora: (But maybe not if you write it in C#)

[12:50] Qie Niangao: Lucy: it *is* very handy that llCastRay can be cast to/from anywhere in the sim, not limited to the scripted object's location. So that might be nice for sensors, too.

[12:51] Lucia Nightfire: castray is limited

[12:51] Lucia Nightfire: its a wildcard

[12:51] Qie Niangao: heh, no kidding

[12:51] Rex Cronon: it like poking a stak of hay with a needle trying to find another needle:)

[12:52] Qie Niangao: it's good at finding haystacks, though. :p

[12:52] Adamburp Adamczyk wishes Baker a very Happy Birthday

[12:52] Kitto Flora: Anyone noticed error in castray return when a small prim close surface is just above a 'megaprim' top surface? It sees the lower megaprim rather than the closer small prim?

[12:54] Andrew Linden: how small is "small"? and how close do the two faces have to be for the cast to fail against the small prim?

[12:55] Meeter: Timecheck : User Group is almost over

[12:55] Kitto Flora: Small is 5,2,0.01 meters, and close in 0.02M surface difference

[12:56] Simon Linden: that's getting down near the collision tolerances, I think

[12:56] Andrew Linden: I see, so the small prim is more thin than small.

[12:56] Andrew Linden: I'll ask Falcon if he knows why that might fail.

[12:57] Kitto Flora: The red cone represets the raycast, firing down, It will see the grey floor here before the track

[12:58] ANSI Soderstrom: then raycast from the other side ;)

[12:58] Kitto Flora: How do I know when to raycast from other side?

[12:58] Andrew Linden: Interesting.

[12:58] Nalates Urriah: ..or even where the other side is?

[12:58] Andrew Linden: Kitto, what happens when you make the track thicker?

[12:59] Andrew Linden: It doens't actually have to be so thin, does it?

[12:59] Kitto Flora: I Think maybe nothig - I think this has failed on thick track too

[12:59] Andrew Linden: ok

[12:59] Kitto Flora: The problem is when the track top surface uis only just above the large floor prim beneath

[13:00] Meeter: Thank you for coming to the Server User Group

[13:00] Andrew Linden: right, I wonder if there is a object scale discrepancy between what is seen and the physics collision shape that si causing this

[13:01] Lucia Nightfire: also, with castray, theres still the issue with physics prims turnign non-physical and still being treated as physical with castray

[13:01] Andrew Linden: I think we've got a tool that can view the collision shapes, so we might be able to use that.

[13:02] Lucia Nightfire: and for some damn reason I can't find that jira

[13:02] Simon Linden: I know we used to trim the physics shape a bit by the collision tolerance, which allowed stacked cubes to look right. But I think we changed some of that with a much lower collision tolerance

[13:02] Andrew Linden: Lucia, you mean the transition is done via something like llSetStatus()?

[13:02] Kelly Linden: bye everyone. o/

[13:02] Mona Eberhardt: bye Kelly

[13:02] Rex Cronon: tc kelly

[13:02] Kallista Destiny: Bye all, and thanks

[13:02] Kitto Flora: Well - this was a problem at one Owner's build - he put a lot of track barly above his megaprim floor - the track updater could not find it using castray. When I move the track up a little it worked correctly.

[13:02] Petr Topaz: gye kelly

[13:02] Simon Linden: Thanks everyone for coming today and the good conversation

[13:02] Lucia Nightfire: no, if you are filtering for physics with castray, you can return a hit from a non-phys prim that was once phys

[13:03] Andrew Linden: Yes Simon, however I'd expect that collision shape shaving to actually help the thin box to be seen by the castray.

[13:03] Mona Eberhardt: Bye Simon, thanks for having us.

[13:03] Kitto Flora: BUT no other user has yet reported same problem

[13:03] Andrew Linden: That is, if the object is thin enough we don't shave it, but we do shave the big objects.

[13:03] Rex Cronon: tc all those leaving

[13:03] Whirly Fizzle: Lucia?

[13:03] Andrew Linden: So the floor should actually collide lower than what is visible, while the thin prim would collide higher than visible.

[13:04] Lucia Nightfire: yes, Whirls, thanks

[13:04] ANSI Soderstrom: thanks all for watching && thanks for Baker for the permisson to decorate a (busy) Linden

[13:04] Baker Linden: haha, no problem! I enjoyed the food pellets.

[13:04] Andrew Linden: Kitto, does the cast ray still fail when you pull the caster back a bit?

[13:05] Andrew Linden whispers: That is, does it depend on the proximity of the device initiating the castray?

[13:05] Andrew Linden: That is, does it depend on the proximity of the device initiating the castray?

[13:05] Kitto Flora: Andrew - that I dont know - would need some mods to the track-layer object's scripts so I could vary its height

[13:06] Andrew Linden: I'm just wondering if your cast ray start point is not quite where you think it is.

[13:07] Andrew Linden: Ok, I've got to go. Thanks for coming.

[13:08] Rex Cronon: tc andrew

[13:08] Yuzuru Jewell: Thank you, Andrew.

[13:08] ANSI Soderstrom: next week, ill decorate andrew Linden

[13:08] Petr Topaz: bye

[13:08] Kitto Flora: Andrew Ray start is the center of this object, who's wheels' rest' on the track top surface(by script llSetPos()0, the end is 2M lower

[13:08] Yuzuru Jewell: Thank you, Simon.

[13:08] Petr Topaz: this is not good

[13:08] Mona Eberhardt: Thanks Simon. :)


Prev 2013.08.13 Next 2013.08.27