Chat log from the meeting on 2010-08-31

From OpenSimulator

Jump to: navigation, search

[17:49] Frank Northmead is Online
[17:54] Snoopy Pfeffer is Online
[17:56] Andrew Hellershanks: Hey, nebadon
[17:56] Nebadon Izumi: hello
[17:57] Penny Lane is Online
[17:58] Richardus Raymaker is Online
[17:58] Master Dubrovna: Hello everyone
[17:59] Nebadon Izumi: hello
[17:59] Penny Lane: 'Morning :-)
[17:59] Andrew Hellershanks: Hello
[17:59] Aaron Duffy: Hello all
[18:01] Penny Lane: I haven't been paying much attention to #vwrap channel recently, but I did notice some odd remarks posted, like references to "SimianSim" and "Opensim can do whatever they want". :-)
[18:02] Richardus Raymaker: hi
[18:02] Penny Lane: Hi Rich :-)
[18:03] paulie Flomar is Online
[18:03] Nebadon Izumi: ya Simian simulator is old
[18:03] Nebadon Izumi: its nothing new
[18:03] Nebadon Izumi: its a very minimal implementation of opensimulator
[18:03] Nebadon Izumi: mostly for dev purposes is my understanding
[18:03] Nebadon Izumi: not something you would want to run an actual simulator from
[18:03] Justin Clark-Casey is Online
[18:03] Penny Lane: Ah righto, thanks Neb
[18:04] Nebadon Izumi: but ya i havent had time for Simiangrid much lately myself
[18:04] Nebadon Izumi: to be honest im not even sure why that channel is named #vwrap
[18:04] Justin Clark-Casey: hello folks
[18:05] Penny Lane: Hi Justin :-)
[18:05] Nebadon Izumi: i rarely if ever have seen any vwrap discussion there
[18:05] Nebadon Izumi: hehe
[18:05] Andrew Hellershanks: hey, justin
[18:05] Aaron Duffy: hi Justin
[18:05] Nebadon Izumi: hello
[18:05] Justin Clark-Casey: ah, dontcha love those monster thread s:)
[18:05] Richardus Raymaker: IEEP !, sound with video ??
[18:05] Richardus Raymaker: whooo
[18:05] Nebadon Izumi: heh ya if your running RC2 Rich you will have sound now
[18:05] Nebadon Izumi: Imprudence RC2 that is
[18:05] Richardus Raymaker: yes
[18:05] Penny Lane: I'me getting sound here
[18:06] Nebadon Izumi: finally
[18:06] Penny Lane:
[10:05 AM] Playing: 4tthieves - dreamflight
[18:06] Nebadon Izumi: oh thats the audio track
[18:06] Nebadon Izumi: not the video
[18:06] Nebadon Izumi: thats my radio station Penny
[18:06] Nebadon Izumi: im also playing video on the main screen too
[18:07] Nebadon Izumi: heh but ya justin that email threads a bit crazy
[18:07] Nebadon Izumi: good discussion though
[18:08] Nebadon Izumi: not sure i have any take on any of the sides myself, though I cant say i feel very supportive of creating standards that have no working examples
[18:08] Justin Clark-Casey: yeah, I have to agree with Diva on that
[18:08] Justin Clark-Casey: better to get standards out of working implementations than try the other way around
[18:08] Nebadon Izumi: not that i am against vwrap, i would just like to see some examples
[18:08] Nebadon Izumi: and know that these standards actually work
[18:09] Richardus Raymaker: I heared the osgrid ToS is now arrivied at the right location and personbs ???
[18:09] Nebadon Izumi: ya we have had a TOS for a few weeks
[18:09] Nebadon Izumi: though its not ready for posting yet
[18:09] Nebadon Izumi: we have to discuss some things with the lawyers still
[18:09] Nebadon Izumi: make some changes, were not 100% happy with the wording just yet
[18:09] Nebadon Izumi: some things right now would probably just start an all out war
[18:10] Richardus Raymaker: ai
[18:10] Penny Lane: Well that's what Hurli is doing, making an example implementation of VWRAP services. Show it working before the standards are set. Makes sense to me.
[18:10] Justin Clark-Casey: well, don't do a digg v4 :)
[18:10] Richardus Raymaker: and new discussion seems the crawlers. but i think thats not for dev meeting
[18:10] Nebadon Izumi: ya I think johns doing a good job
[18:11] Nebadon Izumi: but some others from the VWRAP initiative seem to think its better to document 1st
[18:11] Nebadon Izumi: then make working examples later
[18:11] Nebadon Izumi: i just can not agree with that mentality
[18:11] Richardus Raymaker: documentation first, sounds usefull, saves lots of time with answhere questions
[18:11] Nebadon Izumi: not if the methods are not proven
[18:11] Nebadon Izumi: they are talking about making official standards
[18:12] Nebadon Izumi: before having a single working example
[18:12] Penny Lane: Neb: who thinks that? I don't see anyone advocating that. We're leaving it to John to make the prototype VWRAP services. I guess we trust him.
[18:12] Richardus Raymaker: hmm, how can you make something official if you have not test it ?
[18:12] Andrew Hellershanks: Standards are wonderful things because there are so many to choose from.
[18:12] Justin Clark-Casey: andrew :)
[18:13] Penny Lane: I don't get this anti-VWRAP FUD. John *is* VWRAP. If you have an issue with VWRAP, complain to Hurli, he's working for us.
[18:13] Andrew Hellershanks: I'm getting close to completing my work on ossearch
[18:13] Justin Clark-Casey: nice
[18:13] Andrew Hellershanks: I'm fixing the query routines now
[18:13] Nebadon Izumi: well Meadhbh seems to take offense to some of the things said in the mailing list
[18:13] Penny Lane: He even called his IRC channel #vwrap. You'd think people would realize why.
[18:14] Nebadon Izumi: to the point of just flat out saying that the methods being employed by opensim will just not work out
[18:14] Nebadon Izumi: but ya I think what Johns doing is the right thing
[18:14] Penny Lane: Mead takes offence at a lot of things, don't let it worry you.
[18:14] Nebadon Izumi: but even john doesnt really have a working version yet either
[18:14] Nebadon Izumi: i would just hate to see vwrap be pushed through without any real world testing is all
[18:14] Nebadon Izumi: not that im against it
[18:15] Justin Clark-Casey: if it doesn't work in the real world it will just become a historical document
[18:15] Nebadon Izumi: exactly
[18:15] Nebadon Izumi: and lots of peoples time wasted
[18:15] Penny Lane: At the end of the day, VWRAP is an IETF working group, and standards are decided by consensus, not by one person, and not by a company, nor by a pressure group
[18:15] Justin Clark-Casey: but it doesn't seem that linden themselves are particularly interested in vwrap now which leaves.... opensim?
[18:15] Nebadon Izumi: and simiang
[18:15] Nebadon Izumi: rid*
[18:15] Nebadon Izumi: heh
[18:15] Justin Clark-Casey: I guess
[18:15] Andrew Hellershanks: what is vwrap for?
[18:16] Nebadon Izumi: but ya simiangrid sorta falls under the opensim umbrella
[18:16] Nebadon Izumi: its not that differnt overall
[18:16] Justin Clark-Casey: well, opensim is the only thing that works with it right now, but other people are sure to come along with simulators once the smoke clears
[18:16] Penny Lane: Lindens are officially out of VWRAP. The co-chair Joshua Linden is working on VWRAP only in his spare time, not funded by Linden. And all the other Lindens who were on VWRAP have been fired (Mead and Zero)
[18:17] Nebadon Izumi: the lindens have more important things to worry about that interop
[18:17] Nebadon Izumi: if they dont fix there grid interop wont be a factor anyway
[18:17] Penny Lane: So forget LL. It's in John's hands on a practical front, and the rest of us on VWRAP are supportive of John.
[18:17] Nebadon Izumi: maybe in a couple years LL will be more interested in interop standards if they are still around
[18:17] Penny Lane: Aye
[18:18] Mojito Sorbet: Like AOL suddenly had to discover the WWW
[18:18] Nebadon Izumi: im still a bit skeptical SL is going to last in the long run
[18:18] BlueWall Slade: AOL owns Compuserver
[18:18] Justin Clark-Casey: I think we need LL to remain around to invest in viewer technology
[18:18] Mojito Sorbet: I do not see them doing that
[18:18] Nebadon Izumi: ya well SL2 seems to me to be falling in the fail zone
[18:18] Mojito Sorbet: The viewer innovations are NOT coming from LL
[18:18] Justin Clark-Casey: they appear to be doing that right now, with the renewed focus on sorting the viewer out
[18:18] Justin Clark-Casey: mojito: then from where?
[18:18] Nebadon Izumi: i do see some of the merits of SL2 but so far it sounds to me from their blogs and forums that SL2 is sorta a failure
[18:19] Richardus Raymaker: hi justin, at least for a year of 2
[18:19] Mojito Sorbet: Sure, they have a very innovatinve viewer that can handle user-created content. But what have they come up with RECENTLY? MoaP? lol
[18:19] Justin Clark-Casey: I believe LL can turn it around.
[18:19] Nebadon Izumi: but ya i dont see LL failing in the next 2 years anyway
[18:19] Richardus Raymaker: the LL viewer slow down new things. for my idea
[18:19] Nebadon Izumi: ya i think so too Justin
[18:19] BlueWall Slade: if they do meshes, it bight bring them around in that area
[18:19] Nebadon Izumi: but right now it could go either way
[18:19] Penny Lane: On a personal front, I'm supportive of John as long as he is working towards full interoperability between worlds, and he's always maintained that he is. (Of course individual worlds can be a walled garden if they wish, but the mechanism for interop would be there for those worlds that want it)
[18:20] Nebadon Izumi: ya i think he is, I think he and Diva are pretty set on making things compatible
[18:20] Nebadon Izumi: whatever the name in the end
[18:20] Penny Lane nods
[18:20] Penny Lane: As Rich says, the flaming viewer is slowing down progress.
[18:21] Nebadon Izumi: long road ahead though in terms of a secure and interoperable global grid network
[18:21] Justin Clark-Casey: I think progress depends on more people actually using this stuff
[18:21] Justin Clark-Casey: requires investment
[18:21] Justin Clark-Casey: nebadon: yeah v long and interesting :)
[18:21] Penny Lane: But LL's sources are in public Hg now, so we just need the change of licensing from Opensim and, in principle, the viewer would no longer hold us back.
[18:22] Penny Lane: Talking of which, any news?
[18:22] Justin Clark-Casey: I'm not sure how much it would change
[18:22] Key Gruin is Offline
[18:22] Nebadon Izumi: ya im not sure either especially if LL has no interest
[18:22] Justin Clark-Casey: penny: things are slowly turning in the background
[18:22] Penny Lane: JCC: cool :-)
[18:22] BlueWall Slade: it seems like the LGPL is for LL to be able to use others closed libraries
[18:22] BlueWall Slade: like Vivox, etc.
[18:22] Mojito Sorbet: OpenSim added the new WindLight support, that the LL viewer does not have. That did not hold anyone back form going ahead and doing it.
[18:22] Nebadon Izumi: I do think LL has interest, its just not a priority
[18:23] Nebadon Izumi: the last thing LL has time for right now is interoperability outside their network
[18:23] Penny Lane: What LL's plans are or not isn't of any interest to us.
[18:23] Justin Clark-Casey: yeah, LL are concentratining on survival, I imagine
[18:23] Nebadon Izumi: but the fact that Phillip said in his mind its ok for people to take thier creations outside of LLs grid is a good sign they are not against it
[18:23] Mojito Sorbet: I di dnot get that at all
[18:24] Mojito Sorbet: That is not what I got form what he said
[18:24] Nebadon Izumi: at SLCC he flat out said people can take their creations into OpenSim worlds
[18:24] Mojito Sorbet: They make it as difficult as possible
[18:24] Nebadon Izumi: if your the creator
[18:24] Nebadon Izumi: its not ok if your not the creator
[18:24] BlueWall Slade: not many people create all of the components
[18:24] Mojito Sorbet: Yeah, if you created every little bit of it, regardless of what the intent of the original crator of the doorknob on your house.
[18:24] Penny Lane: I'm just worried about Opensim being held back by the viewer. Not worried about LL at all, they're never going to catch up once the brakes are off viewer developent for open worlds.
[18:25] Justin Clark-Casey: yeah, practically enforcing creatorship makes things difficult
[18:25] Richardus Raymaker: yes, hi blue.
[18:25] Richardus Raymaker: manty things are made with textures form others
[18:25] BlueWall Slade: hi Richardus
[18:25] Nebadon Izumi: well honestly for a secure Vwrap and HG the LL viewer is not going to cut it
[18:25] Penny Lane: Aye, Neb
[18:25] Nebadon Izumi: the reason we can not implement HG2 is because of the LL architecture
[18:25] Nebadon Izumi: really we need something like naali for that
[18:26] Justin Clark-Casey: penny: Do you think open viewer development might lead to more chaos? We already have problems getting things documented in opensim If people stick arbitrary bits on the viewer and manage to get those patches into opensim it could get considerably worse....
[18:26] Nebadon Izumi: either way its a long way off
[18:26] Nebadon Izumi: literally years
[18:26] Nebadon Izumi: even if opensim changed its view on allowing viewer devs
[18:26] Nebadon Izumi: were talking about years of work
[18:26] Penny Lane: That's why I want to see the change of licensing, so that we can hack both sides and bring the viewer into a modern world, able to work with Hurli's assets from multiple worlds. LL will never deliver that.
[18:26] Justin Clark-Casey: yeah, you're probably right
[18:27] Richardus Raymaker: and some worse case thing, what if LL quite sooner. then everybody is stuck with the existing LL code.
[18:27] Nebadon Izumi: well the thing is Penny
[18:27] Nebadon Izumi: if we make it work on Naali
[18:27] Nebadon Izumi: then the LL viewer devs can copy that work
[18:27] Penny Lane: Yep
[18:27] Nebadon Izumi: without violating any licenses
[18:27] Penny Lane: Naali is certainly a good horse to bet on
[18:27] Richardus Raymaker: and without donate something :))
[18:27] Nebadon Izumi: we really need to get more people working on Naali
[18:27] Richardus Raymaker: uhmm :((
[18:27] BlueWall Slade: seem that would be the same circumstance to me?
[18:27] Nebadon Izumi: and worry less about the LL viewers
[18:28] Nebadon Izumi: well ya the LL viewer devs cant work on Naali either
[18:28] Nebadon Izumi: but they can look at naali code
[18:28] Nebadon Izumi: and copy it
[18:28] BlueWall Slade: if their liscense is viral
[18:28] Nebadon Izumi: BSD can be put into GPL
[18:28] BlueWall Slade: right
[18:28] Nebadon Izumi: so if naali can establish a working standard all the other viewers can follow suite
[18:28] Penny Lane: realXtend hasn't made it easy to work on Naali, unfortunately. What a mess of non-standard libraries, jeez.
[18:28] BlueWall Slade: we still have a GPL viewer thoen
[18:28] Justin Clark-Casey: I doubt they could usefully copy much code.
[18:28] Nebadon Izumi: oh im not saying it would be easy
[18:29] Nebadon Izumi: but a working standard helps move it along
[18:29] Justin Clark-Casey: well, the standard really has to be documented.
[18:29] Justin Clark-Casey: yeah
[18:29] BlueWall Slade: I'm guessing that the LL meshes will need to be seen too
[18:29] Nebadon Izumi: whatever we do its years of work
[18:29] Justin Clark-Casey: nebadon: probably
[18:29] Nebadon Izumi: ya probably BlueWall
[18:29] BlueWall Slade: hopefully they are close to naali
[18:29] Nebadon Izumi: but from what i heard from Jhurliman thats not going to be hard
[18:29] BlueWall Slade: yeah
[18:29] Justin Clark-Casey: well, ultimately opensim should be able to accomodate multiple mesh formats
[18:29] Nebadon Izumi: ive heard jhurliman talk about wanting meshes working in opensim almost immediatly after LL releases it
[18:29] BlueWall Slade: ++ jcc
[18:29] Penny Lane: JCC++
[18:30] BlueWall Slade: SVG rendering in html5
[18:30] Andrew Hellershanks: I would love to see meshes
[18:30] Justin Clark-Casey: andrew: have you tried naali on taiga?
[18:30] Andrew Hellershanks: No
[18:31] Justin Clark-Casey: I haven't tried the mesh stuff. Only used naali on opensim
[18:31] Nebadon Izumi: ya Naali on opensim still needs tons of work
[18:31] Penny Lane: Since LL postponed meshes by 6 months for the beta, which means a year for release, perhaps Opensim should introduce their own mesh system. I'm sure Imprudence would be happy to add the new mesh-prim rendering
[18:31] BlueWall Slade: how far along is the SL avatar on it?
[18:32] Richardus Raymaker: I hear more and more now things. but lets nogt forget to optimize sometimes opensim
[18:32] BlueWall Slade: ohhh, they pushed it back?
[18:32] Justin Clark-Casey: well, anybody in OpenSim could propose to put meshes in today, just nobody is
[18:32] Nebadon Izumi: its supposed to be in beta by end of year
[18:32] BlueWall Slade: that is close enough
[18:32] Justin Clark-Casey: yeah, there are lots of opensim problems that need fixing
[18:32] Penny Lane: Yeah. They fired Qarl and Pastrami and pushed it back.
[18:32] Nebadon Izumi: ya personally i dont think opensim is optimzed enough for meshes yet
[18:32] BlueWall Slade: that will tell the direction they are heading - and jh has the inside scoop on it anyway
[18:33] Nebadon Izumi: my 0.7 testing so far has gone horribly bad
[18:33] Penny Lane: Gotta love a company that fires its main developers on a feature about to be released
[18:33] BlueWall Slade: lol, he started working with emerald?
[18:33] Justin Clark-Casey: nebadon: still having problems?
[18:33] Nebadon Izumi: we have alot more important things to worry about right now
[18:33] Andrew Hellershanks: One problem I can see happening is the level of detail on some meshes getting out of hand.
[18:33] Nebadon Izumi: ya Justin, most modules just dont work
[18:33] Nebadon Izumi: Groups is 100% broken in 0.7
[18:33] Nebadon Izumi: that is going to be a big setback for OSgrid moving on
[18:33] Penny Lane: He says he's not *actually* working on Emerald. He's just sitting with them and making snide remarks ;-)
[18:33] BlueWall Slade: yeah, all my things are broken
[18:33] Nebadon Izumi: and Mcortez is MIA for months now
[18:34] Justin Clark-Casey: nebadon: really? I thought that was working - apepars to be okay for me though I haven't tested extensively
[18:34] Justin Clark-Casey: huh
[18:34] Nebadon Izumi: if OSgrid movied to 0.7 right now it would be a huge disaster
[18:34] Andrew Hellershanks: nebadon, is groups module broken or is in the external support files?
[18:34] Justin Clark-Casey: nebadon: what aspect of groups isn't working?
[18:34] BlueWall Slade: the API for things
[18:34] Nebadon Izumi: so far the only thing about groups that works is you can make a group
[18:34] BlueWall Slade: it just takes porting
[18:34] Nebadon Izumi: pretty much nothing else works right
[18:34] Richardus Raymaker: bad 0.7 in what way ?
[18:34] Justin Clark-Casey: nebadon: what do you mean, group chat?
[18:34] Nebadon Izumi: no modifying groups
[18:34] Richardus Raymaker: aha broken groups. thats not good yes.
[18:34] Nebadon Izumi: it doesnst show owners
[18:35] Nebadon Izumi: you can not make it secure
[18:35] Nebadon Izumi: you have to turn off authentication to make it work at all
[18:35] Justin Clark-Casey: mmmmmmm
[18:35] Nebadon Izumi: if you turn authentication on like it is on OSgrid right now
[18:35] Justin Clark-Casey: how about the simiangrid implementation?
[18:35] Nebadon Izumi: it doesnt function al all
[18:35] Nebadon Izumi: simiangrid implementation wont work with Robust
[18:35] Nebadon Izumi: its all on the simiangrid side
[18:35] Nebadon Izumi: uses the simiangrid tables
[18:35] BlueWall Slade: since we are on git, we should take on projects like that and get a core to work as a contact for them
[18:36] Justin Clark-Casey: ah, this is a mess
[18:36] BlueWall Slade: I think mcortez got busy IRL and hasn't had time to update them
[18:36] Nebadon Izumi: also there is no working version of landtool.php in 0.7 yet
[18:36] BlueWall Slade: ?
[18:36] Penny Lane: I don't get the "SimianGrid won't work with ROBUST" angle. I thought the whole idea of ROBUST was to allow external services to be added easily?
[18:36] Nebadon Izumi: so basiclly no one can do land transfers
[18:37] Nebadon Izumi: Penny SimianGrid groups is totally reliant on SimianGrid being functional
[18:37] BlueWall Slade: when Adam commits the SOG refactor, it should brek things too, I suppose
[18:37] Nebadon Izumi: its not a module
[18:37] Nebadon Izumi: its integrated 100% into the siminagrid back end
[18:37] WhiteStar Magic is Online
[18:37] Justin Clark-Casey: all sog refactor work has to be done in the open so we can see how it will break things
[18:37] Nebadon Izumi: without simiangrid being 100% functional simiangrid groups wont work
[18:38] Nebadon Izumi: you cant just run its groups implementation on its own
[18:38] Justin Clark-Casey: do you know if simiangrid makes internal references to its user table from its group module? I was thinking that might be the case
[18:38] Nebadon Izumi: right now SimianGrid is way less functional that RObust too
[18:38] Nebadon Izumi: there is no way OSgrid could run simiangrid, its just not ready yet
[18:38] Justin Clark-Casey: they talk it up, though :)
[18:39] Nebadon Izumi: soo much stuff doesnt work right, i am actually kind of suprised Intel is making such a push so soon to move onto it
[18:39] Nebadon Izumi: only thing i can think is Intel has patches they have not pushed into simiangrid SVN yet
[18:39] Richardus Raymaker: but osgrid cannot ru without groups to
[18:39] Nebadon Izumi: which is probably very likely
[18:39] Penny Lane: Something's wrong. We're trying to get away from walled gardens, not create a new one. Groups are meant to be independent of login services and of asset services.
[18:39] Justin Clark-Casey: I don't think that's the case, but who knows
[18:39] BlueWall Slade: is anyone working to update the groups module to 0.7 ?
[18:39] Nebadon Izumi: simiangrid is very integrated Penny
[18:39] Nebadon Izumi: they seem to think everything should be built directly into simiangrid
[18:40] Penny Lane: Neb: well that's wrong.
[18:40] Nebadon Izumi: where opensim is moving towards modularization
[18:40] Justin Clark-Casey: bluewall: doesn't look like it
[18:40] BlueWall Slade: I have some changes here for prebuild
[18:40] Nebadon Izumi: but ya, honestly ive not had time for simiangrid in about 2 months now
[18:40] Justin Clark-Casey: I didn't even know it was broken until 10 mins ago, though
[18:40] Nebadon Izumi: ive done very little testing or development there lately
[18:40] Penny Lane: And it's not the VWRAP way either. So if Hurli is making everything dependent on SimianGrid, he's screwed up. Get him to talk to us.
[18:40] Nebadon Izumi: i dont even have a working Simiangrid setup at this point
[18:40] Frank Northmead is Offline
[18:40] Richardus Raymaker: hmm, that sounds bad nebadon. only becasue mangement etc. seperate is easy to do updates
[18:41] Nebadon Izumi: oh god updating simiangrid is a nightmare
[18:41] BlueWall Slade: seems that nant is obsolete
[18:41] Nebadon Izumi: thats another reason i dont have time for it
[18:41] Justin Clark-Casey: bluewall: how is nant obsolete?
[18:41] Nebadon Izumi: xbuild is replacing it
[18:41] Richardus Raymaker: maby not the latest beta installed
[18:41] BlueWall Slade: and mono ships with a build system , xbuild
[18:41] Justin Clark-Casey: The last I saw, xbuild wasn't very ready
[18:42] Richardus Raymaker: oh.. i still use nant.. hmm intressting
[18:42] Nebadon Izumi: it works pretty good, just not with opensimulator
[18:42] BlueWall Slade: also, our prebuild screws up monodevelop
[18:42] Nebadon Izumi: oh btw
[18:42] Nebadon Izumi: Mono GIT Trunk has sgen enabled by default
[18:42] BlueWall Slade: well, it is our current prebuild's fault
[18:42] Nebadon Izumi: but opensim runs horrible in mono 2.8.1
[18:42] Nebadon Izumi: it runs ok if you have 1 simulator isolated with zero neighbors
[18:42] BlueWall Slade: so,l jhurliman helped me get as new versio of prebuild in
[18:42] Justin Clark-Casey: one day I'm going to rewrite opensim in java
[18:42] Nebadon Izumi: but as soon as you have 2 sims next to each other that have to talk.. it instantly goes to hell
[18:43] Nebadon Izumi: but i was actually suprised it ran rather well on 1 single sim completely isolated from the rest of the grid
[18:43] Justin Clark-Casey: nebadon: which version of opensim is that?
[18:43] Nebadon Izumi: teleporting sucks though
[18:43] BlueWall Slade: the GC does that neb ?
[18:43] Nebadon Izumi: Master OpenSim
[18:43] Nebadon Izumi: no its not GC i dont think bluewall
[18:43] Nebadon Izumi: i think something has massively changed in Mono
[18:43] Justin Clark-Casey: bluewall: nant is our current build system so we need consultation to change it.. It also does struff like run the testsuite and create an install package
[18:43] Simulator Version v0.5 shouts: OpenSim 0.6.9 (Post_Fixes) 2b4af70: 2010-08-26 14:50:54 -0700 (Unix/Mono)
[18:43] BlueWall Slade: ok
[18:44] BlueWall Slade: I will submit the work I have on it
[18:44] Nebadon Izumi: I was suprised though even OpenSim compiled ok in x64 on mono 2.8.1
[18:44] Nebadon Izumi: they seemed to fix whatever problem we were having in 2.6.7 to 2.7 versions
[18:44] Richardus Raymaker: lopoks scary... opensim dont like mono. ok its some very enw one
[18:44] Justin Clark-Casey: nebadon: how long ago did you test?
[18:44] BlueWall Slade: I can add some more work to the prebuild to fix the nant build
[18:44] Nebadon Izumi: but otherwise opensim is unusable, i suspect it has something to do with HTTP
[18:44] Nebadon Izumi: whenever sims have to talk to each other
[18:44] Nebadon Izumi: 2 days ago Justin
[18:45] BlueWall Slade: hmm, they just use REST
[18:45] BlueWall Slade: no more reflection right?
[18:45] Nebadon Izumi:
[18:45] Nebadon Izumi: im not sure all the changes BlueWall
[18:45] Justin Clark-Casey: bluewall: that would be nice - if we're going to change it would be nice to have a smoother transition.
[18:45] Nebadon Izumi: mono is as good at documentation as we are
[18:45] BlueWall Slade: lol, they work fast too
[18:45] Key Gruin is Online
[18:46] Justin Clark-Casey: nebadon: was that two regions on the same machine or separate machines?
[18:46] Nebadon Izumi: i tried both
[18:46] Nebadon Izumi: even tried mega regions and multi instance
[18:46] Justin Clark-Casey: same machine doesn't involve any network comms, of course
[18:46] Nebadon Izumi: the only instance that worked more than a few minutes without exploding was a completely isolated server
[18:46] Justin Clark-Casey: what do you mean by exploding, core dump?
[18:46] Nebadon Izumi: well even a multi instance sim and mega region did really horrible
[18:46] Nebadon Izumi: no timeouts
[18:46] Nebadon Izumi: it seemed to just spew timeouts constantly
[18:47] Nebadon Izumi: to the point you just get logged out
[18:47] Nebadon Izumi: and then can never relog
[18:47] Nebadon Izumi: it seemed to happen most when sims had to talk to each other
[18:47] Nebadon Izumi: and sometimes when talking to robust also
[18:47] Nebadon Izumi: but to robust wasnt nearly as bad as when sims had to talk to each toher
[18:47] Nebadon Izumi: i couldnt teleport at all
[18:47] Nebadon Izumi: even if the sims were not next to each other
[18:48] WhiteStar Magic is Offline
[18:48] BlueWall Slade: did you try going 3 sims away from the destination?
[18:48] Justin Clark-Casey: You might want to try 75e2a2b3ceeb67d3124bc8b4534b6927d288fd57, in case there are locking issues with the latest code
[18:48] Justin Clark-Casey: or
[18:48] Nebadon Izumi: ok i even tried 0.6.9
[18:48] Nebadon Izumi: had virtually exact same results
[18:48] Justin Clark-Casey: oh okay, perhaps that's not an issue then
[18:49] Justin Clark-Casey: that's kind of bad....
[18:49] Nebadon Izumi: ya personally i think its mono
[18:49] Danny Flyte: <-- smxy
[18:49] Richardus Raymaker: Hrr
[18:49] Nebadon Izumi: ever since mono 2.6.7 things have gone way down hill for opensim
[18:49] Andrew Hellershanks: Hey, Danny
[18:49] Justin Clark-Casey: hi danny/smxy
[18:49] Andrew Hellershanks: <-- Plugh
[18:49] Nebadon Izumi: really mono 2.6.x in general
[18:49] Nebadon Izumi: you really cant run opensim on 2.6+
[18:49] BlueWall Slade: I'm on 2.6.4
[18:49] Nebadon Izumi: not unless you like torturing yourself constantly
[18:49] Justin Clark-Casey: ubuntu will move to mono 2.6 in october so then I'll suffer the pain :)
[18:49] Justin Clark-Casey: nebadon: seems to be okay for me
[18:49] Nebadon Izumi: you dont have any issues with Xengine Bluewall
[18:49] Richardus Raymaker: and 0.6.7
[18:49] Justin Clark-Casey: but then my uses don't have lots of avatars
[18:49] Danny Flyte: My SimianGrid is running on it ok.
[18:49] Nebadon Izumi: i had god awful xengine problems in 2.6.x
[18:50] BlueWall Slade: one or two
[18:50] Mojito Sorbet: I observe that the people trying to run OpenSim in a commercial way are preferring Windows Servers
[18:50] Nebadon Izumi: sims constantly deadlocking with no recovery
[18:50] Richardus Raymaker: sorry nebadon. i run it on 1 server with mono 2.6.7
[18:50] Andrew Hellershanks: justin, I'll be sharing your pain as I'll be updating my Ubuntu install in Oct.
[18:50] Justin Clark-Casey: if you've got the moneyt that's probably sensible
[18:50] Nebadon Izumi: i am unable to move beyond still
[18:50] BlueWall Slade: I took that code Melany put up for the compiling and did some things with it
[18:50] Nebadon Izumi: plazas lock up every hour or less with 2.6.x
[18:50] BlueWall Slade: it improved things
[18:50] Nebadon Izumi: ah ok
[18:50] Nebadon Izumi: ya the stock opensim is bad
[18:50] Justin Clark-Casey: what?
[18:50] Richardus Raymaker: but i cant trust the windows security. i think all the dictonary scans come from windows servers
[18:50] Nebadon Izumi: but even melanie refuses to use 2.6.x
[18:51] BlueWall Slade: the issues seems to be when things are in the compile queue, attachments. And the avatar leaves the scene it crashes
[18:51] Nebadon Izumi: with Mono 2.6.x Justin
[18:51] BlueWall Slade: a deadlock
[18:51] Nebadon Izumi: opensim runs like poo
[18:51] Justin Clark-Casey: it seems okay for me on mono 2.6
[18:51] Nebadon Izumi: ya your probably not doing alot of scripting
[18:51] BlueWall Slade: right now, I see a little more resource useage
[18:51] Nebadon Izumi: most of the problems are with Xengine and prims with multiple scripts inside
[18:51] Justin Clark-Casey: no, that's true - I do a lot of stuff server-side
[18:51] Nebadon Izumi: when you delete them
[18:51] Nebadon Izumi: LSL scripting?
[18:52] BlueWall Slade: that is over the past month oe two
[18:52] Nebadon Izumi: with xengine?
[18:52] Aaron Duffy: mono 2.6.1 works alright for me. Very few avatars but large (>5000) numbers of scripts
[18:52] Nebadon Izumi: AAron try deleting a prim with multiple scripts inside
[18:52] Nebadon Izumi: the sim will likely deadlock and never recover
[18:52] BlueWall Slade: some of these things, collars, are destructive anyway
[18:52] Nebadon Izumi: the plazas were freezing up every 30-60 minutes
[18:52] Justin Clark-Casey: if it deadlocks then it should be possible to find out where the daedlock is by getting a thread dump
[18:52] BlueWall Slade: 35+ scripts w/750 or more lines each
[18:52] Nebadon Izumi: Melanie says its in Mono Justin
[18:52] Aaron Duffy: Ah. okay, my use case NEVER deletes anything due to past issues with llDie
[18:52] Justin Clark-Casey: a deadlock is different from just locking up due to some internal mono problem
[18:53] Justin Clark-Casey: aaron: :) How do you get away with never deleting anything?
[18:53] Nebadon Izumi: and to be honest i didnt actually try that in mono 2.8.1
[18:53] Nebadon Izumi: i will though
[18:53] BlueWall Slade: I would get messages about the heartbeat
[18:53] Nebadon Izumi: hey there Mic
[18:53] Aaron Duffy: I recycle my 'plants'
[18:53] Nebadon Izumi: heh
[18:53] Mic Bowman: hi
[18:53] Aaron Duffy: A region module tracks and reuses them
[18:53] So Crates: hey there
[18:53] Justin Clark-Casey: aaron: how very hackish :)
[18:53] So Crates: how's it going :)
[18:54] Aaron Duffy: tnx :>
[18:54] Justin Clark-Casey: hi mic
[18:54] Justin Clark-Casey: hi so
[18:54] Frank Northmead is Online
[18:54] Mic Bowman: hi justin
[18:55] Nebadon Izumi: Justin if you want i can setup a sim running mono 2.8.1 with sgen again if your intereted in poking on it at all
[18:55] Nebadon Izumi: i know your busy
[18:55] Justin Clark-Casey: what is sgen?
[18:55] Nebadon Izumi: but if you want
[18:55] Justin Clark-Casey: horribly :)
[18:55] Nebadon Izumi: its the Next Generation Garbage Collector service
[18:55] Nebadon Izumi: they actually have compacting GC now by default
[18:55] Nebadon Izumi: in mono Trunk
[18:55] Richardus Raymaker: oh, can the empty my trash :)
[18:55] Nebadon Izumi: hehe
[18:56] Justin Clark-Casey: Oh I see. btw, I'm now thinking about submitting that 0.6.9 compiler bug to mono when I get a chance
[18:56] Nebadon Izumi: hang on let me find you a article i read that caught my attention on it
[18:56] Nebadon Izumi:
[18:56] Nebadon Izumi: its supposed to improve performance of mono by like 10 fold
[18:56] Nebadon Izumi: but in terms of opensim so far its been a bit downhill
[18:56] Richardus Raymaker: whoo
[18:56] Danny Flyte: doesn't it have to freeze the world to do the gc though?
[18:57] Richardus Raymaker: thats OOps nebadon
[18:57] Alejandro Acanthus: :O
[18:57] Alejandro Acanthus: EVERYONES SITTING IN A HALF CIRCLE
[18:57] Nebadon Izumi: not really sure Danny, that doestn necessarily mean opensim will freeze
[18:57] Richardus Raymaker: ok, seems to fail inworld
[18:57] Alejandro Acanthus: Your with the illuminati
[18:57] Justin Clark-Casey: hmm
[18:57] Alejandro Acanthus: :o
[18:57] BlueWall Slade: we are them
[18:57] Nebadon Izumi: heh
[18:58] Justin Clark-Casey: If only
[18:58] Nebadon Izumi: why were not on the dark side of the moon!
[18:58] Nebadon Izumi: hehe
[18:58] BlueWall Slade: lol, they have to ask us if they can go to McDonnalds for a shake.
[18:58] Richardus Raymaker: but is mono faulty or opensim with newer versions ?
[18:59] Nebadon Izumi: both Richardus
[18:59] Nebadon Izumi: hehe
[18:59] Richardus Raymaker: lol
[18:59] Richardus Raymaker: best anmswhere. you can never be wrong
[18:59] Nebadon Izumi: neither are anywhere near perfect
[18:59] Justin Clark-Casey: just shows how difficult it is to get a vm right
[18:59] Nebadon Izumi: but the fact it runs at all is encouragine
[18:59] Qandy Saw: good evening/day
[18:59] Nebadon Izumi: encouraging*
[18:59] Nebadon Izumi: to be honest when i tested it i wasnt expecting opensim to even start or compile
[18:59] Richardus Raymaker: make a .net crosscompiler to C++ or so. thats more painfull i think
[18:59] Nebadon Izumi: the fact i was able to even log in at all, is huge i think
[19:00] Justin Clark-Casey: nebadon: well that would have been bizarre, since it's all compilant c# code
[19:00] Justin Clark-Casey: nebadon: it really shouldn't be.
[19:00] Nebadon Izumi: opensim doesnt compile on mono 2.6.7
[19:00] Nebadon Izumi: so i wouldnt call it too strnage
[19:00] Nebadon Izumi: strange*
[19:00] Justin Clark-Casey: nebadon: I think that in itself is strange :)
[19:00] Nebadon Izumi: it didnt compile on mono 2.7 either
[19:00] Richardus Raymaker: opensim compile on mono 2.6.7
[19:00] BlueWall Slade: it wouldn't compile under monodevelop until the build system was fixed either.
[19:00] Justin Clark-Casey: nebadon: that's strange too :)
[19:00] Nebadon Izumi: but they seemed to fix whatever that was
[19:00] Justin Clark-Casey: The problem is, Java doesn't have a lot of the nice c# features
[19:01] Richardus Raymaker: or do you mena build new mono on mono 2.6.7 ?
[19:01] Richardus Raymaker: sorry nebadon. forgot mono 2.6.7 i have from repro. i made self
[19:01] Nebadon Izumi: we should have more people testing mono 2.8.1 though
[19:01] Nebadon Izumi: i would love for you guys who are looking for something to do to give it a shot
[19:01] BlueWall Slade: I will give it a shot
[19:01] Nebadon Izumi: i can certainly explain what i did if anyone wants instructions
[19:02] Richardus Raymaker: Give me a bigger desk here :) then i have more test space
[19:02] Nebadon Izumi: i basiclly just used the standard mono compile instructions i posted on the osgrid forums
[19:02] BlueWall Slade: since I went to a package system, I dont' ventory out much. But I can set it in it's own place
[19:02] Nebadon Izumi: mono is now using git, so thats nice
[19:02] So Crates: \o/ git converts
[19:02] Richardus Raymaker: sure ?
[19:03] BlueWall Slade: yes
[19:03] Justin Clark-Casey: ok, I need to go. Good talk chaps, if a little deperessing :)
[19:03] Nebadon Izumi: ive only spend about 6 hours testing it so far, it needs wider testing for sure
[19:03] So Crates: take care jcc
[19:03] BlueWall Slade: hehe, so long Justin
[19:03] Nebadon Izumi whispers: later Justin, thanks for swinging by :)
[19:04] Justin Clark-Casey waves
[19:04] Justin Clark-Casey is Offline

Personal tools
About This Wiki