Chat log from the meeting on 2008-11-11

From OpenSimulator

Revision as of 07:02, 10 June 2011 by MakoBot (Talk | contribs)

Jump to: navigation, search


Tuesday Meeting @ Wright Plaza on OSGrid


[10:57] Nebadon Izumi: hey Adam
[10:58] Ruud Lathrop: hello all
[10:58] Nebadon Izumi: Hello Ruud
[10:59] UUID Speaker: Jeffronius Batra: 814eb358-dfd9-479b-a7c8-b83307748e68
[10:59] Adam Frisby shouts: Hey Jeff!
[10:59] letifera Burtova: hm
[11:00] letifera Burtova: tak jsme v prdeli
[11:00] letifera Burtova: :D
[11:00] Jeffronius Batra shouts: Hi Adam!
[11:00] Adam Frisby shouts: We're inside for a change
[11:00] Jeffronius Batra: Ok
[11:01] Adam Frisby: You know, Neb - might be a good idea to put a sign if we move the location. heh.
[11:01] Nebadon Izumi: i did
[11:01] Nebadon Izumi: lol
[11:01] Nebadon Izumi: theres been a sign for weeks
[11:01] Adam Frisby: Aaaahh
[11:01] Nebadon Izumi: and we have had meetings here for well over a month
[11:01] Fly Man: Yup
[11:01] Adam Frisby: I see it. >_>
[11:01] Fly Man: On the outside :)
[11:01] Hiro Protagonist: sorry back
[11:02] Nebadon Izumi: welcome back
[11:02] Nebadon Izumi: uhg man they are painting the condo below me
[11:02] Nebadon Izumi: so if i start ranting about purple squirrels
[11:02] Nebadon Izumi: thats why
[11:02] Hiro Protagonist: Hi Adam, nice 'do
[11:02] Hiro Protagonist: LOL
[11:02] Hiro Protagonist: ...purple squirrels
[11:02] Nebadon Izumi: lol
[11:02] Jeffronius Batra: I will be lurking as much as possible. I am getting ready to conduct some meetings and training sessions.
[11:02] Nebadon Izumi: smells so bad
[11:02] Hiro Protagonist: we'd think you had just 'gone furr'
[11:03] Fly Man: So, did someone already see Whump
[11:03] Fly Man: or Charles ?
[11:03] Adam Frisby: He was on IRC
[11:03] Adam Frisby: I havent seen Charles though
[11:03] Website Display Panel v1.0: Loading URL, Please wait..
[11:03] Hiro Protagonist: no, but it is a little early yet
[11:03] Nebadon Izumi: yea no response from charles this morning
[11:03] Justin Clark-Casey is Online
[11:04] Charles Krinkeb is Online
[11:04] Fly Man: Ahh, there is Charles :)
[11:04] Alby Damden is Online
[11:05] Website Display Panel v1.0: Loading URL, Please wait..
[11:06] Charles Krinkeb: Morning, Jeffronius. Great to see you here!!
[11:06] Nebadon Izumi: hello Charles
[11:06] Nebadon Izumi: have you seen Yang?
[11:06] Jeffronius Batra: Hi Charles, great to be here. I'm trying to do two things at once but will do my best to follow along.
[11:06] Nebadon Izumi: looks like a terrain accident
[11:06] Charles Krinkeb: I'll look laer.
[11:06] Nebadon Izumi: k
[11:07] Nebadon Izumi: i tried fixing it, but it will take to long with tools
[11:07] Fly-Man- parts #osgrid-wp
[11:07] Charles Krinkeb: I can just flatten it and dig the little creek again.
[11:07] Nebadon Izumi: yea
[11:08] Charles Krinkeb: Part of the deal when one is allowing experimentation, but thats ok.
[11:08] Justin Clark-Casey is Online
[11:08] justincc_ joins #osgrid-wp
[11:08] ChrisD1 joins #osgrid-wp
[11:08] AdamZaius: ohai
[11:08] daTwitch joins #osgrid-wp
[11:08] Nebadon Izumi: hey Justin
[11:09] Chris D is Online
[11:09] Justin Clark-Casey: hello nebadon
[11:09] paulie Flomar is Online
[11:09] Justin Clark-Casey: oh my god - frame time tastic
[11:09] Charles Krinkeb: Shall we get going, or wait for more?
[11:09] paulie Flomar: hey. sorreh. :)
[11:09] Mags Portello: hi there
[11:10] Adam Frisby: ooh nice.
[11:10] Nebadon Izumi: we can get started i think
[11:10] Hiro Protagonist: Hey, Paulie, long time no se bub
[11:10] Poppet McGimsie: howdies everyone
[11:10]
[SilverFox] joins #osgrid-wp
[11:10] Hiro Protagonist: Hiya Poopet
[11:10] Adam Frisby: We need better frame time reporting I think
[11:10] whump joins #osgrid-wp
[11:10] paulie Flomar: boo! :)
[11:10] whump: hello
[11:10] Justin Clark-Casey: in what sense?
[11:10] Nebadon Izumi: hey whump
[11:10] Hiro Protagonist: Heya WHump :)
[11:10] Adam Frisby: Justin: our frame time only shows us what the linden server has stats for
[11:10] Poppet McGimsie: howdie whump
[11:10] SachaMagne joins #osgrid-wp
[11:10] SachaMagne: hi all
[11:10] Pixelcat joins #osgrid-wp
[11:11] Charles Krinkeb: Ah, Whump just arrived. We have about a dozen folks in WP and some on IRC. We traditionally start with a discussion of OGP and interop first.
[11:11] Adam Frisby: But our main Update() event fires a whole bunch of subfunctions (about 6-10) which we could get 'timing statistics' for.
[11:11] Poppet McGimsie: hi
[11:11] paulie Flomar: Howdeh, all.
[11:11] Justin Clark-Casey: yep that's true - Teravus at one time was stuffing those in other numbers
[11:11] Justin Clark-Casey: but we ran out of slots
[11:11] Adam Frisby: Stats module could record those and we could get some better idea about what specific part of the frame is taking so long.
[11:11] Justin Clark-Casey: yes, that's a good point
[11:11] Adam Frisby: oooh. yeah looking nice today. >_>
[11:12] Hiro Protagonist: architecture and measurements - the best of tools
[11:12] Charles Krinkeb: We released 0.6.0 on Saturday and it is SVN r7176. Are there any interop issues with the release we need to be concerned with?
[11:12] Adam Frisby: I'm still amused how we can maintain a frametime of 1300ms yet still have 53fps.
[11:13] Fly Man: Whump, any problems with the latest 0.6 with OGP ?
[11:13] Justin Clark-Casey: heh
[11:13] Adam Frisby: and if there arent any problems with OGP, we can make some. :D
[11:14] Fly Man: *lol*
[11:14] coyled joins #osgrid-wp
[11:14] Hiro Protagonist chuckles
[11:14] whump: Fly Man: other than the issue on our side, no, I'm actually connected to my region using top of trunk.
[11:14] Hiro Protagonist: awesome Whump
[11:14] Charles Krinkeb: Coolness, Whump.
[11:14] Fly Man: Nice :)
[11:15] Charles Krinkeb: Are there any significant issues from others with 0.6.0 that we need to talk about?
[11:15] whump: We are in the process of making things "do the right thing" with respect to caps on our sims, so the AD has been borked.
[11:15] Hiro Protagonist: I wish Diva would maintain a slightly higher profile
[11:15] Hiro Protagonist: her hypergrid work is really showing some serious promise
[11:15] Fly Man: uhm, besides everyone waiting for 0.6 to be announce to put in all their patches ??
[11:16] Adam Frisby: Yeah, I'm already working on a few things.
[11:16] Adam Frisby: I've got three projects on the boiler:
[11:16] Charles Krinkeb: I announced it in several places.
[11:16] Nebadon Izumi: yea we are beyond announcing
[11:16] Charles Krinkeb: I even put a note on the bathroom door at my office.
[11:16] Nebadon Izumi: hehe
[11:16] Adam Frisby: 1. Refactoring IClientAPI to IClientCore - as mentioned on the ML and Wiki -- that's already happening (you can look at IClientCore and how it works in the Chat Module as an example)
[11:16] Fly Man: Ghehehe
[11:17] Adam Frisby: 2. Refactoring Login to make it so we can just use Scene.AddClient(IClientX client) rather than needing to do anything else (almost done)
[11:17] Charles Krinkeb: I dont want to tempt Mr.Murphy other then to pay homage to him for allowing WP to work well this morning.
[11:17] Adam Frisby: 3. Refactoring groups after a discovery this morning since it has some stuff in some places it shouldnt IMO.
[11:17] Justin Clark-Casey: well, we are still here with 17 of us
[11:17] Justin Clark-Casey: even if frametmie leaps up every time I move about
[11:18] Charles Krinkeb: Nobody open the map or edit avatar appearance, please.
[11:18] Fly Man: Adam, as long as I can connect to the right parts with OSS and OSG
[11:18] Adam Frisby: Yeah, to your modules - the interface wont change
[11:18] Adam Frisby: I'm just moving some of the 'glue' out of LLCLientView and into it's own module.
[11:18] Adam Frisby: I'm just speccing out what the new events/calls should look like this moment.
[11:18] Fly Man: But Charles
[11:18] Fly Man: as was already yelled on the IRC
[11:19] Fly Man: Will we be making a 0.6.0.1. version
[11:19] Fly Man: as there's a big fix after 0.6 announce made by Adam
[11:19] Fly Man: that makes the sim less crashing
[11:19] whump: nice
[11:19] Adam Frisby: We had a discussion about version numbers on -dev
[11:19] Melanie Milland: adam, there will be some more
[11:19] Charles Krinkeb: I would suggest that a 0.6.1 this next weekend might be helpful if everyone thinks this is good.
[11:19] Justin Clark-Casey: which one was that?
[11:19] Adam Frisby: Justin: the memory leak
[11:19] Hiro Protagonist: a couple actually
[11:19] Adam Frisby: and asset cache
[11:19] whump: I had a question about the Child Agent issue: http://opensimulator.org/mantis/view.php?id=2375
[11:19] Adam Frisby: Still havent found the problem with EventQueueGet
[11:19] Justin Clark-Casey: ah right. Well, to be honest, big fixes come along quite often :)
[11:19] Adam Frisby: But eventqueueget is just broken full stop
[11:20] Hiro Protagonist: and I think Mel cleaned up the more severe effects of a third, if she didnt manage to fix the underlying issue altogether
[11:20] Justin Clark-Casey: in what way, Adam?
[11:20] Adam Frisby: The client handles it badly.
[11:20] Adam Frisby: If there's even the slightest issue the client wont reconnect, for example.
[11:20] Adam Frisby: It'll just abort.
[11:20] Charles Krinkeb: Its really a matter of minimizing the overhead on the developers part and the packegers of the release (i.e. Nebadon and daTwitch).
[11:20] Hans Betsen: can I ask a question about upgrading server side do you lose your sim? (UUID to osgrid server)
[11:20] Justin Clark-Casey: nasty
[11:20] Adam Frisby: So any early timeout, any malformed response, anything causes the client to disconnect.
[11:20] Adam Frisby: and because of the way it works that's really fatal.
[11:20] Adam Frisby: It also has some other issues where the client makes mistakes and leads to us disconnecting the client.
[11:20] Website Display Panel v1.0: Loading URL, Please wait..
[11:20] Hiro Protagonist: packing a new release is a fairly trivial operation
[11:21] Hiro Protagonist: corrdinating on it with everyone concerned is the tough part
[11:21] Adam Frisby: I'm nearly 100% positve that our 'unable to teleport / connect to regions' bug is EQG related
[11:21] Homer_Horwitz joins #osgrid-wp
[11:21] Melanie Milland: can't we tell the client to connect again?
[11:21] Hiro Protagonist: Adam: Mel did some recent work, which I tested pretty severely, that eliminates the problem with coming and going from various regions
[11:21] Justin Clark-Casey: ho hum.. Just out of interest, is there documentation for these intricacies of the protocol? Or is it all worked out by inspection?
[11:22] Fly Man: Whump, a quick question
[11:22] Fly Man: Is this still happening with the latest trunk ?
[11:22] whump: Fly Man: yes?
[11:22] Hiro Protagonist: I *suspect* it may be related to the EGQ problem
[11:22] whump: Fly Man: yes
[11:22] Teravus Ousley is Online
[11:22] Justin Clark-Casey: man, there's a big delay on in world chat
[11:22] Justin Clark-Casey: er, sometimes
[11:22] Adam Frisby: Whump: can you guys just ditch EQG?
[11:22] Hiro Protagonist: *EQG
[11:22] Homer Horwitz: Hi
[11:23] Hiro Protagonist: Hiya Ter, Homer
[11:23] Adam Frisby: What happened to R-HTTP?
[11:23] Justin Clark-Casey: Hey Mr T
[11:23] whump: Adam Frisby: good question, I don't know the answer for that off the top of my head.
[11:23] whump: ah, that was a rhetorical question
[11:23] Adam Frisby: Hehe
[11:23] Adam Frisby: It was serious actually
[11:23] Adam Frisby: EventQueueGet is horrible
[11:24] Teravus Ousley: We don't really like EventQueueGet
[11:24] Adam Frisby: Making it a requirement on SimulatorEnable/etc was a bad move.
[11:24] Hiro Protagonist: too many tracks, only a few trains
[11:24] Melanie Milland: adam, can't we make the client reconnect?
[11:24] Adam Frisby: Melanie: no - once it's disconnected, it's lost forever.
[11:24] Melanie Milland: tell it "hey. we're still here, connect"?
[11:24] Homer Horwitz: It was set mandatory when several UDP packets were blacklisted for security reasons (in SL, the eventqueue runs over HTTPS)
[11:25] Teravus Ousley: You have to get the simulator to send the disable simulator packet
[11:25] Melanie Milland: adam, there has to be a recovery mechanism
[11:25] Adam Frisby: There's not - the same bug happens on the maingrid too.
[11:25] Melanie Milland: i'm sure it wasn't designed to break when there are network issues
[11:25] Teravus Ousley: .. if you keep the event queue sync'ed with the enablesimulator and disable simulaator packet, it'll reconnect
[11:25] Adam Frisby: Teravus: oh?
[11:25] Teravus Ousley: Yeah. But it's not easy
[11:25] Teravus Ousley: and, we don't like it :D
[11:25] Homer Horwitz: At the moment, we sadly don't have a way to sync UDP with Eventqueue.
[11:26] Adam Frisby: Yeah it's bad protocol all around.
[11:26] Adam Frisby: Well, that could be partly fixed if we moved EventQueue into LLClientView as I suggested a while back (since it is a protocol layer)
[11:26] Teravus Ousley: well, it'll always execute in the thread context of the HTTP server regardless of where you have it.
[11:27] Homer Horwitz: In that code region, yes. But you could put the packets in via the normal packet sending algorithm.
[11:27] Melanie Milland: what needs to be synced?
[11:27] Melanie Milland: i'm sure it can be done
[11:27] phrearch joins #osgrid-wp
[11:27] Homer Horwitz: Right. Just a matter of how.
[11:27] Melanie Milland: we need that EQG reconnect
[11:27] Chris D is Offline
[11:27] Teravus Ousley: Well, the big important item.. is if you are disabling the event queue, send the disable simulator packet at about the same time.
[11:28] Homer Horwitz: We could do it in the packet layer. Send a packet; whether it's sent via UDP or EVQ is determined in the packet layer, not by the, say, scene.
[11:28] Adam Frisby: Teravus: hrrm - that strikes me as unrelated to the ConnectingToSim bug though
[11:28] Teravus Ousley: . that takes it out of the client's 'known regions'
[11:28] Adam Frisby: We might have a second problem here then
[11:28] Teravus Ousley: .. the client only ignores 'known caps'
[11:28] Adam Frisby: Homer: Yes, I agree - it should be that way.
[11:28] Melanie Milland: lots of times, when you log in, you get no event queue
[11:28] Adam Frisby: SendXYZ gets routed to EventQueue if appropriate, but we dont know about it.
[11:29] Melanie Milland: it's the most common login issue for us
[11:29] Adam Frisby: Right.
[11:29] Homer Horwitz: Right.
[11:29] Teravus Ousley: I suspect our HTTP threads are exausted
[11:29] Adam Frisby: Teravus: hrrm, we're probably leaking then
[11:29] Melanie Milland: that needs fixing
[11:29] Homer Horwitz: With a backlog of 5, that's not surprising, I think.
[11:29] Adam Frisby: Has anyone tried a higher backlog?
[11:29] Melanie Milland: it means you have to reset sims full of active people
[11:29] Homer Horwitz: Not yet. But I don't have any load on my (local) grid.
[11:29] Melanie Milland: if we have 20 ppl in the welcome
[11:29] Teravus Ousley: no, when I set 5, it was a SWAG
[11:30] Melanie Milland: and that starts happening
[11:30] Melanie Milland: we have to dump 20 users or live with no one else being able to login
[11:30] Adam Frisby: SWAG?
[11:30] Teravus Ousley: (SWAG = Super Wild A$$ Guess)
[11:30] Melanie Milland: and int hat case, even fallback redirect isn't working
[11:30] Adam Frisby: OK, let's bump 5 up
[11:30] Adam Frisby: Anyone opposed to me raising that to 64?
[11:31] Melanie Milland: go ahead
[11:31] Teravus Ousley: well, think of it in terms of IIS threads
[11:31] Homer Horwitz: If that maps to the normal C listen() backlog, then linux can cope with up to 128, more if you reconfigure your system (via sysctl)
[11:31] Teravus Ousley: .. because it's essentially the same 'control'
[11:31] Melanie Milland: not opposed to anything that will fix our #1 issue
[11:31] Adam Frisby: I dont know that's a fix though
[11:31] Charles Krinkeb: go ahead and if it doesnt work, blame ckrinke
[11:31] Adam Frisby: because it's probably having issues because of a thread leak
[11:31] Melanie Milland: if it makes it survive 5 days rather than 1 hour, it's a fix
[11:31] Justin Clark-Casey: hey, blame ckrinke get out of jail free card!
[11:31] Fly Man: but won't that leak more then ... ?
[11:31] Adam Frisby: We do need to find that leak and fix it.
[11:31] Homer Horwitz: Ter: That's why I stumbled over it: If I use Apache Tomcat for any serious work, I start with 50.
[11:31] Melanie Milland: because within 5 days, ti will crash from other causes
[11:31] phrearch quits saying ":Remote closed the connection"
[11:33] Homer Horwitz: Suddenly so quiet... Is chat still working?
[11:33] Adam Frisby: Yeah
[11:33] Fly Man: Yes\
[11:33] paulie Flomar: yep
[11:33] Adam Frisby: OK
[11:33] Homer Horwitz: Ah ok :)
[11:33] Adam Frisby: So we've resolved to up that to a higher value
[11:33] Teravus Ousley: I obstain
[11:33] Adam Frisby: 64 is probably fine.
[11:33] Fly Man: and the sim is still working ;)
[11:33] Adam Frisby: Next order of business:
[11:33] Fly Man: *knocks on wood*
[11:33] Adam Frisby: I want to add another client to OpenSim.
[11:33] Teravus Ousley: *knocks on prim*
[11:33] Adam Frisby: Now that we've got IClientCore coming in it's a lot easier
[11:33] Teravus Ousley: .
[11:33] Fly Man: Adam, will that client be on the Gforge ?
[11:34] Adam Frisby: I'm thinking something already out there and popular
[11:34] Adam Frisby: like Wonderland, Croquet, Forterra, ActiveWorlds, OSMP, something.
[11:34] whump: WoW? :)
[11:34] Adam Frisby: We'd like to not get sued by Blizzard. ;) Having LL over our shoulders is enough. :P
[11:34] Justin Clark-Casey: oh, so you don't have a definite candidate in mind yet?
[11:34] paulie Flomar: DigitalSpace Traveller
[11:34] Adam Frisby: Justin: Nope.
[11:34] Teravus Ousley: the only issue with WoW is it makes use of 'known assets'
[11:34] Adam Frisby: VastPark is a contendor, but I need Bruce to get me protocol documents and that's taking a bit longer than planned.
[11:35] Adam Frisby: and I dont know I'll get it anytime soon which means I'm open to other options.
[11:35] Melanie Milland: an alternative client would be so hot
[11:35] Teravus Ousley: each MOB has a mesh ID that translates to a 'known model'
[11:35] Poppet McGimsie: a game would be cool, I know you were joking
[11:35] Justin Clark-Casey: Yeah, something like Croquet would be real tough if possible at all, I imagine, as twith their very different architecture
[11:35] Melanie Milland: especially if it can view the same objects
[11:35] Davy Jones: wonderland has source available i think
[11:35] Adam Frisby: Yeah, Croquet is probably out of the picture since it's P2P not ClientServer.
[11:35] Adam Frisby: Wonderland has a small problem that the client is GPL'd
[11:35] Justin Clark-Casey: wonderland is GPL, but it's probably a good vandidate
[11:35] Adam Frisby: which means we go through all the fun we've had with the SL Viewer all over again.
[11:35] Adam Frisby: But if we can get someone to give us protocol docs
[11:35] Justin Clark-Casey: unless they are willing to write specification documents....
[11:35] Adam Frisby: then we at least have a libsecondlife equivilent.
[11:35] Justin Clark-Casey: yeah
[11:36] Adam Frisby: Wonderland may be a good choice though otherwise.
[11:36] Davy Jones: but we are only allowing client to connect
[11:36] Fly Man: and how far is the Rex client ?
[11:36] Teravus Ousley: haboo hotel?
[11:36] Teravus Ousley: j/k
[11:36] Master Zephyr: we test all platforms lol
[11:36] Justin Clark-Casey: yeah,wonderland would be very interesting - would stir them up a bit too :)
[11:36] Nebadon Izumi: lol
[11:36] Adam Frisby: Fly: the Rex client work is just about done - the problem is structurally it's not too different.
[11:36] Justin Clark-Casey: heh,
[11:37] Adam Frisby: Well, 'done' by I mean you'll be able to connect, get into Scene, etc correctly.
[11:37] Adam Frisby: We can sort of do that already now actually (the modrex code is on forge)
[11:37] Justin Clark-Casey: so what happens if the LL client comes across a rex mesh?
[11:37] Master Zephyr: and also you cant test but you need users to feedback as to what thay see too
[11:37] Adam Frisby: Justin: it's not displayed.
[11:37] Master Zephyr: just loacl testing isnt the same
[11:38] Adam Frisby: Justin: this is where IClientCore shows it's usefulness - because each client only subscribes to stuff they can understand.
[11:38] Justin Clark-Casey: okay. yeah.
[11:38] Master Zephyr: i found the problem btw
[11:38] Justin Clark-Casey: do they see any placeholder object, or does it simply not appear?
[11:38] Adam Frisby: Justin: in theory a placeholder, I havent actually tested though.
[11:38] Adam Frisby: They would probably see our default prim.
[11:38] Master Zephyr: it has todo w. users browsers , blcoking scipts
[11:38] Justin Clark-Casey: plywood cube tastic?
[11:38] Adam Frisby: Way we're doing RexPrims is inheriting from SOG/SOP.
[11:38] Homer Horwitz: How is physics handled if you don't see half of the world you might collide with?
[11:38] Nebadon Izumi: ..
[11:38] Justin Clark-Casey: hmm, although I guess it might not be that
[11:39] Adam Frisby: So, RexSOP : SOP - which has the extra .Mesh properties
[11:39] Fly Man: Adam, and how about the OpenLife viewer thingy ?
[11:39] Adam Frisby: Fly: No idea - I dont use it, and AFAIK it's just either the SL viewer or Rex viewer with rebranding
[11:39] Adam Frisby: nothing special about it.
[11:40] Adam Frisby: I'm testing the rex stuff with the rex viewer, SL stuff with SL viewer, and the occasional vice-versa to make sure no-one's recieving stuff they shouldnt.
[11:40] Melanie Milland: it's rex
[11:40] Fly Man: Adam, does the client need to be graphical ?
[11:41] Adam Frisby: Nope.
[11:41] Justin Clark-Casey: that's a good point - a text client would be interesting too
[11:41] Adam Frisby: I was thinking a jabber IM or something would be cool
[11:41] Adam Frisby: But a lot of work.
[11:41] Hiro Protagonist: SLim?
[11:41] Homer Horwitz: The regions as chat rooms?
[11:41] Justin Clark-Casey: most of them are probably quite LL like
[11:41] Adam Frisby: I dunno, I'm hoping for something "not-SL" (Sorry whump!)
[11:41] Justin Clark-Casey: isn't that the irc bridge? :)
[11:41] Hiro Protagonist: that wouldnt be all bad
[11:41] Fly Man: Well, how about the one that's already out there
[11:41] Fly Man: Uhm, let me find the name
[11:41] Adam Frisby: I think wonderland may not be a bad option, but we need protocol docs, or someone on the other side who can help
[11:42] Adam Frisby: since we face the same GPL contamination issues there.
[11:42] Justin Clark-Casey: yeah, that might be more possible since they have permanent Sun guys I believe
[11:42] Teravus Ousley: yay, GPL! We love you!
[11:42] Justin Clark-Casey: depends how interested they are interop I guess
[11:42] Hiro Protagonist: Any progress on a erinterpretation of law in re: GPL vs. BSD?
[11:42] Adam Frisby: Hiro: Not really, another lawyer is taking a look but initial reports were, I quote: "It's a pain in the arse"
[11:42] Fly Man: AjaxLife
[11:43] Adam Frisby: AjaxLife is libsl - same protocol
[11:43] Master Zephyr: lol
[11:43] Ruud Lathrop: Java/Flash/Silverlight client
[11:43] Hiro Protagonist: LOL@Adam
[11:43] Hiro Protagonist: it could only be, what with no real precedence
[11:43] Adam Frisby: Java/Flash/Silverlight would be cool
[11:43] Adam Frisby: But we'd probably need to build a new client from scratch there, no?
[11:44] Poppet McGimsie: have any big game companies ever expressed interest in interoperability?
[11:44] Master Zephyr: thay wount bother as theres no money in it for them anyway
[11:44] Ruud Lathrop: A light protocol, only chat first
[11:44] Adam Frisby: Yeah, you need an incentive for interop - only VW companies have that incentive righ tnow.
[11:44] Master Zephyr: vant get blood ouut a rock
[11:44] Justin Clark-Casey: user information - they could gather data about what people are doing ;)
[11:45] Teravus Ousley: I know, let's implement mpeg-v! haha
[11:45] Hiro Protagonist: that was my thinking with SLim - well known protocol, friendly partners, different paradigm
[11:45] Teravus Ousley: j/k
[11:45] Fly Man: https://simtk.org ??
[11:45] Master Zephyr: i for one, am dammed proud of anything done here
[11:45] Master Zephyr: sue me lol
[11:45] Poppet McGimsie: I think there is money in it for games like wow, because of how much people want a level of customization they can't get in them
[11:46] Hiro Protagonist: it's really perfect as a canditate for a 'first additional' client
[11:46] Master Zephyr: what money.. lol
[11:46] Master Zephyr: nothing to get
[11:46] Poppet McGimsie: create the need and people will pay
[11:46] Master Zephyr: there just mad cuz thay see what we know
[11:46] Adam Frisby: Hiro: I'm not sure it's a great candidate TBH - there's no protocol docs or source for example.
[11:47] Master Zephyr: and thay dont like it
[11:47] Master Zephyr: screw em
[11:47] Hiro Protagonist: for SLim?
[11:47] Adam Frisby: Yeah
[11:47] Adam Frisby: I'm coming back to Wonderland or Vastpark, depending on which we can get protocol docs for first.
[11:47] Hiro Protagonist: I may be mislablelling my reference
[11:47] Teravus Ousley: slfat vs slim
[11:47] Justin Clark-Casey: vastpark is closed source?
[11:47] Hiro Protagonist: I'm speaking of the new text-only LL client
[11:47] Justin Clark-Casey: oh no hang on, it's some kind of toolkit, right?
[11:47] Adam Frisby: Vastpark is GPL with a 'compatibility exception'
[11:48] Adam Frisby: Which allows us to use the code for making interop purposes.
[11:48] Teravus Ousley: isn't FAT some kind of legacy file system used at some point?
[11:48] Teravus Ousley: :D
[11:48] Justin Clark-Casey: ah, so that allows us to actually loko at it?
[11:48] Adam Frisby: Yes
[11:48] Adam Frisby: Do speak to me first before you do since I've got the details on that one
[11:49] Justin Clark-Casey: sounds like quite a good candidate actually - as they're not 'a single virtual world'
[11:49] Adam Frisby: It was something myself and the owner of VastPark are in the process of working out.
[11:49] Justin Clark-Casey: sure - cool
[11:49] Master Zephyr: if were gonna be dammed for something then let them damm us for a good reasion
[11:49] Master Zephyr: make it work
[11:50] Justin Clark-Casey: it sounds like they wuld have more naturally interest than Wonderland....
[11:50] Adam Frisby: Yes, VP is very interested
[11:50] Justin Clark-Casey: oh god, the words keep coming out slightly wrong - you wouldn't beleive that English was my first language
[11:50] Adam Frisby: The only criticism I have is that the client is missing some key stuff like chat/IM for instance - which was partly on purpose with their goals.
[11:50] Hiro Protagonist: LO@Justin
[11:50] Hiro Protagonist: LO*
[11:50] Master Zephyr: thay cant
[11:50] Hiro Protagonist: me neither
[11:50] Adam Frisby: So our 'first possible thing to implement' is a scenegraph which is adding to our complexity levels.
[11:50] Master Zephyr: its outta the box ll opend it and no way OSG is gonna stop
[11:51] Master Zephyr: we shure arnt
[11:51] Justin Clark-Casey: I should go and have a look at this stuff myself sometime
[11:51] Justin Clark-Casey: yeah. So how do they do chat?
[11:51] Melanie Milland: no chat and IM? how are you supposed to communicate?
[11:51] Adam Frisby: I was kinda hoping we'd have a 3D virtual world which supports avatars/chat/IM/etc but we can just initially implement chat/IM and add more as we go - so we dont need to implement a lot of stuff to get a scene loaded.
[11:51] Hiro Protagonist: Ballons!
[11:51] Adam Frisby: Melanie: It's more abstracted than a conventional VW
[11:51]
[SilverFox] parts #osgrid-wp
[11:51]
[SilverFox] joins #osgrid-wp
[11:52] Master Zephyr: correct
[11:52] Teravus Ousley: Microsoft ESP(tm) melanit
[11:52] Ruud Lathrop: multiverse?
[11:52] Adam Frisby: The idea is you use their toolkit to describe your world, and it's rules - which means they dont even define avatars in their spec, they require you do that yourself.
[11:52] Teravus Ousley: .
[11:52] Master Zephyr: all of it
[11:52] Hiro Protagonist: interesting
[11:52] Justin Clark-Casey: where do they make their money?
[11:52] Reena Bismuth is Offline
[11:53] Justin Clark-Casey: Ah, I see
[11:53] Master Zephyr: im not gonna sit worring about ll and every lil thing , just do it and let them sort it out
[11:53] Adam Frisby: Good question, but I think a lot of consulting and they also have a mechant system in for people selling content in theor formats.
[11:53] Ruud Lathrop: Did you look at multiverse, Adam?
[11:53] Adam Frisby: Multiverse may be an option
[11:53] Hiro Protagonist: Multiverse would be good
[11:53] Adam Frisby: It's not open source though is it?
[11:53] Adam Frisby: (I havent paid attention to it in the last year really)
[11:54] Ruud Lathrop: Dont know, see they have a dev license
[11:54] Hiro Protagonist: I want to say it is
[11:54] Master Zephyr: its open .. the would has it all
[11:54] Master Zephyr: world
[11:54] Adam Frisby: How open? what's the license?
[11:54] Hiro Protagonist: I've watched this place become viable while waiting on them to produce anything like linux support
[11:54] Master Zephyr: what matters here is who has the best time using it
[11:54] Master Zephyr: not who has it
[11:55] Ruud Lathrop: Ask Corey Bridges, think he can answer that :)
[11:55] Adam Frisby: Multiverse does have one big advantage to it.
[11:55] Adam Frisby: It uses the same mesh and avatar formats as Rex
[11:55] Fly Man: http://gigaom.com/2008/04/08/here-comes-the-open-source-metaverse/
[11:55] Hiro Protagonist nods
[11:55] Adam Frisby: So we could have Rex objects shared between em
[11:55] Master Zephyr: and others :)
[11:55] Hiro Protagonist: that's a pretty strobg plus
[11:55] Fly Man: For the people that want to see all the clients in 1 list:
[11:55] Adam Frisby: Would be way cool to have that.
[11:55] Fly Man: http://gigaom.com/2008/04/08/here-comes-the-open-source-metaverse/
[11:56] Hiro Protagonist: the sixtyfour thousand dollar question is, has rex improved their (nonexistent) linux support?
[11:56] Master Zephyr: hy fly
[11:56] kane hart: Does this Grid run latest or almost latest svn revs? runs so well :O
[11:56] Master Zephyr: dident see yah there
[11:56] Fly Man: That page lists some of the viewers that were used during Virtual Worlds 2008 conference in New York
[11:56] Jasmin Summers: It keeps saying physical objects would bounce out of the sim and cause issues and give out this error:
[PHYSICS]: Object reference not set to an instance of an object, Single Simulate(Single), System.NullReferenceException: Object reference not set to an instance of an object at OpenSim.Region.Physics.OdePlugin.OdeScene.Simulate (Single timeStep)
[0x00000]
[11:56] Adam Frisby: OSGrid is always the latest version, yeah.
[11:56] Jasmin Summers: Hello, I am currently testing out opensim but i keep getting an odd error I have not been able to fix.
[11:56] Justin Clark-Casey: oof
[11:56] Adam Frisby: Hiro: they hired a linux guy recently if that helps -- apparently it compiles, but there's a bit of work to be done. Might be best asking sempuki.
[11:56] Master Zephyr: so we all do
[11:57] Adam Frisby: OK
[11:57] Hiro Protagonist: thanks Adam, that -is- progress
[11:57] Master Zephyr: but id love a better way to feed back
[11:57] Hiro Protagonist: do they need some help?
[11:57] Adam Frisby: I dont know - speak to Sempuki
[11:57] Fly Man: Jasmin, visit the Q&A hour after this
[11:57] Master Zephyr: input things easyer ..
[11:57] Adam Frisby: He got hired by them and he's a linux fanatic. ;)
[11:57] Jasmin Summers: ty
[11:57] Fly Man: in 1 hour I will be having the Q&A hour here
[11:57] Master Zephyr: k
[11:58] Hiro Protagonist: for those who didnt know, this is an office hours meeting, where current issues are being visited and plans for the immediate future are being discussed
[11:58] Nebadon Izumi: i would think working on rex viewer would automaticly disclude you from working on OpenSIM again
[11:58] Justin Clark-Casey: I miss the days when we would bounce constantly
[11:58] Adam Frisby: Nebadon: the viewer, yes.
[11:58] Nebadon Izumi: their server works on linux
[11:58] Adam Frisby: Nebadon: that's why they have two companies - one doing server, one doing viewer.
[11:58] Nebadon Izumi: has for a while
[11:58] Hiro Protagonist: does it work, or just hobble along?
[11:58] Master Zephyr: and on vista and xp pro btw
[11:58] Nebadon Izumi: the server runs
[11:58] Teravus Ousley: Increase the avatar standup tensor.. you'll get a bounce for sure :D
[11:58] Nebadon Izumi: viewer doesnt at all that i know of
[11:58] Teravus Ousley: .. and the PID strength..
[11:59] Justin Clark-Casey: heh - bouncy bouncy
[11:59] Melanie Milland: so if i decrease them it will bounce less?
[11:59] Justin Clark-Casey: beautiful bouncing sims
[11:59] Nebadon Izumi: ive not had much luck with Rex myself though, im not expert
[11:59] Nebadon Izumi: it just crashes for me in about less than 5 minutes
[11:59] Fly Man: But back to the part were we talked about 0.6.0.1 ;)
[11:59] Fly Man: Will this be a "mayor" release version
[11:59] Teravus Ousley: it'll overshoot, and compensate.. but overshoot the compensation.. and compensate.. and overshoot that compensation.....
[11:59] Fly Man: or just a "minor"update
[12:00] Melanie Milland will have 0.6.0.1.2.0.0.4.5
[12:00] Justin Clark-Casey: the mayor of Oulo!
[12:00] Master Zephyr: i see no issues to it on our grid
[12:00] Justin Clark-Casey: We will have to ask the mayor
[12:00] Justin Clark-Casey: humph, and now chat comes out in the wrong order
[12:00] Master Zephyr: no need to fix it for windows servers that i see
[12:00] Hiro Protagonist: I think it safe to say, in light of recent ML discussions, it will likely be 0.6.2, as it will address fixes for stability
[12:01] Master Zephyr: oln thing we see is when rezing linked prims on the ground
[12:01] Master Zephyr: there under the ground
[12:01] Hiro Protagonist: sorry to steal y'alls fun :D
[12:01] Adam Frisby: Re: the mailing list stuff.
[12:01] Nebadon Izumi: yea it will always rez the root prim on ground
[12:01] Nebadon Izumi: so if its at the top of your build
[12:01] Adam Frisby: I'd like to propose we go ahead and tag 0.6.1 - but obviously not release
[12:02] Melanie Milland: fix freezes on login
[12:02] Adam Frisby: Odd numbers are 'development versions'
[12:02] Melanie Milland: fix EQG reconnect
[12:02] Master Zephyr: center prim of linked set
[12:02] Melanie Milland: well, list of things that need to go into it: fix eventqueue not connectong on login
[12:02] Adam Frisby: Hrrm, I'm getting chat out of order again
[12:02] Adam Frisby: We do releases on even numbers
[12:02] Nebadon Izumi: yea chat is lagging bad
[12:02] Justin Clark-Casey: does that really mean anyuthing?
[12:02] Master Zephyr: but no bouncy problems under windows ..
[12:02] Melanie Milland: fix unable to move/folded legs
[12:02] Melanie Milland: :)
[12:02] Melanie Milland: i am, too
[12:02] Nebadon Izumi: no doub tits all the running around
[12:02] Hiro Protagonist: so we roll forward on 0.6.2 as a stability fix on even numbers without skipping a release in that sequence?
[12:02] Nebadon Izumi: making lag
[12:02] Homer_Horwitz: I'd wait with tagging 0.6.1 until we got at least the regressions out and the things that worked with 0.6 work again.
[12:03] Adam Frisby: So to repeat on one line to avoid the order: We tag 0.6.1 now - we dont release it. We release 0.6.2 later - therefor odd numbers are development versions, even are releases.
[12:03] Nebadon Izumi: memory is pretty good
[12:03] Master Zephyr: todays build is way good
[12:03] Hiro Protagonist: homer, 0.6.1 would be a dev release, and so indicated by it's odd minor version number
[12:03] Fly Man: Homer, how's Friends ?
[12:03] Homer_Horwitz: What's the benefit of tagging 0.6.1?
[12:03] Nebadon Izumi: the entire meeting
[12:03] Nebadon Izumi: still 300meg of ram free
[12:03] Homer_Horwitz: Fly: Failing.
[12:03] Justin Clark-Casey: it just means a long delay until 0.6.2
[12:03] Fly Man: Uhm, so what's your suggestion ?
[12:03] Justin Clark-Casey: yeah, I don't see much benefit
[12:03] Hiro Protagonist: that is awesome Neb
[12:03] Teravus Ousley: yay adam
[12:03] Adam Frisby: Homer: the idea being that when people say "I'm running X.Y.Z"
[12:03] Fly Man: Will this be fixable before the weekend ?
[12:04] Hiro Protagonist: the benefit is to bring the code into sync with the release numbering scheme
[12:04] Adam Frisby: Actually more than people saying it
[12:04] Hiro Protagonist: dev vs. stable
[12:04] Adam Frisby: It's so that we can make certain releases 'stable' releases then provide patches to them later if we want.
[12:04] Adam Frisby: It's similar to how the linux kernel used to work
[12:04] Adam Frisby: 2.5 was became 2.6 when it was stable.
[12:04] Homer_Horwitz: Well, that was after pre-alpha stage.
[12:04] justincc_: well, that implies someone to make sure even point releasses are stable
[12:05] Homer_Horwitz: Far after beta, actually
[12:05] Justin Clark-Casey is Offline
[12:05] Hiro Protagonist: well, a major version number of zero implies alpha
[12:05] Adam Frisby: Justin, yeah it does - but some people have expressed interest in making hardened releases, this enables it - and it doesnt require much work on our behalf to support.
[12:05] Master Zephyr: well said :)
[12:06] justincc_: well, what stops them hardening 0.6.1 in that case?
[12:06] Homer_Horwitz: So, after a "even" release, we immediately tag an "odd" release, too?
[12:06] Adam Frisby: 0.6.1 is in active development.
[12:06] Adam Frisby: 0.6.0 is a fixed point in time
[12:06] Hiro Protagonist: nothing - but we would hoe that patches would be submuitted in the process
[12:06] Adam Frisby: Homer: yes.
[12:06] justincc_: aren't all releases? At least without any actual effort on our part to do stablity before doing one
[12:06] Homer_Horwitz: I see.
[12:06] Hiro Protagonist: *hope
[12:06] Ruud Lathrop: No fixed time line?
[12:06] Homer_Horwitz: Nope.
[12:06] Hiro Protagonist: nope
[12:06] Ruud Lathrop: target time?
[12:07] Hiro Protagonist: release early and often
[12:07] Hiro Protagonist: tbd at office hours
[12:07] Adam Frisby: Uh oh
[12:07] Adam Frisby: Sim die/
[12:07] Adam Frisby: Justin: well, 0.6.0 is really a range
[12:07] Hiro Protagonist: ?
[12:07] Hiro Protagonist: a little slow
[12:07] justincc_: yeah, it's become unworkable for me now
[12:07] Hiro Protagonist: seems ok
[12:07] Adam Frisby: oh n/m
[12:07] whump: stepping away for food
[12:07] Adam Frisby: Justin: 0.6.0 is from r7176 to rXYZ
[12:08] Adam Frisby: Vs being just r7176
[12:08] Adam Frisby: So then we can harden that revision specifically, rather than hardening a range 'as we go'.
[12:08] Hiro Protagonist: Neb, I wonder if reducing script engine threads here to 10 is causing the increase in frametime
[12:08] justincc_: sure. But I'm not sure I see the value of tagging odd releases as development releases since without any action on our part, they are no different from the even releases
[12:08] Charles Krinkeb: When enough folks wish, we can tag, package, offer binary of 0.6.1. Perhaps this weekend, perhaps next. It depends on when enough folks feel comfortable.
[12:08] Poppet McGimsie: think tetravus needs some ritalin
[12:09] Nebadon Izumi: have we seen a huge change in frametime?
[12:09] Hiro Protagonist: He probably has some LOL
[12:09] Nebadon Izumi: i thought it was lower
[12:09] Fly Man: Charles, i'm not comfortable with the 0.6.1
[12:09] Charles Krinkeb: Teravus is doing his "Testing Physics" dance.
[12:09] Fly Man: as friends was broken AFTER 0.6
[12:09] Adam Frisby: Justin: Hrrm, I guess we could do that - of course if we hardened 0.6.0 right now, it's got the same version number as the unstable 0.6.0 we're on now
[12:09] Poppet McGimsie: : ))
[12:09] Charles Krinkeb: Ok, good point.
[12:09] Nebadon Izumi: last night we tested and not only is freinds broken
[12:09] kane hart: yeah friends was a big issue here too :(
[12:09] Nebadon Izumi: but giving invbentory
[12:09] Nebadon Izumi: is also broken
[12:09] kane hart: No friends no osgrid
[12:09] justincc_: wouldn't it get labelled 0.6.0.1 or similar?
[12:09] Nebadon Izumi: giving inventory
[12:09] Hiro Protagonist: yup
[12:09] Teravus Ousley: someone mentioned that the sim might be down..... so I jump! ;D
[12:09] Adam Frisby: justin: but what about a feature release on 0.6.0?
[12:10] Hiro Protagonist: had to drop something +copy this morning in order to give it away
[12:10] Charles Krinkeb: I would urge us to get trunk to the point where it is more stable then 0.6.1 and then tag the next release.
[12:10] Nebadon Izumi: yea same
[12:10] Master Zephyr: theres workaround for it though
[12:10] Nebadon Izumi: it is mantised
[12:10] Fly Man: Homer ??
[12:10] justincc_: hmm, difficult to tell if there's out of order chat now :)
[12:10] Adam Frisby: Yeah I dont know.
[12:10] Teravus Ousley naps
[12:10] Adam Frisby: It seems OK ish now
[12:10] Hiro Protagonist: how we doin?
[12:10] Homer Horwitz: Yes?
[12:10] Master Zephyr: the issue isant just about the code
[12:10] Fly Man: Do you have any clue if the Friends part is fixable for the weekend ?
[12:11] Charles Krinkeb: We can use 0.6.0 for the more average users until trunk is ready to become 0.6.1 or whatever. Unless we wish to use the experimental/stable notions expressed by Adam now.
[12:11] Homer Horwitz: No idea.
[12:11] Hiro Protagonist slaps Fly-Man around with a large trout
[12:11] Master Zephyr: harwere were finding can sort out alot
[12:11] Charles Krinkeb: The trick is going to be staying away from branching our source if possible.
[12:11] Master Zephyr: some we cant
[12:11] Fly Man grabs a whale for Hiro
[12:11] justincc_: well, I just want to see what value there is in having odd releases as stable if we aren't making any special effort to make the even ones stable
[12:12] Hiro Protagonist: I'm +1 for modelling releases after the old linux kernel model, more or less
[12:12] justincc_: sorry, having odd releases as development, I meant
[12:12] Hiro Protagonist: :)
[12:12] Hiro Protagonist: even ones should be stable Mr. Justin
[12:12] Charles Krinkeb: Doesnt that imply that we need to have two trees? One for even, one for odd?
[12:12] justincc_: the linux kernel had enough maturity and stability to support both dev and stable branches
[12:12] Master Zephyr: one thing i do see is when 2 computers hosting many sims on each tend to tippl the packet loop
[12:12] Hiro Protagonist: Nah, stable is just snapshots from dev
[12:13] Master Zephyr: sometimes casuing a a grid crash
[12:13] Homer_Horwitz: Well, if you want to stabilize them more (as with Linux), you need a branch for it
[12:13] justincc_: well, the alternative is to do proper release candidates
[12:13] daTwitch: bad chat lag
[12:13] justincc_: and have the release manager stop people putting anything in the tree in the run up to a release that isn't a bug fix
[12:14] daTwitch: I still got a one-liner floating out there somewhere
[12:14] Teravus Ousley: IC chat lag!
[12:14] Nebadon Izumi: hehe
[12:14] Teravus Ousley: 123456789
[12:14] Nebadon Izumi: memory and cpu are low on the box
[12:14] daTwitch: just got logged out
[12:14] Nebadon Izumi: yea i wonder whats causing this
[12:14] Homer_Horwitz: Yep,. got to slow; I logged off. And it didn't notice yet ;)
[12:14] Adam Frisby: Justin: I'm not a big fan of proper releases with OpenSim.
[12:15] AdamZaius: OK
[12:15] AdamZaius: I'm moving to here.
[12:15] Adam Frisby: For the reason of exactly that - we need to stop work
[12:15] Master Zephyr: its other things , trust me i spent the last 5 days looking into it
[12:15] AdamZaius: since it's working
[12:15] Fly Man: I'm not a fan of splitting the tree into a dev and stable part
[12:15] daTwitch: I dont think it's time for that level of complexity yet, but it is time for us to communicate better about the level of stability of our releases
[12:15] AdamZaius: daTwitch, we could use a scale.
[12:15] daTwitch: this is a simple step in the right direction I think
[12:15] justincc_: yeah, at this level of maturity it might not be worth it
[12:15] AdamZaius: 0.0 = Unstable as anything, 10.0 = worse than that.
[12:15] Davy Jones: still early days but many using it so releases would be welcome but too early to start branching
[12:15] daTwitch: yes. we could
[12:15] Nebadon Izumi: i just killed LBsa plaza
[12:15] Nebadon Izumi: that was running this whole time as well
[12:15] daTwitch: Wow
[12:16] Poppet McGimsie: would be nice for people to know what you mean by "stable" also
[12:16] Nebadon Izumi: hehe
[12:16] daTwitch: that's not bad, we had 21 agents
[12:16] Ruud Lathrop: But do you want to do nightly builds with stability levels?
[12:16] Davy Jones: controlled checkins to allow a stable point would be good
[12:16] Charles Krinkeb: Can we consider just trying to gain stability and features back into trunk like the friends stuff and then tagging 0.6.1?
[12:16] Nebadon Izumi: get some threads back
[12:16] Teravus Ousley: ping
[12:16] justincc_: stable = 1 avatar sitting completely still doesn't crash the serve r;)
[12:16] daTwitch: I think nightly builds are a bad idea right now
[12:16] Fly Man: +1 Charles
[12:16] Homer_Horwitz: Hey, it noticed :)
[12:16] Master Zephyr: todays build is god one, well done
[12:16] Davy Jones: yay
[12:16] Poppet McGimsie: justin cc, I suspected that
[12:17] Charles Krinkeb: Lets use 0.6.0 as our baseline and recommend it for new users and work to get trunk a bit more stable again.
[12:17] Teravus Ousley: ping!
[12:17] justincc_: yeah, there is no benchmark for stability right now, apart from a vague feeling
[12:17] daTwitch: we have seen over 100 patches applied since friday from all over the globe - the 24 hour time period is just not relevant for this project
[12:17] justincc_: pong
[12:17] daTwitch: pango
[12:17] Charles Krinkeb: I would say that when the majority of *this* group feels comfortable, that we have a modicum of stability.
[12:18] daTwitch: :)
[12:18] Fly Man: Charles, is this a good idea:
[12:18] Fly Man: You decide when it's stable with all the other ppl
[12:18] Melanie Milland: EQG stability, logins
[12:18] Fly Man: and then CLOSE the submissions
[12:18] Melanie Milland: well, i said what i have issues with
[12:18] Fly Man: so we have a "clean" stable
[12:18] Master Zephyr: were not shy
[12:18] Master Zephyr: well if yah want feedback from our grid , let us know what yah what to have
[12:18] Charles Krinkeb: That is what I have been trying to do.
[12:18] justincc_: AdamZaius: well you know, if everybody else is happy with saying even branches will be more stable then I'm happy to go with that
[12:19] Master Zephyr: im not a coder, but i do get it working and many sims
[12:19] justincc_: perhaps just saying it will encourage it - I just hate saying such a thing without a plan to actually achieve it
[12:19] AdamZaius: justincc_, maybe that is the way to go.
[12:19] daTwitch: we *do* get to pick the time
[12:19] justincc_: sorry, I meant even minor releases rather than branches
[12:19] Poppet McGimsie: I was just going to say, I don't care about the stable/unstable thing as long as the bin files that new people use skip the unstable ones
[12:19] AdamZaius: We only tag an even branch once we've got it relatively stable.
[12:19] Nebadon Izumi: well what an improvement though this week, great meeting
[12:19] daTwitch: and we are in the best position, generally sopeaking, to know what stable is
[12:19] AdamZaius: That way we alternate between cycles of major dev work, and patching
[?]
[12:19] Charles Krinkeb: Zephyr: The key right now is to confirm 1) Things did not break that where working and 2) Help us test new features to get a feeling of stability for trunk again.
[12:19] daTwitch: as CKrinke says, when folks here are comfortable...
[12:20] Melanie Milland: as long as we don't need to stop trunk
[12:20] justincc_: yeah, I guess my fear is that having an even branch might lead to the 'development' branch being more readily broken or made gratuitously unstable. But that might be just me
[12:20] Master Zephyr: id love to offer the feedback to OSG
[12:20] Melanie Milland: i think stopping innovation is a BadThing(tm)
[12:20] Melanie Milland: but stable bvranches, great
[12:20] daTwitch: I think we have sufficient ability wrt SVN to not to have to 'freeze' the repositiory
[12:20] AdamZaius: There is a lot of folks with a stable branch FYI
[12:20] AdamZaius: Tribal has one, we've got one, 3Di has one I believe.
[12:20] justincc_: I think your scheme implies that up to the odd release, people can put major changes in. THen after that release, there should be more emphasis on stability and people should hold back
[12:20] Melanie Milland: i need to brb
[12:20] daTwitch: we just identify the revision and check it out to do our packaging
[12:20] Melanie Milland: staying here to log chat
[12:20] daTwitch: the rest of you can roll right on
[12:20] Fly Man: justincc, exactly
[12:21] justincc_: yeah, you guys also have commercial discipline and paid programmers :)
[12:21] AdamZaius: We do try push the stable patches back, but it would be nice to have that public that other people can contribute stability work to if nessecary.
[12:21] Charles Krinkeb: I agree. Right now, I think we have a reasonable feeling that 0.6.0 is mostly stable. We also think that trunk has new features, but broke some things and is less sstable.
[12:21] AdamZaius: Maybe. hehe.
[12:21] Fly Man: use the "dev" odd numbers to build mayor things
[12:21] Fly Man: and use the "stable" even numbers to do "cleanups"
[12:21] Charles Krinkeb: This has been the most useful meeting we have had in weeks.
[12:21] Master Zephyr: nodds too
[12:21] Charles Krinkeb: Good job, everyone.
[12:21] Master Zephyr: truth is im amazed it works at all
[12:21] justincc_: well, people could in principle start patching off the 0.6.0 line right? My understanding is that every tag in SVN potentially leads to a new branch anyway?
[12:22] paulie Flomar: :)
[12:22] Ruud Lathrop: How will you communicate where you are with youre releases?
[12:22] AdamZaius: justincc_, yeah it does.
[12:22] daTwitch: it has been a good meeting eh, Charles?
[12:22] AdamZaius: justincc_, we do need some marker though on checkout
[12:22] Poppet McGimsie: it's because I came, crinks
[12:22] Teravus Ousley: N E 1 got any Lindens?
[12:22] Master Zephyr: lol
[12:22] justincc_: some marker?
[12:22] AdamZaius: Well, consider:
[12:22] AdamZaius: in the viewer - they see:
[12:22] AdamZaius: OpenSimulator 0.6.0.SVN
[12:23] AdamZaius: We released stable om 6176
[12:23] Fly Man: I would like that to be
[12:23] daTwitch gives ter_afk a few hundred linden
[12:23] AdamZaius: err 7176
[12:23] AdamZaius: But if someone puts a patch on /tags/release-0.6.0
[12:23] AdamZaius: that will change to 0.6.0.8000
[12:23] AdamZaius: or whatever
[12:23] Charles Krinkeb: In reality, what we have now is a snapshot of a reasonably stable 0.6.0. Next we need to get trunk as stable.
[12:23] Master Zephyr: make a opin or out for real time feedback from grid providers
[12:23] Fly Man: OpenSimulator STBL 0.6.0.SVN
[12:23] AdamZaius: But 8001 could be a patch on /trunk which shows 0.6.0.8001
[12:23] AdamZaius: There's no way to tell if a version is the release patched version, or the trunk
[12:23] justincc_: that's true, if they are building from source. But then wouldn't one expect the version info to be changed to 0.6.0.1?
[12:23] Charles Krinkeb: Perhaps we should be running 0.6.0 on Wright Plaza and Lbsa Plaza for a while?
[12:23] Master Zephyr: get the data better that way i think
[12:23] justincc_: so yo uwould get 0.6.0.1.SVN
[12:23] AdamZaius: justincc_, so we add another version digit on?
[12:24] justincc_: that's one way, though it is a bit messy
[12:24] AdamZaius: When patches occur to the release version, we add a .1 or something to indicate it?
[12:24] AdamZaius: What about an "R" for Release, or "S" for stable ?
[12:24] justincc_: yes
[12:24] AdamZaius: 0.6.0R
[12:24] Master Zephyr: <-- looks for my pills
[12:24] daTwitch: do we HAVE to fiddle with versioning on SVN (yet)?
[12:25] AdamZaius: daTwitch, there's growing demand for it.
[12:25] daTwitch nods
[12:25] justincc_: but what about the second release in that stable series?
[12:25] daTwitch: I suppose we should at least be talking about it then
[12:25] justincc_ is now known as justincc
[12:25] AdamZaius: justincc_, we've still got the SVN number there.
[12:25] Fly Man: 0.6.1R
[12:25] justincc: not on the actual released versions, I don't think
[12:25] Ruud Lathrop: I think you slowly have to as people starting to use it for business
[12:25] justincc: since they won't have the .svn/Entries file to read it from
[12:25] AdamZaius: It should still be there.
[12:25] AdamZaius: oh true
[12:25] AdamZaius: date ?
[12:26] Charles Krinkeb: Are we at the stage where the .svn/Entries file needs to be inside the /bin directory?
[12:26] AdamZaius: Can we do date of last modification automatically somehow? Maybe a script running on a cron to add it?
[12:26] justincc: well, I think jamming another digit on the end might be more readily understandable
[12:26] Master Zephyr: there going to do that no matter what ..
[12:26] Fly Man: Uhm, why are we still using .svn/
[12:26] AdamZaius: justincc, I agree, it's just extra maintainence work.
[12:26] Richardus Raymaker is Online
[12:26] justincc: though it could be a little against what that fourth digit traditionally means, I'm not sure
[12:26] Fly Man: to find the version
[12:26] Fly Man: why not have a file update
[12:26] AdamZaius: Fly Man: because there's no other way.
[12:26] justincc: not too much though - just change a number in VersionInfo :)
[12:27] Ruud Lathrop: If you have one version info file then it is not a big problem
[12:27] Nebadon Izumi: no way to have bamboo add a file
[12:27] Nebadon Izumi: after it succefully builds?
[12:27] AdamZaius: Fly Man: find a way to do it in a cross-platform way that works efficiently and I'm sure lots of people would love you long time.
[12:27] Fly Man: Adam, if Ch1cken his bot can clean up
[12:27] AdamZaius: Nebadon: doesnt work for SVN checked out versions
[12:27] Nebadon Izumi: k
[12:27] Fly Man: then why can't it update a file ;)
[12:27] Charles Krinkeb: Cannot we have our binary read VersionInfo.cs and get the version string?
[12:28] justincc: yes, that's exactly what its doing
[12:28] justincc: I don't think it's too bad just to twiddle one number in VersionInfo.cs
[12:28] AdamZaius: Chi11kens bot might not be a bad idea, but I have no idea where he is these days
[12:28] Charles Krinkeb: I need to go pretend I am working. I will leave my avatar to read the log a bit later.
[12:28] justincc: yes, I need to get on with something too
[12:29] Nebadon Izumi: ok i'll be posting the logs to website shortly too
[12:29] Master Zephyr: its also locking things from backup now fails
[12:29] Fly Man: Adam, then let's ask Ch1cken when he's on IRC
[12:29] Ruud Lathrop: AssemblyInfo shoud use that versioninfo to get the version
[12:29] Charles Krinkeb: Thanks again, all. See you all next Tuesday at 1900 UTC if not on IRC before then.
[12:29] Fly Man: as I think using the svn for version info is bad ;)
[12:29] Richardus Raymaker is Online
[12:30] AdamZaius: Oh, I dont disagree.
[12:30] AdamZaius: Using that for svn version isnt the worlds greatest decision.
[12:30] Fly Man: Then let's clean up things ;)
[12:30] AdamZaius: But IIRC it was the only method that worked practically - an svn post-commit hook was the othe roption, but I dont know anyone was capable of writing it at the time.
[12:30] daTwitch: I'm ducking out too y'all - great meeting
[12:30] AdamZaius: anyway
[12:30] AdamZaius: I do have to run too
[12:30] daTwitch waves
[12:30] daTwitch parts #osgrid-wp
[12:31] AdamZaius: I've got to pack up some stuff because I'm moving to Sydney today
[12:31] Richardus Raymaker is Online
[12:31] Hiro Protagonist is Online
[12:31] justincc parts #osgrid-wp
[12:31] Homer_Horwitz parts #osgrid-wp
[12:32] Jeffronius Batra: I've been lurking here and have learned a lot. Thanks, and I'll be back!
[12:32] Ruud Lathrop: Ok bye all
[12:32] Fly Man: Bye
[12:32] Fly Man: Nebadon, how was the pressure on WP ?
[12:33] Fly Man: as we could still move
[12:33] Richardus Raymaker is Online
[12:34] Serina Lacava: i have a question.i need to transfer money by paypal or any others way too?
[12:34] kane hart: when does QA start?
[12:34] Fly Man: Serina, ??
[12:34] Fly Man: Kane, it will start in 26 mins
[12:34] kane hart: ty
[12:34] kane hart: <silverfox :D
[12:35] Master Zephyr: no and yes to the pay pal
[12:35] Fly Man: Kane, I know ;)
[12:35] Master Zephyr: i wouldnt go there yet
[12:35] Serina Lacava: ok, thank you.
[12:35] Master Zephyr: let them cook a bit 1st lol
[12:36] Fly Man: Hey Jenni
[12:36] Poppet McGimsie: bye everyone -- see you later
[12:36] Richardus Raymaker: hello al
[12:36] Jenni Briggs: hi rich
[12:36] Richardus Raymaker: hi jenni
[12:36] Master Zephyr: hi jenni
[12:36] Jenni Briggs: hi master hi every one
[12:38] Master Zephyr: ohh you can sit by me jenni, i wount bite
[12:38] Master Zephyr: lol
[12:38] Jenni Briggs: i not a big sitter
[12:38] Jenni Briggs: ]
[12:38] Master Zephyr: lol

Personal tools
General
About This Wiki