Simulator User Group/Transcripts/2012.06.26

From Second Life Wiki
< Simulator User Group
Revision as of 17:28, 12 July 2012 by Andrew Linden (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search


Prev 2012.06.22 Next 2012.06.29

List of Speakers

Andrew Linden Arawn Spitteler Baker Linden
Jonathan Yap Kallista Destiny Kelly Linden
Kennylex Luckless Melvin Starbrook Nalates Urriah
Qie Niangao Rex Cronon Sahkolihaa Contepomi
Simon Linden Widely Linden Yuzuru Jewell


[12:00] Rex Cronon: lagging

[12:00] Meeter: Welcome to the Server User Group

[12:00] Melvin Starbrook: hi andrew, hi widely

[12:00] Kallista Destiny: Hi Andrew, Widely

[12:00] Andrew Linden: hello

[12:00] Rex Cronon: this thing doesn't work simon. it doesn't register the word lag

[12:00] Rex Cronon: hello andrew

[12:01] Kallista Destiny: That is because there is not lag, except in your viewer.

[12:01] Simon Linden: it's on 25 seconds, so it must have seen something

[12:01] Simon Linden: like lag ?

[12:01] Simon Linden: it caught that one

[12:01] Rex Cronon: lag

[12:01] Rex Cronon: didn't get las lag

[12:01] Rex Cronon: oh. it has to be in a sentence, not by itself

[12:02] Rex Cronon: lag

[12:02] Rex Cronon: talking about lag

[12:02] Jonathan Yap: mute Rex

[12:02] Rex Cronon: lol

[12:02] Yuzuru Jewell: Hello Baker.

[12:02] Baker Linden: Hi

[12:02] Andrew Linden: Ok so news...

[12:02] Simon Linden: oh, yeah, there's some extra code around the searching so it doesn't trigger on "flag" and such

[12:03] Rex Cronon: hello baker

[12:03] Baker Linden: Hi Rex

[12:03] Andrew Linden: We've got a main channel deploy happening this morning.

[12:03] Andrew Linden: I'll bet Simon could fill us in on that.

[12:04] Kallista Destiny: Wow we're blessed by a plethora of Lindenii

[12:04] Simon Linden: It's rolling along -- we're using new re-imaging and deploy methods which are a bit slower than before, but hopefully will improve as we iron out the kinks

[12:04] Kelly Linden: How many is a plethora of Linden?

[12:05] Andrew Linden: This is the migration from debian/etch to debian/squeeze

[12:05] Kallista Destiny: ratio of Lindens to others > 1:5

[12:06] Andrew Linden: The RC channels will all make the migration to debian/squeeze tomorrow, if they have not done so already.

[12:06] Simon Linden: hmmm, that rule needs some limit checks, as any gathering with one linden and 4 or less people really shouldn't be a plethora

[12:06] Kallista Destiny: Oh your not just updating the Server code but the hosting OS?

[12:06] Simon Linden: Yes Kallista

[12:06] Andrew Linden: Yes Kallista, and GCC version along with it.

[12:06] Simon Linden: Machines are getting re-imaged today

[12:07] Kallista Destiny: Oh dear Be afraid, be very afraid.

[12:07] Andrew Linden: The reimaging deploys take a little longer than the previous deploy methods.

[12:07] Kallista Destiny: Is this going to be the usualy deploy procedure from now on?

[12:08] Andrew Linden: Actually, the pathfinding beta has been using a newer version of GCC since before it was ever put on the beta grid.

[12:08] Andrew Linden: Yes Kallista, however I don't think the subsequent reimaging deploys will take as long as this one.

[12:08] Andrew Linden: but I'm not sure what the times will be yet

[12:09] Simon Linden: fwiw the deploy info in the forums is available at

[12:09] Kallista Destiny: So you can keep the OS uptodate

[12:09] Melvin Starbrook: oopsie lol

[12:09] Andrew Linden: yes Kallista, it will be much easier to update the OS going forward

[12:09] Nalates Urriah: GCC? Did I miss the defintion?

[12:09] Andrew Linden: each deploy will actually be a complete OS reinstall

[12:10] Andrew Linden: GCC - Gnu C Compiler

[12:10] Nalates Urriah: Thx

[12:10] Kallista Destiny: nods nods

[12:10] Andrew Linden: which is used to compile the C/C++ code we use for many of our servers

[12:11] Andrew Linden: Pathfinding is also getting very close to done (the project I've been working on with Falcon)

[12:11] Andrew Linden: It should go out into a real RC this week (tomorrow)

[12:11] Nalates Urriah: there a compiler on each machine? Or you mean just your work machines used for building code?

[12:12] Andrew Linden: otherwise, I don't have much maintenance news myself. I haven't been able to work on the LSL syntax info as a service download, or some of the other maintenance issues I wrote down in my notes recently

[12:13] Andrew Linden: the server code is compiled on build machines, and deployed to the simhosts as part of the OS image

[12:13] Simon Linden: I've been wrapping up a bit of code adding the parcel names to "Top Scripts" and "Top Colliders"

[12:13] Simon Linden: ... and fixing the name filters

[12:13] Kallista Destiny: Do you distribute compiles across the servers?

[12:13] Nalates Urriah: Thx Andrew

[12:13] Andrew Linden: the simhosts have compilers installed I believe, but they are not used

[12:13] Kallista Destiny: nods

[12:14] Kallista Destiny: althought you could use an idle Sim slot to compile I suppose.

[12:14] Simon Linden: It's better used doing the back-end work and maybe cutting down on lag

[12:15] Melvin Starbrook: willl there be anything changed viewer sided?

[12:15] Andrew Linden: Melvin, you mean related to the OS upgrade or the pathfinding changes?

[12:15] Melvin Starbrook: os upgrade

[12:16] Andrew Linden: The pathfinding changes will have a bunch of viewer changes associated, but not the os upgrade.

[12:16] Melvin Starbrook: ok thank you

[12:16] Baker Linden: I don't have much to report, other than I'm still trying to get SVC-4968 fixed. This bug hates me. :(

[12:16] JIRA-helper:

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

[12:16] Jonathan Yap: A new viewer just came out, does that mean people will have to use a Project viewer when PF goes into RC?

[12:17] Kallista Destiny: My understanding is that passive use of Pathfinding does not require a special viewr, is this correct?

[12:17] Kelly Linden: I did a fix yesterday for sorting resident names in the about land and estate access/ban lists. I'm not quite up on the viewer release process to say with any certainty when it would hit a publicly visible repository, but it is done I swear!

[12:17] Jonathan Yap: yes, it is only when you want to work on nav meshy things

[12:17] Simon Linden: Yes, Kallista

[12:17] Kallista Destiny: nods

[12:18] Andrew Linden: Right, passive use of the pathfinding feature will not require a new viewer.

[12:18] Nalates Urriah: Is it just me or is there more of a delay rezing a first object in a PF region than a non-PF regions?

[12:18] Melvin Starbrook: mmm means when using an older viewr i cannot make pathfinding things.. or does it also mean i cannot see them?

[12:18] Andrew Linden: You can even make pathfinding characters without the new viewer

[12:19] Melvin Starbrook: hihi you allready answer andrew :)) tank you :)))

[12:19] Melvin Starbrook gigles

[12:19] Andrew Linden: however if you want to try to debug your characters and figure out why they are not navigating through areas that appear to be navigable then you would find the new pathfinding viewer's features to be helpful

[12:19] Qie Niangao: Speaking of pathfinding changes: suddenly vehicles can't cross from PF RC-channel sims into non-PF sims. Any chance that incompatibility gets fixed in the new main server release?

[12:20] Andrew Linden: Qie, all vehicles? or just vehicles that use mesh shapes?

[12:20] Qie Niangao: all vehicles, physical or NPVs

[12:20] Andrew Linden: I don't know of a restriction for region crossling of vehicles in general

[12:20] Widely Linden: have you filed a Jira?

[12:20] Qie Niangao: not even vehicles, in fact. my phantom single-prim probes get stuck

[12:21] Qie Niangao: not yet, I thought it might be a known thing...

[12:21] Andrew Linden: but I think mesh objects can only pass one way (old regions to new?) when encountering a region boundary between a pathfinding region and non-

[12:21] Qie Niangao: (started a thread about it earlier this morning)

[12:21] Rex Cronon: eventually all sims will be using PF, so it might not matter

[12:21] Qie Niangao:

[12:21] Andrew Linden: Ok Qie, we'll try to look into that today.

[12:22] Qie Niangao: It has the balloons and trolleys in Bay City *very* confused, among other things, hehehe

[12:22] Andrew Linden: It will matter while Pathfinding is in RC, which will be at least a week, probably more.

[12:22] Jonathan Yap: I'd like to bring up this issue which affects people weekly (unexpected disconnect from sim) and which is probably pretty easy to fix: svc-7759

[12:22] JIRA-helper:

[#SVC-7759] Scheduled restart notices no longer noticeable

[12:23] Andrew Linden: Hrm...

[12:23] Jonathan Yap: Maybe wrong message type is being sent, or no "/" at start of message?

[12:23] Andrew Linden: Simon, do you know if SVC-7759 could be a symptom of the OS-upgrade-deploy process?

[12:24] Jonathan Yap: That date on that jira is in March

[12:24] Kallista Destiny: Happens with EM restarts too

[12:24] Andrew Linden: Oh ok, probably not releated then.

[12:24] Nalates Urriah: The problem has been around awhile.

[12:25] Simon Linden: It might be, if they're using a different way to shut down teh region, but I'd suspect it's really just the region doesn't sent the right type of alert

[12:25] Andrew Linden: Yeah, sounds like a bug in the deploy process glue code, probably not a bug in the simulator itself.

[12:25] Kallista Destiny: That was my understanding from yesterday's scripting meeting.

[12:26] Andrew Linden: We'll have to raise the issue with the #deploy team

[12:26] Andrew Linden: but someone in the maint team will probably have to search for and find the fix

[12:26] Simon Linden: It's been around for a while, I think it's a simple sim change if we can verify how they shut down regions

[12:26] Kallista Destiny: Inany case it is a PITA as the old Bright Yellow region going down nolonger shows up, and so it's easy to miss.

[12:27] Jonathan Yap: I'm surprised there is more than one way a sim sends out a "shutting down" message

[12:27] Simon Linden: It does sent a notification, but it's not a type that gets in your face enough to demand to be noticed

[12:27] Kallista Destiny: Especally if your wrapped up in a build or scripting loop

[12:27] Nalates Urriah: I've been caught a couple of times and kicked off

[12:27] Jonathan Yap: The Product Team is going to consider adding a sound when one of those messages arrives

[12:28] Simon Linden: Right, I've been hit by it too -- it's easy if you get heads-down into other things and ignore it

[12:28] Andrew Linden: Oh I see, the notices are indeed being sent, but they are not visible enough.

[12:28] Jonathan Yap: yea, it is not in the normal Alert type format

[12:28] Simon Linden: Right, they just pile up in the upper right notices, I believe

[12:29] Kallista Destiny: Or lower right for TPV

[12:29] Jonathan Yap: Which makes me think it is just a simple problem with wrong type or lack of leading "/" character

[12:29] Jonathan Yap: All Alert messages start with a "/"

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

[12:30] Andrew Linden: ok, well I've taken a note of SVC-7759

[12:31] Jonathan Yap: Thank you Andrew

[12:31] Kallista Destiny: Thank you very much

[12:33] Andrew Linden: Anybody have a new topic for discussion?

[12:33] Arawn Spitteler hadn't been able to sit on blocks, last time here.

[12:34] Arawn Spitteler: Today, I sit on this railing first try, but last I was here, I couldn't sit on a block, nor this railing.

[12:34] Nalates Urriah: I'd like to mention Loki's Linden Thank you for SL9B, which has to be about the cutest thank you I've ever seen. And add my thinks for all the stuff you guys put up with.

[12:34] Yuzuru Jewell: Do you know about bump road problem? Now I am searching the jira.

[12:35] Andrew Linden: Nalates, that Thank You note was in the forums?

[12:35] Nalates Urriah: Its in SL9B the LL building he made.

[12:35] Rex Cronon: would the cloud party be an interesting topic:)

[12:36] Simon Linden: oh the Linden office - that is great

[12:36] Andrew Linden: How long is SL9B going to be up?

[12:36] Nalates Urriah: It is, Very well done and so cute.

[12:36] Nalates Urriah: About 24 more hours I think

[12:37] Nalates Urriah: Loki has a copy in his region

[12:37] Nalates Urriah: I found the off SL9B copy in the Events and Destinations thing

[12:39] Simon Linden: Yuzuru - what is the bump problem you were talking about? Is that new?

[12:40] Yuzuru Jewell: I think it is since the bigining of June.

[12:41] Arawn Spitteler: On roads, where prims join? Hasn't that been solved?

[12:41] Nalates Urriah: Yuz haas been fighting with it for awhile. Nyx has been talking to Falcon (?) about it,

[12:41] Yuzuru Jewell: I told about it on mesh UG.

[12:42] Nalates Urriah: Its where mesh sections of road come together. Vehicles seem to launch. Sigificant bump.

[12:42] Andrew Linden: The symptoms of the "road bump" problem is that vehicles hit a bump at prim seams? or is it something else?

[12:42] Andrew Linden: ah ok

[12:42] Rex Cronon: the roads r not "meshing" right;)

[12:43] Kallista Destiny groans

[12:43] Andrew Linden: I didn't think that bug was quite fixed yet, although I thought it had been reduced, since the Havok7 upgrade

[12:43] Andrew Linden: I can't remember the state of our possible fixes for that bug. I'll have to ask Falcon what he remembers

[12:43] Yuzuru Jewell: and one more jira.

[12:43] JIRA-helper: [#VWR-28416] Mesh: Default convex hull is smaller than it should be.

[12:45] Andrew Linden: that's right... there are two distinct features/bugs that contribute to that "bump at road seams" problem

[12:46] Andrew Linden: (1) Sometimes the physics engine computes collisions with moving objects and the "hidden" faces between the prims

[12:47] Andrew Linden: which have collision normals that are pointing against the objects motion, and cause the jump

[12:47] Yuzuru Jewell: So we can't make mesh road.

[12:47] Andrew Linden: and (2) the effective collision shape of the object is slightly smaller than the apparent shape

[12:48] Rex Cronon: i think u might be able, if u incread the thickness of the meshes

[12:48] Andrew Linden: so road prims do indeed have a gap at the seams, as far as the collision engine is concerned

[12:48] Yuzuru Jewell: We covered the joints on normal prim. Then the car jumps twice.

[12:48] Rex Cronon: if u increase*

[12:48] Andrew Linden: Right, I'd guess that item (1) is the major cause

[12:49] Andrew Linden: very hard to fix because it is a quirk of the physics engine itself

[12:49] Yuzuru Jewell: I want to know how to make mesh road to avoid the bump.

[12:49] Arawn Spitteler recalls a remedy for prim seam bump;; it had something to do with a script that would adjust a characteristic to its default, rather than leave it at default.

[12:49] Andrew Linden: in fact, we have some hackery on the vehicles that tries to detect and reduce the problem, but that system doen't catch all fo the cases

[12:49] Nalates Urriah: Is it possible to eliminate the wrong facing normals on the road mesh? Could one reverse or remove the face with the wrong pointing normal? I suppose a track would then have to 'one-way' only.

[12:50] Yuzuru Jewell: To delete vertical face of load?

[12:50] Nalates Urriah: yes or reverse it...

[12:50] Andrew Linden: I believe people discovered that they can eliminate the bumps by using box prims with a tapered bottom face. That was a workaround from a long time ago.

[12:51] Yuzuru Jewell: Thank you. We will try it.

[12:52] Arawn Spitteler: You may want road prims at less than 10 meters, if you want to taper the normals away.

[12:52] Andrew Linden: although that solution was from the era before we shrunk the collision shape so that dynamic objects really do appear to touch the surface on which they are resting

[12:52] Andrew Linden: which is what item (2) is all about

[12:52] Yuzuru Jewell: not 64m but 10m?

[12:53] Simon Linden: I'd try big ones first and see what you can do

[12:53] Arawn Spitteler: Are we allowed 64, now?, then a little less than that.

[12:53] Andrew Linden: why limit to 10m Arwan?

[12:53] Arawn Spitteler: I haven't been building in that long

[12:53] Jonathan Yap: 64 became the new limit when mesh came out

[12:54] Andrew Linden: which was... one or two years ago? I've lost count

[12:54] Jonathan Yap: last August(?)

[12:54] Kallista Destiny: Not quite that long

[12:54] Qie Niangao: that sounds right

[12:54] Yuzuru Jewell: It is important to make physics shape of mesh roads.

[12:54] Nalates Urriah: crashed

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

[12:55] Nalates Urriah: Since a race track is one-way, do small steps down reduce the bump?

[12:55] Andrew Linden: Nalates, I would guess that that would help, yes.

[12:56] Rex Cronon: if u go from the one the is a little higher to the one that is a little lower than the bump might not be so drastic?

[12:56] Andrew Linden: Other solutions people have used are giant megaprims for flat tracks

[12:56] Rex Cronon: one that is a*

[12:56] Yuzuru Jewell: If it is only one-way, we can't make mesh street roads.

[12:57] Yuzuru Jewell: Andrew, can I make megaprims?

[12:57] Nalates Urriah: Sure just put up OneWay signs... :)

[12:57] Arawn Spitteler: The fix I recall was a script that would adjust the taper, and then adjjust back, so the script did nothing, but the problem was fixed. I suppose Havoc never figured out why.

[12:57] Andrew Linden: No, we still limit new prims to 64m

[12:57] Arawn Spitteler: 64 meters used to be megaprims.

[12:58] Arawn Spitteler: Can I make a disk of 120 meters?

[12:58] Andrew Linden: nope

[12:58] Arawn Spitteler: Fixed that?

[12:59] Kennylex Luckless: I think it is posible to torture a prim so it can cover a sim.

[13:00] Andrew Linden: I wouldn't recommend tortured prims to be used for racetracks.

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

[13:00] Yuzuru Jewell: Megaprim.. I don't use it.

[13:00] Rex Cronon: can't u apply a mesh to a "real" megaprim(256^3m)?

[13:00] Andrew Linden: no you can't apply mesh to an existing object

[13:00] Qie Niangao: apply a mesh? this is new magic

[13:01] Rex Cronon: right

[13:01] Andrew Linden: you must upload the mesh, and it is stuck to the obejct in inventory

[13:01] Rex Cronon: at least u can do it with a sculptie:)

[13:02] Andrew Linden: Ok, I've got to go get some food.

[13:02] Andrew Linden: Thanks for coming everyone.

[13:02] Rex Cronon: tc andrew

[13:02] Sahkolihaa Contepomi: Enjoy.

[13:02] Jonathan Yap: Thank you everyone

[13:02] Yuzuru Jewell: Thank you Andrew.

[13:02] Qie Niangao: Thanks, Andrew, everybody... have fun!

[13:02] Simon Linden: Thanks everyone for the good discussion

[13:02] Rex Cronon: tc everybody. and have a nice day:)

[13:02] Kennylex Luckless: Yet another Linden adicted to food.

[13:02] Rex Cronon: tc simon

[13:02] Kallista Destiny: Thank you all

[13:02] Nalates Urriah: bye all

[13:02] Arawn Spitteler: Some of them are even addicted to Oxygen

[13:02] Melvin Starbrook: have a nice day all :)

[13:03] Kennylex Luckless: :-9


Prev 2012.06.22 Next 2012.06.29