Chat log from the meeting on 2009-11-03

From OpenSimulator

Revision as of 06:08, 10 June 2011 by MakoBot (Talk | contribs)

Jump to: navigation, search


Tuesday Meeting @ Wright Plaza on OSGrid


[10:52] paulie Flomar is Online
[10:53] Nebadon Izumi: hello
[10:53] Friendly Harbour: hi neb
[10:53] Nebadon Izumi: almost forgot about the meeting
[10:53] Nebadon Izumi: this server totally died last night
[10:53] Friendly Harbour: ... and everyone else
[10:53] Nebadon Izumi: we had HD failure
[10:53] Friendly Harbour: ouch
[10:53] Friendly Harbour: never fun
[10:53] Nebadon Izumi: i was up until 3am reinstalling SuSe
[10:53] Justin Clark-Casey: Hi Nebadon. My sympathies on the hd failure
[10:53] Justin Clark-Casey: nebadon: yeuck
[10:54] Nebadon Izumi: ya wasnt so bad i had most everything backed up
[10:54] Nebadon Izumi: but it was a self install server
[10:54] Justin Clark-Casey: that's something at least
[10:54] Nebadon Izumi: so i had to KVM in again and re-install suse
[10:54] Nebadon Izumi: seems to be running great though
[10:54] Nebadon Izumi: its on 11 hours uptime right now
[10:54] Adelle Fitzgerald: if anything its good practice ;)
[10:54] Justin Clark-Casey: shall we continue holding the meeting at this time?
[10:55] Nebadon Izumi: ya this works for me
[10:55] Friendly Harbour: are you not using RAID to protect from HD failures?
[10:55] Nebadon Izumi: im open for whatever is best
[10:55] Justin Clark-Casey: cool - works much beter for me too - an hour later cuts into my dinner time
[10:55] Nebadon Izumi: ok
[10:55] Justin Clark-Casey: so the sims have been running well?
[10:55] Nebadon Izumi: ya running awesome
[10:55] Nebadon Izumi: this server has been up just about 11 hours now
[10:55] Nebadon Izumi: this simulator i mean
[10:56] Nebadon Izumi: 550m on show stats
[10:56] Justin Clark-Casey: nice.
[10:56] Nebadon Izumi: oh btw
[10:56] Nebadon Izumi: i lost the SSH keys
[10:56] Nebadon Izumi: can you email me your key again to michael@osgrid.org justin
[10:56] Justin Clark-Casey: sure
[10:57] Hiro Protagonist is Offline
[10:57] Nebadon Izumi: probably good i clear em all out anyway and get everyone to send me them again
[10:57] Justin Clark-Casey: heh. One thing I am noticing is that automatic landing seems nearly impossible here
[10:57] Nebadon Izumi: ya
[10:57] Hiro Protagonist is Online
[10:57] Nebadon Izumi: i think jhurliman noted that in the logs at one point as currently broken
[10:58] Justin Clark-Casey: ok
[10:58] Adelle Fitzgerald: you can land on terrain fine, its just prims
[10:58] Justin Clark-Casey: ah I see
[10:59] Justin Clark-Casey: well, I spent a large chunk of my day wrestling with the Ubuntu 9.10 upgrade
[10:59] Justin Clark-Casey: as usual my x server broke, and even samba went a bit funny this time
[10:59] Justin Clark-Casey sighs
[10:59] Nebadon Izumi: heh that sucks
[10:59] Nebadon Izumi: thats sorta what drives me nuts about linux
[11:00] Justin Clark-Casey: not as bad as restoring a server at 3am ;)
[11:00] Nebadon Izumi: hehe
[11:00] Justin Clark-Casey: well, I still much prefer it to Windows. At least in Linux I have a fighting chance of working out for myself what is going wrong
[11:00] Nebadon Izumi: hehe ya im not knocking it
[11:00] Nebadon Izumi: just drives me crazy
[11:00] Charles Krinke: +1 Justin and good morning
[11:00] Justin Clark-Casey: heh
[11:00] Justin Clark-Casey: hello Charles, salutations!
[11:01] Penny Lane: Hiya CK, long time no see. All well?
[11:01] Nebadon Izumi: server is running awesome at the moment, its been up for over 11 hours now
[11:01] Hiro Protagonist: Heya Charles, Justin, all
[11:01] Penny Lane: Woot Neb!
[11:01] Justin Clark-Casey: Hi Hiro
[11:01] Charles Krinke: coolness, Nebadon
[11:01] Hiro Protagonist: a little distracted, prepping out a server
[11:02] Penny Lane: Wow, group notices. Sounds like society is on the way here ;-)
[11:03] Justin Clark-Casey: well, looks like we've managed to get rid of all the lightweights on this meeting with all the shenanigans over the past month. Either that or every one else expects to be here an hour later
[11:03] Justin Clark-Casey grins
[11:03] Snoopy Pfeffer: hi all :)
[11:03] Charles Krinke: Probably an hour later, I would suspect
[11:04] Friendly Harbour: not all the lightweights have been scared away ... i'm still here :)
[11:04] Charles Krinke: Can I tossout a subject?
[11:04] Snoopy Pfeffer is Online
[11:04] Justin Clark-Casey: Friendly: heh
[11:04] Nebadon Izumi: sure
[11:05] Snoopy Pfeffer: Justin your avatar is completely back for me
[11:05] Snoopy Pfeffer: all others look ok
[11:06] Nebadon Izumi: probably on 2nd monitor?
[11:06] Nebadon Izumi: what did you want to talk about Charles?
[11:06] Penny Lane: This is the fewest number of greys I've ever seen here, things are definitely improving.
[11:07] Charles Krinke: Can someone consider experimenting with the difference in a 2-5 linked prim set and editing between SL and OS to see if we can intuit some direction on linked set editing issues?
[11:07] Justin Clark-Casey: yeah, it's the second monitor. I'll move this and hide my irc channels instead :)
[11:07] Snoopy Pfeffer gave you Snapshot : OSGrid Meeting Center, Wright Plaza (140, 145, 32).
[11:07] Snoopy Pfeffer: ah ok lol
[11:07] Penny Lane: Only Hiro is grey to me ... special text client?
[11:07] Friendly Harbour: justing is black and hiro is grey, everyone else are here in full colour :)
[11:07] Justin Clark-Casey: better?
[11:07] Hiro Protagonist: nope meerkat Penny
[11:07] Snoopy Pfeffer: not for me
[11:07] Hiro Protagonist: I just rebaked just now tho
[11:07] Kitto Flora: Charles, is that editing physical link sets?
[11:08] Nebadon Izumi: cam onto justin
[11:08] Nebadon Izumi: cam around him
[11:08] Nebadon Izumi: he will rez in
[11:08] Snoopy Pfeffer: ah now
[11:08] Penny Lane: Whatever you did, it worked JCC
[11:08] Snoopy Pfeffer: better :)
[11:08] Arthur Valadares: I don't think rebaking helps much.. My guess is avatar textures are somehow stuck behind other textures in line
[11:08] Arthur Valadares: I was just looking into this :)
[11:08] Justin Clark-Casey: yep, the old second monitor issue - which has existed from when I started using the linden lab client
[11:08] Justin Clark-Casey: hey arthur!
[11:08] Arthur Valadares: Hey!
[11:08] Justin Clark-Casey: though in fairness it may be an nvidia windows driver issue
[11:09] Arthur Valadares: you showed up btw (no longer black)
[11:09] Charles Krinke: No, not physical. I think looking for the simplest set that shows a difference might let us intuit some linked set editing problems with non-physical.
[11:09] Melanie Milland is Online
[11:09] Charles Krinke: I understand that editing a complicated linked set in OS "blows" up and am hoping to find a way to make it better.
[11:10] Nebadon Izumi: ya ive had some recent issues with linking myself
[11:10] Snoopy Pfeffer: I did try linking huge link sets 2 days ago
[11:10] Nebadon Izumi: as some other testers / builders i work with
[11:10] Snoopy Pfeffer: linking two 2000 prim link sets did crash the viewer 3 times
[11:10] Charles Krinke: But, making it better depends on finding the simplest test case that 'blows up" and how it "blows up"
[11:10] Nebadon Izumi: mostly editing linksets
[11:10] Snoopy Pfeffer: but it was a viewer problem
[11:11] Mark Spencer: last time I tried there was a 256? max
[11:11] Nebadon Izumi: say you want to move or adjust 1 prim in a linkset
[11:11] Nebadon Izumi: it tends to explode
[11:11] Arthur Valadares: Ive seen weird behaviors linking 200-300 linksets, but didnt crash
[11:11] Nebadon Izumi: not crash
[11:11] Nebadon Izumi: the linkset is destroyed
[11:11] Nebadon Izumi: it explodes or rotates all funky
[11:11] Snoopy Pfeffer: well finally I had a 4000 prim link set
[11:11] Arthur Valadares: Yea, I sasw some pieces of it dissapear
[11:11] Charles Krinke: There may be viewer differences also, so we have to be careful with viewer versions in this sort of testing.
[11:11] Snoopy Pfeffer: but moving and rotating was slow when I had that :)
[11:11] Nebadon Izumi: theres a bunch of differnt results you can get
[11:11] Snoopy Pfeffer: yes Neb I saw that already months ago
[11:11] Justin Clark-Casey: Does LL have a 256 max linkset size?
[11:12] Snoopy Pfeffer: that creates ghost prims and phanotoms
[11:12] Snoopy Pfeffer: phantoms*
[11:12] Kally Rakan: if the linkset is an attachment the editor sometimes switched randomly to editing the hud window and the linkset vanishes
[11:12] Hiro Protagonist: yes, for instance meerkat doesnt do 'undo' on current revisions
[11:12] Kitto Flora: There is a limit in SL
[11:12] Charles Krinke: nice red shirt, Justin
[11:12] Melanie Milland: LL has a max size of 256 anf complicated linkability rules
[11:12] Hiro Protagonist: hippo and no doubt the SL viewer do
[11:12] Snoopy Pfeffer: up to 2000 it did work well with Hippo
[11:12] Justin Clark-Casey: charles: it's all very minimalist. I also seem to have lost my hair some time back
[11:12] Justin Clark-Casey: hi melanie
[11:12] Hiro Protagonist: Hi Mel thanks for coming
[11:12] Snoopy Pfeffer: between 2000 and 4000 is a viewer problem
[11:13] Charles Krinke: morning,Melanie
[11:13] Kitto Flora: There is also alink limit dependong o na ratio of size to center-wpacing of the prims.
[11:13] Nebadon Izumi: ya its not so much the prim limit
[11:13] Nebadon Izumi: ive had 100 prim linksets explode
[11:13] Nebadon Izumi: editing just 1 prim
[11:13] Nebadon Izumi: rotate a prim inside the linkset one of the child prims
[11:13] Nebadon Izumi: and it goes boom
[11:13] Snoopy Pfeffer: did you experience wrong phantoms and ghost prims Neb?
[11:13] Charles Krinke: Well, I was just trying to find a way to simplify the problem and be able to make some progress, thats all.
[11:14] Nebadon Izumi: usually for me its 100% destroyed at that point
[11:14] Nebadon Izumi: no recovery
[11:14] Nebadon Izumi: unless i have a copy in inventory
[11:14] Nebadon Izumi: also Undo
[11:14] Nebadon Izumi: god undo sucks
[11:14] Snoopy Pfeffer: did that happen suddenly without editing?
[11:14] Nebadon Izumi: no while editing it
[11:14] Snoopy Pfeffer: I thought it has to do with rotations and movements of child prims
[11:15] Nebadon Izumi: editing a child prim in linkset rotate specificily
[11:15] Melanie Milland: there are issues with local/global conversion
[11:15] Snoopy Pfeffer: yes Melanie that is what I thought
[11:15] Melanie Milland: they can cause a prim to be moved to an absolute position corresponding to it's offset
[11:15] Melanie Milland: which again can trigger a region corssing
[11:15] Charles Krinke: The trick is always going to be to find the simplest test that reproduces an issue in order to let others work with a 'recipe'. And this probably means a maximum of 10 or 20 prims
[11:15] Kitto Flora: There is a known issue with script rotation of non-physical not correctly updating the collision position
[11:16] Snoopy Pfeffer: interestingly it can happen that the visual representation of physics can get out of sync
[11:16] Snoopy Pfeffer: creating ghosts
[11:16] Melanie Milland: i believe i have once heard that the issue occurs when the root prim is not rotated to 0,0,0
[11:17] Penny Lane: Once you find that minimal test, it really needs to be captured for regression testing.
[11:17] Hiro Protagonist: Melanie, what's the solution
[11:17] Snoopy Pfeffer: I did try to find that month ago and never succeeded :(
[11:17] Nebadon Izumi: an easy test is probably just link 100 cubes
[11:17] Nebadon Izumi: and rotate any one of the child prims
[11:17] Justin Clark-Casey: yeah, incorporating it into the automatic unit tests would be heaven
[11:17] Snoopy Pfeffer: it seems to be somehow random
[11:18] Nebadon Izumi: ah in automatic testings yea
[11:18] Nebadon Izumi: hmm
[11:18] Justin Clark-Casey: there are already some linking tests (though obviously incomplete)
[11:18] Mark Spencer: do the automatic tests check performance?
[11:19] Charles Krinke: My experience is that 'random' when testing software usually means we dont understand all the variables in play.
[11:19] Justin Clark-Casey: no. But that's much more difficult to do
[11:19] Snoopy Pfeffer: yep
[11:19] Snoopy Pfeffer: it was more than moving and rotationg child prims
[11:19] Hiro Protagonist: Ok
[11:19] Snoopy Pfeffer: I did try all such things and it was still ok
[11:19] Hiro Protagonist: just in this simple experiment I've caught a bug
[11:20] Hiro Protagonist: rotating one prim affects a different prim in the linkset 50% of the time
[11:20] Nebadon Izumi: yep
[11:20] Nebadon Izumi: occasionaly it does work
[11:21] Kitto Flora: Oh. Does it maybe involve a list that was converted to a hashset?
[11:21] Hiro Protagonist: it stopped when I said it
[11:21] Hiro Protagonist: probably 1 in 10 now
[11:21] Melanie Milland: no linking test can be considered valid unless it is executed on a drag copy of a lnk set, or a link set rezzed from inventory
[11:21] Nebadon Izumi: heh problem is it only takes 1 to run a days worth of work if your not careful
[11:21] Melanie Milland: any link set that was just linked inworld is broken
[11:21] Nebadon Izumi: run=ruin
[11:22] Melanie Milland: and it stays broken until one of the following becomes true:
[11:22] Melanie Milland: - the link set is drag-copied
[11:22] Melanie Milland: - the link set is taken and re-rezzed
[11:22] Dahlia Trimble is Online
[11:22] Melanie Milland: - the pbserver relogs
[11:22] Snoopy Pfeffer: hm I cannot build here - otherwise I could show you a simple thing that doesnt work
[11:22] Snoopy Pfeffer: likd two rotated prims
[11:22] Melanie Milland: observer*
[11:23] Snoopy Pfeffer: then rotating a child prim also rotates the other child lol
[11:23] Melanie Milland: this is because the viewer and the sim disagree about the LocalID of the prims in a linkset
[11:23] Melanie Milland: so rotating a prim will send a message to the server, which will rotate a prim with that local ID
[11:23] Dahlia Trimble: hi :)
[11:23] Snoopy Pfeffer: yep Melanie and there is some inconsistency between viewer and physics engine
[11:23] Melanie Milland: but the server's notion of which prim that is differes from that of the viewer
[11:23] Penny Lane: Hi Dahlia :-)
[11:23] Charles Krinke: Well, my hope is that someone will make a simple recipe, post it as a Mantis and let the developers work the issue until folks are happy
[11:23] Snoopy Pfeffer: creating gosts
[11:24] Snoopy Pfeffer: ghost*
[11:24] Justin Clark-Casey: melanie: Why do they differ? Is this simply buggy code or is there some deeper reason?
[11:24] Justin Clark-Casey: the local ids, that is
[11:24] Hiro Protagonist is Online
[11:24] Melanie Milland: something in OS seems to reorder things
[11:24] Melanie Milland: the viewer is not updated
[11:25] Kitto Flora: Does it maybe involve a list that was converted to a hashset?
[11:25] Justin Clark-Casey: I see
[11:25] Mark Spencer: who is the lead for building issues?
[11:25] Melanie Milland: there are no hashsets in the call path for linking
[11:25] Nebadon Izumi: ya if i rotate the root prim on Hiros blocks
[11:25] Nebadon Izumi: all the prims rotate accordingly
[11:25] Hiro Protagonist: meerkat ate me window manager lols
[11:25] Nebadon Izumi: like its being fed down to all the child prims too
[11:26] Dahlia Trimble: Is this the issue jhurlian was discussing yesterday on IRC?
[11:26] Hiro Protagonist: I love it when I can say the regions server is running tighter than the client
[11:26] Nebadon Izumi: hehe
[11:26] Snoopy Pfeffer: :)
[11:26] Justin Clark-Casey: mark: I suspect that nobody is currently claiming the lead on these issues
[11:26] Nebadon Izumi: see
[11:27] Nebadon Izumi: i rotated the root prim on this light cycle only 90 degreees
[11:27] Nebadon Izumi: the engire thing went cockeyed
[11:27] Snoopy Pfeffer: lol
[11:27] Dahlia Trimble: This has been a problem for quite a while hasn't it?
[11:27] Nebadon Izumi: ya
[11:27] Nebadon Izumi: forever
[11:27] Nebadon Izumi: hehe
[11:27] Snoopy Pfeffer: yes dahlia
[11:27] Snoopy Pfeffer: since at least 9 months
[11:27] Dahlia Trimble: so possibly not a regression?
[11:27] Nebadon Izumi: oh definatly not
[11:28] Melanie Milland: no regression
[11:28] Kitto Flora: Is this rotating problem recent? Did it used to work OK 6 months back?
[11:28] Hiro Protagonist: nope
[11:28] Nebadon Izumi: just discussing things that bother us the most
[11:28] Nebadon Izumi: hehe
[11:28] Melanie Milland: I have rewritten llinking twice
[11:28] Snoopy Pfeffer: never Kitto hehe
[11:28] Justin Clark-Casey: hey dahlia
[11:28] Melanie Milland: I would not call myself a lead on it
[11:28] Dahlia Trimble: hi jcc :)
[11:28] Melanie Milland: but of those present I have probably poked sticks at that code the most
[11:28] Hiro Protagonist: Melanie: you're probably the closest thing we have to a lead on it
[11:28] Charles Krinke: The reason I brought it up today is I hear from several folks that this is a deal-breaker for getting experienced SL builders to use OpenSim
[11:28] Dahlia Trimble: has it ever worked? I dont remember
[11:28] Nebadon Izumi: i dont know honestly
[11:28] Hiro Protagonist: Dahlia: it hasnt
[11:28] Nebadon Izumi: i want to say no it neve has
[11:28] Kitto Flora: This indeed was an issue with Physical lonksets and scripted rotation - I did fix that for physical a while back.
[11:28] Melanie Milland: dalia: no
[11:29] Snoopy Pfeffer: true Charles
[11:29] Melanie Milland: it's never worked as intended
[11:29] Hiro Protagonist: most of us have adopted workflows that involve unlinked builds
[11:29] Hiro Protagonist: linking then becomes a sort of final-stage packaging of the build
[11:29] Snoopy Pfeffer: a friend I know also asked me about link set problems
[11:29] Nebadon Izumi: but its a lesson in pain
[11:29] Melanie Milland: linking used to be risky
[11:29] Hiro Protagonist: but that is quite different from the workflows of secondlife builder types
[11:29] Nebadon Izumi: most of us do that becuase of how painful linking is
[11:29] Melanie Milland: i was able to remove the risk of a linkset exploding in the first rewrite
[11:29] Nebadon Izumi: most are learning the hard way
[11:29] Snoopy Pfeffer: workarounds are no solution :)
[11:30] Melanie Milland: my second rewrite involved unlinking individual child prims
[11:30] Hiro Protagonist: really, we should have at least as good support for building workflows as SL, and I'd really like to see better
[11:30] WhiteStar Magic: linksets being unlinked still can and do on occasion explode all over the place
[11:30] Hiro Protagonist: Melanie: that actually works quite well
[11:30] Hiro Protagonist: .
[11:30] Hiro Protagonist: Thanks
[11:30] Hiro Protagonist: ooc
[11:31] Charles Krinke: I suspect there is more then one problem and am merely advocating getting to where we can find *one* of the issues and solve it.
[11:31] Dahlia Trimble: I dont mean to belittle the issue, but it seems to me if an experienced SL builder considers it a dealbreaker, then they probably arent as good as they claim or they're using it as an excuse
[11:32] Melanie Milland: dahlia, it's because of muscle memory
[11:32] Hiro Protagonist: well, it's less about can object X be built in opensim
[11:32] Hiro Protagonist: yes, exactly
[11:32] Dahlia Trimble: muscle memory? lol
[11:32] Melanie Milland: it's just too tedious for them to relearn
[11:32] Melanie Milland: they can't be bothered
[11:32] Hiro Protagonist: you dont do a lot of building do you Dahlia?
[11:32] Nebadon Izumi: its just disheartening when your used to something and you build for hours then destroy it
[11:33] Nebadon Izumi: ive done it
[11:33] Nebadon Izumi: several times
[11:33] Mark Spencer: reliable building tools are critical to builders
[11:33] Dahlia Trimble: I do some, but I'm constantly relearning my 3d skills. Most content creators I know do so also
[11:33] Melanie Milland stage-whispers "She's a blonde..."
[11:33] Melanie Milland: :)
[11:33] Snoopy Pfeffer: yes Neb :(
[11:33] Hiro Protagonist: lol Melanie
[11:33] Nebadon Izumi: lol
[11:33] Dahlia Trimble: it's not a static field
[11:33] Hiro Protagonist: true Dahlia
[11:33] Kitto Flora: If you cant be bothered to support less-that-fantastically -flexible builders in OS then you will have few sim-load issues.
[11:33] Hiro Protagonist: but think if someone kept rewriting the interface to visual studio over and over again
[11:33] Dahlia Trimble: If you cant re-learn in the 3d world, may as well give up
[11:34] Kally Rakan: it's kinda like having a programming editor that randomly, irretrievably overwrights part of you code
[11:34] Hiro Protagonist: yep
[11:34] Hiro Protagonist: it's got to be fixed
[11:34] Melanie Milland: hiro, someone does
[11:34] Melanie Milland: they're called icrosoft
[11:34] Hiro Protagonist: not because its impossible to live without
[11:34] Nebadon Izumi: ok so your saying we will never be able to rotate child prims then and we should give up trying to fix it?
[11:34] Melanie Milland remembers VC++ 6
[11:34] Hiro Protagonist: but because its the right thing to do
[11:34] Dahlia Trimble: like I said, I dont mean to belittle the issue.
[11:34] Melanie Milland: well, it is an issue that needs to be addressed
[11:35] Justin Clark-Casey: I do think that making things easier for builders is the best path. But it does take people motivated to do the hard work to fix linking
[11:35] Melanie Milland: it is a small part of a larger issue, though
[11:35] Nebadon Izumi: its not about relearing i dont think anyone should have to relearn something because it doesnt work as it should either though
[11:35] Kitto Flora: OK - what a bout Mantis #4336?
[11:35] WhiteStar Magic: this also impacts scripted prim movement as well, doesn't it
[11:35] Hiro Protagonist: it does not impact movement
[11:35] Snoopy Pfeffer: bugs that frustrate even good builders we should get rid off
[11:35] Hiro Protagonist: it impacts the appearance of movement
[11:35] WhiteStar Magic: rotations
[11:35] Hiro Protagonist: the appearance of rotations
[11:36] Hiro Protagonist: it's a scene update bug
[11:36] Nebadon Izumi: its screwed me up a bunch
[11:36] Hiro Protagonist: not a scripting or physic bug
[11:36] WhiteStar Magic: MmmmHmmm
[11:36] Nebadon Izumi: all i can say
[11:36] Melanie Milland: 4336 is a s howstopper
[11:36] Melanie Milland: raising prio on it to urgent
[11:36] Hiro Protagonist: 4336 has seen some improvement in the past 24 hours after some of JHurlicorn's work yesterday
[11:36] Justin Clark-Casey: needs to be brought to jhurliman's attention
[11:37] Dahlia Trimble: so is the higher level protocol between the viewer and the server during object editing well understood?
[11:37] Penny Lane: Hurli needs to come in world a bit. It brings you face to face with bugs.
[11:37] WhiteStar Magic: I've got Head running right now and was doing some tests on it and it's a heck of a lot better then most all previous versions
[11:37] Nebadon Izumi: he comes in all the time
[11:37] Justin Clark-Casey: penny: who are you calling bugs? :)
[11:37] Dahlia Trimble: yep he's here a lot
[11:37] Penny Lane: Hehe
[11:38] Snoopy Pfeffer: lol
[11:38] Nebadon Izumi: he is not on IRC either
[11:38] Nebadon Izumi: or he would probably be here
[11:38] Nebadon Izumi: i think i recall intel guys have a prior meeting
[11:38] Nebadon Izumi: at the same time we do every week
[11:38] Justin Clark-Casey: oh
[11:38] Kitto Flora: Hiro: Do you claim that 4336 is fixed in some version of OS? If so, what region is running that version?
[11:38] WhiteStar Magic: correct neb... he mentioned that the other day
[11:39] Kitto Flora: Is he at the LL meeting?
[11:39] WhiteStar Magic: INtel Meeting
[11:39] Nebadon Izumi: no they have some kind of Intel weekly meeting
[11:39] Nebadon Izumi: something internal to intel
[11:40] Justin Clark-Casey: appearance is loking good today. I see no grey at all
[11:40] Hiro Protagonist is Online
[11:40] Nebadon Izumi: ya not too bad i saw some greys a bit ago, but now everyone is rezzed fully for me
[11:40] Melanie Milland: kitto: 4336 is most certainly and emphatically NOT fixed
[11:40] Nebadon Izumi: just delayed probably with this level of load
[11:40] Dahlia Trimble: I seem to remember some linkset rotation problems cropping up after a libomv update a while back, maybe a year or so ago
[11:41] Arthur Valadares: Im trying to figure out why people remain grey and some don't
[11:41] Kitto Flora: OK Melanie :(
[11:41] Dahlia Trimble: is this the same?
[11:41] Melanie Milland: and I see no grey people
[11:41] Nebadon Izumi: show stats reads 804mb memory
[11:41] Nebadon Izumi: 23 avatars
[11:41] Kitto Flora: lag?
[11:41] Charles Krinke: Let me toss out another idea and completely derail the meeting, since you gave me an inch, I'll take a mile.
[11:41] Kitto Flora: OK Melanie :(
[11:41] Arthur Valadares: I see just one grey
[11:41] Nebadon Izumi: feels pretty smooth to me
[11:41] Kitto Flora: Hum - chat packet lost?
[11:41] Mark Spencer: nope
[11:42] Dahlia Trimble: I count 27 avies
[11:42] Charles Krinke: Can we consider holding one or more plazas at more stable revisions and updating less frequently in an attempt to gain some understanding on our underlying progress.
[11:42] WhiteStar Magic: hell it even ran on Nebs P3 Dual CPU Server with 1 GB Ram
[11:42] Orion Hax: what is the network stats
[11:42] Nebadon Izumi: ya i did a 1000 prim drop on my p3 server
[11:42] Nebadon Izumi: didnt crash
[11:42] Nebadon Izumi: was lagged to death but did not crash
[11:42] Justin Clark-Casey: I see there is some rubberbanding now with this many people
[11:42] Melanie Milland: charles: as long as it doesn't get turned into a means to block/stall URM updates, that seems to be a good idea
[11:43] Nebadon Izumi: not so bad here
[11:43] Adelle Fitzgerald: i usually do keep teravus and seaprior plazas at what I consider to be 'as stable as'
[11:43] Adelle Fitzgerald: even if it means running a revision for two weeks
[11:43] Melanie Milland: maybe osg can offer those plaza versions as download alternativess?
[11:44] Hiro Protagonist: anyone can checkout the same hash, if they/we know what that is
[11:44] Nebadon Izumi: our releases are usuually as close to the master release as we can stay and still be stable
[11:44] Charles Krinke: Oh, no, I dont suggest blocking any updates, but merely finding a way to run a "stable" and a "bleeding edge" and gain some performance understanding of the differences on an ongoing basis.
[11:44] Nebadon Izumi: i can do another release again today
[11:44] Penny Lane: Opensim's mor stable than any viewer here currently, so you're winning ;-)
[11:44] WhiteStar Magic: Current Head is pretty decent neb
[11:44] Nebadon Izumi: ya
[11:45] Nebadon Izumi: we are only a few revs behind here right now
[11:45] Justin Clark-Casey: we do seem to be on a peak of the sine wave of stability, ignoring the physical prim rotation and linking issues :)
[11:45] Nebadon Izumi: 12 hours uptime here now
[11:45] WhiteStar Magic: git# 82554e9 - r11432 is the one I have been beating on for a bit
[11:45] Snoopy Pfeffer: yes the current version works great
[11:45] Server has been running since Tuesday, 11/3/2009 3:07:18 AM
[11:45] Hiro Protagonist: over 24 ~on lbsa plaza~
[11:45] Mark Spencer: how about 'I can Win' and 'Bring it on' levels?
[11:45] WhiteStar Magic: and teh crossings seems to have stabalized
[11:45] That is an elapsed time of 11:37:58.0315290
[11:45] Snoopy Pfeffer: even my sims with much traffic run 1 or 2 days without restarts
[11:46] Dahlia Trimble: anyone notice any new collision issues if they've updated in the last 8 hours or so?
[11:46] Hiro Protagonist: as far as releases and binaries go, we want to keep that as simple as possible so as not to kill anyone who is keeping up with that task
[11:46] Nebadon Izumi: i will do some testing in a bit Dahlia and let you know
[11:46] Nebadon Izumi: ive not tried any of todays updates yet
[11:46] Dahlia Trimble: ty nebadon :)
[11:46] Snoopy Pfeffer: was something changed how the http listener port responds?
[11:47] WhiteStar Magic: sorry Dahlia, I did not do any collision tests... what would l you like for tests ?
[11:47] Snoopy Pfeffer: suddenly with the new head version it does not respond to Monit tests anymore
[11:47] Hiro Protagonist: another problem, albeit very freakin minor, is the bent legs are back
[11:47] Snoopy Pfeffer: does anybody know why and how to fix that?
[11:47] Hiro Protagonist: now there's a sticky regression for ya
[11:47] Dahlia Trimble: it's not critical to do any additional testing, but if you notice any new collision issues pop up. please let me know
[11:47] WhiteStar Magic: will do
[11:48] Dahlia Trimble: ty :)
[11:48] Melanie Milland: hiro: with avatar_capsule_tilted off, it doesn't happen
[11:48] WhiteStar Magic: the only thing I have noticced in teh past few versions is that the avatar is sinking into the prim by roughly .2m when standing on it
[11:48] Melanie Milland: it's only the tilted capsule that makes it do that
[11:49] Hiro Protagonist: is on the default?
[11:49] Penny Lane: Stack trace end in OpenJPEG in Meerkat, regular as clockwork. How hard can it be to decode a format without dying, grrr ...
[11:49] Melanie Milland: on is still the default, but I would recommend changing it to off
[11:49] Hiro Protagonist: aha
[11:49] Justin Clark-Casey: +1 melanie
[11:49] Kally Rakan: whitestar, my avi tends to float above prims
[11:49] Hiro Protagonist: thanks Mel
[11:49] Hiro Protagonist: maybe we could make off the default?
[11:49] Melanie Milland: on is, as far as I remember, the old (fixed tilt) behavior
[11:49] WhiteStar Magic: Last night we ran some other tests which if you walk across region lines and back too fast while on prims, you fall through on the 3rd shot
[11:49] Dahlia Trimble: if you are sinking, your capsule may be compressed. Try opening the appearance editor and closing it again to reset it
[11:49] Melanie Milland: off is the new, slight variable tilt behavior
[11:49] Arthur Valadares: Hiro, by making tilted off, animation going uphill looks weird
[11:50] Hiro Protagonist: you'll fall through on the first pass in SL WS
[11:50] Arthur Valadares: So it's a tradeoff for now
[11:50] Hiro Protagonist: ahh
[11:50] Arthur Valadares: We've been observing the same issue and not sure what we want either
[11:50] Melanie Milland: well, i'd opt for a real fix
[11:50] Hiro Protagonist: +1
[11:50] WhiteStar Magic: Actually Hiro, on first criss cross it's fine, it's after teh 2nd crossing things get weird.
[11:50] Melanie Milland: like, dumping the capsule and meshing the avatars' bounding box
[11:50] Arthur Valadares: Last change was from nlin, physics is a very dangerous place
[11:51] Hiro Protagonist: where'd Dahlia go lol
[11:51] Justin Clark-Casey: arthur: any opinion on whether tilting default would be better on or off?
[11:51] Dahlia Trimble: I see a negative number for total frame time in my stats
[11:51] Arthur Valadares: I'ld say off is better, since you stand more then walk uphill
[11:51] Justin Clark-Casey: dahlia: cool - that must mean that we're going back in time :)
[11:51] Justin Clark-Casey: arthur: good point
[11:51] Dahlia Trimble: lol
[11:51] Kitto Flora: Falling thru prims at border crossings is a result of the fundimental way the system works. Its easily fixed in SL by suitable construction practices. Likeley here too.
[11:51] Snoopy Pfeffer: is the new ODE version already used by head?
[11:52] Dahlia Trimble: I dont think border crossing problems are related to collisions specifically, it's a more global issue
[11:52] Justin Clark-Casey: ping
[11:53] Hiro Protagonist: pang
[11:53] Kitto Flora: how is falling through prims not a collision issue?
[11:53] WhiteStar Magic: the prims do NOT cross teh bouundaries, they abutt the line on both ends
[11:53] Kitto Flora: Whitestar: Yes. Thats the problem
[11:54] Dahlia Trimble: if it's during a border crossing, odds are that physics isnt even involved yet.
[11:54] Dahlia Trimble: a prim only exists in one region
[11:54] Charles Krinke: Its always a good feeling when the meeting runs for the while hour. Kudos to Justin, Melanie, Dahlia and others for their efforts in the last month
[11:54] WhiteStar Magic: exactly
[11:54] Nebadon Izumi: ya arent you supposed to overlap prims from each region
[11:54] Kitto Flora: physics-collision for the avatar is always involved when the Av is walking
[11:54] Nebadon Izumi: extend the prims like 5m into the nieghbor sim on both sides
[11:54] Hiro Protagonist: Dahlia, how difficult do you suppose it would be to make a child agent for active physical obkects in the interest of retaining physics state when physical objects cross
[11:55] Justin Clark-Casey: yes, I'm very pleased that we now have a stable enough region again to actual do an hour's meeting (or at all :)(
[11:55] Justin Clark-Casey: it does provide a good feel for where we're currently at
[11:55] Nebadon Izumi: hehe ya its very nice indeed Justin
[11:55] WhiteStar Magic: but the overhang of the prim from Region-A into B, B will be phantom
[11:55] Dahlia Trimble: Hiro, Im not sure allowing one region to manipulate the scene of another is a good idea
[11:55] Nebadon Izumi: it was rought there for a while, but well worth the pain
[11:55] Kitto Flora: Nebadon: If you dont overlap them yo uwill surely fail. Unless the fundimental system is entirely revised.
[11:55] Nebadon Izumi: ya
[11:55] Nebadon Izumi: exactly Kitto
[11:55] Nebadon Izumi: hehe
[11:55] Justin Clark-Casey: nebadon: yeah, it can get quite frustrating in the dips inbetween
[11:55] Nebadon Izumi: mega regions
[11:55] Nebadon Izumi: . :P
[11:56] Mark Spencer: yays mega regions!
[11:56] Omar Abdelrahim loves megas
[11:56] Mark Spencer pines for the days of open AW regions
[11:56] Nebadon Izumi: though it seems there is some kind of mono bug on x64 that is preventing me from running mega regions on this particular server
[11:57] Dahlia Trimble: gotta go, bye all :)
[11:57] Kitto Flora: Is the region switch for avatars at +2M in OS?
[11:57] Justin Clark-Casey: bye Dahlia
[11:57] Dahlia Trimble is Offline
[11:57] Omar Abdelrahim: I spoke yesterday to Melanie about FreeSwitch.
[11:57] Omar Abdelrahim: which seems to be a bit of an orphan
[11:58] Omar Abdelrahim: is there anyone actively working on it now
[11:58] Omar Abdelrahim: cuz I need it
[11:58] Snoopy Pfeffer: it works well on my regions since months
[11:58] Nebadon Izumi: its a pain with NAT
[11:58] Nebadon Izumi: ive never gotten it working at my home regions
[11:58] Snoopy Pfeffer: ah ok
[11:58] WhiteStar Magic: a ton of people are constantly asking how to set it up and get it going
[11:58] Nebadon Izumi: only dedicated servers with no NAT walls
[11:58] Snoopy Pfeffer: I use it on a server on the internet
[11:58] Kitto Flora: Did anyone here work on region border crossing?
[11:58] Snoopy Pfeffer: not behind NAT
[11:59] Justin Clark-Casey: omar: not afaik. Though you might want to ask Teravus
[11:59] Omar Abdelrahim: My freeswitch server is on the internet, but my sims are natted
[11:59] Justin Clark-Casey: afair he was working on it last
[11:59] Omar Abdelrahim: Does anyone have that working?
[11:59] Melanie Milland: in my opinion, opensim should not even try to support NAT beyond what it does now.
[11:59] Snoopy Pfeffer: btw a friend works on a new voice implementation with speaker indication, lip sync and spreaker animations, all that sl has
[11:59] Melanie Milland: servers on the intenet have become dirt cheap
[11:59] Omar Abdelrahim: Nat is critical to the bulk of the region operators on OSG
[12:00] Melanie Milland: and home lines are not suited to the load a busy sim puts on them
[12:00] Snoopy Pfeffer: I will tell you as soon as it is ready for testing
[12:00] Snoopy Pfeffer: he will make it open source he said
[12:00] Omar Abdelrahim: ANd servers on the internet are dirt cheap unless you need performance or memory
[12:00] Omar Abdelrahim: and are willing to cede control;
[12:00] Adam Frisby is Online
[12:01] Adam Frisby: Aloha
[12:01] Snoopy Pfeffer: hi Adam :)
[12:01] Nebadon Izumi: hey there Adam
[12:01] Omar Abdelrahim: Home line include, in my case, 22 x 5 burstable to double that
[12:01] Justin Clark-Casey: Hi Adam
[12:01] Omar Abdelrahim: If I can get Freeswitch working in my ocnfig I'd be happy
[12:01] Omar Abdelrahim: If not, I'm willing to put some work in
[12:01] Melanie Milland: omar, that is still not comparable to 100M bidi
[12:01] Snoopy Pfeffer: Omar it is probably just NAT
[12:02] Justin Clark-Casey: ping
[12:02] Nebadon Izumi: heh did server croak?
[12:02] Snoopy Pfeffer: I got it working without problems on a server not behind a firewall
[12:02] Omar Abdelrahim: But in that case I'd like some help with how theing should fit together arcitecturally with the SO growth path
[12:02] Melanie Milland: freeswutch needs inbound connectivity. it requires special router setup
[12:02] Nebadon Izumi: see if it recovers or not
[12:02] Melanie Milland: that is beyond most home users
[12:02] Nebadon Izumi: there are no console errors
[12:02] Melanie Milland: which is why I say we should not support that
[12:02] Melanie Milland: and leave it up to packagers
[12:02] Justin Clark-Casey: it's definitely flaky. I lagged right outside the building for a moment there
[12:02] Omar Abdelrahim: Snoopy, may I talk to you later about your config?
[12:02] Nebadon Izumi: wow
[12:03] Nebadon Izumi: 289 threads
[12:03] Melanie Milland: because that also shofts the service calls to the packager
[12:03] Nebadon Izumi: in use on show threads
[12:03] Snoopy Pfeffer: yes sure
[12:03] Omar Abdelrahim: preciate it
[12:03] Justin Clark-Casey: mmmmm, maybe that's why I suddently cannot move
[12:03] Snoopy Pfeffer: it is the config descibed on the wiki page
[12:03] Nebadon Izumi: i dont think its going to recover
[12:03] Nebadon Izumi: hehe
[12:03] Justin Clark-Casey: alright, well, we got to an hour anyway
[12:03] Nebadon Izumi: ya
[12:03] Hiro Protagonist: did we lose it?
[12:03] Nebadon Izumi: strange the console looks nice and quiet
[12:03] Kitto Flora: No motion
[12:03] Justin Clark-Casey: 289 threads is a bit disturbing though
[12:03] Omar Abdelrahim: Melanie, OK if we talk later, 1 to 1
[12:03] Charles Krinke: Omar: If you can define a "DNS" for one server on your gateway, you can sidestep NAT issues
[12:03] Kitto Flora: We will be here for ever :)
[12:04] Omar Abdelrahim: I'd like advice
[12:04] Nebadon Izumi: ya i will get a kill -3 in a bit
[12:04] Nebadon Izumi: if it doesnt crash on its own before then
[12:04] Nebadon Izumi: we can keep chatting for now
[12:04] Nebadon Izumi: hehe
[12:04] Justin Clark-Casey: anyhow, I'm off - good meeting folks - sounds like people are going to continue chatting until the sim blows up anyway
[12:04] Nebadon Izumi: maybe it will receover you never know
[12:04] Hiro Protagonist: it's not like we planned to dance or anything rofl
[12:04] WhiteStar Magic: are osDynamicTextures working now ?
[12:04] Nebadon Izumi: ya its back on
[12:04] Nebadon Izumi: it never stopped working
[12:04] Nebadon Izumi: it was just crashy
[12:04] Mark Spencer: chat until the sim goes down he he
[12:04] WhiteStar Magic: how are you finding it ?
[12:04] Nebadon Izumi: and probably might be still
[12:05] Kally Rakan: is there a way to find out how many pending http calls are in progress?
[12:05] paulie Flomar is Online
[12:05] WhiteStar Magic: ...
[12:05] Melanie Milland: ##sim has stopped sendign textures
[12:06] Nebadon Izumi: ya threads are very high
[12:06] Nebadon Izumi: 284 threads on show threads
[12:06] WhiteStar Magic: Thiis is now running on teh new box, right neb ? with Suse 11.1 ?
[12:06] Server has been running since Tuesday, 11/3/2009 3:07:18 AM
[12:06] Snoopy Pfeffer: may I ask again why Monit suddenly has problems getting responses from the httpports?
[12:06] That is an elapsed time of 11:59:23.3693060
[12:07] Snoopy Pfeffer: something was changed during the last week
[12:07] workers: 286 (500); ports: 0 (250)
[12:07] Snoopy Pfeffer: before it did work
[12:07] Mark Spencer: thus spoke Zarathustra
[12:07] Snoopy Pfeffer: it simply sends a request to the http listener ports and expects some answer
[12:07] paulie Flomar is Offline
[12:07] Snoopy Pfeffer: but since today it does not work
[12:08] Nebadon Izumi: ..
[12:08] Nebadon Izumi pokes the simulator
[12:08] Snoopy Pfeffer: lol
[12:08] Snoopy Pfeffer: well Monit also tried that and did not get a reponse
[12:08] Hiro Protagonist: forget pokin it lol
[12:08] Snoopy Pfeffer: and started to do many restarts lol
[12:08] Hiro Protagonist: eeps
[12:09] paulie Flomar is Online
[12:09] Allocated to OpenSim : 899 MB
[12:09] Snoopy Pfeffer: how good is the heart beat function?
[12:09] Nebadon Izumi: ok well good meeting none the less
[12:09] Snoopy Pfeffer: does it detect most crashes?
[12:10] Nebadon Izumi: not really sure
[12:10] Charles Krinke: yes, good progress in the last month
[12:11] Snoopy Pfeffer: yes in some areas opensim is already better than sl and whenever I am in sl I see how bad it is there lol
[12:11] paulie Flomar is Online
[12:11] Snoopy Pfeffer: here is less lag
[12:11] Snoopy Pfeffer: far less
[12:12] Hiro Protagonist: yep
[12:12] WhiteStar Magic: I spent some time in SL yesterday and it was.... welll a good reminder of how nasty some things can be
[12:12] Kitto Flora: Client console here showing a lot of comms errors
[12:12] Snoopy Pfeffer: yes WhiteStar lol
[12:12] Snoopy Pfeffer: repeated packet sends?
[12:12] Snoopy Pfeffer: I also see that often since about 2 weeks, when that was introduced
[12:12] Hiro Protagonist: I can hear ya Paulie
[12:14] WhiteStar Magic: partial crash state
[12:14] Snoopy Pfeffer: really?
[12:14] WhiteStar Magic: meeting has been over an hour long and we are still "alive" such as it is
[12:14] Hiro Protagonist: nobody can move right now Paulie, we're just chattin
[12:14] Hiro Protagonist: heh
[12:14] Kally Rakan: be interesting to see if it recovers
[12:15] Nebadon Izumi: ya the region has been running for about 12 hours now too
[12:15] Snoopy Pfeffer: hehe
[12:15] WhiteStar Magic: LOL Snoopy
[12:15] Nebadon Izumi: oh
[12:15] Nebadon Izumi: i can chair hop
[12:15] Nebadon Izumi: lol
[12:15] Snoopy Pfeffer: sitting and standing up works
[12:15] Snoopy Pfeffer: yes
[12:15] Hiro Protagonist: hey, we're compatible with bluemars now
[12:15] Nebadon Izumi: lol
[12:15] Snoopy Pfeffer: lol
[12:15] paulie Flomar: lolz
[12:15] Melanie Milland: lol
[12:15] WhiteStar Magic: They updated, I was in tehre 2 days ago
[12:15] Nebadon Izumi: any difference?
[12:16] Nebadon Izumi: i already feel like i wasted too much time there
[12:16] WhiteStar Magic: yeah... I think they are getting into teh nuts of it
[12:16] Nebadon Izumi: lol
[12:16] WhiteStar Magic: I ran into several folks meandering about
[12:16] Nebadon Izumi: its hard to download it again
[12:16] WhiteStar Magic: yeah... it is a monster
[12:16] Snoopy Pfeffer: well I need to find out why Monit cannot monitor opensim regions anymore
[12:16] Nebadon Izumi: ya i never installed it again after reloading my machine
[12:16] Snoopy Pfeffer: see you all later :)
[12:16] Snoopy Pfeffer: bye bye
[12:16] Nebadon Izumi: see you snoopy
[12:16] Hiro Protagonist: good luck Snoopy
[12:16] WhiteStar Magic: See Ya Snoopy
[12:17] Snoopy Pfeffer: thanks :)
[12:17] WhiteStar Magic: 1 sec
[12:17] WhiteStar Magic: Ohh 1 Secx
[12:17] WhiteStar Magic: did you get that updated ?
[12:17] Snoopy Pfeffer: yes WhiteStar?
[12:17] Snoopy Pfeffer: opensim?
[12:17] Snoopy Pfeffer: yes
[12:17] Nebadon Izumi shouts: anyone stuck at the landing zone you can probably cam inside and sit ona chair
[12:17] WhiteStar Magic: and able to run teh bash script
[12:17] Snoopy Pfeffer: ah ok
[12:17] Snoopy Pfeffer: updating git did not work
[12:17] Snoopy Pfeffer: well I didnt want to compile git myself
[12:17] WhiteStar Magic: Hmmmm
[12:18] WhiteStar Magic: I hear that
[12:18] Snoopy Pfeffer: I use the lastest stable version of git
[12:18] WhiteStar Magic: I jumped up to 1.6.5 and it's still working fine
[12:18] WhiteStar Magic: for teh versioning
[12:18] Snoopy Pfeffer: but it should be possible to implement that script without the --oneline option
[12:19] Snoopy Pfeffer: like by using "head -n 1"
[12:19] Charles Krinke: I'm off for lunch. See you guys & gals a bit later
[12:19] WhiteStar Magic: there is a previous version resoltion thing but I have no idea where I hid it
[12:19] Hiro Protagonist: tc ckrinke
[12:19] Snoopy Pfeffer: bye Charles
[12:19] Snoopy Pfeffer: well we should not depend on trunk versions of all tools ;)
[12:19] Snoopy Pfeffer: Mono is enough we additionally have to compile from time to time, beside opensim :)
[12:20] WhiteStar Magic: Mono is hard enough for the NIX people to keep ontop of
[12:20] Snoopy Pfeffer: alpha software on an alpha platform using alpha tools sounds risky hehe ;)
[12:21] WhiteStar Magic: Tripple A
[12:21] Melanie Milland: it's fun
[12:21] Snoopy Pfeffer: yep

Personal tools
General
About This Wiki