User:Andrew Linden/Office Hours/2008 01 24
Jump to navigation
Jump to search
Transcript of Andrew Linden's office hours:
| [17:03] | Kitto Flora: | Whats the bug? |
| [17:03] | Ryozu Yamamoto: | It's a known bug but internal only |
| [17:03] | Ryozu Yamamoto: | Because it's super easy to exploit and crash nearly any script |
| [17:04] | Kitto Flora: | OH |
| [17:04] | Kitto Flora: | OK |
| [17:04] | Kitto Flora: | I've had a couple other engins lock up like that |
| [17:04] | Andrew Linden: | That vehicle session was hectic |
| [17:04] | Andrew Linden: | not nearly enough time |
| [17:04] | Gaius Goodliffe nods. | |
| [17:04] | Kitto Flora: | Good turn out then! |
| [17:04] | Ryozu Yamamoto: | Did I miss something? |
| [17:04] | Gaius Goodliffe: | I wrote up notecards ahead of time because I was half expecting that... |
| [17:04] | Ryozu Yamamoto: | Heya Simon! |
| [17:04] | Kitto Flora: | Hello, hello, hello, hello |
| [17:05] | Ryozu Yamamoto: | I miss 1 meeting, I miss everything huh? |
| [17:05] | Simon Linden: | Hi all! |
| [17:05] | Gaius Goodliffe: | Luckily, my problems are easy to explain. :) |
| [17:05] | Andrew Linden: | We held a special "vehicle debugging" session over in the Havok4 preview during the hour before this hour |
| [17:05] | Ryozu Yamamoto: | No kidding |
| [17:05] | Arawn Spitteler is sure Ryo got the memo, that everyone but he was to atend the vehicle tests | |
| [17:05] | Ryozu Yamamoto: | Did Cubey Terra show up? |
| [17:05] | Andrew Linden: | we'll have to hold another one |
| [17:05] | Andrew Linden: | after we fix some vehicle bugs |
| [17:06] | Andrew Linden: | yes, Cubey was there |
| [17:06] | Ryozu Yamamoto: | Awesome |
| [17:06] | Ryozu Yamamoto: | Cubey has a lot riding on getting vehicles fixed |
| [17:06] | Gaius Goodliffe: | I suspect some bugs may be obscuring others... so, yeah, another session will probably be needed after this set of bugs gets fixed... |
| [17:06] | Ryozu Yamamoto: | That's usually the case Gaius.. such was true with a few other vehicle issues |
| [17:06] | Ryozu Yamamoto: | Such as odd gravity |
| [17:07] | Ryozu Yamamoto: | Err, not gravtiy |
| [17:07] | Ryozu Yamamoto: | Center of blance |
| [17:07] | Gaius Goodliffe: | Buoyancy is still a bit off... |
| [17:07] | Ryozu Yamamoto: | What do you mean Gaius? |
| [17:07] | Andrew Linden: | vehicle buoyancy or llSetBuoyancy()>? |
| [17:07] | Gaius Goodliffe: | Should have mentioned that in my notecard. Ah well, that one's minor compared tot he ones I mentioned. |
| [17:07] | Gaius Goodliffe: | Vehicle buoyancy. |
| [17:08] | Rex Cronon: | hi everybody |
| [17:08] | Ryozu Yamamoto: | Heya |
| [17:08] | Rex Cronon: | what happenend on the beta, with the boats? |
| [17:08] | Rex Cronon: | it crashed me |
| [17:08] | Rex Cronon: | i had to turn power off manually |
| [17:08] | Andrew Linden: | I've got a boat with a buoyancy problem that I'm going to try to fix |
| [17:08] | Andrew Linden: | could be the same issue |
| [17:09] | Gaius Goodliffe: | Mine is, all my airships slowly descend now. Under H1, they either are stationary or actually rise a bit when you aren't moving. |
| [17:09] | Andrew Linden: | Gaius, htey continually descend? |
| [17:09] | Gaius Goodliffe: | Yup. |
| [17:10] | Andrew Linden: | That is... tha sag just a little bit and stop? Or slowly fall? |
| [17:10] | Gaius Goodliffe: | As soon as physics goes on, they start slowly falling... |
| [17:10] | Andrew Linden: | ok |
| [17:10] | Andrew Linden: | I wonder if thatis related to what simon found in that there were two different constants being used for gravity |
| [17:11] | Andrew Linden: | although, they were only off by less than 1% |
| [17:11] | Andrew Linden: | Simon, you fixed that in the current version of Havok4? |
| [17:11] | Sidewinder Linden: | /long time no see guys ;) |
| [17:11] | Simon Linden: | I fixed that on Tuesday, I think |
| [17:11] | Gaius Goodliffe: | hehe |
| [17:11] | Andrew Linden: | Ok |
| [17:12] | Simon Linden: | Let me look - maybe I can figure out if it made it into the latest push |
| [17:12] | Rex Cronon: | hi andrew |
| [17:12] | Andrew Linden: | I wonder if that was related to the problem with llSetBuoyancy(). I'll have to retest that. |
| [17:12] | Rex Cronon: | btw what happend with that boat on beta, it crashed me hard |
| [17:12] | Andrew Linden: | Hello Rex |
| [17:13] | Andrew Linden: | I'm not sure Rex. I'm not sure which boat was yours. |
| [17:13] | Rex Cronon: | i didn't have a boat |
| [17:13] | Rex Cronon: | i just saw one tilting to one side than i crashe |
| [17:13] | Simon Linden: | Andrew - the gravity fix went in about an hour before you did the version number, so it should be in the build we just pushed, right? |
| [17:13] | Rex Cronon: | sorry, i have lage |
| [17:14] | Rex Cronon: | lag* |
| [17:14] | Andrew Linden: | That sounds right Simon. |
| [17:14] | Simon Linden: | I can confirm easily - brb |
| [17:15] | Andrew Linden: | ok, well the only announcement I've got is that I think I've got better avatar speed limits in place for the next update. |
| [17:15] | Andrew Linden: | I've raised it to 256 for all modes |
| [17:15] | Ryozu Yamamoto: | Awesome |
| [17:15] | Sidewinder Linden: | cool |
| [17:15] | Sidewinder Linden: | /<- has a couple when we're done with everything else |
| [17:15] | Andrew Linden: | but use a gradually increasing backpressure to help balance the avatar's speed -- much more like Havok1 |
| [17:16] | Ryozu Yamamoto: | Sounds good |
| [17:16] | Andrew Linden: | however I only had one jetpack to test -- I tuned it for that one which normally hit 64 m/sec on Havok1 |
| [17:16] | Simon Linden: | Nope, the gravity fix doesn't apper to be in this version |
| [17:16] | Andrew Linden: | The omniphase appears to work right |
| [17:16] | Andrew Linden: | at least, the older one I had on hand to test |
| [17:16] | Andrew Linden: | Well, that is kinda good news Simon. |
| [17:17] | Ryozu Yamamoto: | Indeed, the OmniPhaze works great now |
| [17:17] | Gaius Goodliffe nods. | |
| [17:17] | Andrew Linden: | There is hope that it may actually fix something. |
| [17:17] | Ryozu Yamamoto: | There's a few quirks but I'm more than happy with them |
| [17:17] | Andrew Linden: | That is, it may fix some of the known vehicle bugs. |
| [17:17] | Simon Linden: | yes, it might be behind some of that slow drifting |
| [17:17] | Gaius Goodliffe: | If it fixes the slow sinking, I'll be happy. :) |
| [17:18] | Andrew Linden: | Well, the changes I'm talking about Ryozu have not yet been depoyed. So you'll be needing to test that omniphaze at least once more. |
| [17:18] | Ryozu Yamamoto nods. | |
| [17:18] | Andrew Linden: | So, that wasthe only announcement I had. Except for the fact that I'll probably have another vehicle debug/demo session next week, same time. |
| [17:18] | Sidewinder Linden: | sry guys sort of afkish for a few |
| [17:19] | Sidewinder Linden: | i'll brb |
| [17:19] | Andrew Linden: | Assuming I fix some bugs first. |
| [17:19] | Ryozu Yamamoto: | Thankfully, with the fix to hard landing mode, speed issues seem to be a lot less of a problem for me -personally- |
| [17:19] | Sidewinder Linden: | ok back |
| [17:19] | Rex Cronon: | wb |
| [17:19] | Andrew Linden: | Oh yeah.... kitto I was working on that performance problem today that you've been talking about |
| [17:19] | Ryozu Yamamoto: | Ouch |
| [17:19] | Andrew Linden: | oops... where did kitto go? |
| [17:20] | Ryozu Yamamoto chuckles | |
| [17:20] | Sidewinder Linden: | he must have found the new black hole feature :) |
| [17:20] | Gaius Goodliffe: | hehe |
| [17:20] | Gaius Goodliffe: | Is that was sucks you from 670 to 500m in a second? :p |
| [17:20] | Andrew Linden: | Kitto, did you hear what I just said about the vehicle performance problem? |
| [17:20] | Rex Cronon: | a singularity in sl? |
| [17:21] | Sidewinder Linden: | / :) |
| [17:21] | Gaius Goodliffe: | Rex: Apparently. I can't fly around inside my castle. If I try to fly horizontally, I get thrown through the floor... |
| [17:21] | Kitto Flora: | I didnt hear stuff that seemed significant to me |
| [17:21] | Gaius Goodliffe: | (It should be noted I built my castle at 640m) |
| [17:21] | Rex Cronon: | interesting |
| [17:21] | Andrew Linden: | I was saying that I've started to track down the vehicle performance problem |
| [17:21] | Sidewinder Linden: | high altitude flight feather misbehavior is on the docket |
| [17:21] | Kitto Flora: | Well yes |
| [17:21] | Andrew Linden: | I've seen the smoke, and located a small spark of a fire. |
| [17:22] | Sidewinder Linden: | hehe |
| [17:22] | Arawn Spitteler noticed Kitto crash, just before Andrew said: "Oh yeah... kitto" | |
| [17:22] | Kitto Flora: | Heard thatAnd will have another test session |
| [17:22] | Andrew Linden: | I'm not sure it is the real source of all the lag |
| [17:22] | Kitto Flora: | if more fixed |
| [17:22] | Andrew Linden: | but I'm going to investigate further |
| [17:22] | Andrew Linden: | and I have a few more clues to go on |
| [17:22] | Andrew Linden: | it appears to NOT be in the vehicle action code |
| [17:22] | Kitto Flora: | I spect the vehicles show more up |
| [17:22] | Andrew Linden: | it is probably deeper down in the physics engine or our own abstract |
| [17:22] | Gaius Goodliffe nods. | |
| [17:22] | Rex Cronon: | btw, why is crash still offline? |
| [17:23] | Gaius Goodliffe: | It seems to affect vehicles that don't use llSetVehicleType, so yeah... |
| [17:23] | Andrew Linden: | ? Rex. Are you talking about "crash me"? |
| [17:23] | Kitto Flora: | Whats deper? The extra cpu use? |
| [17:23] | Rex Cronon: | yes, crash me |
| [17:23] | Andrew Linden: | kitto, the source of the lag... it appears to be time spent in the physics step, but NOT in the vehicle code specifically |
| [17:24] | Andrew Linden: | so I expect I'll have to look harder |
| [17:24] | Kitto Flora: | OK. It does appear when one sets the object physical. But it does not vary with motion |
| [17:24] | Ryozu Yamamoto chuckles | |
| [17:24] | Gaius Goodliffe nods. | |
| [17:25] | Ryozu Yamamoto points: Might want to grab a copy of that one Andrew | |
| [17:25] | Gaius Goodliffe: | I noticed that too -- the lag is there even when stationary. |
| [17:25] | Andrew Linden: | kitto, it is related to the complexity of the object |
| [17:25] | Ryozu Yamamoto: | If you haven't already |
| [17:25] | Kitto Flora: | I dont think so |
| [17:25] | Andrew Linden: | at least, the source of lag I was looking at |
| [17:25] | Kitto Flora: | I just see the jump when the object is set physical |
| [17:26] | Andrew Linden: | Yes, but I copied the offending vehicle script and put it on a simpler object |
| [17:26] | Andrew Linden: | and faile to repro the lag |
| [17:26] | Kitto Flora: | The F20 is not very complicated, it has same load as steam engine. |
| [17:26] | Andrew Linden: | and i was able to put it on a few different vehicles |
| [17:26] | Kitto Flora: | Oh |
| [17:26] | Andrew Linden: | er... vehicle bodies |
| [17:26] | Kitto Flora: | Less prims? |
| [17:26] | Andrew Linden: | more and less, in varying degrees |
| [17:27] | Kitto Flora: | Are you saying that prim count seems to affect the lag? |
| [17:27] | Andrew Linden: | but mostly I was checking multi-concave prims against much simpler objects |
| [17:27] | Gaius Goodliffe: | Ah. Shape... |
| [17:27] | Andrew Linden: | yes, in the vehicle lag I was looking at |
| [17:27] | Ryozu Yamamoto: | Or more precisely, collisions |
| [17:27] | Kitto Flora: | I can check if I see such an effect |
| [17:27] | Gaius Goodliffe: | Oh dear... maybe I shouldn't hollow out my prims so much... |
| [17:28] | Andrew Linden: | no, it was not caused by collisions |
| [17:28] | Andrew Linden: | that is, it wasn't caused by any collision callbacks |
| [17:28] | Ryozu Yamamoto: | No? That's what I would have assumed |
| [17:28] | Andrew Linden: | I had the lag when running a hover vehicle high above flat terrain |
| [17:28] | Andrew Linden: | and I had breakpoints in the collision handlers |
| [17:28] | Ryozu Yamamoto: | So not actual collisions |
| [17:29] | Andrew Linden: | Right. In any case, that is what I'm currently working on. |
| [17:29] | Andrew Linden: | But tomorrow I have to move my desk and computers over to the other building |
| [17:29] | Andrew Linden: | I don't expect to get much done tomorrow |
| [17:29] | Ryozu Yamamoto nods. | |
| [17:29] | Andrew Linden: | and it has been a distracting week, for moving stuff |
| [17:29] | Andrew Linden: | but I did fix the Galaga arcade cabinet |
| [17:30] | Sidewinder Linden: | hehe :) |
| [17:30] | Ryozu Yamamoto: | :O |
| [17:30] | Sidewinder Linden: | i have a few things, if that's ok |
| [17:30] | Sidewinder Linden: | hi kitto |
| [17:30] | Sidewinder Linden: | again |
| [17:30] | Gaius Goodliffe: | :D I haven't played that since the 80's(I think... |
| [17:30] | Andrew Linden: | go ahead Sidewinder |
| [17:30] | Sidewinder Linden: | i figured i'd do a quick update on the early adopter program and what we're learning so far |
| [17:30] | Kitto Flora: | Andrew, it appears to me that prim count makes a big difference to the Sim Time (Physics) |
| [17:30] | Sidewinder Linden: | we now have over 320 regions running havok4 |
| [17:31] | Andrew Linden: | Ryozu, I want to be able to kick my feet up on the table. |
| [17:31] | Ryozu Yamamoto: | Kitto: I'm guessing complexity over count |
| [17:31] | Sidewinder Linden: | hehe kitto is now my official straight man |
| [17:31] | Sidewinder Linden: | so... |
| [17:31] | Gaius Goodliffe: | Wow nice. |
| [17:31] | Ryozu Yamamoto: | XD |
| [17:31] | Sidewinder Linden: | we're not quite yet pushing on performance optimization |
| [17:31] | Gaius Goodliffe: | How many sims are there these days anyhow? |
| [17:31] | Sidewinder Linden: | given as you've just seen the number of vehicle issues to straighten out |
| [17:31] | Sidewinder Linden: | we have over 320 regions up with havok4 |
| [17:31] | Andrew Linden: | it could just be prim-count. I can test and verify that |
| [17:31] | Sidewinder Linden: | out of those only a very few, i think under six |
| [17:31] | Kitto Flora: | around 15,000 total sims? |
| [17:31] | Sidewinder Linden: | had problems with content on startup |
| [17:32] | Sidewinder Linden: | total's around 16k i think |
| [17:32] | Rex Cronon: | oh the main grid. right? |
| [17:32] | Sidewinder Linden: | right |
| [17:32] | Gaius Goodliffe: | So about 2% then eh? cool. |
| [17:32] | Sidewinder Linden: | in general people are pretty happy, although as you'd expect with these scripting weirdies somet hings are notworking |
| [17:32] | Sidewinder Linden: | some regions that are historically griefed really heavily are very happy |
| [17:32] | Sidewinder Linden: | as they're not crashing |
| [17:33] | Rex Cronon: | i wonder which H4 sims, are public? |
| [17:33] | Sidewinder Linden: | one of our governance team folks im'd me late last night and said... "have griefing attacks gone way down on these regions (name witheld ;) " |
| [17:33] | Sidewinder Linden: | and i said umm no, but they're running havok4, and the residents are saying that most of the toys have little effect any more and are easy to clean up ;) |
| [17:33] | Andrew Linden: | must... deliver... Havok4... |
| [17:34] | Gaius Goodliffe: | heheh |
| [17:34] | Sidewinder Linden: | we're not publishing a full list of who's running havok4 since many current early adopters do not want to be over-run with folks playing with it |
| [17:34] | Gaius Goodliffe nods. "I've just been using this sim as my sandbox." ;) | |
| [17:34] | Sidewinder Linden: | so the good news is that the crash rate does seem to be subjecively much lower with havok4 |
| [17:34] | Kitto Flora: | How about H4 on MG on OI Public and HI Public? |
| [17:34] | Ryozu Yamamoto gave you Legs Up. | |
| [17:34] | Sidewinder Linden: | and our data warehouse folks are tracking so that we can measure this in an objective way |
| [17:34] | Kitto Flora: | They are frequet griefing tragets |
| [17:34] | Ryozu Yamamoto: | Ah nuts |
| [17:34] | Sidewinder Linden: | we aren't putting h4 on the main grid yet |
| [17:35] | Sidewinder Linden: | umm see my post #59 on the last deploy announcement for why (just had to answer that question a few minutes ago :) |
| [17:35] | Sidewinder Linden: | there are some performance issues that we've seen in high load regions,and will be looking at that |
| [17:36] | Sidewinder Linden: | and there is a problem with "falling through floor" under high load, that used to exist in havok1, and seems a bit more pronounced now... it's being looked at |
| [17:36] | Rex Cronon: | it seems it is still possible to crash h4 sims, as the crash me is down |
| [17:36] | Sidewinder Linden: | so we've got a good amount to go, but are starting to see something like light :) |
| [17:36] | Sidewinder Linden: | umm i think it's back up - or was |
| [17:36] | Sidewinder Linden: | there is a crash me |
| [17:36] | Sidewinder Linden: | and a havok4 sandbox dmg, and a havok4 sandbox nodmg |
| [17:37] | Sidewinder Linden: | there are still ways to crash a havok4 region, but it takes much more work, and leaves big fingerprints ;) |
| [17:37] | Sidewinder Linden: | and they're not physics crashes |
| [17:37] | Sidewinder Linden: | another team will be taking care of those asap |
| [17:37] | Gaius Goodliffe: | Ah... my favorite kind of problem: sombody else's... ;) |
| [17:37] | Andrew Linden: | That means fewer crashes by accident |
| [17:37] | Sidewinder Linden: | btw there are very few ways known... and not many folks will find those ways |
| [17:38] | Sidewinder Linden: | as andrew said, many fewer incidental crashes... |
| [17:38] | Sidewinder Linden: | i don't think we've seen many if any incidental crashes ina while |
| [17:38] | Sidewinder Linden: | the early adopter program will tell us a lot about that |
| [17:38] | Sidewinder Linden: | hopefully within a week or two |
| [17:38] | Sidewinder Linden: | that's what i've got... any other questions? |
| [17:39] | Ryozu Yamamoto: | I just have two things to bring up |
| [17:39] | Ryozu Yamamoto: | The first, is just a little bug you guys might want to fix for fear of being lynched ;) |
| [17:39] | Sidewinder Linden: | uhhhhh |
| [17:39] | Ryozu Yamamoto: | A little while back this "Misfeature" was fixed and there was quite an uproar |
| [17:40] | Andrew Linden: | I'd have expected the lynch mobs to be forming already |
| [17:40] | Sidewinder Linden: | this sounds like a good one, kitto :) |
| [17:40] | Andrew Linden: | but go on |
| [17:40] | Kitto Flora waits for it | |
| [17:40] | Sidewinder Linden: | drum roll |
| [17:40] | Ryozu Yamamoto: | In current H1 behavior, llSetLinkPrimParams can affect the position of an avatar sitting on the linkset |
| [17:40] | Ryozu Yamamoto: | Which appears to be broken in H4 now |
| [17:40] | Silicon Plunkett: | I thought that was removed? |
| [17:40] | Andrew Linden: | Ah, that is one of the bugs reported for one of the boats |
| [17:40] | Ryozu Yamamoto: | They did a rolling restart to fix that after it was "fixed" the first time |
| [17:41] | Andrew Linden: | Owen said it was broken then fixed back in August, but now Havok4 has the bug |
| [17:41] | Ryozu Yamamoto: | When it was found out that a -lot- of people use that |
| [17:41] | Gaius Goodliffe nods. "Creem mentioned that bug earlier today, too. Apparently a lot of people are noticing it, so it'll be a good sized lynch mob." | |
| [17:41] | Silicon Plunkett: | since setting the PRIM_TYPE of an agent crashed the sim |
| [17:41] | Andrew Linden: | which kinda makes sense, since much of that code was rewritten |
| [17:41] | Andrew Linden: | the fix probably went into the old code |
| [17:41] | Andrew Linden: | we'll have to track it down |
| [17:41] | Ryozu Yamamoto: | Right Silicon, it was never intended to be a feature of SetLinkPrimParams |
| [17:41] | Ryozu Yamamoto: | It just happened to be a widely used bug |
| [17:42] | Sidewinder Linden: | but like any thing that gets used whether it was meant to be or not, like those large "m" things *wink* |
| [17:42] | Silicon Plunkett: | yep, and "fixing" it broke a lot of products.. plenty of outrage |
| [17:42] | Ryozu Yamamoto chuckles | |
| [17:42] | Sidewinder Linden: | well it seems that "Unfixing" it would fix several open vehicle bugs.. so... |
| [17:42] | Ryozu Yamamoto: | So it's supposed to be working now. That blue teleport pad thingy there uses it |
| [17:43] | Silicon Plunkett: | hmm |
| [17:43] | Silicon Plunkett: | so it is |
| [17:43] | Simon Linden: | hey folks, sorry but I have to run ... see you next time |
| [17:43] | Andrew Linden: | later simon |
| [17:43] | Rex Cronon: | bye simon |
| [17:43] | Simon Linden: | bye |
| [17:43] | Kitto Flora: | Byebye Simon |
| [17:43] | Sidewinder Linden: | thanks simon... see you tomorrow |
| [17:43] | Ryozu Yamamoto: | It just moves you up or down a few meters after choosing which way to go in a dialog |
| [17:44] | Andrew Linden: | Wait, lemme get this clear... this was a "misfeature" |
| [17:44] | Ryozu Yamamoto: | That's the term Aric likes to use ;) |
| [17:44] | Andrew Linden: | and then we fixed it (broke a lot of content) |
| [17:44] | Andrew Linden: | so we rolled it back? |
| [17:44] | Ryozu Yamamoto: | It wasn't so much rolled back, as it was made to work again sans crash |
| [17:44] | Silicon Plunkett: | well, it was "fixed" but killing the use of the function on agents in general |
| [17:45] | Ryozu Yamamoto: | Limiting it's use to PRIM_ROTATION and PRIM_POSITION I believe |
| [17:45] | Andrew Linden: | yes, a "misfeature" is a bug that users consider a feature |
| [17:45] | Ryozu Yamamoto: | on Agents |
| [17:45] | Silicon Plunkett: | whereas a check could've been done against setting an agents PRIM_TYPE |
| [17:45] | Andrew Linden: | so now it is a real feature |
| [17:45] | Gaius Goodliffe: | hehe |
| [17:45] | Andrew Linden: | and hence a real bug in Havok4 |
| [17:45] | Ryozu Yamamoto nods. | |
| [17:45] | Ryozu Yamamoto: | The other question is a bit of a policy thing |
| [17:45] | Ryozu Yamamoto: | I know Philip is eager to get H4 out, as I'm sure you guys are as well |
| [17:46] | Sidewinder Linden: | go for it ryozu ;) |
| [17:46] | Arawn Spitteler rolls a drum | |
| [17:46] | Sidewinder Linden: | *waits with baited breath* |
| [17:46] | Ryozu Yamamoto: | I know with the way things are going, retuning scripts won't be needed much, the team is doing great to get things in line with Havok1 behavior |
| [17:46] | Ryozu Yamamoto: | However, there's sure to be some cases where it'll be needed |
| [17:46] | Sidewinder Linden: | true |
| [17:47] | Sidewinder Linden: | and... |
| [17:47] | Andrew Linden: | If havok4 does not ship this quarter I may go insane |
| [17:47] | Ryozu Yamamoto: | What I'd like to ask is that once the Havok4 code is "frozen" and ready for gridwide rollout |
| [17:47] | Gaius Goodliffe: | You mean you aren't already? I thought we were all mad here. ;) |
| [17:47] | Sidewinder Linden: | *turning blue waiting* |
| [17:47] | Ryozu Yamamoto: | If it's not going out in small chunks, that we have some kind of grace period |
| [17:47] | Andrew Linden: | ... more insane |
| [17:47] | Gaius Goodliffe nods. | |
| [17:47] | Sidewinder Linden: | ok this was asked in the vehicle session |
| [17:47] | Ryozu Yamamoto: | Basically, just have it on the beta grid for a week, and announce in the blog that this is THE release version |
| [17:48] | Ryozu Yamamoto: | Ah, |
| [17:48] | Sidewinder Linden: | here's my take on this... |
| [17:48] | Ryozu Yamamoto: | Bugger, I miss everything |
| [17:48] | Sidewinder Linden: | and andrew please feel free to chime in... |
| [17:48] | Sidewinder Linden: | we will hopefully have most of the vehicles issues resolved at some point not too far from now |
| [17:48] | Ryozu Yamamoto: | I just think many creators (And their customers for that matter) would be greatful for a little time to tune what they have to |
| [17:48] | Sidewinder Linden: | there are still some performance tuning and other boundary case issues |
| [17:48] | Sidewinder Linden: | my assumption was that we'd have a build that is functionally complete from a vehicle scripting perspective |
| [17:49] | Sidewinder Linden: | and still need to maybe do some performance tuning |
| [17:49] | Rex Cronon: | how is volumedetect on attachments, doing? |
| [17:49] | Silicon Plunkett: | any plans on fixing object interpolation issues? because at the moment.. things just seem to.. explode |
| [17:49] | Sidewinder Linden: | so that may give you a week or so of clearance anyway...without a "policy" to support it... ? |
| [17:49] | Andrew Linden: | It was my intention to give the content creators time to test. There are many who cannot be bothered to test on the preview, but when we start making noise about deploying it some of them will take the time. |
| [17:49] | Ryozu Yamamoto nods. | |
| [17:49] | Gaius Goodliffe: | Did we ever get some way to tell in a script what Havok version is running? |
| [17:50] | Andrew Linden: | This whole early adoptor process will stretch out the transition |
| [17:50] | Ryozu Yamamoto: | Gaius: I think that was determined to be outside the scope of the changes needed |
| [17:50] | Gaius Goodliffe nods. | |
| [17:50] | Andrew Linden: | and once we think we're ready for the final version, we'll still have to test it internally for a week or so |
| [17:51] | Andrew Linden: | I expect there to be at least a week, probably more, where content creators can test and possibly pull a veto string |
| [17:51] | Ryozu Yamamoto: | I just come from the standpoint that, I stressed over needing to tune my products toward the beginning. Had I done a lot of changing around to the scripts, it would have been wasted time since at this point, no tuning would have been needed |
| [17:51] | Rex Cronon: | if a major bug is found, will it be rolled back, and the bug will be fixed, and that will be realeased again? |
| [17:51] | Andrew Linden: | ah I see, a tuning period, not a debugging perios |
| [17:51] | MagicExpandingTable: Hello, Avatar! | |
| [17:51] | Andrew Linden: | period |
| [17:52] | Rex Cronon: | and after that* |
| [17:52] | Ryozu Yamamoto: | I think a lot of creators would rest easier if they knew they had a grace period, where they knew for sure what needed tuning and what didn't |
| [17:52] | Sidewinder Linden: | i understand why that would be so... |
| [17:52] | Sidewinder Linden: | so... |
| [17:52] | Sidewinder Linden: | here's my proposal... |
| [17:52] | Sidewinder Linden: | hmm actually... |
| [17:53] | Sidewinder Linden: | when this goes to be folded in to the standard simulator |
| [17:53] | Sidewinder Linden: | i'm sure we'll want to go through it pretty carefully |
| [17:53] | Andrew Linden: | well, suppose we raised the Havok4 deployment to 5% of the regions in the main grid. I'd expect quite a few content owners to start retuning their content |
| [17:53] | Sidewinder Linden: | and the likelihood is that like other big releases we'd do an internal code freeze |
| [17:53] | Sidewinder Linden: | at that point we could mention this publicly as teh notice that "now's the time for your final tuning"... does that work? |
| [17:54] | Gaius Goodliffe: | Once we know the code is frozen, we can tune, yeah... |
| [17:54] | Ryozu Yamamoto: | I think that would be appropriate |
| [17:54] | Sidewinder Linden: | btw we all know that a code "freeze" is more like a valiant attempt, and that there can be changes to fix things after that, right? :) |
| [17:54] | Rex Cronon: | . |
| [17:54] | Ryozu Yamamoto: | Hehe, right ;) |
| [17:54] | Gaius Goodliffe: | lol |
| [17:54] | Gaius Goodliffe: | Code never freezes solid, it just gets a little slushy... |
| [17:54] | Ryozu Yamamoto: | But a freeze is the intent to have things working as they are |
| [17:54] | Ryozu Yamamoto: | tuning and featurewise |
| [17:55] | Sidewinder Linden: | but by that time, weshould know which functions in lsl are supposed to work when shipped for instance |
| [17:55] | Sidewinder Linden: | i think we should revisit this in 2-3 weeks, ok? |
| [17:55] | Ryozu Yamamoto nods. | |
| [17:55] | Sidewinder Linden: | by then we'll hopefully be much closer and see what is really left to work on |
| [17:55] | Ryozu Yamamoto: | That's all I had on my plate |
| [17:55] | Sidewinder Linden: | as i've said before, i think a lot of these issues are tied to a smaller number of bugs... and i really hope i'm right |
| [17:56] | Sidewinder Linden: | which makes it likely that we'll have a much lighter queue in a couple of weeks |
| [17:56] | Ryozu Yamamoto: | Aye, just one bug can have a large number of side effects |
| [17:56] | Sidewinder Linden: | yup |
| [17:57] | Sidewinder Linden: | if you listen to the issues raised in the vehicle session, you hear the same sorts of things over many times... couple dozen issues, but sounds like only a few root causes to me |
| [17:57] | Gaius Goodliffe nods. | |
| [17:57] | Gaius Goodliffe: | I saw a number of problems that look for all the world to me to be related to my bump the ground issues... |
| [17:57] | Andrew Linden: | 3 or 4 vehicle bugs I would guess |
| [17:57] | Sidewinder Linden: | cool |
| [17:57] | Sidewinder Linden: | that's in the range i thought, but hoped i wasn't being hopelessly optimistic :) |
| [17:58] | Andrew Linden: | I take that back... at least 4. |
| [17:58] | Sidewinder Linden: | hehe |
| [17:58] | Sidewinder Linden: | well i'm really guessing somewhere under six |
| [17:58] | Kitto Flora: | Did looking at vehicles in Beta Grid today include the GSLR trolley? |
| [17:58] | Sidewinder Linden: | fwiw :) |
| [17:58] | Andrew Linden: | No Kitto. It was rather hectic. Not nearly enough time. |
| [17:58] | Sidewinder Linden: | nope - it turned out we had 11 people and 60 mintues :) |
| [17:59] | Kitto Flora: | Wow, thats a lot |
| [17:59] | Sidewinder Linden: | 12 atually |
| [17:59] | Sidewinder Linden: | yeah the attendance was about 2x what i thought it would be when i woke up today |
| [17:59] | Andrew Linden: | I took about 6 bugged vehicles to inventory. |
| [17:59] | Gaius Goodliffe: | Oh, I couldn't really demo my elevator issue [SVC-1202] but I assume you can pop over to the Atoll when you get around to it. It's easy enough to see. |
| [18:00] | Arawn Spitteler: | My primary issue is a known misfeature, but I'm wondering how my physics toy sank into the ground, if it was to be regurgitated. |
| [18:01] | Arawn Spitteler: | It looked more like a Havoc4 issue, than havoc 4 Vehicle issue |
| [18:01] | Sidewinder Linden: | /recounting... 14 people were there |
| [18:01] | Andrew Linden: | Arawn, your primary issue is a Havok1 misfeature that is fixed in Havok4? Or a Havok4 misfeature you hope won't be fixed? |
| [18:02] | Arawn Spitteler: | llSetStatus( STATUS_ROTATE_Y, FALSE) doesn't work |
| [18:02] | Kitto Flora: | Owww |
| [18:02] | Ryozu Yamamoto: | Ah yeah,those |
| [18:02] | Rex Cronon: | oops |
| [18:02] | Rex Cronon: | sorry |
| [18:02] | Kitto Flora: | Thats *weird* |
| [18:02] | Rex Cronon: | i turned off my shield |
| [18:02] | Andrew Linden: | Huh... Simon fixed a STATUS_ROTATE_ bug more than a week ago, I think |
| [18:02] | Sidewinder Linden: | arawn isthere a jira on that |
| [18:03] | Sidewinder Linden: | i'm not sure, but i'm sure someone else complained of that as well |
| [18:03] | Arawn Spitteler: | Appearently, it's a Linden Side Hack, but a documented use for Users |
| [18:03] | Sidewinder Linden: | *not sure who |
| [18:03] | Arawn Spitteler: | I did before, so I might be the only one |
| [18:03] | Andrew Linden: | That sounds like a "bug" rather than a "misfeature". It should be working. |
| [18:03] | Andrew Linden: | And is not very usefull when it doesn't work. |
| [18:03] | Andrew Linden: | a misfeature is a bug that is useful |
| [18:03] | Arawn Spitteler: | I haven't been able to use Vehicles, yet, and the Physical Uprightness doesn't work for Stools |
| [18:04] | Sidewinder Linden: | arawn - do you by chance have a jira #? |
| [18:04] | Andrew Linden: | one thing about the STATUS_ROTATE_* feature... |
| [18:04] | Arawn Spitteler: | I was tumbling all over, and rolling on my head. I thought I looked a nuisance |
| [18:04] | Andrew Linden: | you almost never want to prevent rotations about one axis |
| [18:04] | Andrew Linden: | you usually want to disable rotations about two axes |
| [18:05] | Andrew Linden: | or more suscinctly... only enable rotations about one axis |
| [18:05] | Ryozu Yamamoto: | Or all 3 |
| [18:05] | Arawn Spitteler: | I've been using llSetStatus(12,FALSE) a lot |
| [18:05] | Gaius Goodliffe: | Hmm... usually I want to stop just one... |
| [18:05] | Andrew Linden: | the problem is that rotations do not commute, in general |
| [18:05] | Andrew Linden: | so enabling rotations about only two axes causes some very weird behavior |
| [18:05] | Gaius Goodliffe: | A turret, for example, needs to pitch and yaw, but never roll. |
| [18:05] | Ryozu Yamamoto: | Such as this pyramid on the table here (which is failing a lock on all 3) |
| [18:05] | Kitto Flora: | Well thats got issues |
| [18:05] | Sidewinder Linden: | andrew - i'm almost sure that someone else said they needed that to work with one axis in conversation inthe last couple of days - i just can't remember who or what product it was... |
| [18:06] | Gaius Goodliffe: | That reminds me of something I never tested... |
| [18:06] | Gaius Goodliffe: | Oh that's right, this thing isn't physical anyway... |
| [18:06] | Arawn Spitteler notices the Asset Servers are laggy in this world also | |
| [18:07] | Andrew Linden: | the pitch and yaw of a gun turret is not a case where only two axes of rotation are enabled |
| [18:07] | Andrew Linden: | I'm going to have to go soon. |
| [18:07] | Arawn Spitteler: | So, I should Jira the feature change? |
| [18:08] | Andrew Linden: | but I'll look into that bug Arawn. STATUS_ROTATE_Y does not work. |
| [18:08] | Ryozu Yamamoto: | Indeed, it's a pity rotations have to be so, painful |
| [18:08] | Sidewinder Linden: | i just entered the llSetStatus( STATUS_ROTATE_Y, FALSE) doesn't work issue |
| [18:08] | Sidewinder Linden: | on internal |
| [18:08] | Ryozu Yamamoto: | Andrew: none of the Status Rotate false work |
| [18:08] | Sidewinder Linden: | and also noted that we should check all three axes ;) |
| [18:08] | Andrew Linden: | ok cool, no need to enter it Arawn |
| [18:08] | Sidewinder Linden: | ryo |
| [18:08] | Arawn Spitteler: | This has a work around, but I'd rather be a purist |
| [18:09] | Arawn Spitteler: | It's using llRotLookAt() |
| [18:09] | Arawn Spitteler: | Not much control about how much turning it'll decide to do. |
| [18:09] | Ryozu Yamamoto: | Hmm? |
| [18:09] | Electromagnetodynamic Vehicle whispers: CCC Ready. | |
| [18:10] | Andrew Linden: | ah hrm... llRotLookAt() combined with two axes of rotation enabled might provide good behavior... if you fed it the right rotations all the time. |
| [18:10] | Ryozu Yamamoto: | Well that's odd |
| [18:10] | Ryozu Yamamoto: | Now it IS working |
| [18:10] | Andrew Linden: | I took a copy of the stool Arawn. |
| [18:10] | Arawn Spitteler: | Can you read the script, or should you take another? |
| [18:10] | Ryozu Yamamoto: | I swear it wasn't working a moment ago |
| [18:11] | Rex Cronon: | i don |
| [18:11] | Andrew Linden: | Arawn I'll be able to read the script, I think |
| [18:11] | Arawn Spitteler: | I'm thinking of publishing these scripts with llSetText |
| [18:11] | Rex Cronon: | i don't know if rifght now is really a good time to test tings, as searching for sims in map doesn't seem to be working |
| [18:12] | Arawn Spitteler: | Vehicles seem to complicated |
| [18:12] | Sidewinder Linden: | well i'm going to have to run guys... as well |
| [18:12] | Ryozu Yamamoto: | Take care guys |
| [18:12] | Rex Cronon: | bye sidenwinder |
| [18:12] | Andrew Linden: | I have to run everyone. Thanks for coming. |
| [18:12] | Kitto Flora: | Byebye Sidewinder |
| [18:12] | Sidewinder Linden: | thanks again for all of your help |
| [18:12] | Rex Cronon: | bye andrew |