Simulator User Group/Transcripts/2011.03.25

From Second Life Wiki
Jump to navigation Jump to search


Prev 2011.03.22 Next 2011.03.28

List of Speakers

Andrew Linden Arawn Spitteler Ardy Lay
Ashiri Sands Bitterly Sweet dantia Gothly
DogWomble Dollinger Dora Gustafson Draconis Neurocam
Flip Idlemind Hope Dreier Horuzlna Basiat
Kadah Coba Liisa Runo Maggie Darwin
Opensource Obscure Pauline Darkfury Qie Niangao
Rex Cronon Script Counter Simon Linden
Siobhan McCallen Sling Trebuchet Stickman
Stickman Ingmann TankMaster Finesmith Void Singer


[16:00] Hope Dreier: Hi Andrew

[16:00] Flip Idlemind: I've arrived!

[16:01] Draconis Neurocam: hello andrew

[16:01] Ardy Lay: My only outstanding service issue was placed into SEC-720 and thanks for getting attention to SVC-6859.

[16:01] JIRA-helper:

[#SVC-6859] llTeleportAgentHome on an agent not over your land shouldn't result in the _target_ getting the message

[16:01] Pauline Darkfury: It gives the can't create plants error on SG 1.5 (the Jira says it doesn't on 1.4), or at least does on Phoenix

[16:01] Andrew Linden: Hey everyone.

[16:01] DogWomble Dollinger: hi andrew, hi flip

[16:01] Pauline Darkfury: Afternoon, Andrew :)

[16:01] Ashiri Sands: hi Andrew

[16:01] Stickman Ingmann: Hi Andrew.

[16:01] Andrew Linden: Simon, it's still taking a long time for the floor of yoru house to show up when I arrive.

[16:01] Ardy Lay: Pauline, when I tried that I was on land where I can place Linden Plants.

[16:02] Hope Dreier: Ok Maggie, that is indeed an ostentatious entry

[16:02] Ardy Lay: Hi Andrew

[16:02] Andrew Linden: Perhaps it is showing up faster? But it is still one of the last things to appear.

[16:02] Pauline Darkfury: ahh, ok, yeah, did wonder if it might be different with landowner privs

[16:02] Hope Dreier laughs

[16:02] Maggie Darwin: Hey, only works where you can rez.

[16:02] Andrew Linden: I don't visit very many regions. I wonder if the object cache system is broken again.

[16:03] Stickman Ingmann: I thought there was no object cache. Just textures.

[16:03] Maggie Darwin: At least I'm not standing in the table and going idle/away like some bim did last time.

[16:03] Simon Linden: Hmm, are you on the latest 2.6.1? We got a few comments that it was working better

[16:03] Siobhan McCallen: I'm having a hard time loading textures today...but I *am* running an alt in Kowloon right now

[16:03] Andrew Linden: No, I'm on 2.5.2.

[16:03] Andrew Linden: I'll upgrade next week and see.

[16:04] Draconis Neurocam: Second Life 2.6.2 (225035) Mar 24 2011 20:33:31 (Second Life Development) is the latest viewer-development anyway

[16:04] Andrew Linden: Ok so news...

[16:04] Simon Linden: 2.5 still has a bug where the viewer initially reports it's first location as 0,0,0 ... now that the server is more efficient, it sends down a bunch of data early from that corner of the sim

[16:04] Andrew Linden: First of all, I left on Tuesday with some action items. I actually did some work on them, but all last minute of course.

[16:05] Andrew Linden: (1) Talk with Richard about animations for Stickman -- this I did. I'll come back to this one.

[16:05] Stickman Ingmann: Yay!

[16:05] Andrew Linden: (2) SVC-6859 llTeleportAgentHome() sends message to target

[16:05] Stickman Ingmann: Thanks.

[16:05] Andrew Linden: Turns out Josh Linden fixed that one this week. It is sitting in my andrew/maint-server repository

[16:06] Ardy Lay: \o/

[16:06] Andrew Linden: so it should be included in the next maint-server RC, whenever that is

[16:06] Hope Dreier: excelent progress

[16:06] Ardy Lay: THanks

[16:06] Andrew Linden: (3) Ask a Linden to weigh in on the per-parcel IP ban feature request

[16:06] Andrew Linden: (which most here considered a bad idea)

[16:07] Andrew Linden: I did approach one Linden who would know more about the feasibility of this idea.

[16:07] Andrew Linden: I learned (not suprisingly) that this is an idea that we've kicked around internally already.

[16:07] Arawn Spitteler: By the time you have to use IP Ban, the guy knows the counter measures

[16:07] Andrew Linden: Proposals with more details had been written.

[16:08] Void Singer: thanks qie

[16:08] Andrew Linden: And that is where they sit. As far as I know no one is working on it.

[16:08] Qie Niangao: :)

[16:08] Rex Cronon: greetings all

[16:08] Horuzlna Basiat: counter measures against what?

[16:09] Pauline Darkfury: The pros and cons are fairly well set out in the Jira for it, at least from those of us that care

[16:09] Hope Dreier: IP bans

[16:09] Andrew Linden: No one wants to read all the comments of the jira issue.

[16:09] Horuzlna Basiat: IP bans for what?

[16:09] Arawn Spitteler: Didn't it used to be possible to contact the ISP, about griefers?

[16:09] Rex Cronon: against griefers i presume

[16:09] Andrew Linden: Sorry, I mean I didn't have time to read all the comments.

[16:09] Horuzlna Basiat: griefers who crash your client?

[16:09] Andrew Linden: There are many

[16:09] Horuzlna Basiat: the whole entire crashing-clients thing can be fixed in 4 lines of code

[16:10] Horuzlna Basiat: I did it and never crashed again from any kind of prim bomb

[16:10] Andrew Linden: (4) check to see if SVC-1443 is easy. I took a peak and I think it is easy but...

[16:10] JIRA-helper:

[#SVC-1443] Avatar attachments should always listen from the avatar's position and not from any object the avatar may be sitting on.

[16:10] Void Singer: I take it weretalkin about the exended ban controls for land owners to include IP/hardwareID on LL's side?

[16:10] Andrew Linden: I believe that linked prims listen from the center of their root prim

[16:10] Andrew Linden: and we probably don't want to change that behavior now... I'm sure something would break

[16:10] Pauline Darkfury: Yeah, pretty sure normal in-world objects listen from the root pos

[16:10] Liisa Runo: jira is about attachments

[16:11] Hope Dreier: @Horuzlna perhaps for one type of client crash

[16:11] Horuzlna Basiat: Not for one

[16:11] Maggie Darwin: Seems to me that "listening from root prim" was deemed a bug and fixed.

[16:11] Andrew Linden: so we'd have to pivot the new behavior on whether it was an attachment or not (which is what is being asked for anyway)

[16:11] Draconis Neurocam: a person sitting on an object becomes a link effectively

[16:11] Horuzlna Basiat: It works for all that I've encountered

[16:11] Maggie Darwin: I recall some spaceships that didn;t work proprly because of it

[16:12] Andrew Linden: Oh yeah... (4) was actually ask Dan about CrashMe region... I didn't get in touch with him yet.

[16:13] Andrew Linden: (5) was SVC-1443

[16:13] Flip Idlemind: Basically if you sit on a big enough megaprim and say something, your attachments won't ehar it

[16:13] Liisa Runo: llGetPos from attachment return the agent pos, so to my logic the listen fromt he same attachments should also listen from agent pos

[16:13] Andrew Linden: (6) was to look into the auto-return behvior to see if it really was making assets that were too big (too big to rez)

[16:13] Andrew Linden: and whether it is easy to fix. I didn't get to this one.

[16:14] Andrew Linden: (7) fin

[16:14] Ardy Lay: We can test that one. ;-)

[16:14] Pauline Darkfury: fair enough, it's a "very nice to have", but not all that frequent problem, the auto-return object size

[16:14] Horuzlna Basiat: I find it hilarious that the Linden's wouldn't even ask what the fix is

[16:14] Horuzlna Basiat: bye

[16:14] Andrew Linden: What I've been working on this week is the mesh project... but not really the meshy parts

[16:15] Maggie Darwin: Ah....SVC-3710 was autoclosed because the whining stopped

[16:15] JIRA-helper:

[#SVC-3710] I would like to know which mainland continent I am on.

[16:15] Maggie Darwin: SVC-3170 rather

[16:15] JIRA-helper:

[#SVC-3170] Listeners in child prims get positioned at root prim position first, then switch to child prim position after re-rez (resulting in wrong listener / whisper radius)

[16:15] Draconis Neurocam: haha

[16:15] Andrew Linden: There is some stuff in mesh that was done quickly and isn't quite working right, such as parcel prim accounting for linked objects

[16:15] Andrew Linden: I'm trying to clean that up

[16:15] Andrew Linden: and also trying to clean up some other code while I'm at it.

[16:16] Andrew Linden: I started to try to make progress toward heirarchical objects with this clean up (modest but non-zero progress)

[16:16] Andrew Linden: but by the time I really got into it I had to pare back my ambitions

[16:16] Hope Dreier: Progress is Progress. Tall Oaks and all that.

[16:16] Andrew Linden: I'm going to just focus on cleaning up parcel resource accounting

[16:17] Pauline Darkfury: what's the 1 line summary for what hierarchical objects gain for us?

[16:17] Andrew Linden: ok that's all for my news.

[16:17] Hope Dreier: Did you get a chance to look into SVC-5927 ?

[16:17] Maggie Darwin: I guess we must pass though "no longer defective" on the way to "written the way it should have been"

[16:17] JIRA-helper:

[#SVC-5927] Temp on Rezzed objects get queued

[16:17] Andrew Linden: Simon if you have any news go ahead.

[16:18] Andrew Linden: No, I haven't looked into SVC-5927.

[16:18] Simon Linden: I don't have much news ... I've been working on our testing system for loading down regions and thus measuring performance

[16:19] Maggie Darwin: "hierarchical objects" is such an insanely deep architectural change that I'm stunned it's even being considered.

[16:19] Maggie Darwin: As good an idea as it is.

[16:20] Andrew Linden: Maggie, I think we can do it, but it is going to be a long haul.

[16:20] Andrew Linden: Stickman, I did talk to Richard about your requests for animation improvements

[16:20] Andrew Linden: I'll list your topics:

[16:20] Andrew Linden: (1) Better accuracy for animations

[16:20] Maggie Darwin: I can well imagine. We have them in OpenWonderland, and it's definately the way to go, openis up so many possibilities

[16:20] Andrew Linden: (2) Feedback about how expensive/resource-intensive animations are

[16:21] Andrew Linden: (3) More control over default animations (disable entirely, or set custom priority levels)

[16:21] Andrew Linden: and (4) Preloading of animations

[16:22] Stickman Ingmann: Hearing them listed, I almost feel like I should have listed the problems they were intended to solve, rather than offering solutions.

[16:22] Andrew Linden: Ok so the general response from Richard is that we need to decide if we want to either (a) hire a dev or a team to work on animations or (b) buy a third party animation engine

[16:23] Stickman Ingmann: Oh. Exciting.

[16:23] Andrew Linden: which would address (1) and (3)

[16:23] dantia Gothly: new animation engine would be nice.

[16:23] Maggie Darwin: And allow answering any further complaints with "it doesn't work that way". :-)

[16:23] Andrew Linden: I'm not saying we are doing (a) or (b), but either of those efforts would allow us to make progress

[16:24] Arawn Spitteler: Preloading is an important feature

[16:24] Andrew Linden: Richard isn't working on animations right now, his plate is full with User Interface work

[16:24] Stickman Ingmann: Whatever happens, the long-term has to be considered. Custom skeletons are potentially in the future, and whatever animation engine exists is going to need to handle that.

[16:24] Andrew Linden: and it is a full plate

[16:24] Stickman Ingmann: The current format is set up to handle it, in a way. Dunno about the engine.

[16:24] Draconis Neurocam: i think the engine is more robust than you think stickman, we are just given little access to it

[16:24] Qie Niangao: umm. so there's also, maybe someday, a revisiting of "expressive puppeteering"... dunno if that's "engine" or deeper

[16:24] Andrew Linden: I believe we interviewed a dev candidate who knows something about building animation engines.

[16:24] Stickman Ingmann: Preloading is useful -- but syncing is better. Even if the animation takes longer to download, everyone would still see it the same when it was done.

[16:24] Andrew Linden: Dunno the status of that.

[16:25] Stickman Ingmann: Hm. Could be fun to have an animation usergroup if it comes to that.

[16:25] Hope Dreier: Hmmm .. I have a Jira about AOs sometimes not being trigered, The Linden who answered it agreed bu said that was something being bandied about wher you could 'change' the default Animation fo Say walking.

[16:25] Stickman Ingmann: Guess part of it is how serious LL wants to take the concept.

[16:25] Andrew Linden: Richard and I had an idea on how we could make suggestions to the viewer of lists of asset_id's it might want to preload -- it would then be up to the viewer to preload them or not

[16:25] Andrew Linden: so (4) doesn't seem to hard to do

[16:26] Andrew Linden: and might also work for some sounds and textures... maybe.

[16:26] Stickman Ingmann: That's good. If preload is easier, it's better than nothing.

[16:26] Arawn Spitteler: Most of animation is viewer side

[16:26] Stickman Ingmann: We do have llPreloadSound or something like that.

[16:26] Pauline Darkfury: does llPreloadSound actually work tho?

[16:26] Arawn Spitteler: Setting a ballet to music, would seem to require preload anim

[16:26] DogWomble Dollinger: that's a point actually

[16:26] Stickman Ingmann: But having any references in a script download automatically would be nice.

[16:26] Andrew Linden: As to (2)... Richard said that the feedback probably wouldn't be necessary if we could get all the others

[16:27] DogWomble Dollinger whispers: could we add llPreloadAnimation to that as well?

[16:27] Siobhan McCallen: better control of anims would help with machinima creation

[16:27] DogWomble Dollinger: that might be really useful for those that write AOs and such to have all the animations loaded

[16:27] Andrew Linden: he's got some picture of an animation state machine UI for making anims, rather than the system we use now

[16:27] Andrew Linden: that is, as the "new and improved" anim creation system

[16:28] Rex Cronon: a creation system?

[16:28] Arawn Spitteler: Preloading any reference might not work. "animstage"+(string)n

[16:28] Siobhan McCallen: I use Phate Shepherd

[16:28] Stickman Ingmann: An inworld creation tool, you mean? Like puppeteering? Or something else, like an advanced animation preview and touch-up tool?

[16:28] Siobhan McCallen: Shepherd's "Anypose" system

[16:29] Siobhan McCallen: It's an inworld pose creation system and it works quite well

[16:29] Andrew Linden: hrm... I think I'd rather improve the automagic preloading and viewer-side cache system rather than add llPreloadAnimation()

[16:29] Siobhan McCallen: and it now makes anims

[16:29] Stickman Ingmann: I agree, Andrew. Automatic, when it works, is much better than manual. For something like this, where you can just throw useful data down the pipe rather than trying to manually predict it.

[16:29] Andrew Linden: yeah Stickman, I think it would be an in-world anim tool, with some save-as-asset features

[16:30] Stickman Ingmann: New user enters the range? Call the preload function again to throw all the animations at them! Just gets complicated.

[16:30] Andrew Linden: but it is currently an idea that Richard has been kicking around for years with no time to work on

[16:30] Void Singer: I was also told to ask about an odd change that seems to have happened on the server side some time ago... calling card permissions being ignored and blocked from transfer?

[16:31] Stickman Ingmann: An inworld tool would solve a lot of problems regarding accuracy. But reproducing the features that exist in the major 3D animation programs would be expensive in both time and resources.

[16:31] Andrew Linden: Void... more detail on that? calling cards are transfering when they should not? or they are not when they should?

[16:31] Stickman Ingmann: Which is why I'm for better importing tools. Make the default client accept .anims, and encourage developers (or have LL make an example) to write exporters for the major programs to the .anim format.

[16:31] Void Singer: NOT, when they should....

[16:32] Draconis Neurocam: stickman on that, i could not agree more

[16:32] Ashiri Sands: Understanding where the inaccuracy has crept into an animation has always been a issue for me

[16:32] Siobhan McCallen: the current .bvh implementation is not very complete.

[16:33] Void Singer: namely in the circumstance of physical giving a copy of your own card, or recieving a copy back from someone that you've given it to

[16:33] Stickman Ingmann: There's a lot of reasons, Ashiri. Everything from avatar height, to inaccurate skeletons in the external tools, to BVH conversion and upload compression causing all sorts of horrors.

[16:33] Ashiri Sands: and minimum movements

[16:33] Andrew Linden: Well if I hear any updates on animations I'll try to mention them here.

[16:33] Stickman Ingmann: Hah. Yes, the complete lack of subtle animations. Geeze.

[16:34] Ashiri Sands: that's why so many walk animations look wrong... no subtlety

[16:34] Arawn Spitteler pretends to twiddle his thumbs

[16:34] Stickman Ingmann: Please do. Thank you for contacting Richard about it. I'm more concerned that the ideas be present within LL so that information is there when the choices are made.

[16:34] Andrew Linden: Since we interviewed a potential anim dev I believe the question about whether to work on anims is in the thoughts of several Lindens

[16:34] Pauline Darkfury: the anims are fine, it's just our AVs that are wrong, if we all wore duck AVs, it would look fine ;)

[16:34] Andrew Linden: but I don't know what decision will be made, if any

[16:35] Stickman Ingmann: I would be happy to explain the problems I encounter with animations to any dev that's interested. I'm sure several other SL animators would be, too.

[16:35] Andrew Linden: I haven't worked on calling cards in a long time. Do they still work?

[16:35] Andrew Linden: You can only give out your own calling card right? You can't transfer others' cards?

[16:35] Maggie Darwin: I've never found a use for them.

[16:35] Qie Niangao: Kinda. In fact there was a recent Snowstorm thing about calling cards.

[16:35] Stickman Ingmann: Calling cards were recently redone in the viewer, so they're useful again

[16:35] Arawn Spitteler: There's a slight problem, with pulling up your own profile

[16:36] Stickman Ingmann: This one?

[16:36] JIRA-helper: [#STORM-28] As a User, I want the ability to send my calling card to others.

[16:36] Ardy Lay:

[16:36] JIRA-helper: [#SVC-6853] Calling Cards as inventory assets, need LSL support

[16:36] Qie Niangao: Right

[16:36] Void Singer:

[16:36] JIRA-helper: [#VWR-18977] Viewer 2.0: Cannot send a Calling Card

[16:36] Hope Dreier: Lol I have, or had, a number of OLD Calling cards that when accessed gave me MY profile.

[16:37] Stickman Ingmann: That would be useful. "Click for a calling card." Use 'em like business cards.

[16:37] Opensource Obscure: ability to send our calling card to others has recently been implemented in development branch

[16:37] Void Singer: that's the one that started it for me, but the additional brokenness is what has me asking after it (just hunted down the one I commented on)

[16:37] Andrew Linden: Yeah, I haven't heard of any calling card related changes in the server. This sounds like a viewer issue.

[16:38] Siobhan McCallen: there was a problem with being able to give out OTHER peoples' cards, as I recall

[16:38] Void Singer: but lsl support would be nice... if I could drop it and give it just like a notecard asset

[16:38] Arawn Spitteler: We used to have Give Calling Card in our pie menu

[16:38] Bitterly Sweet: its a viewer issue. my roommate has been following that allot

[16:38] Draconis Neurocam: andrew since you are working on mesh parcel management stuff, do you have any idea how meshes that are theoretically worth a fraction of a prim will be handled by parcel prim limits

[16:38] Void Singer: Andrew, not just the viewr,... tested across three viewers, can no longer directly give from inventory

[16:39] Arawn Spitteler: Whose card are you trying to give?

[16:39] Andrew Linden: Draconis, you mean whether the prim accounting will go to floating point numbers instead of integers?

[16:39] Draconis Neurocam: yeah

[16:39] Void Singer: which was the only workaround to the V2 issue

[16:39] Andrew Linden: I don't know yet, but I probably will know next week.

[16:39] Andrew Linden: I'm in the thick of it.

[16:39] Pauline Darkfury: yeah, at least getting 1 decimal place would be nice there

[16:40] Andrew Linden: If we have lots of prims worth < 1 then we'll need it

[16:40] Ashiri Sands: Though I can imagine fractional prim counts confusing many

[16:40] Pauline Darkfury: idea: if floating point is somehow hard, could make the limit 150,000, where 1 legacy prim is 10 new units

[16:40] Andrew Linden: but I think maybe the system doesn't allow < 1. It may clamp to 1, or just use ciel()

[16:41] Andrew Linden: er... I mean ceil()

[16:41] Void Singer: lol I think we al speak typonese here =)

[16:42] Maggie Darwin: spellchk(ciel())

[16:42] Draconis Neurocam: i hope then, what will be considered the maximum faces for a mesh worth a single prim will be within some reasonable limit then

[16:42] Rex Cronon: u become very well versed in that around here:)

[16:42] Flip Idlemind: Its a requirement to get your internet license

[16:42] Arawn Spitteler: Typonese is native to the internat Relay Chat, and SL remembers it's ancestry.

[16:42] Draconis Neurocam: faces verts, not sl faces, i know the difference

[16:42] Arawn Spitteler: What'sthe problem, with giving your calling card? it works line of sight in Snowglobe

[16:43] Andrew Linden: Draconis, I believe that "well made" mesh objects will be cheaper in general.

[16:43] Andrew Linden: However... there are some regular prims that are much more expensive than they should be

[16:43] Void Singer @ Arwan: only from pie menu, not from inventory

[16:43] Andrew Linden: not all prims are really equal when it comes to CPU costs (server and viewer sides)

[16:44] Kadah Coba: Can something be done about these group chat bet spam bots?

[16:44] Arawn Spitteler: lHave Calling Cards ever come from Inventory?

[16:44] Draconis Neurocam: andrew do you think sculpts prim costs will be refactored at all, considering exactly that, or will it stay the same

[16:44] Liisa Runo: it will stay the same or it will break content

[16:44] Andrew Linden: so... it may be that if you make a mesh for a hollow/cut/twisted torus you'll find it much more expensive than a regular prim that looks identical

[16:44] Arawn Spitteler: Sculpties, Twisted Torus comes to mind

[16:45] Void Singer @ Arawn: I used to use it that way all the time, but since I've cut back my shop hours I missed the change

[16:45] Andrew Linden: Draconis, I believe we are trying to maintain backward compatibility for sculpty

[16:45] Pauline Darkfury: Falcon said there would be some form of opt-in/out for existing stuff, to the new accounting, so that we can get some gains without having to completely rebuild old stuff.

[16:45] Draconis Neurocam: makes sense

[16:45] Rex Cronon: sculpties could become very efficient if any size greater than 2x2 would be allowed:)

[16:45] Arawn Spitteler: llGiveOwnerCallingCard?

[16:46] Andrew Linden: but if you use any new meshy feature you'll be subject to the new "prim cost" metrics

[16:46] Andrew Linden: which are complicated (I don't know the details of any of the formulas)

[16:46] Draconis Neurocam nods

[16:46] Pauline Darkfury: yeah, we get to choose between old and new on a per-object basis, old gives is 1/15,000 model but no new features, if I understood what was said a while back

[16:46] Pauline Darkfury: default will be old model for existing stuff

[16:47] Qie Niangao: (Void, Arawn, do you get any calling cards if you touch the blue plywood box?)

[16:48] DogWomble Dollinger: seems to work for me qie

[16:48] Liisa Runo: (works for me, snowglobe)

[16:49] Hope Dreier: ((worked for me Phoenix 1000, but both are Yours))

[16:49] Arawn Spitteler: I seem to have a folder of two calling cards

[16:49] Qie Niangao: oh!

[16:49] Qie Niangao: that's not good, hehehe... one should be Fleep Tuque's. :p

[16:49] Void Singer: I got it in phonix.... which amazes me since I couldn't do it any way I tried yesteday in V2, Cool, or Phoenix

[16:50] Sling Trebuchet: (Imprudence - Gives me two names in folder, but both open as Qie)

[16:50] Hope Dreier: It says Fleep but when opened it's yours

[16:50] Dora Gustafson: viewer 2 got it

[16:50] Void Singer: yup same thing... says fleep, but it links to your profile

[16:50] Qie Niangao: well... tis a mystery.

[16:50] Andrew Linden: so calling cards can be given via LSL script, but not by viewer-2 User Interface?

[16:51] Bitterly Sweet: about the size of it

[16:51] Hope Dreier: Well for some value of Given...

[16:51] Void Singer: I couldn't even get one into a prim yesterday, so I'm kinda suprised myself

[16:51] Ardy Lay: Looks like Viewer 2.x uses the UUID in the description field to select profile to open.

Total Avatar Scripts - 1528

Highest - Bitterly Sweet: 367 (24%)

Lowest - Andrew Linden: 0 (0%)

[16:52] Opensource Obscure:

[16:52] JIRA-helper: [#STORM-28] As a User, I want the ability to send my calling card to others.

[16:52] Hope Dreier: ((OMG Bitterly 367?))

[16:53] Opensource Obscure: works in Second Life 2.6.1 (224674)

[16:53] Ardy Lay: Drag from inventory, right?

[16:53] Arawn Spitteler: 2.6.1 works for somebody?

[16:53] Void Singer can't even find a copy of my own calling card any more... odd

[16:53] Opensource Obscure: i currently have my own card in inventory

[16:54] Opensource Obscure: i can share it by rightclicking it

[16:54] Bitterly Sweet: i had a quesion about inventory

[16:54] Opensource Obscure: not only by dragging on avatars.

[16:54] Void Singer @ hope: that's nothing.... I've got avies running around my region eating 40+MiB

[16:54] Bitterly Sweet: i would like to ask quickly

[16:54] Ardy Lay: Right. Viewer now creates that inventory item if it's missing.

[16:54] Andrew Linden: Go ahead Bitterly

[16:55] Bitterly Sweet: why is it that there are allot of people unable to fully fetch inventoy

[16:55] Bitterly Sweet: it seems to stall out when it tries to fetch the links in the outfit folders

[16:55] Bitterly Sweet: and will not resume

[16:55] Bitterly Sweet: its really bad on firestorm, but its also all v2 viewers

[16:56] Andrew Linden: Bitterly, are these people using the new inventory API (in viewer 2.5.something and higher)? or the old UDP protocol (viewer-1.23 and most 3rd party viewers)?

[16:56] Bitterly Sweet: new

[16:56] TankMaster Finesmith: firestorm is 2.4 based, we still used the old API

[16:56] Bitterly Sweet: well maybe not then

[16:57] Bitterly Sweet: but its been reported on all v2 viewers i know of

[16:57] TankMaster Finesmith: we will rebase to 2.5.2 which has the new API in a week or 2

[16:57] Bitterly Sweet: my roommate and i are always clouds

[16:57] Andrew Linden: I've heard rumors that the new inventory API is working well for most people

[16:57] Andrew Linden: however none of those rumors are from these User Groups

[16:58] Stickman Ingmann: Heh.

[16:58] Andrew Linden: anyone here have experience there?

[16:58] Stickman Ingmann: Not I.

[16:58] Draconis Neurocam: ive not had any problems with it, though i only have about 18k items

[16:58] Ardy Lay: No, and I see no clouds here.

[16:58] TankMaster Finesmith pokes kadah

[16:58] Bitterly Sweet: for me right now im a cloud.

[16:58] Pauline Darkfury: Quick thing before the hour is over. There's some feedback on the abandoned land change at:

[16:58] Bitterly Sweet: my outfit folder has two of the 13 links it should have

[16:58] Andrew Linden: Bitterly, and your viewer + version is...?

[16:59] Flip Idlemind: Im on Phoenix and I never have any inventory problems. But I'm also on the extreme low end for inventory count. 6000-ish items

[16:59] Ashiri Sands: v2.5+ have had other issues for me

[16:59] Bitterly Sweet: 2.5.3

[16:59] Opensource Obscure: 18k as well, and inventory is not very fast. i'm in italy. it usually completes after a few minutes

[16:59] Void Singer cant comment... only has <5k invetory and it always loads fine

[16:59] TankMaster Finesmith: 2.4.3 is the latest version of firestorm...

[16:59] Opensource Obscure uses 2.5 and 2.6 various releases

[16:59] TankMaster Finesmith: and thats only if you self compile it

[16:59] Bitterly Sweet: im not on firestorm right now tank :)

[16:59] TankMaster Finesmith: ok

[16:59] Pauline Darkfury: I think the only major concern in that thread not already raised here is the lack of a replacement for Jack's OH to discuss such things

[16:59] Kadah Coba: What tank?

[16:59] Hope Dreier: I'm at 35K Phoenix and I've nver had a serious Inventory issue

[17:00] TankMaster Finesmith: they wer asking about AIS2 experence

[17:00] Pauline Darkfury: I'm on 40k inventory, Phoenix, no major issues, just takes a while to load if I need to do a fresh load of it after a crash

[17:00] Bitterly Sweet: I think my rommate emailed you or is planning to email you about this

[17:00] Kadah Coba: The new API is a bit slower but it doesn't seem to fail to return the whole inventroy like the old system, so thats win.

[17:01] Andrew Linden: Well, the short answer Bitterly is that I don't know what is going on with inventory yet.

[17:01] Andrew Linden: I have seen some real cases where particular resident inventories can NOT load and may cause the viewer to crash.

[17:02] Bitterly Sweet: now if i load into a v1 viewer and kill all links in the uotfit folders and log in naked it works

[17:02] Kadah Coba: The real question is does Tank's 10million mega prim filled inventory fully load on AIS2? :P

[17:02] Andrew Linden: when this happens it could be caused by some corrupted inventory rows in the database

[17:02] Pauline Darkfury: Have you got separate cache folders if you're using both V1 and V2? I believe there's some known issues trying to share cache between them

[17:02] Stickman Ingmann: I've gotta run. Thanks for the office hours, Simon, Andrew!

[17:02] Andrew Linden: I filed an internal jira issue about it but I don't think we've got someone working on it right now.

[17:03] Siobhan McCallen: good to see everyone again

[17:03] Draconis Neurocam: thanks simon and andrew

[17:03] Siobhan McCallen: thanks for listening to us, simon and andrew

[17:03] Hope Dreier: Actually you shold have different Cachers for EACH different Viewer

[17:03] Bitterly Sweet: i do

[17:03] Andrew Linden: If anyone is crashing for inventory load they can send me email and I'll try to look into it

[17:03] Simon Linden: Thanks everyone for coming and the good ideas about SL workings

[17:03] Bitterly Sweet: differant drives even

[17:03] Kadah Coba: Andrew, the only fequent problem I've seen caused by inventories not fully loading on v2 are bake-fail and empty friends list.

[17:03] Qie Niangao: Thanks for your time, Andrew and Simon.

[17:03] Rex Cronon: tc everybody

[17:03] Andrew Linden: I'm curious to know more about that problem, but am pretty swamped.

[17:03] Pauline Darkfury: Thanks Lindens, have a good weekends :)

[17:03] Bitterly Sweet: ty andrew and simon

[17:04] Maggie Darwin: Good weekend to all!

[17:04] Andrew Linden: I was hoping that another team would pick up that problem.

[17:04] Ashiri Sands: Thank You Andrew & Simon

[17:04] TankMaster Finesmith: thx andrew, simon

[17:04] Simon Linden: Have a good weekend, eveyrone

[17:04] Opensource Obscure: thanks andrew and simon - bye all

[17:04] Andrew Linden: Thanks for coming everyone.

[17:04] Simon Linden: ah, also ... scripting user gruop is Mondays at 9

[17:04] Simon Linden: Kelly is running them .. check the wiki for the location

[17:05] Andrew Linden: Yes. All of the Simulator User Group meetings are mentioned here:

[17:05] Andrew Linden: and transcripts from earlier meetings are being published there

[17:05] Hope Dreier: Thank you Andrew, do you think you can find a bit of time to look at SVC-5927 ?

[17:05] JIRA-helper:

[#SVC-5927] Temp on Rezzed objects get queued

[17:06] Simon Linden: Bye everyone

[17:06] Andrew Linden: Hope, not next week, but maybe after mesh completes.

[17:06] TankMaster Finesmith: tc simon

[17:06] Hope Dreier: any estimate on when that will be?

[17:07] Liisa Runo: 2014

[17:07] Ardy Lay: Still a moving target

[17:07] Arawn Spitteler thx the lindens, and dives back into a Swedenborg Reader.

[17:07] Andrew Linden: Hehe, definitely 2011.


Prev 2011.03.22 Next 2011.03.28