User:Andrew Linden/Office Hours/2008 04 10

Transcript of Andrew Linden's office hours:

[17:01]	Arawn Spitteler would like very simple trowsers, in the Plywood pattern: It shouldn't be impossible to texture an H4 T-Shirt
[17:01]	Areth Gall:	Nah. Andrew's Army comes to mind. But then you get into the gay jokes.
[17:01]	Tsuno Soyinka:	Woot! Here comes entertainment!
[17:01]	Areth Gall:	Speaking of the devil.
[17:01]	Tsuno Soyinka:	Maybe I should break out my seduction clothes too!
[17:02]	Arawn Spitteler:	Hi, Simon
[17:02]	Simon Linden:	Hi !
[17:02]	Kitto Flora:	Lookit, we are surrounded
[17:02]	Andrew Linden:	Hey everybody
[17:03]	Holger Gilruth:	hey you :)
[17:03]	Kitto Flora:	Hi Andrew, Hi Simon
[17:03]	Tsuno Soyinka:	Hi!
[17:03]	Areth Gall:	Hello Andrew.
[17:04]	Andrew Linden:	We've got a new simulator update locked and loaded, but I'm not sure when exactly it i going out
[17:04]	Kitto Flora:	What uesful fixes does it have?
[17:05]	Andrew Linden:	I just asked the deploy team and they're testing some functionality that is somewhat independent of Havok.
[17:05]	Andrew Linden:	They're calling it "the conductor". It is some sort of tool for helping with deploys and general region management, I think.
[17:06]	Andrew Linden:	Anyway, this is the same update that I was talking about on Tuesday, and largely the same that we started testing last Thursday.
[17:06]	Andrew Linden:	So, hopefully soon.
[17:07]	Andrew Linden:	Today I fixed (er... greatly improved) the oscillating camera location problem.
[17:07]	Andrew Linden:	Not in the update, sadly. It will have to be in the next.
[17:07]	Cinthya Loveless:	Andrew was the anit griefing thing you were talking about work for exploding barrals? that explode into alot of parts that use physics?
[17:07]	Andrew Linden:	I think Kelly already fixed it so that the camera is not affected by avatars that walk between your avatar and your camera. I don't think that is in this current update either.
[17:08]	Andrew Linden:	I'm starting to work on the llTargetOmega() problem.
[17:08]	Cinthya Loveless:	yesterday someone almost took one of my islands down with those
[17:08]	Areth Gall:	The llTargetOmega() problem?
[17:08]	Andrew Linden:	I think I've found some "smoke", but haven't found the fire yet.
[17:09]	Andrew Linden:	the llTargetOmega() problem is that sometimes objects/prims that are supposed to be spinning sometimes just stop.
[17:09]	Andrew Linden:	And cannot easily be restarted.
[17:09]	Andrew Linden:	It ranks as one of the most annoying bugs right now, as far as we can tell.
[17:09]	Tsuno Soyinka:	I wonder what precipitated the start of said problem.
[17:09]	Kitto Flora:	More so than all the boats sinking?
[17:09]	Areth Gall:	Ah. That doesn't include the decrease in angular velocity over time of an object that has been pushed into motion by llApplyRotationalImpulse(), is it?
[17:10]	Areth Gall:	does it* sorry.
[17:10]	Andrew Linden:	No Areth, this is a distinct bug.
[17:10]	Areth Gall:	Okay. And I'll let you continue with announcements.
[17:11]	Andrew Linden:	Yes, the activity in the public jira list suggests that the llTargetOmega() problem outweighs sinking boats.
[17:11]	Tsuno Soyinka:	I have a clue for you reguarding things sinking.
[17:11]	Andrew Linden:	Also, I was hoping to knock off a few "quick and easy" bugs before I sank into the dismal realm that is vehicle code.
[17:12]	Kitto Flora:	Odd - cause its a recent bug. H4 used to be good for boats.
[17:12]	Andrew Linden:	Heh, I started the camera motion bug as a little snack to get me back in the coding mood on Tuesday, but it ended up taking me about two days to fix :-(
[17:13]	Andrew Linden:	anyway, I think I'm done with announcements. You got any Simon?
[17:13]	Les White:	(sneaks in late)
[17:13]	Simon Linden:	No, I've spent the past few days trying to reproduce that rotation problem as well as our memory issues, but no good fixes to mention
[17:14]	Andrew Linden:	Alright then. Open topics... anyone have something to talk about?
[17:14]	Tsuno Soyinka:	May I?
[17:14]	Andrew Linden:	Please.
[17:14]	Andrew Linden:	There are no rules of order here.
[17:15]	Tsuno Soyinka:	I have noticed that hovering avatars no longer hold altitude reliably. There is a loss of altitude of about 1 meter every 20 seconds. This can be seen in this sim and many others.
[17:15]	Arawn Spitteler wonders if weird flight paths of ground vehicles has been noticed.
[17:15]	Tsuno Soyinka:	This phenomenon has been going on for , three weeks maybe? I am not sure.
[17:15]	Simon Linden:	Tsuno - we know about that bug, and do want to fix it sometime.
[17:16]	Tsuno Soyinka:	Ok.
[17:16]	Tsuno Soyinka:	I res.
[17:16]	Tsuno Soyinka:	*rest.
[17:16]	Simon Linden:	We realize it's annoying, but it's less of a problem than some of the broken content we have
[17:16]	Tsuno Soyinka:	I understand/
[17:16]	Andrew Linden:	Ah yes, that is a known bug Tsuno. I was considering it for my "quick and easy" bug list, but decided to pick up llTargetOmega() today instead.
[17:16]	Arawn Spitteler has also had an instance of drifting while standing, but only once.
[17:16]	Andrew Linden:	Simon, were you planning on taking that sinking avatar problem?
[17:17]	Simon Linden:	It occurs when you stop flying, but it seems like the AV retains a little bit of downward speed ... we counteract gravity as you hover, but don't give the extra push to stop you
[17:17]	Simon Linden:	I've looked at it before, and understand what's going on, but the fix isn't simple. You and I should talk later about it ... it probably would be good for you to take a look
[17:17]	Andrew Linden:	hrm... I wonder if we've lost ALL the friction in the Z direction during hover that would have brought the avatar to a stop eventually
[17:18]	Tsuno Soyinka:	I wish I could take a look at it.
[17:18]	Andrew Linden:	Alright.
[17:18]	Arawn Spitteler notes Sidewinder
[17:18]	Les White:	hiya sidewinder
[17:18]	Kitto Flora waves to 'winder
[17:18]	Tsuno Soyinka:	Hello.
[17:18]	Sidewinder Linden:	hi guys
[17:18]	Simon Linden:	Possibly - fwiw, you can often compensate by pressing the space bar a few times. That may stop you and you'll be frozen in place until you move again (or something bumps you)
[17:18]	Sidewinder Linden:	hmm panera connection is pretty wonky - i might get bounced more than once
[17:19]	Andrew Linden:	Hey Simon, I just had an idea as to what would cause that hover-down problem. I'll make a note and check it later.
[17:19]	Tsuno Soyinka:	The hover down does not always occur.
[17:20]	Simon Linden:	Sounds good - yes, I haven't figured out why you sometimes have that little downward speed after you should have stopped
[17:20]	Tsuno Soyinka:	It is also altitude independent.
[17:20]	Andrew Linden:	how does it depend on altitude?
[17:20]	Simon Linden:	Other than stopping / hovering (with the little sink) does flying work OK for everyone? Various flight assists seem OK?
[17:20]	Andrew Linden:	oh sorry... independent
[17:21]	Tsuno Soyinka:	It does not depend on altitude, however it does stop at the distance that physical bounding boxes display.
[17:21]	Les White:	i never noticed any issues myself
[17:21]	Tsuno Soyinka:	So no worries about sinking through objects.
[17:22]	Andrew Linden:	alright... a bug I heard Sidewinder talking about this morning is avatars getting trapped inside big rezzed builds
[17:22]	Tsuno Soyinka:	I figured it was a change on how buoyancy of an avatar was handled, but I probably have the wrong model of how the system works.
[17:22]	Andrew Linden:	that is... when standing where a big rez suddenly happens you get trapped
[17:22]	Andrew Linden:	anyone here have that happen to them?
[17:22]	Tsuno Soyinka giggles. "Isn't that when a person gets lost?"
[17:22]	Just Cattaneo:	can get stuck inside prims
[17:22]	Cinthya Loveless:	I've had that problem my self getting trapped like that
[17:23]	Just Cattaneo:	it is as if the edges exist, but not the filling
[17:23]	Arawn Spitteler:	There's an armored roadway, from Horisme to Spini, where we can fall through the blocks and get trapped
[17:23]	Just Cattaneo:	can move freely inside the box
[17:23]	Just Cattaneo:	but edges stop you
[17:23]	Kitto Flora:	Not buildings - But Avs trapped in hollow vehicles - thats a big problem. Same thing? (Vehicle is non-phys)
[17:23]	Andrew Linden:	ah Just, that was probably inside a concave prim
[17:23]	Tsuno Soyinka:	Kool.
[17:23]	Just Cattaneo:	was regular box
[17:23]	Les White:	possibly related...when you unsit from a moving vehicle both you and the vehicle go flying
[17:24]	Andrew Linden:	it turns out that concave prims use Havok's "MeshShape" data structure
[17:24]	Kitto Flora:	Les: Unrelated and an old ccommon 'feature'
[17:24]	Just Cattaneo:	unsit is bout 5x more than it was
[17:24]	Les White:	it's new with h4
[17:24]	Andrew Linden:	and the mesh shape collides like a collection of triangles... but collide both on the inside and outside
[17:24]	Just Cattaneo:	20 m and 100m are big diffrence, takes while to fly back to the vehicle
[17:26]	Kitto Flora:	There appears to be an enhancement to the ejection of an av fro ma phys-vehicle, if the Av is also slightly embedded in a prim of the vehicle.
[17:26]	Areth Gall:	Bye bye prim...
[17:26]	Tsuno Soyinka:	Makes sense.
[17:26]	Just Cattaneo:	happens to me when in motion
[17:26]	Just Cattaneo:	if still no effect
[17:26]	Simon Linden:	FWIW we found some bugs with objects that would unsit the avatar, and then give them a little push. Unfortuately, the effectiveness of the push was limited by the energy system, and there's a range of distances where a small change can have a huge effect on how far the AV gets pushed
[17:26]	Kitto Flora:	As the Train says *do not stand up while the train is in motion*
[17:26]	Just Cattaneo:	no push, took out with h4
[17:26]	Les White:	i make crashing race vehicles. before the unsit would make the av travel with the vehicle. now they go boing! to space or even under the track prims.
[17:26]	Just Cattaneo:	i did, tryin to figure out why i was goin 100+m
[17:27]	Kitto Flora:	Simon: I found that one in practice, however it turns out that llPush is generally of limited use - cant be relied on.
[17:27]	Tsuno Soyinka:	Do you have one I can try here?
[17:27]	Kitto Flora:	I'm trying to abandon use of llPushObject()
[17:28]	Tsuno Soyinka:	I never managed to get the push right.
[17:28]	Les White:	they need the track environment to do the magic. just set any veicle to not turn off physics on unsit. drive. stand up. boing!
[17:28]	Les White:	no push involved. just unsit
[17:29]	Andrew Linden:	Ah yes, I tried to fix the vehicle dismount problem
[17:29]	Kitto Flora:	Not physics off --- trick is to get the Av to intesect a prim and the prim will push em out
[17:29]	Andrew Linden:	but broke a whole bunch of content in the process
[17:29]	Les White:	it's not a show stopper. i have some ideas..offset av animation and such, but if there is a cure. great
[17:29]	Andrew Linden:	I'd like to take another swing at it
[17:29]	Areth Gall:	I actually wonder if it would be possible to make a llUnsitAvatarAt() function.
[17:29]	Andrew Linden:	but it will take at least a day, maybe two
[17:29]	Kitto Flora:	Actually Av trapping is a big content breaker
[17:30]	Les White:	it's not a killer andrew. i'm sure you have a full plate
[17:30]	Just Cattaneo:	yes
[17:30]	Kitto Flora:	Continuous complaints... and customers wont buy when they get trapped in the demo
[17:30]	Simon Linden:	You might experiment with shifting the sit pos before the unsit - not sure if it will help or look right, but it's an idea
[17:30]	Just Cattaneo:	yep
[17:30]	Les White:	yes. though of using prim param to shift the av out and then unsit. need to test!
[17:30]	Kitto Flora:	Simon: Woould that actually DO anything?
[17:30]	Andrew Linden:	turns out a lot of scripters have implemented workarounds for bad dismount -- they shift the position of the avatar right before it gets off or something
[17:31]	Arawn Spitteler:	llSetLinkPos() should get them out, if you ask for exit
[17:31]	Simon Linden:	Kitto - dunno, it's just an idea :)
[17:31]	Areth Gall:	But doesn't that only work if the vehicle is not physical?
[17:31]	Tsuno Soyinka:	Personally I have not had much trouble with dismounts.
[17:31]	Just Cattaneo:	no can shift pos when physical
[17:31]	Kitto Flora:	Simon: I think not... one cannot scit an action *Before* sitting or standing.
[17:31]	Les White:	mostly from vehicles that are still pysical when you unsit
[17:32]	Simon Linden:	The problem may not really be in the unsit, but in our code that moves physical objects apart when the are interpenetrating
[17:32]	Kitto Flora:	Seems to me the only good solution is a limited llTeleportAgent() - one thats limited to say 5M max range
[17:32]	Kitto Flora:	To stop griefing use
[17:32]	Tsuno Soyinka:	I believe it is the combination of the two which brings fourth such a reaction.
[17:32]	Simon Linden:	I've been stressing the server with hundreds of objects in the same spot, and sometimes they get thrown really, really far
[17:33]	Andrew Linden:	unfortunately, the code that separates penetrating objects is the Havok "solver" and is kinda hard to override in a smart way
[17:33]	Tsuno Soyinka:	I can feel your headache already!
[17:34]	Andrew Linden:	I've got an idea that could help the problem with residual velocity on separation from penetration, but it looks like one of those mult-day projects.
[17:34]	Les White:	when you get some spare time :)
[17:35]	Areth Gall:	On the bright side, with summer coming up, I'll be able to make everyone one of these meetings in a few weeks :P
[17:35]	Tsuno Soyinka:	I don't envy that one.
[17:35]	Arawn Spitteler wonders if there's a way to get college credit here.
[17:35]	Andrew Linden:	Actually, I was just pondering (again) the value of weekly and bi-weekly office hours
[17:36]	Andrew Linden:	Anybody here tired of talking about Havok4 yet ? ;-)
[17:36]	Les White:	amost!
[17:36]	Areth Gall:	Tired? I'm addicted. Of course, I don't really do much talking.
[17:36]	Kitto Flora:	Well the fixes are coming so slowly (now) - one has to wonder about 2 meets a week
[17:36]	Just Cattaneo:	im finding the call for groundcollision doesnt occur as consistently and requently as regular collision, mono seemed to improve, but the regular collsions were still heavier than the ground collision activity
[17:36]	Les White:	i wouldnt have an issue with cut backs
[17:36]	Kitto Flora:	We are chewing over the same old stuff and none of its getting fixed
[17:36]	Kitto Flora:	(in the meetings timeframe)
[17:37]	Areth Gall:	How many people generally come to the Tuesday Meetings?
[17:37]	Andrew Linden:	about the same... I don't notice a big attendance difference between the two hours
[17:37]	Simon Linden:	Just - can you make up an object that shows that? Is there a Jira bug?
[17:37]	Just Cattaneo:	is one event
[17:37]	Just Cattaneo:	say ground, in ground collision
[17:37]	Just Cattaneo:	can make in 5 sec guess
[17:38]	Just Cattaneo:	tried with 2 independant scirpts, one ground one reg collision
[17:38]	Just Cattaneo:	and in same script
[17:39]	Areth Gall:	Well, I dunno. If different people attend the Tuesday meeting, it would make sense to keep it. I can only make the Thursday one and I'd hate to lose my meeting time, and I bet they'd feel the same. So maybe 1 meeting a week, but alternate between Tuesdays and Thursdays every week?
[17:39]	Les White:	just, you said something about that with mono?
[17:39]	Just Cattaneo:	yes, mono is faster, so it caught more ground collisions than withought, but it still was more reg collisions, bout 5x as much
[17:39]	Kitto Flora:	Which meet I can make varies.
[17:39]	Tsuno Soyinka:	Doesn't it just check for object collisions more often than ground collisions?
[17:40]	Just Cattaneo:	yes, seems torus doesnt hit as well.. and sim lag
[17:40]	Andrew Linden:	interesting... that would suggest that the bottleneck is in the script engine and not in the simulator's collision callback handler
[17:40]	Simon Linden:	There's code that throttles the various collision events, since we get many more from the physics engine than LSL could handle
[17:40]	Just Cattaneo:	well, i understood that h4 was gonna reduce the quantity, and does when i test on beta..
[17:41]	Andrew Linden:	right and that trottle is in the simulator's collision callback
[17:41]	Simon Linden:	Just to be clear, you're talking about the land_collision events, right?
[17:41]	Just Cattaneo:	yes
[17:41]	Kitto Flora:	Fixing the collision range would cut down on the number of events :)
[17:42]	Andrew Linden:	If the script engine were slow to handle the collision events, then the "integer count" variable passed in should be a high number
[17:42]	Just Cattaneo:	its less now.. but there is more lag, and ground doesnt call as much, im more intersted if it is a problem, maybe easy fix, and if not, i can work around
[17:42]	Andrew Linden:	that is, the collisions would pile up and you'd bet multiple events in the handler.
[17:42]	Andrew Linden:	Just, have you noticed the value of that count?
[17:42]	Andrew Linden goes to look at the scripting wiki
[17:43]	Just Cattaneo:	no clue really, i just check with a timer if it has been called, setting an integer true
[17:43]	Just Cattaneo:	didnt so i put straight chat in the collisions itself, and got the same results
[17:44]	Andrew Linden:	oh... land_collision() only has a vector
[17:44]	Just Cattaneo:	right
[17:44]	Just Cattaneo:	but it is an event
[17:44]	Tsuno Soyinka:	yea... got me goin there.
[17:44]	Just Cattaneo:	yes, i dont do anythign with the vector, just say ground
[17:44]	Andrew Linden:	meanwhile the normal collision() event has an "integer num_detected"
[17:45]	Simon Linden:	Are you looking at the start and end events as well?
[17:45]	Just Cattaneo:	sure, but i am not using it, to verify
[17:45]	Just Cattaneo:	yes
[17:45]	Les White:	he's just saying ground triggers 5 times less offten then normal col event
[17:45]	Just Cattaneo:	was same
[17:45]	Andrew Linden:	no, just collision() and land_collision()
[17:45]	Just Cattaneo:	with start as wo
[17:45]	Tsuno Soyinka:	Hmmm.
[17:45]	Just Cattaneo:	both have start option event
[17:46]	Andrew Linden:	and what was the Havok1 behavior comparison?
[17:46]	Andrew Linden:	between collision() and land_collision()?
[17:46]	Just Cattaneo:	both are less
[17:46]	Just Cattaneo:	think was more reg before as well
[17:46]	Just Cattaneo:	but now.. hardly any ground
[17:46]	Andrew Linden:	frequency of the events is less, and less regular?
[17:47]	Just Cattaneo:	yes
[17:47]	Andrew Linden:	oh, you mean more collision() events than land_collision()'s
[17:47]	Tsuno Soyinka:	Sporadic triggering?
[17:47]	Just Cattaneo:	yes, i can hardly get ground to fire
[17:47]	Andrew Linden:	well, if it is that bad, shouldn't be too hard to reproduce and figure out what is going on
[17:48]	Just Cattaneo:	thx, isnt killing me, and mono was better
[17:48]	Andrew Linden:	btw, a really good way to get your bug's picked up is to provide a clear "repro recipe"
[17:48]	Just Cattaneo:	more regular
[17:49]	Just Cattaneo needs to find the jira tutorial, idiot
[17:49]	Andrew Linden:	in this case, that would mean a small script in the bug to copy and paste into the object, and instructions on how to make the bug happen
[17:49]	Just Cattaneo:	easy nough ;)
[17:49]	Andrew Linden:	often just figuring out how to reproduce the bug is half or more of the time it takes to fix
[17:50]	Areth Gall:	No kidding
[17:50]	Andrew Linden:	so I've noticed the behavior in myself -- I'll prefer to work on bugs that appear easy to repro over bugs that have no repro
[17:51]	Andrew Linden:	sometimes, right after fixing a big long hard-to-fix bug I'll try to knock off a bunch of easy-but-annoying ones
[17:51]	Andrew Linden:	and those that have the example script in the comments are likely candidates for 'easy'
[17:52]	Les White:	makes sense to me!
[17:52]	Sidewinder Linden:	sorry to have been a bit afkish.. but in light of easy repro... if any one has a formula script or action that causes llTargetOmega objects to stop spinning spontaneously in a region.... ?!
[17:52]	Tsuno Soyinka:	I am making it right now.
[17:52]	Sidewinder Linden:	cool
[17:52]	Sidewinder Linden:	the team has been wrestling with trying to build a theory about that one for days
[17:52]	Tsuno Soyinka:	I meant the land collision thing, not the omege
[17:52]	Sidewinder Linden:	oh well
[17:53]	Sidewinder Linden:	got my hopes up and everything :)
[17:53]	Areth Gall:	I was trying to reproduce that. But I couldn't get it to appear to slow down. Do you guys have any info on how it appears to even slow down?
[17:53]	Tsuno Soyinka:	Sorry. I haven't seen either happen.
[17:53]	Sidewinder Linden:	that's the thing
[17:53]	Sidewinder Linden:	all the bugs read essentially
[17:53]	Just Cattaneo:	me either, sounds gfx card related
[17:53]	Sidewinder Linden:	"it stops after a while, and with no apparent trigger"
[17:53]	Sidewinder Linden:	@just - well yes, except that when it stops, it stops for all viewers
[17:53]	Andrew Linden:	I've got some clues on the llTargetOmega() problem that I'm going to look into tomorrow. I don't think it is an easy-to-reproduce bug.
[17:53]	Les White:	that targetomega is issue prone
[17:53]	Sidewinder Linden:	cool
[17:54]	Areth Gall:	Hm. Maybe it's cutting off with a timer type function. Just shuts down after time x.
[17:54]	Simon Linden:	... and sometimes various things will or won't fix it - selecting, restarting the region, etc
[17:54]	Tsuno Soyinka:	Just, will you put somethign into this test script to make for .. ahh nm
[17:55]	Andrew Linden:	Sidewinder, do you have any updates on the ETA of the next update?
[17:55]	Andrew Linden:	I think we're calling it 1.21.0, right?
[17:55]	Sidewinder Linden:	boy i wish i could be sure...
[17:55]	Sidewinder Linden:	i believe so
[17:55]	Sidewinder Linden:	it was being held in qa for resolution of a few open issues, then deploy started
[17:56]	Sidewinder Linden:	i'll check now - i *think* they were going to roll 3 racks tonight, but i've been offline for a couple of hours
[17:56]	Sidewinder Linden:	brb
[17:56]	Andrew Linden:	Oh yeah, btw, I heard that the SL client 1.20-RC0 was release recently, is that correct?
[17:56]	Les White:	yes. the dazzle delusion
[17:56]	Arawn Spitteler:	I got motion sickness with that one
[17:56]	Kitto Flora:	Yes, I Was using it earlier. Crashed in 30 secs.
[17:56]	Andrew Linden:	If so... the edit limit should be bumped from 768 to 4096 or so in that client.
[17:57]	Les White:	i really cant understand how something so ugly and annoying makes it so far (but that's another meeting!)
[17:57]	Andrew Linden:	Oh... sounds like there will be an RC1 for that...
[17:57]	Sidewinder Linden:	back
[17:57]	Sidewinder Linden:	so it looks like a bit later tonight they're planning to do 3 racks (around 500 regions)
[17:57]	Andrew Linden:	Sorry Les, are you talking about the bad UI design in "dazzle"?
[17:57]	Les White:	yes
[17:57]	Simon Linden:	Ah, yes, the fun of visual design 'discussions' :)
[17:58]	Kitto Flora:	Someone at that other meeting compained about 'avatars sliding down slopes' and was directed to come to *this* meeting
[17:58]	Sidewinder Linden:	see how that does anddo either 110 rack partial deploy tomorrow or the rest of the grid at that point
[17:58]	Sidewinder Linden:	per update timing question
[17:58]	Sidewinder Linden:	oops 10 rack - not 110
[17:58]	Andrew Linden:	Yeah, that is for another Linden's office hour ;-)
[17:58]	Les White:	hehe
[17:58]	Arawn Spitteler:	Yeah, in my last Spiral Stair Class, I noticed a bit of odd sliding.
[17:58]	Sidewinder Linden:	tell ya what andrew - you handle the viewer questions here and i'll handle'm on the blog ok? :)
[17:59]	Arawn Spitteler:	I do a lot with Hollow Tubes; are hollows getting iffy?
[17:59]	Andrew Linden:	awwww Sidewinder, can't you take both?
[17:59]	Andrew Linden:	"are hollows getting iffy"...
[17:59]	Simon Linden:	Arawn - shouldn't be. They usually are mesh shapes, so are more costly than a normal cylinder, but there isn't any particular badness with them
[18:00]	Simon Linden:	Concave shapes are more costly than convex ones
[18:00]	Arawn Spitteler:	i heard something about their filling in laggy sims.
[18:01]	Simon Linden:	Well, when the sim gets laggy, we will simplify the shape we use for some objects. Instead of the full concave mesh, it would just be a box
[18:01]	Arawn Spitteler:	I was wondering if that might be why it seems easier to walk off my stairs, or if it's just my magination
[18:01]	Andrew Linden:	Simon, I thought we didn't do that for static objects anymore.
[18:01]	Just Cattaneo:	well, i have lot more trouble jumping on top of a pole now personally
[18:01]	Simon Linden:	Right, this is only physical objects
[18:01]	Andrew Linden:	Or is it just done at a higher threshold?
[18:01]	Simon Linden:	And smaller ones - if the AV can fit inside, we shouldn't be doing it
[18:02]	Harleen Gretzky:	Are you talking about http://jira.secondlife.com/browse/SVC-1531 Arawn?
[18:02]	Andrew Linden:	If nothing else, these office hours serve to remind me about bugs I meant to fix. --> /me takes notes on the side
[18:03]	Simon Linden:	If it does happen to a staircase, it wouldn't be a simple walk-off the stair. You'd likely get ejected dramatically :)
[18:03]	Kitto Flora:	Which reminds me - there was suggestions at a previus meet that enclosing a multi-prim vehicle in one big box prim *might* reduce the Sim Time (Physics) load. I tested - it does not.
[18:03]	Simon Linden:	http://jira.secondlife.com/browse/SVC-1531 is fixed ... I think that's in the update going out
[18:03]	Sidewinder Linden:	i think the other value of these, is that we get to hear the background on the bugs.. why some of them are more important than others, and what the impact is.. tha'ts really helpful in juggling priorities... just my 2c
[18:03]	Les White:	yes. some kinda convex hull for vehicles was mentioned as a possible future project. that would be hot
[18:04]	Andrew Linden:	right... I guess I'll still hold them then... ;-)
[18:04]	Les White:	stay in touch andrew! :)
[18:04]	Sidewinder Linden:	hehe :)
[18:04]	Sidewinder Linden:	oh
[18:05]	Sidewinder Linden:	there was a bug with top colliders not including all of the items it should... do you (andrew, simon) remember if that got into the qa'd build, or was after that?
[18:05]	Kitto Flora:	And as a by-product of the testing, I discovered that a vehicle going around a curve creates much more load that one going in a straight line. 3x or 4x by the look of it.
[18:05]	Sidewinder Linden:	i'll look it up later if not sure - just got a query on that
[18:05]	Andrew Linden:	I think it did not make it time for that last QA pass
[18:05]	Sidewinder Linden:	wow kitto
[18:05]	Simon Linden:	I think that didn't make it
[18:05]	Sidewinder Linden:	ok thanks
[18:05]	Sidewinder Linden:	maybe based on collision rate?
[18:06]	Les White:	could be banking parms making parts drag?
[18:06]	Sidewinder Linden:	i wonder if that's also related to the relative sizes of track and wheel/torus/ring device
[18:06]	Sidewinder Linden:	?
[18:06]	Kitto Flora:	Its a flat sled vehicle sliding on a flat prim. No banking.
[18:06]	Tsuno Soyinka:	No wonder I had trouble creating something that orbited and stayed pointed at the center of orbit!
[18:06]	Sidewinder Linden:	hmmm
[18:06]	Kitto Flora:	No torii
[18:06]	Sidewinder Linden:	andrew or simon - any ideas on that?
[18:07]	Andrew Linden:	No. I'll have to verify./
[18:07]	Kitto Flora:	'wheels' are disks. Tapered
[18:07]	Arawn Spitteler:	The given issue SVC-1531 seems to be a failure of dimple
[18:07]	Sidewinder Linden:	i'm wondering about interpenetration... seems like the going around corners might hit that piece harder than straight line
[18:07]	Sidewinder Linden:	i've seen a few roller coaster problems that seem to fit that general description as well.. hmm
[18:08]	Kitto Flora:	interpenetration of what?
[18:08]	Sidewinder Linden:	well if it's wheels (maybe i'm not undersatnding the train construction)
[18:08]	Sidewinder Linden:	agains the sides of rails
[18:08]	Simon Linden:	Arawn - it was a bug in our code that decided if a shape was convex or concave. It wasn't taking full account of all the prim parameters like dimple, cut, etc
[18:08]	Areth Gall:	Well, I gotta go take care of some sim-admin stuff. But I there was something I wanted to say before I left. I got my job last Spring in part because I knew a little bit about programming. And I learned that programming from learning how to Script on Second Life (which gave me a jump start on programming in the courses I took in college). So, thanks to you, I get to try and figure out why the heck things keep breaking when I make fixes. I mean, really. It's all your fault that you pulled me into having to fix bugs like you guys do. But really, I know you get a lot of complaints and whatnot. Thought you might like to know that some people actually benefit from Second Life in actual manner to improve themselves, and I wanted to say thanks.
[18:08]	Kitto Flora:	Theres no 'rails'
[18:08]	Sidewinder Linden:	as it turns wheels would maybe in between steps end up in rails
[18:08]	Sidewinder Linden:	ahh ok
[18:08]	Sidewinder Linden:	for the roller coasters there are guides and collars
[18:09]	Les White:	see you Areth
[18:09]	Tsuno Soyinka:	I got a similar benefit!
[18:09]	Kitto Flora:	Sidewinder - thats a mechanically steered vehicle. None of mine are like that
[18:09]	Andrew Linden:	Well the buggyness of SL is not really a "feature".
[18:09]	Simon Linden:	Thanks everyone - that's really nice to hear. We want to make this all work much better so you can build cool stuff
[18:09]	Tsuno Soyinka:	I thought the old bug where insideout spheres were outlined was a feature. sad to see it go.
[18:10]	Andrew Linden:	You get buggy code and content even without the extra SL bugs... that is a normal aspect of programming.
[18:10]	Simon Linden:	I have to run as well .... see you next time
[18:10]	Les White:	see you SImon. thanks!
[18:10]	Just Cattaneo:	yes, most programming is fixin bugs :)
[18:10]	Just Cattaneo:	thx guys
[18:10]	Andrew Linden:	yeah, time to go. It's been a long day.
[18:10]	KFengine2-1 14.0.4 Black whispers: Reading card, please WAIT....
[18:10]	KFengine2-1 14.0.4 Black whispers: Set to channel 1
[18:10]	Les White:	good stuff Andrew. see you
[18:10]	Les White:	thanks Sidewinder!
[18:10]	Arawn Spitteler:	Most of fixing bugs is finding out what blithering idiot was yourself.
[18:11]	Tsuno Soyinka:	I want to help with creating the reproduction things.
[18:11]	Les White:	bai all
[18:11]	Kitto Flora:	Byebye Andrew
[18:11]	Andrew Linden:	seeya Kitto
[18:11]	Kitto Flora:	Only in SL it seems.
[18:11]	Sidewinder Linden:	see you kitto
[18:11]	Jansen Miles:	Since Babbage doesn't have office hours, has anyone heard how Mono is coming along?
[18:12]	Kitto Flora:	Sidewinder gone too?
[18:12]	Sidewinder Linden:	hmm i thought they did have office hours jansen
[18:12]	Tsuno Soyinka:	Well, we could kiss to help spread mono
[18:12]	Sidewinder Linden:	i'm here kitto
[18:12]	Sidewinder Linden:	just wsn't sitting
[18:12]	Sidewinder Linden:	i've been blog shepherding and time split a bit otherwise so wasn't at the table
[18:12]	Harleen Gretzky:	He attends the Mono Office hours
[18:12]	Arawn Spitteler:	Isn't there a Prospero Hour, in Beta?
[18:12]	Arawn Spitteler:	I hope they've gone to H4
[18:13]	Jansen Miles:	Hmm. Not Babbage.
[18:13]	Ellla McMahon:	bye everyone ... thanks :)
[18:13]	Andrew Linden:	I thought babbage held hours.
[18:13]	Andrew Linden:	Yes, I have one little update on H4+MONO...
[18:13]	Jansen Miles:	Harleen, do you have a link to the Mono office hours?
[18:13]	Andrew Linden:	one of the devs on it was having a crash bug...
[18:13]	Harleen Gretzky:	8 AM Wednesday and 3 PM Fridays on the Beta Grid are the Mono OHs with Babbage and usually Scouse
[18:13]	Andrew Linden:	turns out it is caused by "misaligned memory"
[18:14]	Tsuno Soyinka:	I'd do so much more, but I just can't get into looking at a few things.
[18:14]	Andrew Linden:	Havok requires that its objects be aligned in memory... starting at addresses that are zero on the last nibble
[18:14]	Tsuno Soyinka:	Now I wanna go find people who are reporting bugs and help make reproduction directions.
[18:14]	Jansen Miles:	Thanks. I'll google it.
[18:14]	Andrew Linden:	but under MONO the memory allocation is getting skewed a few bits to the side
[18:15]	Tsuno Soyinka:	Too bad I get hurt when I try to help
[18:15]	Andrew Linden:	that is the main blocker for H4+MONO
[18:15]	Arawn Spitteler:	That puts MONO under an alpha stage of testing.
[18:15]	Sidewinder Linden:	thanks guys...
[18:15]	Tsuno Soyinka:	hmm where was the unsit thing there?