Simulator User Group/Transcripts/2012.09.07

From Second Life Wiki
Jump to: navigation, search

Simulator_User_Group

Prev 2012.09.04 Next 2012.09.11

List of Speakers

Andrew Linden Baker Linden Cheesey Linden
Chieron Tenk Dahlia Trimble Iain Maltz
Jane1 Bookmite Lomoco Binder MartinRJ
Moundsa Mayo Nalates Urriah Rex Cronon
Simon Linden Sniper Siemens Spoof Kelberry
SprSrsNoFun TankMaster Finesmith Toysoldier Thor

Transcript

[16:01] Jane1 Bookmite: Hi Simon !

[16:02] Lomoco Binder: And suddenly, purple

[16:02] Jane1 Bookmite: Welcome to, oops wrong sim,,,,,

[16:02] MartinRJ: Hello Andrew!

[16:02] Sniper Siemens: hello Andrew

[16:02] Andrew Linden: Hello.

[16:02] Toysoldier Thor: So Lindens.... to expose the elephant in the room - what was Rod thinking with the move of "gagging" the community JIRA from residents? How does LL think they will be able to provide better support and response to SL countless problems? And direct to these meetings, how will we even be able to point to JIRAs when talking to you about progress of issues?

[16:02] Rex Cronon: hi andrew

[16:03] Rex Cronon: toysoldier u get right to the point:)

[16:03] Simon Linden: I have my pre-recorded spiel ready for that one ...

[16:03] Chieron Tenk: hi andrew, simon

[16:03] Andrew Linden: We'll consider that question after the news.

[16:03] Jane1 Bookmite feels the love

[16:03] Toysoldier Thor: yes because this is just plain stupid

[16:03] Toysoldier Thor: ok

[16:03] Toysoldier Thor: thanks

[16:03] Andrew Linden: Simon goes first with the news today.

[16:04] Simon Linden: ok, let's do the news and then deal with the elephant-sized pile that is Jira ...

[16:04] Toysoldier Thor: ohh noo - THE TREE"

[16:04] MartinRJ: hello Baker: )

[16:04] Spoof Kelberry: A wild tree has appeared.

[16:04] Chieron Tenk: a linden tree..

[16:04] Simon Linden: We had a little excitement with the rolls this week and now all 3 RC channels have the same version. That is scheduled to be rolled to the main channel next Tuesday, barring any newly discovered issues

[16:05] Simon Linden: There will be yet another bug fix version going to all 3 RC channels on Wednesday of next week

[16:05] Andrew Linden: That excitement was the L$ transaction notification failure issue, as I recall.

[16:05] Simon Linden: Yes, exactly that

[16:06] Simon Linden: That bug was cosmetic, but balances that don't update are definitely not good so we had to fix it ASAP

[16:06] Nalates Urriah: SCV-8208

[16:06] Simon Linden: I'll bet the jira bots are all totally broken now

[16:06] Moundsa Mayo: You win!

[16:06] Toysoldier Thor: lol

[16:06] Simon Linden: oh rhm, at least they should make links, and those would be broken

[16:07] Nalates Urriah: Mostly broken... some ID's still work

[16:07] Toysoldier Thor: This is the jira we are working on.... But you can no longer see it

[16:07] Moundsa Mayo: So. More news?

[16:08] Andrew Linden: My news is that I think I'm done with my first pass on the server's interestlist code.

[16:08] Simon Linden: The RC update has two ex-public bug fixes ...

[16:08] Iain Maltz: (nice SLURL in the ocean :P)

[16:08] Andrew Linden: I now need to package up my work for QA. I might have a few regions up on aditi next week if all goes well.

[16:09] Nalates Urriah: ex-public bug?

[16:09] Simon Linden: one was SVC-8146, where llRezAtRoot() wasn't setting some parameters (for sale) correctly

[16:09] Nalates Urriah: nvr mind

[16:09] MartinRJ: Andrew physical spheres are somehow broken

[16:09] Andrew Linden: MartinRJ, broken how?

[16:09] MartinRJ: https://jira.secondlife.com/browse/SVC-8212?

[16:10] Simon Linden: The other was SVC-8119, where llApplyRotationalImpulse() was giving a script error. Now it will be capped at the max rotational velocity

[16:10] MartinRJ: they stop rolling over surfaces immediately

[16:10] MartinRJ: they only slide. it's like there is a immense huge friction

[16:10] MartinRJ: applied

[16:10] MartinRJ: to every sphere

[16:10] Rex Cronon: can't access the jira. i get error:(

[16:10] Simon Linden: There are a few others that will hopefully help stability and some TP problems

[16:10] MartinRJ: (on main server channel)

[16:11] Andrew Linden: Interesting MartinRJ. I'll have to ping Falcon to see if he has any theories about that one.

[16:11] MartinRJ: thank yous!

[16:11] Chieron Tenk: it only happens with unlinked prims, tho

[16:11] Simon Linden: That's it for news ... so for jira ...

[16:11] Baker Linden: My news is that my group member changes will soon be deployed on aditi. You'll need a development viewer to test (on aditi). You can find them here: Windows:

http://automated-builds-secondlife-com.s3.amazonaws.com/hg/repo/Viewer-Development-Baker/rev/264495/arch/CYGWIN/installer/Second_Life_3-4-1-264495_Developer_Setup.exe

OS X:

http://automated-builds-secondlife-com.s3.amazonaws.com/hg/repo/Viewer-Development-Baker/rev/264495/arch/Darwin/installer/SecondLife_3_4_1_264495_DEVELOPER.dmg

Linux:

http://automated-builds-secondlife-com.s3.amazonaws.com/hg/repo/Viewer-Development-Baker/rev/264495/arch/Linux/installer/SecondLife-i686-3.4.1.264495.tar.bz2

[16:12] Andrew Linden: Wait... I've got more news...

[16:12] Simon Linden: That should be on Aditi, I'll find the regions ...

[16:12] Nalates Urriah: Simon's gone deaf

[16:12] Toysoldier Thor: and what about the jira formerly known as.....

Second Life Viewer - VWR (Inactive)

VWR-28409

Texture animations ignore repeats per face (llTextureAnim())

[16:12] Baker Linden: I'll be posting in the JIRA (SVC-4968) on which regions this will be enabled (on aditi)

[16:13] Simon Linden: go ahead, Andrew, sorry....

[16:13] Andrew Linden: While working on the interestlist stuff I spent some time optimizing how updates to the terrain and wind are prepared for the viewer.

[16:13] Andrew Linden: While I was at it I decided to see if I could tweak the wind, which has been described here at these user groups as "useless"

[16:13] MartinRJ: hehe

[16:13] Andrew Linden: so that it might be a little more what people are looking for

[16:14] Cheesey Linden: what does wind look like?

[16:14] Andrew Linden: probably not useful for the sailing community, but I think I've managed to slow down its variation

[16:14] Andrew Linden: so that it doesn't get so wild sometimes

[16:14] Moundsa Mayo: Ask Baker - he's a tree ...

[16:14] Cheesey Linden: he'll cut me

[16:14] Moundsa Mayo: He'll cut the Cheesey?!!

[16:14] Baker Linden: Nevar!

[16:15] Moundsa Mayo: That's good news, Andrew. 'Scuse my digressions.

[16:15] Andrew Linden: That's all. It will eventually be up on aditi for people to play with if they like. I know wind isn't really used for much except for some flexi prims that bend to the wind.

[16:15] Simon Linden: will that change how flexis behave?

[16:15] Andrew Linden: Only the wind varation rate

[16:15] Andrew Linden: so fundamentally flexies will still blow in the wind, but perhaps not so wildly.

[16:16] Simon Linden: I guess we'll find out how many people notice :)

[16:16] Andrew Linden: At least the wind speed and direction will vary at a slower rate.

[16:16] Nalates Urriah: Is variation you fixed speed and direction or just direction?

[16:16] Nalates Urriah: nvr mind

[16:17] Moundsa Mayo: I use it for windsocks, windmills, and .. I forget.

[16:17] Moundsa Mayo: Well, lots of trees sense the local wind and react.

[16:17] Andrew Linden: fin

[16:17] Toysoldier Thor: and what about the jira formerly known as.....

Second Life Viewer - VWR (Inactive)

VWR-28409

Texture animations ignore repeats per face (llTextureAnim())

[16:17] Chieron Tenk: will the size of the vortices/curls/eddies increase. it seems that the structures are pretty small (like 100m diameter for a vortex)

[16:17] Simon Linden: ok ... any more news before we open up the jira topic?

[16:17] Chieron Tenk: ?

[16:17] Toysoldier Thor: any status on it?

[16:18] Toysoldier Thor: i cant even properly point out a jira for discussion other than point out the title

[16:18] Rex Cronon: does the wi

[16:18] Simon Linden: No news on that one, Toysoldier

[16:18] Iain Maltz: i've had a SEC bug open related to group functions for 2 weeks that has had no attention, is there some way i can get someone poked to look at it (i'll admit to having not verified it since, but it was independantly verified at the time)

[16:18] Rex Cronon: i mean will the wind affect physical prims or pathfinding

[16:19] Toysoldier Thor: and Cheesey - and progress on assessment on the VLM jira?

[16:19] Andrew Linden: No Chieron, I don't think the size of the vortices will change.

[16:19] Chieron Tenk: hm, ok.

[16:19] Andrew Linden: That is determined by the fidelity of the fluid simulation, which didn't change.

[16:19] Baker Linden: oh, something I wanted to mention

[16:19] Cheesey Linden: Toy it's been vetted and is not in products hands to prioritze

[16:20] Toysoldier Thor: is that good news Cheesey?

[16:20] Cheesey Linden: yes.

[16:20] Simon Linden: So for jira ... this is the policy change: http://community.secondlife.com/t5/Tools-and-Technology/JIRA-Update-Changes-to-The-Bug-Reporting-Process/ba-p/1660981

[16:20] Toysoldier Thor: great

[16:20] Cheesey Linden: that is progress

[16:21] Baker Linden: if you decide to test large group management, and find an issue with it while testing, email me at baker@lindenlab.com (as you can no longer comment in the JIRA)

[16:21] Andrew Linden: I think he meant to say: "... it is *now* in product's hands..." where "product" is our word for an internal team

[16:21] Simon Linden: If you have feedback, please fill out the survey there ... give clear, objective information on what you think about the change and how it affects you. It WILL be looked at.

[16:21] Cheesey Linden: woops yes i meant now

[16:21] Toysoldier Thor: yup i picked that up - thanks Andrew

[16:21] TankMaster Finesmith: and please done fill out " the jira is broken"

[16:22] TankMaster Finesmith: those will just get cclsoed

[16:22] Dahlia Trimble: looked at by whom?

[16:22] TankMaster Finesmith: lindens will still look at them

[16:22] Lomoco Binder: Will this affect how feature requests are handled?

[16:22] Toysoldier Thor: ok explain for Rod WHY

[16:22] Simon Linden: The people here are not making that policy and are not a good way to get feedback to Linden Lab - the survey and forum comments are teh way to go

[16:22] Rex Cronon: what forum?

[16:22] Toysoldier Thor: we know full well this idea did not come from the LL staff

[16:23] Toysoldier Thor: this was a Rod move

[16:23] Toysoldier Thor: but why

[16:23] Simon Linden: I thought that post had a feedback area, but I may be wrong

[16:23] Toysoldier Thor: and how will your support staff keep up and properly support problem reslution

[16:23] Lomoco Binder: "Ideas about how we can continue to improve the bug reporting process can be shared here."

[16:23] Lomoco Binder: A link is included in the post.

[16:23] Moundsa Mayo: Link at the end of this line: deas about how we can continue to improve the bug reporting process can be shared here.

[16:23] Iain Maltz: and how are we supposed to figure out how the platform is actually working as opposed to how its *supposed* to work?

[16:24] Rex Cronon: that is the survey. is not a forum

[16:24] Toysoldier Thor: so will SL residents now have to use the Forums to communicate and collaborate on SL bugs we are all experiencing?

[16:24] Moundsa Mayo: It's 'supposed' to work the way it used to, only without all the whining, kvetching, and personal attacks.

[16:24] Simon Linden: I understand there are concerns and you have questions, but we're not the people who can provide those answers. If you want to talk about why you lag when crossing regions, let's get into the fun tech stuff

[16:24] Toysoldier Thor: it also major impacts these UserGroup meetings

[16:25] Rex Cronon: moundsa. goverment in rl is supposed to work the same way:0

[16:25] Iain Maltz: not the jira, the bugs in second life which we know are plenty but now have no way of looking up previous expereince, fixes, workarounds for etc, which i'm sure i'm not alone in doing regularly

[16:25] Lomoco Binder: Simon, does this mean open table?

[16:25] Dahlia Trimble: there are no bugs, only undocumented features! :D

[16:25] Andrew Linden: This user group is still happening, and is a good resource for discussing bugs however...

[16:25] Andrew Linden: we can't expect everyone to have time to attend.

[16:25] Simon Linden: Sure LOmoco, I'd love to talk about anything besides Jira policy :)

[16:25] Iain Maltz: (or the timezone to attend)

[16:25] MartinRJ: Still I have a question about how we will handle these things in this meeting in future

[16:25] Spoof Kelberry: With the integration of navmesh in the viewer, you guys have removed the -loginuri agrument from it. Will you be supporting this feature in another binary version? I'm asking this because I've spent quite a lot of time on things that depend upon this feature and it is a bit saddening if it will be removed from the official viewer as all my work will have gone to waste.

[16:25] MartinRJ: What do we do instead of posting Jira-links in future, Simon?

[16:26] Rex Cronon: were will i have to look now for workaround different issues?

[16:26] Toysoldier Thor: my question exactlu Martin

[16:26] TankMaster Finesmith: loginuri will not be reimplmented ever

[16:26] Simon Linden: I really have no idea, MartinRJ

[16:26] TankMaster Finesmith: that was removed along with the other comand line options to do restrictions put in place by intel on the havoc licence

[16:27] MartinRJ: = D

[16:27] Toysoldier Thor: I got an idea - we can all file SL Grid bugs on the Phoenix JIRA

[16:27] Nalates Urriah: Other than sending our use cases in via the Survey, we are pretty well stuck with what we have. Oz and Alexa were pretty clear.

[16:27] Andrew Linden: I certainly found jira's useful for our user group discussions here, as reference links that went into the record for future lookup.

[16:27] Simon Linden: The forums are supposed to be a platform for SL discussions, and they are still open, so perhaps discussions will migrate there

[16:28] Nalates Urriah: Andrew, would you send Alexa a note to that affect?

[16:28] Toysoldier Thor: i agree simon

[16:28] Spoof Kelberry: I'm assuming that not everyone will be given a sublicence with havok if they want to compile their own viewer, will they be allowed to reimplement the -loginuri agrument?

[16:28] Toysoldier Thor: we now need a new forum group call "Post-JIRA Forums"

[16:28] TankMaster Finesmith: i agree, andrew, it now kinda defets the purpus of hacing a jira parcer to auto profice lints whever someone pose the name of a jira (such as STORM-123)

[16:28] JIRA-helper: http://jira.secondlife.com/browse/STORM-123

[#STORM-123] Resizing ability of nearby chatbar in bottomtray appears only after click. - Second Life Bug Tracker

[16:28] Rex Cronon: hard to find in the forums a workaround to a problem u r having:(

[16:28] Andrew Linden: We're still going to be fixing bugs, and communicating about bugs, so we'll need some sort of forum, issue database, or other thingy.

[16:29] Dahlia Trimble: Spoof if you compile your own viewer they cant stop you from implementing whatever you want

[16:29] TankMaster Finesmith: yes, spoof

[16:29] Toysoldier Thor: create a new section in the fpurms

[16:29] Andrew Linden: But I'm not sure what that will look like going forward. I expect that the new system will be updated as we try to improve it.

[16:29] Iain Maltz: set up a 3rd jira....

[16:29] TankMaster Finesmith: as long as theres no havoc code, it can be used

[16:29] Andrew Linden: Hence the survey will be a good place for now to propose improvements.

[16:30] Meeter: Timecheck : User Group is half over

[16:30] Spoof Kelberry: I just wish there was a binary that Linden Labs offered as a download that contained no Havok code alongside the one that does.

[16:30] Simon Linden: Yes - PLEASE give your feedback in that survey. That is what will be read and considered. Our talk here will not reach the people making decisions

[16:30] TankMaster Finesmith: theres a physics stub

[16:30] MartinRJ: yes

[16:30] TankMaster Finesmith: does just that, spoof

[16:31] Toysoldier Thor: Simon - whats the link where we can provide feedback (not that i suspect the decision makers will read them)

[16:31] Lomoco Binder: http://lindensurvey.force.com/support/JIRASurvey

[16:31] Simon Linden: It's here : http://lindensurvey.force.com/support/JIRASurvey

[16:31] Toysoldier Thor: thanks

[16:33] Andrew Linden: Years ago we used Bugzilla for our bug database system

[16:33] MartinRJ: Baker's avatar is funny :)

[16:33] Andrew Linden: and it worked ok, and then got hard to manage

[16:33] Baker Linden: wait til you see it dance

[16:34] MartinRJ: hahahah

[16:34] Lomoco Binder: Hey, SprSrsNoFun and I have been playing with llCastRay and have encountered some interesting behavior.

[16:34] SprSrsNoFun: Gentlementlemen

[16:34] Andrew Linden: so we threw up our hands and said, let's get something better

[16:34] Andrew Linden: and tried JIRA

[16:35] Andrew Linden: JIRA, in its first instance was indeed better, but eventually became cluttered

[16:35] SprSrsNoFun: Ok, so the thing is, llCastRay seems to not function properly once you get to a certain velocity.

[16:35] Andrew Linden: so we rolled out a new instance, migrated a few of the old bugs over to the new system, and started again with improvements

[16:35] Simon Linden: you mean you're doing casts while moving?

[16:35] SprSrsNoFun: Yeah

[16:35] SprSrsNoFun: I throw a cast at the object itself

[16:35] SprSrsNoFun: ...And it sometimes misses

[16:35] Lomoco Binder: Using llGetPos

[16:35] SprSrsNoFun: ^

[16:36] SprSrsNoFun: I'll send the script to anyone who wants it

[16:36] Simon Linden: well, your llGetPos() might execute on one frame, and the ray cast on another, after things have moved

[16:36] Lomoco Binder: The fastest one has about 44% success rate

[16:36] Toysoldier Thor: and Andrew... how will you LL be able to categorize 100's of similar JIRAs coming in from users but calling them the same thing if they cant even search if the bug has already been reported

[16:36] Andrew Linden: SprSrsNoFun, the problem you're seeing is that the script engine is running asynchronously with the physics engine

[16:36] MartinRJ: if you upload it to the Jira, I'm sure Maestro will look at it, right?

[16:36] Rex Cronon: if both u and the object r moving, it might be that the object is no longer in the same place by the time u call llcastray

[16:37] Andrew Linden: that is to say, the moments your script gets to run is spotty, and not necessarily synched with the physics engine

[16:37] Andrew Linden: basically, the script can be put on hold at any point in its operation

[16:37] Andrew Linden: and revived some indeterminant time later

[16:37] SprSrsNoFun: So wait, llSetVelocity /isn't/ syncronized iwth the physics engine?

[16:37] SprSrsNoFun: with*

[16:38] SprSrsNoFun: (I'm using llSetVelocity)

[16:38] SprSrsNoFun: (To set their velocity)

[16:38] Andrew Linden: so you may compute the object's position to use in your raycast

[16:38] Toysoldier Thor: i cant count how many times the SL Community HELPED LL support staff solve bugs on the grid because several residents on an JIRA were talking together.... LL Marketplace has been saved many times because of the merchants

[16:38] Andrew Linden: but by the time the raycast call is made the object may have moved a lot or only a little bit

[16:38] SprSrsNoFun: Hmm...

[16:39] Andrew Linden: it depends on how much time the script engine gets to run its scripts

[16:39] Andrew Linden: and which scripts get run in which order

[16:39] SprSrsNoFun: The thing I can't wrap my mind around is how it's not just simply 0%, but instead has some miss percentage when its off beat

[16:39] Lomoco Binder: Something I can see that might fix this is if there was a llCastRayLocal, where the input vectors are relative to the object doing the cast, so when the raycast is being performed by the server, the position it's using is always up to date?

[16:39] SprSrsNoFun: That could work, or adding an RC flag that states the position is local rather than global

[16:40] Andrew Linden: Yes Lomoco, that would work. The RC would then be able to compute the world-frame position in real-time.

[16:41] Lomoco Binder: Not sure how useful that would be.. The way I see it, llCastRay on fast objects is most useful for weapons

[16:41] MartinRJ: Andrew I have a feature request for a LSL function "llGetActiveScreenArea() in pixels", please. https://jira.secondlife.com/browse/SCR-410

[16:41] JIRA-helper: [#SCR-410] llGetActiveScreenArea() in pixels - Second Life Bug Tracker

[16:41] Andrew Linden: SprSrsNoFun, the script engine doesn't get the exact same amount of time to run each simulator frame.

[16:42] MartinRJ: Since we can now attach objects to other resident's huds,

[16:42] SprSrsNoFun: ...Makes sense

[16:42] Andrew Linden: It depends on other things that happen during that frame.

[16:42] MartinRJ: so we would like to have a function to properly resize those hud objects

[16:42] Rex Cronon: i can't read this: https://jira.secondlife.com/browse/SCR-410

[16:42] JIRA-helper: [#SCR-410] llGetActiveScreenArea() in pixels - Second Life Bug Tracker

[16:42] Simon Linden: I'm not sure how that would work, Martin - you'd have to ask the viewer and then wait for a reply

[16:42] Rex Cronon: :(

[16:42] Andrew Linden: If something earlier in the frame took up more time than usual then the script system may get fewer milliseconds to run than the previous frame.

[16:43] Spoof Kelberry: Under what conditions does a simulator send the client the capability url? I've marked down the following, successful teleports, enablesimulatormessage, crossing simulator boarder, and upon login. Am I missing anything?

[16:43] Andrew Linden: Hence, the scheduling of your scripts really is randomly associated with the physics steps.

[16:43] MartinRJ: Rex: http://pastebin.com/EEJWt3Nu

[16:44] Andrew Linden: Spoof, those are all that I can think of: basically when you first start to see into a region, or when you cross into it.

[16:44] Spoof Kelberry: If you have not been giving a capability url, will the simulator reject your ability to connect to it?

[16:44] Andrew Linden: The only other reason to resend the the seed capability would be if the list of capabilities suddenly changed

[16:44] Rex Cronon: thanks. i figured it out. i had cookies off:(

[16:44] MartinRJ: oh haha

[16:44] Andrew Linden: but they don't change under normal operations, so the list is not resent.

[16:45] MartinRJ: I thought there are already other LSL functions that query information from the viewer/client, Simon?

[16:45] Simon Linden: Are there? I can't think of any off the top of my head, but I rarely work with HUDs

[16:45] Andrew Linden: Spoof, without seed capability then a lot of operations just won't work -- the server is not expecting them.

[16:46] Andrew Linden: However, the simulator also uses some UDP messages. If you successfully create a UDP circuit then a few operations may still be available

[16:46] Andrew Linden: but HTTP services that require capabilities will not be available.

[16:47] Spoof Kelberry: When I attempt to create a udp circuit, the simulator appears to silently decline my handshake packet.

[16:47] Spoof Kelberry: That has nothing to do with the capability seed correct?

[16:47] Andrew Linden: Hrm... there is some magic that is being missed. It may very well be the seed capability.

[16:48] Andrew Linden: Spoof, are you going through login.cgi first? Or are you hitting a simulator via UDP out of the blue?

[16:48] Spoof Kelberry: The latter, heh.

[16:48] Simon Linden: That'll never work

[16:48] Spoof Kelberry: Okay, just wanted to verify a bug in OpenMetaverse.

[16:49] MartinRJ: what about getcamerapos? Simon

[16:49] MartinRJ: and getcamerarot

[16:49] Lomoco Binder: Isn't that data sent to the server anyways?

[16:49] MartinRJ: idk

[16:49] Andrew Linden: Yeah, the simulator must be told to expect you.

[16:49] Spoof Kelberry: That is exactly what I was looking for.

[16:49] Andrew Linden: So you want to pass through login.cgi first.

[16:49] Spoof Kelberry: Understandable.

[16:49] SprSrsNoFun: What exactly is login.cgi? Is that the login thing to let you connect to SL?

[16:50] Spoof Kelberry: Yes.

[16:50] Andrew Linden: When login.cgi completes a message is sent internally to the region where you're arriving.

[16:50] SprSrsNoFun: How long does a sim store a seed cap after the person has left the sim?

[16:51] Spoof Kelberry: I believe it is aborted after about 40 seconds?

[16:51] Andrew Linden: Yeah, it is a cgi script that authenticates your login username and password, and figures out where you should be placed in-world

[16:51] Andrew Linden: (last location, home location, or other)

[16:51] Dahlia Trimble: if I'm behind a firewall that requires a http proxy, does the request to login.cgi go thru the proxy that I define in prefs?

[16:51] Andrew Linden: that all needs to be known to login.cgi before you can actually show up somewhere

[16:52] Simon Linden: I believe so, Dahlia, but I'm not too famliar with the viewer-to-sim pipeline

[16:52] Andrew Linden: hrm... I'm not sure Dahlia. I've never used an http proxy myself, so I don't really know much about how they interact with login.cgi

[16:53] Simon Linden: That proxy/firewall is going to have to let UDP through too, I think

[16:53] Andrew Linden: but... if all your HTTP traffic needs to go through the proxy then yeah, because login.cgi is definitely speaking HTTP

[16:53] Dahlia Trimble: is it port 80?

[16:53] Andrew Linden: yes, I believe login.cgi is served at port 80 on the login host

[16:54] Andrew Linden: but the udp packets use different ports

[16:54] Andrew Linden: I don't remember which ones.

[16:54] Lomoco Binder: I have a fairly straightforward question related to lag. What problems would there be with imposing a simple resource consumption cap on avatars, sort of like a script grey goo. Not just on what they're wearing, but everything they have rezzed. The way I see it, something like this would put a stop to most sources of intentional sim lag.

[16:55] Meeter: Timecheck : User Group is almost over

[16:55] Rex Cronon: :)

[16:55] Moundsa Mayo: Port Protocol Used For

53 UDP and TCP DNS lookup

80 TCP Second Life web resources

443 TCP Second Life web resources/client authentication

3478 UDP Voice/STUN traffic

3479 UDP Voice/STUN traffic

5060 UDP Voice/SIP traffic

5062 UDP Voice/SIP traffic

12000-29999 UDP Voice/RTP traffic/Core protocol communication ** (see note below)

12043 UDP and TCP Capabilities/map services/simulator communication

12046 TCP Texture downloading

21002 TCP Voice signaling

[16:55] Andrew Linden: Hrm... is Kelly present?

[16:55] Andrew Linden looks around...

[16:55] Chat Hide Detector Public: Andrew Linden said that.

[16:56] Lomoco Binder: Darn that thing.

[16:56] SprSrsNoFun: Aww

[16:56] Andrew Linden: Lomoco, we already have some code in the simulator that tries to keep track of the script resources used by the attachments on an avatar

[16:57] Andrew Linden: however, as far as I know we don't actually enforce a cap yet.

[16:57] MartinRJ: I also have a very straightforward question:

[16:57] Andrew Linden: That code was added a long time ago but the project wasn't finished.

[16:57] MartinRJ: can we have some servers in Europe :D

[16:57] MartinRJ: my ping is catastrophic

[16:58] Spoof Kelberry: Will you guys be adding a checkbox in preferences to disable left click sit?

[16:58] SprSrsNoFun: How many ms is it?

[16:58] MartinRJ: I can't use the premium vehicles we got gifted lately

[16:58] Andrew Linden: However, the regular grey goo fence might be able to be enhanced to catch more forms of grief or resource hogging

[16:58] MartinRJ: (those vehicles for premium members from last month)

[16:58] Dahlia Trimble: Ive seen scripted thingies that estimate the lag an avatar causes by measuring the drop in time dilation as they tp in

[16:58] MartinRJ: 241msec

[16:58] Lomoco Binder: It'd be neat if a "resource cap" could be implemented. Maybe to allow for a bit of sim lag so legitimate things don't break, but for things like physics crashers that try to overload the sim or something, it could be useful.

[16:59] SprSrsNoFun: Beh, that's not too bad. Once you get into the thousands it becomes really terrible

[17:00] Lomoco Binder: Well, that was fun. Looks like the meeting is up.

[17:00] Meeter: Thank you for coming to the Server User Group

[17:01] Baker Linden: DANCE PARTY

[17:01] MartinRJ: hahahahahaha

[17:01] Andrew Linden: No plans for European servers that I know of, unfortunately.

[17:01] Cheesey Linden: this is why we cant have nice things

[17:02] Baker Linden: :(

[17:02] Chieron Tenk: cos you are trees armed with chainsaws dancing?

[17:02] Andrew Linden: Lomoco, the problem with that "resource cap, but don't break anything cool" plan is that it is very hard, if not impossible

[17:02] Simon Linden: We've actually found that multiple server colos cause problems - there are more region crossings and TPs between them, plus chat and other data sharing, and all that extra distance causes lag

[17:02] Andrew Linden: to actually NOT break anything cool that happens to use a lot of resources

[17:03] Iain Maltz: last ditch bump for SEC 1134!

[17:03] Lomoco Binder: Well, it might encourage content creators to use better scripting practices.

[17:03] Andrew Linden: ok thanks for coming everyone

[17:03] Moundsa Mayo: Andrew, Simon, Baker, Cheesey, thanks for your time and hard work!

[17:04] Lomoco Binder: But yeah, like I said, you could still allow for some sim lag. If you just focus on the extremities (like, oh I don't know, 30%+ resource usage?)

[17:04] Cheesey Linden: you're welcome

[17:04] MartinRJ: goodbye Andrew

[17:04] Rex Cronon: tc everybody and have a nice day

[17:04] Baker Linden: Have a great weekend everyone

[17:04] Simon Linden: Thanks everyone for coming today and the, erm, "spirited" discussion. We do value and listen to your feedback

[17:04] Chieron Tenk: bye everyone

[17:04] MartinRJ: Goodbye Simon, Baker, Cheesey :)



Simulator_User_Group

Prev 2012.09.04 Next 2012.09.11