Simulator User Group/Transcripts/2012.07.24

From Second Life Wiki
Jump to: navigation, search


Prev 2012.07.20 Next 2012.07.27

List of Speakers

Acheron Gloom Baker Linden Davido Chrome
Draconis Neurocam Fadeout Razorfen Kallista Destiny
Kelly Linden Motor Loon NeoBokrug Elytis
Qie Niangao Rex Cronon Simon Linden
Tiberious Neruda Visitor Detector


[12:01] Simon Linden: well, let's get rolling ... Andrew is taking the day off

[12:01] Baker Linden: I guess go ahead and start. I don't have much news; yesterday I was helping some QA testers confirm some of my bug fixes were working properly. The rest of this week I'll be re-learning python so I can finally get large groups showing up (for SVC-4968). Hopefully I'll get it done in time for our current bug fixing sprint (group-related issues).

[12:01] JIRA-helper:

[#SVC-4968] Group won't load - too many members

[12:02] Motor Loon: excellent baker

[12:02] Motor Loon: did they work then?

[12:02] Kallista Destiny: Most wonderful Baker

[12:02] Simon Linden: For release news ... we rolled the bug fix code that was on Bluesteel this morning to the main grid

[12:02] Simon Linden: Info in here:

[12:02] Motor Loon: does that solve the "loosing animations" bug?

[12:03] Motor Loon: or rather "sit animation started on simcrossings"

[12:03] Kelly Linden: yes

[12:03] Simon Linden: It should have, yes. The other 2 RCs will need to get updated tomorrow

[12:03] Motor Loon: sweet

[12:03] Simon Linden: ... so unless you're crossing into one of those, it should be OK now

[12:03] Motor Loon: oh so it may still happen crossing to those remaining RCs ?

[12:03] Simon Linden: or out of, I forget if the bug was directional or not

[12:04] Motor Loon: ok, guess I'll keep my workaround in my sailboat a bit longerthen

[12:04] Motor Loon: it wasnt... happened both on entering and leaving

[12:04] Davido Chrome: You were able to mute yourself? What did that do? XD

[12:04] Simon Linden: Yes Motor - after tomorrow morning the bug should be gone

[12:05] Motor Loon: Excellent

[12:05] Visitor Detector: Returning visitor to your land: Rex Cronon

[12:05] Simon Linden: Tomorrow's RCs will turn on the experience tools functions on 2 channels

[12:05] Rex Cronon: hello everybody

[12:06] Motor Loon: Now we just need to get PATHBUG-113 over with too °?°

[12:06] Simon Linden: That's about it for my news ...

[12:07] Baker Linden: davido, nothing good.

[12:07] Simon Linden: So the floor (or roof) is open for topics

[12:07] Tiberious Neruda: okay...

[12:07] Tiberious Neruda: I'd like an explanation for some changed physics I've noticed on some regions

[12:08] Simon Linden: Are they pathfinding regions by any chance?

[12:08] Tiberious Neruda: yes they are

[12:08] Fadeout Razorfen: what are they

[12:08] Motor Loon: SVC-8048

[12:08] JIRA-helper:

[#SVC-8048] vehicles being very iratic and jittery on magnum regions

[12:08] Tiberious Neruda: I use an object on this avatar that intercepts the 'forward' key and doesn't pass it on, instead using an llSetForce on my avatar to propel it at speeds faster than it can normally walk or run

[12:08] Simon Linden: I'm pretty sure those have a new version of the physics engine -- there's at least one bug open about them.

[12:09] Simon Linden: The one I'm thinking of is that some pushes were not behaving the same as before

[12:09] Tiberious Neruda: however, I'm noticing it's very hit-or-miss when it comes to actually moving

[12:09] Simon Linden: Have you filed a jira about it?

[12:09] Tiberious Neruda: no

[12:09] Tiberious Neruda: I don't use the jira anymore...

[12:10] Motor Loon: it will depend alot on the objects collision shape Tib

[12:10] Simon Linden: Well, if you want someone from Linden QA to look at it, that's the way to go

[12:10] Tiberious Neruda: Motor: the object being moved is my avatar

[12:10] Motor Loon: ah

[12:10] Simon Linden: There's one open now that sounds similar, as I mentioned, but more examples would probably help

[12:11] Motor Loon: yeah if avatars generally have more "drag" on the updated havok, it could be a problem

[12:11] Simon Linden: SVC-8076

[12:11] JIRA-helper:

[#SVC-8076] llApplyImpulse and llPushObject does not work properly on Magnum servers

[12:11] Tiberious Neruda: mine uses llSetForce

[12:11] Motor Loon: oh cool, didn't know that jira

[12:12] Fadeout Razorfen: i use exact same thing in my project to adjust running speed, and for some reason avatars start randomly fly up and down. (i'm 100% sure math is correct, and the force vector is always pointing a little bit down to prevent that)

[12:12] Tiberious Neruda: you too?

[12:12] Fadeout Razorfen: but still it starts randomly to jump up/down on some avatars

[12:12] Simon Linden: Right, that's one of the ideas about that one now. Either the friction itself changed, or the way the two surfaces were in contact may have changed, with the same results

[12:13] Tiberious Neruda: there's times when I can tap the 'backward' key, and it kicks in, but sometimes it doesn't

[12:13] Rex Cronon: maybe when u push the ave it hits the ground at an angle and that is what makes the ave jump up?

[12:13] Simon Linden: Do they fly up near the edge of a roadway or some other 2 objects? I know we've had issues with collisions near edges before

[12:14] Tiberious Neruda: oddly enough, mine also has a slight downward angle...

[12:14] Tiberious Neruda: was intended to keep me from flying off the slightest downhill sections

[12:15] Davido Chrome: Maybe it's the same as when you sink into a prim because your speed is too great, you get flung up?

[12:15] Davido Chrome: So it's the downward tilt that's the problem, possibly?

[12:15] Tiberious Neruda: I kinda doubt it

[12:15] Fadeout Razorfen: it looks more like cosine function flying

[12:15] Qie Niangao: hmmm. wonder how swimmer HUDs (with buoyancy) are affected. Would kind of tease-out the friction/collision thing.

[12:16] Fadeout Razorfen: not bumping

[12:16] Motor Loon: I havent notice oddness with my swimming AO on magnum that I can recall

[12:16] Simon Linden: I suspect swimmer HUDs are OK - you're not colliding when swimming

[12:16] Qie Niangao: right.

[12:16] Simon Linden: These problems seem to be with frictions and/or collisions

[12:16] Tiberious Neruda: correct

[12:16] Motor Loon: happens when you walk on ground or prims Tib?

[12:17] Fadeout Razorfen: both

[12:17] Tiberious Neruda: I mean, at least it seems like it to me

[12:17] Davido Chrome: Does it happen when you hit seams in roads perhaps?

[12:17] Tiberious Neruda: both

[12:17] Tiberious Neruda: no

[12:17] Tiberious Neruda: in my case, I just act like I'm rooted to the ground

[12:17] Tiberious Neruda: or not pressing anything

[12:18] Davido Chrome: I don't know, simboards tend to get stuck in plygon edges.

[12:18] Simon Linden: That sounds like the friction is either too high, you somehow picked up more mass, or the push has been decreased. I'm guessing, but I'm suspicious of the friction

[12:18] Davido Chrome: Polygon*

[12:18] Tiberious Neruda: except I know the script is working because I have an indicator number that shows the force I'm using, and it goes up to its maximum

[12:19] Simon Linden: and it works correctly on non-pathfinding regions, I assume?

[12:19] Tiberious Neruda: yes

[12:19] Fadeout Razorfen: mine doesn't XD

[12:19] Fadeout Razorfen: same stuff

[12:19] Tiberious Neruda: and I'm suspecting friction as the culprit because it also jumps via llSetVelocity, and I can move in the air

[12:19] Davido Chrome: It honestly sounds like avatars are getting the same problems as Vehicles have always had.

[12:20] Davido Chrome: To a degree.

[12:20] Tiberious Neruda: I mean my jumping is controlled by the script

[12:20] Tiberious Neruda: so that it's instantaneous, you see

[12:21] Simon Linden: well, I'll pass this along to the pathfinding folks

[12:21] Rex Cronon: let me see if i get this right. u apply force to push u forward but u r not moving?

[12:21] Tiberious Neruda: right

[12:21] Rex Cronon: doesn't matter how much force u r using?

[12:22] Tiberious Neruda: it's variable....

[12:22] Draconis Neurocam: would it ever be possible to have larger brush sizes for llModifyLand?

[12:22] Rex Cronon: maybe it applies ony to attachments?

[12:22] Tiberious Neruda: sometimes I DO start moving when the force gets high enough, sometimes not

[12:23] Simon Linden: and that's testing at the same spot?

[12:23] Tiberious Neruda: yes

[12:23] Tiberious Neruda: same region

[12:23] Simon Linden: well, it certainly should be consistent

[12:23] Tiberious Neruda: I'd kinda prefer that my stuff not get broken, honestly

[12:24] Simon Linden: We don't like to break stuff either

[12:25] Tiberious Neruda: but if it -does- get broken, so be it... as long as I know how to work around it to get the effect I'm intending

[12:26] Simon Linden: I'm most concerned with it being inconsistent. Physics laws should be fairly predictable :)

[12:26] Motor Loon: Now if only you had a Jira on it Tib... mabye Falcon could give you some pointers on how to modify it to work °?°

[12:26] Motor Loon grins

[12:27] Tiberious Neruda: Motor... you know about me and the jira

[12:27] Motor Loon nods

[12:27] Fadeout Razorfen: oh, inconsistency, okay sec let me simplify the formula and give what i use

[12:27] Simon Linden: If you do have a stripped-down version that shows the bug, that helps a lot

[12:27] Rex Cronon: if u have jira is very likely that workingonit linden will start working on it:)

[12:28] NeoBokrug Elytis: workingonit linden is working on too many problems at once. :)

[12:28] Fadeout Razorfen: llSetForce




* mass



( // 120 is the maximum speed decreasement. default walk speed is considered as 100% speed

//(-(1-hp/100.)*60) + 60 - 30*movement_speed_decreasement_multiplier

-( 1.0 - ( base_movement_speed_percentage*health_multiplier + wounded_minimum_base_movement_speed_percentage*(1.0-health_multiplier) ) * (movement_speed_percentage / 100.0) / 100.0 ) // health interpolated base value inverted for decreasement

* 120 // force required to stop walking


- 87.5*!!(agent_info & AGENT_ALWAYS_RUN) // running fix



- <0,0,1>*mass,TRUE


[12:28] Fadeout Razorfen: this is damn dirty

[12:28] Fadeout Razorfen: but with comemnts and explainations

[12:28] Motor Loon:

[12:29] Tiberious Neruda: ...mine is a bit simpler...

[12:29] Fadeout Razorfen:

[12:29] Bear Hunt Start: Bear Quest Start touched by Davido Chrome

[12:29] Simon Linden: a sample object works too

[12:29] Motor Loon: pastebin = less spam ;-))

[12:30] Simon Linden: Are there other topics anyone wants to bring up?

[12:30] Rex Cronon: come on. that is not spame. just a wall of text:)

[12:30] Rex Cronon: spam*

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

[12:30] Motor Loon: thats what that girls keeps saying when she puts ads in my open groups °?°

[12:30] Tiberious Neruda: llSetForce(<(float)cur_speed * .75,0,cur_speed * -0.025>,TRUE); (where 'cur_speed' is an integer that increments as the key is held, from the control() event)

[12:31] Tiberious Neruda: the downward force increases as the forward does, but it's intended to be negligible

[12:32] Simon Linden: Could you use llSetVelocity() ?

[12:32] Tiberious Neruda: no

[12:32] Tiberious Neruda: I experimented with that, but it just doesn't work well

[12:32] Fadeout Razorfen: yea me too

[12:33] Fadeout Razorfen: i don't remember why was that tho

[12:33] Tiberious Neruda: in my case, it's not smooth

[12:33] Fadeout Razorfen: oh, also in my example i just modify the speed instead of giving it completely new value

[12:33] Simon Linden: That function is pretty direct in the way it just sets the velocity value for an object

[12:34] Davido Chrome:

[12:34] JIRA-helper: [#WEB-4113] OR and AND booleans do not work in Marketplace search, nor does the minus sign to exclude keywords from search.

[12:34] Tiberious Neruda: it is

[12:34] Simon Linden: Alve - do they say those things should work?

[12:35] Tiberious Neruda: but it's also a 'burst', and not a stream like we need

[12:35] Simon Linden: It would be great to have ... I'm asking to see if that's a bug or a feature request

[12:36] Davido Chrome: No, but it's strange. Every database worth the name supports booleans, in my opinion.

[12:36] Object: Hello, Avatar!

[12:36] Davido Chrome: Or Database search function I mean.

[12:36] Simon Linden: I could make some really un-professional comments about our databases, but I won't

[12:36] Tiberious Neruda: maybe it's an 'expected feature missing'?

[12:37] Simon Linden: ok - well, it's been passed onto the web and marketplace guys, so hopefully they'll take a look and add it. Like you say, it's pretty common functionality and would definitely be nice to have

[12:37] Motor Loon: damn Simon.. that would have been gold for my quote-book

[12:38] Qie Niangao: Unrelated topic: Do I remember correctly that cached sim state is supposed to get persisted at shutdown? (I'd been scripting for a while on an object when I got the 5-minute restart warning this morning; came back later and the object had vanished, as if the sim state had been rolled-back maybe an hour. If that's normal, no big deal; I just didn't expect it.)

[12:38] Simon Linden: I know :)

[12:38] Fadeout Razorfen: i'l make example and post it on jira

[12:38] Tiberious Neruda: oh, by all means. I never expected my question to get the time it did... so if you wanna jump in, go for it

[12:39] Simon Linden: Qie- that object should persist across the simulator shutdown and re-start. Sometimes it fails, however. We were looking at some of those earlier today, actually, so maybe you got caught in one that failed

[12:39] Qie Niangao: thanks. at least I didn't mis-remember how it was supposed to work. :)

[12:40] Davido Chrome: Hum. Seems like I never filed a jira abour script messages between different objects sometimes failing.

[12:42] Motor Loon: I've seen that happen 100 times Qie

[12:42] Simon Linden: Those "sometimes failing" problems can be tough :P

[12:42] Davido Chrome: Well, I gave Kelly my self setup teleporters. Usually you don't have to rez more than five pairs before you get a fail.

[12:42] Qie Niangao: Thanks, Motor. I guess I've been lucky, mostly.

[12:43] Simon Linden: Do they send a lot of messages, or just a few?

[12:43] Davido Chrome: Well, about two or three.

[12:44] Simon Linden: Repeatedly?

[12:44] NeoBokrug Elytis: This is still a minor annoyance:

[12:44] Davido Chrome: No, just when they setuo,

[12:44] JIRA-helper: [#SVC-3017] Server drops first touch event when a script returns to a state with a touch_start handler

[12:44] Kelly Linden: I have not had a chance to look into that.

[12:45] Kelly Linden: could be a race condition on rez?

[12:45] Davido Chrome: Race?

[12:45] Simon Linden: hmm, ok - I was just trying to figure out if there was enough traffic that something could have been capped, but it doesn't sound like that

[12:45] Kelly Linden: Some chat messages sent from the rezer before the scripts on the rezee are running

[12:46] Davido Chrome: No, it sends a message when you rez it. And if there is one allready rezzed it hears it and they send a couple of messages exchanging location.

[12:46] Davido Chrome: Then they close their listens.

[12:46] Acheron Gloom: Kelly was referring to what a race condition is.

[12:46] Acheron Gloom: Oh sorry

[12:46] Kelly Linden: you always wait a second between the two rezes?

[12:46] Acheron Gloom: Ignore me

[12:46] Davido Chrome: And send no more messages.

[12:46] Simon Linden: Alve - if you can file a jira on that too, it'll help. We'll get it imported and into the queue

[12:47] Davido Chrome: Yeah.

[12:47] NeoBokrug Elytis: object_rez(key blah) may be helpful.

[12:47] Davido Chrome: Simon, okay.

[12:49] NeoBokrug Elytis: I gotta head out, but before I go -- Thank you for fixing the Top Scripts portion of the estate menu.

[12:49] Kelly Linden: Alve: also, I do not have the objects you mention in my inventory.

[12:49] Kelly Linden: oh, nvm I found it.

[12:49] Simon Linden: Glad to hear that's working right, Neo

[12:50] Davido Chrome: Also, I have been playing Logos, a card game played Via HUDs, has been a lot of "HUDs lost sync" messages lately.

[12:51] Davido Chrome: I didn'ty create that game though, so can't give you it.

[12:51] Simon Linden: hmm, I can't think of other bugs or work done on the messaging code recently, but maybe something broke

[12:52] Baker Linden: No problem Neo

[12:52] Baker Linden: Though if it's the one I'm thinking of, it may not be fully fixed

[12:52] Draconis Neurocam: just made this

[12:52] JIRA-helper: [#SCR-377] Add larger brush sizes for llModifyLand

[12:53] Davido Chrome: Larger?

[12:53] Draconis Neurocam: yes the largest size is currently 8x8

[12:53] Davido Chrome: I would rather be able to edit vertex by vertex.

[12:53] Draconis Neurocam: which is really annoying when you are trying to edit large amount of land via script

[12:54] Simon Linden: Do you know how large the viewer allows?

[12:54] Draconis Neurocam: will have to test that, sec

[12:54] Motor Loon: bigger than this roof we're on

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

[12:56] Draconis Neurocam: preliminary testing says the viewer goes up to 24m square

[12:57] Draconis Neurocam: the reason llmodifyland is so useful is because it is the only way to get land to a specific number height

[12:57] Draconis Neurocam: which you cannot do with the land tools without lots of redoing small edits

[12:58] Rex Cronon: somebody could make a bot that shapes the land u want?

[12:58] Motor Loon: Flatten usually works for me

[12:58] Draconis Neurocam: right but say i want to flatten to exactly 32m

[12:58] Draconis Neurocam: for building purposes

[12:58] Fadeout Razorfen: okays, i've got example working, randomly when i use this speed modificator it starts jumping up/down with no reason, sorry for taking so long. you can copy the red cube and attach it to check it out

[12:58] Davido Chrome: Easiest way to terraform that I have found is to use a drawing tablet.

[12:58] Draconis Neurocam: you can only do it with llmodifyland

[12:58] Motor Loon: so make a 8x8 piece go to 32m

[12:58] Motor Loon: then use the flatten too to quickly "copy" that to the rest

[12:59] Motor Loon: can't take more than 5 minutes

[12:59] Motor Loon: "flatten tool"

[12:59] Draconis Neurocam: it would just be nice is all

[12:59] Draconis Neurocam: and the change seems trivial to impliment

[13:00] Motor Loon: not automatic then... but still ... 5 minutes of work °?°

[13:00] Simon Linden: thanks - no promises, but that sounds like a reasonable change.

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

[13:00] Motor Loon: Alrighty... guess that means I have to go make dinner then

[13:00] Simon Linden: Thanks everyone for coming today and the good discussion

[13:00] Davido Chrome: Ok, thank you for your time Lindens!

[13:00] Rex Cronon: tc everybody and have a nice day:)


Prev 2012.07.20 Next 2012.07.27