Chat log from the meeting on 2015-02-17

From OpenSimulator

Jump to: navigation, search
[11:00]  BlueWall.Slade so, basically LL took part of the aniumation process and opened it up to prims
[11:00]  Connecting to in-world Voice Chat...
[11:00]  Connected
[11:00]  Arielle.Popstar Hi Jak
[11:00]  Richardus.Raymaker hello jak
[11:00]  BlueWall.Slade Hi Jack
[11:00]  Jak Daniels: good evening ;)
[11:00]  BlueWall.Slade Jak*
[11:00]  Justin Clark-Casey: hello folks
[11:00]  BlueWall.Slade Hi Justin
[11:00]  Seth.Nygard Hello everyone
[11:00]  Shez.Oyen Hey Justin
[11:00]  Richardus.Raymaker hi justin
[11:00]  Jak Daniels: hi justin
[11:01]  Arielle.Popstar Hi Justin
[11:01]  Richardus.Raymaker maby my memory is wrong. but it where not possible to run scripts on different opensim instance. so its seperated from the simulator ?
[11:01]  Justin Clark-Casey: rich: that is not possible
[11:02]  BlueWall.Slade I thjink in the far distant past it was - or at least partly so.
[11:02]  BlueWall.Slade I have memories of some script server.
[11:02]  Richardus.Raymaker i have a new train. script works fine. expcept smoothness and a mysterious derail today. *never happens in sl)
[11:02]  Aine.Caoimhe seems like there is a pretty severe issue with HG inventory transfers..particularly of scripted objects....hoping very much that this can be tracked down and fixed before 0.8.1 goes live
[11:02]  Arielle.Popstar nod,  it is bad
[11:02]  Seth.Nygard
[11:03]  BlueWall.Slade I think we have a cure, just needs to be don.
[11:03]  BlueWall.Slade done*
[11:03]  Justin Clark-Casey: we do?
[11:03]  Aine.Caoimhe disable HG compeltely? :p
[11:03]  Arielle.Popstar backwards compatible?
[11:03]  BlueWall.Slade The inventory caps
[11:03]  Justin Clark-Casey: I don't think that is connected with this issue
[11:03]  BlueWall.Slade is that the same
[11:03]  BlueWall.Slade Ohh, ok
[11:04]  Justin Clark-Casey: I am awre of the mantis but my time is currently being taken up by other things. After that I will look
[11:04]  Aine.Caoimhe this seems to be something borking with state in the XML
[11:04]  Arielle.Popstar is it a viewer issue?
[11:04]  Justin Clark-Casey: I don't understand yet how this has suddenly started happening
[11:04]  BlueWall.Slade inventory being assets inside the rezzed prim?
[11:04]  Aine.Caoimhe and will happen for anything containing scripts, particle effects, target omega
[11:04]  Dahlia.Trimble is Online
[11:04]  Seth.Nygard This is going to be a majour issuse as people try out the newer versions. I already know of 5 grids that are seeing this.
[11:04]  Aine.Caoimhe it can be any scripted object, boxed or not
[11:05]  Aine.Caoimhe or anything that has some sort of effect applied to it
[11:05]  Justin Clark-Casey: well, this is one reason for release candidates
[11:05]  Dahlia.Trimble hi
[11:05]  Arielle.Popstar it seems fine  here
[11:05]  BlueWall.Slade Hi Dahlia
[11:05]  Justin Clark-Casey: this installation is running older software
[11:05]  BlueWall.Slade Hi Robert
[11:05]  Justin Clark-Casey: hi dahlia
[11:05]  Richardus.Raymaker talking about HG, is it possible ifd a HG sim is private to addd HG users to the accesslist. last time i tried it it failed. or i do it wrong.
[11:05]  Shez.Oyen Hi Dahlia, Robert
[11:05]  Arielle.Popstar but almost any other grid i jump to others will not see my hair
[11:05]  Arielle.Popstar (assumes i have hair)
[11:05]  Robert.Adams hello all
[11:05]  Justin Clark-Casey: most of the grids are going to be running older opensim
[11:05]  BlueWall.Slade RiRa, you can set it up under estate / group management I think
[11:06]  Justin Clark-Casey: hello robert
[11:06]  Robert.Adams waiting for things to rez so I can sit down
[11:06]  Richardus.Raymaker hi robert
[11:06]  Arielle.Popstar Hi Dahlia and Robert
[11:06]  Richardus.Raymaker ok. ned to test that when im sofar
[11:06]  Justin Clark-Casey: THe most useful thing you can do for me is pin down the commit when this started failing
[11:06]  Aine.Caoimhe it's not just a matter of seeing it, Arielle....most of these objects will completely fail to transfer the script
[11:07]  BlueWall.Slade 0007397 ??
[11:07]  Arielle.Popstar nod not sure about the scripted stuff but already a prob for regular attachmnets
[11:07]  Richardus.Raymaker did you checked the perms aine. maby only scripts with some eprms fail ?
[11:07]  Richardus.Raymaker hi dahlia
[11:07]  Aine.Caoimhe all fill perm
[11:07]  Aine.Caoimhe *full
[11:08]  Richardus.Raymaker uch.. and HG is getting so important
[11:08]  Arielle.Popstar it is Rira
[11:08]  Aine.Caoimhe and according to console logs it's the state in the XML that is causing issues
[11:08]  BlueWall.Slade Aine - what is the mantis # for that issue, pleas?
[11:08]  Seth.Nygard
[11:09]  BlueWall.Slade thanks
[11:09]  Aine.Caoimhe I think it might also be related to the one AiAustin opened re issues on inventory transfer as well
[11:09]  Aine.Caoimhe
[11:09]  Seth.Nygard yes, it may very well be related to that
[11:10]  Arielle.Popstar is it possible that it is to do with export/import configs and the various HG versions?
[11:10]  Justin Clark-Casey: no, most likely it's a bug I have introduced
[11:10]  Richardus.Raymaker people need to stop using 0.7.6
[11:10]  Justin Clark-Casey: since I'm the only person who has recentloy done hg related work
[11:11]  Aine.Caoimhe I'm thinking it's most likely something when you did that other script state commit
[11:11]  Seth.Nygard If there are other things taking priority then would you suggest I revert all my regions to older code and not test new code until something gets fixed?
[11:11]  Justin Clark-Casey: no I doubt that's the issue
[11:11]  BlueWall.Slade no worries, sometimes that happens when moving forward.
[11:11]  Justin Clark-Casey: really, you should only be running master in test regions. I would recommend using releases for important regions
[11:11]  Aine.Caoimhe you'll need to revert Robust too, Seth
[11:11]  Justin Clark-Casey: but we do very much appreciate the testing, it's a tricky balance
[11:12]  Aine.Caoimhe yeah, we *wanted* to be running master to ensure it's getting *some* testing in a grid situation
[11:12]  Seth.Nygard While I agree with that in principal, the only way some of these things get tested is for them to be actually used across many grids and users
[11:12]  BlueWall.Slade It doenst cost a lot to run a robust stack and two or three regions in a test setup.
[11:12]  Aine.Caoimhe and we're willing to suffer through the bumps and bruises of doing that as much as possible
[11:12]  BlueWall.Slade I guess that should be part of a good migration workflow
[11:13]  Marcus.Llewellyn I spose a grid may designate certain sims as testing sims if they chose. Sandboxes may be suitable. Run the rest on release.
[11:13]  BlueWall.Slade ++
[11:13]  Aine.Caoimhe this bug would appear to be a robust-related one, though so it's all-or-nothing there
[11:13]  Justin Clark-Casey: I doubt this is robust related
[11:13]  Seth.Nygard Actually, no it is a region one
[11:14]  Richardus.Raymaker sounds logic, region is handling most of the HG stuff
[11:14]  BlueWall.Slade I will need to test it. I have done some HG around on grids running master w/o issues as far as I could tell.
[11:14]  Arielle.Popstar so same characteristics between standalones?
[11:14]  Seth.Nygard Robust is still running an older version. I tried changing Robust versions and it had no impact
[11:14]  Aine.Caoimhe asset transfer between grids is region level?
[11:14]  BlueWall.Slade But, I will go back after reading the mantis and check it out.
[11:14]  Justin Clark-Casey: aine: not completely, but all serialization and deser is handled in the simulators
[11:14]  Aine.Caoimhe ah
[11:15]  Justin Clark-Casey: possibly the changes I did to vastly improve performance there introduced an issue
[11:15]  Justin Clark-Casey: or possibly some other hg related performance chance
[11:15]  Richardus.Raymaker hi marcus
[11:15]  Justin Clark-Casey: change
[11:15]  Justin Clark-Casey: or sometihng else entirely
[11:15]  Arielle.Popstar but how recent was that?
[11:15]  BlueWall.Slade git bisec is our friend :)
[11:16]  Justin Clark-Casey: well, sometimes these bugs don't show up for a while. And then you people are often incapable of actually dientifying which commit caused the issue
[11:16]  Aine.Caoimhe it's something that seems to be new within the last couple of weeks or perhaps month at most
[11:16]  Richardus.Raymaker then you need to go back in commit by the week.
[11:16]  Seth.Nygard Talking to some users it seem to have started around the end of January, although we found it a major issue only last week
[11:16]  BlueWall.Slade checkout a version from 1.5 months ago and see if it works ...
[11:16]  Richardus.Raymaker until you find one thats working and then go up. yiks
[11:17]  BlueWall.Slade then use git bisect to zero in on the break.
[11:17]  BlueWall.Slade it will make fast work of it
[11:17]  Aine.Caoimhe if I knew how to do that and how to read the results, I suppose it would
[11:18]  Aine.Caoimhe I'm just a dumb user though
[11:18]  BlueWall.Slade it might be a two-fold issue since the foreign grid is involved.
[11:18]  BlueWall.Slade Hmmm
[11:18]  Justin Clark-Casey: it's quite easy. git bisect automatically does a divide and conquer and requires you to say whether a commit is 'good' or 'bad'
[11:18]  Justin Clark-Casey: i.e. whether it works or doesn't
[11:18]  Billy.Bradshaw What happens if the issue is due to an XML format change? and is subsequently stored.
[11:19]  Arielle.Popstar is there a wiki how to?
[11:19]  Justin Clark-Casey: bluewall: yeah, this makes it more complicated
[11:19]  BlueWall.Slade it does, but it should fail between 2 standalone installations?
[11:19]  Justin Clark-Casey: arielle;
[11:20]  BlueWall.Slade then use the code to run both installations and create from scratch if needed - load iar/oar to poulate it.
[11:20]  Justin Clark-Casey: billy: it's unliikely to be that, since ppl still report the items working on the original source grid
[11:20]  Justin Clark-Casey: and when taken manually. Can we confirm that buy is the issue or is that misleading?
[11:21]  BlueWall.Slade with the kitely thing, that is an external module that makes delivery, right?
[11:21]  Aine.Caoimhe yes, and transfer of those items between regions in the grid works perfectly
[11:21]  BlueWall.Slade So, we should rule that out by testing/reporting stock OpenSim
[11:22]  Aine.Caoimhe even if they're all running in different instances (which ours all are)
[11:22]  Billy.Bradshaw Error I have seen indicates that the XML format is wrong, or something missing.
[11:22]  Seth.Nygard Taking a copy of a boxed item still does the same thing. And sometimes taking the actual scripted object still results in the problem as well
[11:22]  BlueWall.Slade No matter. We only need to know if the issue is OpenSim itsself
[11:22]  Seth.Nygard I did find that boxed items have the problem with much greater repeatability
[11:23]  Justin Clark-Casey: seth: you are implying this does not occur every single time
[11:23]  Seth.Nygard with boxed items I was finding it every time, taking the actual object it was not 100%
[11:24]  Aine.Caoimhe also if I put the object (unboxed) into my suitcase, travel there, then rez it to ground, it *usually* works
[11:24]  BlueWall.Slade So, leave an item in a box on a Hg region and the visiting avatar gets a broken object?
[11:24]  Aine.Caoimhe right
[11:24]  BlueWall.Slade Ok, please add a 100% duplicatable case on the mantis to help us test it.
[11:25]  Aine.Caoimhe Seth did
[11:25]  BlueWall.Slade hmm, 1 sec
[11:25]  BlueWall.Slade
[11:25]  BlueWall.Slade ?
[11:25]  Seth.Nygard yes
[11:25]  Aine.Caoimhe his "steps to reproduce" will always fail
[11:26]  BlueWall.Slade I only see 2 entries - maybe it's cached.
[11:26]  BlueWall.Slade ohhh, sorry - I see it now
[11:26]  BlueWall.Slade good deal.
[11:26]  Justin Clark-Casey: seth: Could you put on there what opensimulator code the source grids were running
[11:26]  Mal.Burns is Online
[11:26]  Justin Clark-Casey: if you know it
[11:27]  Seth.Nygard Those tests where done between RG and a parallel test grid I setup, both running the version listed
[11:27]  Aine.Caoimhe OpenSim Dev r/25792 dd6f560c05 11-Feb-2015
[11:27]  Justin Clark-Casey: Please could you add that fact to the mantis?
[11:27]  Aine.Caoimhe it's there
[11:28]  Aine.Caoimhe in "git revision or version number" field of the Mantis
[11:28]  Seth.Nygard I was able to reproduce it from other girds running older versions of he Dev branch as well, Gavin's XMIR grid for example, but I don't know specifically what commit level he is running
[11:28]  BlueWall.Slade Seth - did you get what I said about using git bisect with standalones and iar/oar to test it?
[11:29]  Justin Clark-Casey: seth: ok, if it happens between grids at the same level then that suggests it's not a backward compatibility issue at least
[11:29]  BlueWall.Slade although you may not see it if all are on the same host.
[11:29]  Seth.Nygard yes, I have aleady put over 20 hours into testing this but needed to spend my time on other issues as well
[11:29]  Justin Clark-Casey: I think we're only stressing this point because if one can idnetify the commit, a 6 hour fix might become a 3 hour fix
[11:29]  BlueWall.Slade ok
[11:30]  Justin Clark-Casey: or even 2. And that will get attention sooner than the dev having to work this out him/herself. But it will get attention anyway
[11:30]  BlueWall.Slade but, if some of the time is spent using git bisect to identify the breaking point, it is possible to get it fixed quickly ...
[11:30]  BlueWall.Slade because we are all facing the same issue.
[11:30]  Justin Clark-Casey: yes
[11:30]  Seth.Nygard nods
[11:30]  Seth.Nygard That I understand
[11:31]  BlueWall.Slade time is short, and that workflow is absoutley the best and fastest way to identify the breaking point.
[11:31]  Richardus.Raymaker i would not know how bisect works :O but that we see when its ever needed
[11:32]  BlueWall.Slade It might be an easy fix. And it could also be in the middle of seom refinements Justin is working on, in that case he would be able to decide how to go about resolving it.
[11:32]  Justin Clark-Casey: if you build from git then it's not that hard. I don't know how it works with git UI tools though
[11:32]  Seth.Nygard What I have seen so far is that the issue happens if either source or destination are running newer code
[11:32]  Richardus.Raymaker i use git console (windows)
[11:32]  Justin Clark-Casey: well, you know, half the battle is idenifying exactly when it started. That's super useful in narrowing it down
[11:32]  Seth.Nygard nods
[11:32]  BlueWall.Slade
[11:34]  Seth.Nygard I am building from git so that is not an issue
[11:34]  BlueWall.Slade everything else works great - no bugs?
[11:34]  BlueWall.Slade hehe
[11:35]  Justin Clark-Casey: ahhhh
[11:35]  BlueWall.Slade I see Dashlia has been hacking away at physics still
[11:35]  Richardus.Raymaker dont think so, thats more finetuning or rewriting
[11:35]  Marcus.Llewellyn My sims stopped dispensing free beer. Fix it.
[11:35]  BlueWall.Slade Dahlia
[11:35]  Justin Clark-Casey: seth: did you ever get a chance to try the debug commands with the bad network connection?
[11:35]  Dahlia.Trimble moi?
[11:35]  Richardus.Raymaker Marcus, check if the beer barrel is full
[11:35]  BlueWall.Slade si
[11:36]  Seth.Nygard I have not yet had a time where there was enough impact on a region to be able to see any difference.
[11:36]  Dahlia.Trimble ya a few minor issues with the lookat stuff. Seems to be pretty good now
[11:36]  Justin Clark-Casey: seth:ok
[11:37]  Aine.Caoimhe unfortunately the ideal test case was a week ago at Nara's Nook but they didn't have console access to run it
[11:37]  Seth.Nygard The actual trigger seems to be more than just the single user, it is a combination of things that cause a major issue. I am continually watching for the first chance I have however
[11:37]  Aine.Caoimhe there were 3 latent viewers all trying to enter the 'sploded real good
[11:37]  BlueWall.Slade how to get the vehicles to run at the desired speed? The value in the motor
[11:37]  BlueWall.Slade seems that friction overpowers it.
[11:38]  Seth.Nygard use llAddOil() function maybe ;)
[11:38]  BlueWall.Slade (mayby I don't have a full grasp of the vehicles yet too)
[11:38]  Robert.Adams friction from the ground?
[11:38]  BlueWall.Slade lol, maybe
[11:38]  Richardus.Raymaker i where thinking at the w-e-cayote acme way. just add more trusters
[11:38]  Arielle.Popstar llclearsnow function?
[11:38]  BlueWall.Slade I gues so Robert.
[11:39]  Marcus.Llewellyn is still waiting for llOrderPizza().
[11:39]  BlueWall.Slade but, I know applying an impulse to the +Z direction only causes a bump.
[11:39]  Aine.Caoimhe llSecondMortgage(house, toMechanic)
[11:39]  Richardus.Raymaker pizza waiting for llrezz(doorbell)
[11:39]  Arielle.Popstar noit til llfreeBeer is fixed
[11:39]  Robert.Adams there is friction set for the object and the terrain... but setting friction to zero makes everything glass
[11:39]  BlueWall.Slade removing the vehicle and applying the same impulse will launch it 200m into the sky
[11:40]  BlueWall.Slade ok, wil try that.
[11:40]  Richardus.Raymaker atach a bungee coord bluewall
[11:40]  Robert.Adams I've also found that reducing the density of an object makes movement better... I suspect that weight is too much in BulletSim
[11:40]  BlueWall.Slade ok.
[11:40]  Richardus.Raymaker whats a good max. weight robert ?
[11:41]  Arielle.Popstar any of you tried out Shin Ingrens bikes?
[11:41]  Richardus.Raymaker never seen them
[11:41]  Arielle.Popstar take a look Rira
[11:41]  Arielle.Popstar ]they are great
[11:41]  Robert.Adams weight is a function of size,density and gravity.... reducing density by half or gravity by half is something I've played with with good results
[11:41]  Arielle.Popstar on Bullet
[11:42]  BlueWall.Slade They are Arielle
[11:42]  Richardus.Raymaker aha. i where thinking in Land Impact if you talk about weight :O
[11:42]  Robert.Adams I was working with someone that had a game table with pieces that hit each other (pool table like)... reducing density of hte pieces by half greatly improved the sliding ad interaction of the pieces
[11:43]  BlueWall.Slade pool tables FTW :)
[11:43]  Aine.Caoimhe wonders if parts of Blender's physics engine would be any's quite good an I assume is opensource
[11:44]  Marcus.Llewellyn It's a shame that there doesn't seem to be much documentation for Bullet that addresses content creators. All the docs are for programmers. Guidelines for creating content that works with bullet would be brilliant. Things like best collision shapes, etc.
[11:44]  Marcus.Llewellyn Blender uses Bullet physics.
[11:44]  BlueWall.Slade I imagine that bullet is good as you can find on any opensource project.
[11:44]  Aine.Caoimhe ah
[11:45]  Aine.Caoimhe hadn't realized that Blender used bullet
[11:45]  Marcus.Llewellyn They write so much of it themselves, it's not unreasonable to assume they did the physics too. Hehe. But nope.
[11:47]  Arielle.Popstar some good info here
[11:47]  Robert.Adams the best collision shapes are cubes and spheres... these shapes are computed and not mesh comparisons
[11:48]  Marcus.Llewellyn There's other things I've wonders. If one has a large mesh surface, is a single collision shape for it best? Or multiple, smaller ones?
[11:48]  Robert.Adams the best solution in OpenSim is like Nebadon's test car.... the mesh body is phantom and there is a transparent, physical cube for the body and spheres for the wheels
[11:49]  BlueWall.Slade ++
[11:49]  Aine.Caoimhe yeah, that's what I'm now doing too
[11:49]  Marcus.Llewellyn I use cubes whenever possible for the colision shape. But some things need to interact with an avatar more granularly... if there's best practices for that, that's the sort of thing I'd love to learn more about.
[11:49]  Dahlia.Trimble you can upload a special collision shape when you upload a mesh
[11:49]  BlueWall.Slade ++
[11:50]  Robert.Adams if you are editing meshes (in Blender, for instance) you can define a simplified, physical mesh in a addition to the visible mesh
[11:50]  BlueWall.Slade I was thinking about that when working on terrain meshes
[11:50]  Robert.Adams BulletSim will use your physical mesh if specified
[11:50]  BlueWall.Slade that's great
[11:50]  Marcus.Llewellyn I do that all the time. I'm just not sure I'm provideing the best possible shapes.
[11:50]  Seth.Nygard I almost always create a very simplified physics mesh
[11:50]  BlueWall.Slade It is awesome that you have been working on those things.
[11:50]  Marcus.Llewellyn Just go for simple? Is that all we need to aim for?
[11:50]  Richardus.Raymaker it works very good, at least in SL making a mesh physics cube shape as example can drop Li. but.. its hard to know whats real and wahst artificial
[11:50]  Nebadon Izumi: hello everyone, sorry I couldnt make it
[11:51]  Arielle.Popstar Hi Neb
[11:51]  BlueWall.Slade Hey Nebadon
[11:51]  Seth.Nygard Hello Neb
[11:51]  Robert.Adams If a mesh doesn't define a convex mesh, BUlletSim estimates a convex hull meshes for the shape -- inefficient and approximate
[11:51]  Marcus.Llewellyn 40 demerits for House Izumi.
[11:51]  Justin Clark-Casey: hi neb
[11:51]  Aine.Caoimhe hi Neb
[11:52]  BlueWall.Slade I have seen large mesh landscape objects blocking an avatar - I gues that is it estimating the collision surface?
[11:52]  Dahlia.Trimble Hi neb, sleep in?
[11:52]  Robert.Adams did it take a long time to HG from your Pi sim?
[11:52]  BlueWall.Slade hahaha
[11:52]  Aine.Caoimhe I tried to go there the other day...wouldn't let me
[11:52]  Aine.Caoimhe maybe you had it down, though
[11:52]  Robert.Adams most likely explanation, BlueWall
[11:53]  Nebadon Izumi: heh no sorry I had to run out, and it took me a lot longer to get back then I expected
[11:53]  Dahlia.Trimble we tried to get native bulletsim to work on neb's pi but failed :/
[11:53]  BlueWall.Slade Ice storm in your neighborhood?
[11:53]  Nebadon Izumi: yea I have no idea really why it failed
[11:53]  Justin Clark-Casey: I'm impressed one can run opensim on a pi at all, at least with anything in it
[11:54]  Marcus.Llewellyn I'd love to see it running something heavy, like UC. The physics is neat, but how well it handles a lot of content would be interesting.
[11:54]  Richardus.Raymaker but dont hold a part on it justin
[11:54]  Billy.Bradshaw Microsoft will be releasing Win 10 for PI 2
[11:54]  Nebadon Izumi:
[11:54]  Robert.Adams even better, he has a virtual Pi in the world.... so we were standing on a Pi while running on a Pi
[11:55]  Justin Clark-Casey remembers not to crash his laptop yet again witha video link
[11:55]  Robert.Adams doesn't get more meta :)
[11:55]  Justin Clark-Casey: though tbh I think youtube is fine
[11:55]  Marcus.Llewellyn All it needs is a ARM emulator in LSL. >:)
[11:55]  Billy.Bradshaw
[11:55]  Justin Clark-Casey: I think you jmight have more luck creating an arm emulator in minecraft :)
[11:55]  Marcus.Llewellyn hehe
[11:56]  Nebadon Izumi: ya that Windows 10 for arm
[11:56]  Nebadon Izumi: is useles
[11:56]  Nebadon Izumi: its not a GUI version
[11:56]  Nebadon Izumi: and you cant run any apps
[11:56]  Nebadon Izumi: its pretty freakin lame
[11:56]  BlueWall.Slade MS DOS 6.0?
[11:56]  Nebadon Izumi: basically ya
[11:56]  Nebadon Izumi: its dos
[11:56]  Richardus.Raymaker no thats fine bluewall
[11:56]  BlueWall.Slade MS Linux .005
[11:57]  Marcus.Llewellyn Presumable it just supplies a subset of the windows API. Ssorta like running .NET command line apps.
[11:57]  Dahlia.Trimble could you run opensim on it?
[11:57]  Marcus.Llewellyn So neb is really better of with mono. heh
[11:57]  BlueWall.Slade ++
[11:57]  Richardus.Raymaker well, more you dont need for opensim. console
[11:58]  Nebadon Izumi: very doubtful you can do much of anything with Windows 10 on a raspberry pi
[11:58]  Dahlia.Trimble why? windows 10 can run tablets
[11:58]  Richardus.Raymaker if you not dead before its booted by coffee poiseing
[11:59]  Arielle.Popstar Opensim for Android
[11:59]  Billy.Bradshaw Mono have made some significant improvements to memory management this past few weeks.
[11:59]  Marcus.Llewellyn has anyone tried the Mono git with OpenSim lately?
[11:59]  Nebadon Izumi: yea I don't know, it sounded like it will have either no gui or a very limited GUI taht is more like power console only
[11:59]  Billy.Bradshaw I am running it now
[12:00]  BlueWall.Slade I hjave done no hard tests, but it seems to be running pretty solid.
[12:00]  Dahlia.Trimble Im using 3.99.something on my grid
[12:00]  Marcus.Llewellyn Awesome.
[12:00]  Billy.Bradshaw Git master as of 2 hours ago
[12:00]  Justin Clark-Casey: billy: Did you have to patch prebuild.xml to get it to build?
[12:00]  Billy.Bradshaw no
[12:00]  Nebadon Izumi:
[12:00]  BlueWall.Slade I'm just onXamarin releases. Maybe 3.12 or something.
[12:00]  Nebadon Izumi: ya i was running 3.12 on the pi too
[12:00]  Marcus.Llewellyn I tend to wait foor the official tarballs to build from those.
[12:00]  Billy.Bradshaw And Dahlia updated prebuild
[12:01]  Billy.Bradshaw so Opensim runs unadjusted
[12:01]  Dahlia.Trimble ya prebuild works with 3.99+ now
[12:01]  BlueWall.Slade ++
[12:01]  Marcus.Llewellyn Neato.
[12:01]  Nebadon Izumi: yea last i tried it worked
[12:01]  Justin Clark-Casey: oh ok, so one just needed to recompile Prebuild.exe? cool
[12:01]  Marcus.Llewellyn Any tangilble benefits?
[12:01]  Nebadon Izumi: ive mostly been running 3.10 or 3.12 now mysel which are release versions
[12:01]  Billy.Bradshaw I am seeing a 40% improvement in memory allocation
[12:01]  BlueWall.Slade awesome
[12:01]  Seth.Nygard 40% is significant
[12:01]  Dahlia.Trimble memory management much better
[12:01]  Justin Clark-Casey: how do you know it has improved?
[12:01]  Marcus.Llewellyn Zowie. That's excellent if reproducable.
[12:02]  Billy.Bradshaw I log meory usage
[12:02]  BlueWall.Slade :)
[12:02]  Justin Clark-Casey: process usage?
[12:02]  Billy.Bradshaw memory*
[12:02]  Seth.Nygard How is thread handlng? Any differences?
[12:02]  Dahlia.Trimble memory seems to even be better than 2.10.x
[12:02]  Dahlia.Trimble but way better than 3.x
[12:02]  Marcus.Llewellyn 3.2 was better than 2.10 for memory. Heh.
[12:02]  Dahlia.Trimble not for me
[12:02]  Billy.Bradshaw We are just beginning to profile using Dtrace, early days yet
[12:03]  Nebadon Izumi: i didnt see much change from 3.2 to maste git
[12:03]  Nebadon Izumi: it ran about the same for me
[12:03]  Nebadon Izumi: i didnt spend a lot of time testing the git master code for mono though
[12:03]  Nebadon Izumi: its very much in flux at the moment
[12:03]  Nebadon Izumi: likely to break from time to time
[12:04]  Dahlia.Trimble my grid uses about 8.5 gb on 3.99, it used almost 15gb on 3.12
[12:04]  Justin Clark-Casey: I'm interested in when somebody tries opensim on the microsoft clr under linux
[12:04]  Dahlia.Trimble about 9gb on 2.10
[12:04]  Marcus.Llewellyn I assumed that Mono git was pretty unstable with all the code they're moving into it.
[12:04]  Richardus.Raymaker Ok. i need to run.. :(
[12:04]  Richardus.Raymaker bye all
[12:04]  Justin Clark-Casey: bye richardus
[12:04]  Seth.Nygard Have a great evening Richardus
[12:04]  BlueWall.Slade bye RiRa
[12:04]  Dahlia.Trimble we all need to load test Nebadon's pi!!!
[12:04]  Marcus.Llewellyn I'm still fuzzy on whether there will *be* an official MS CLR for linux. I have the impression that MS is allowing Mono to be the official port.
[12:04]  Billy.Bradshaw Justincc, We will be trying that out in the next few weeks
[12:04]  Arielle.Popstar waves@riras
[12:05]  Justin Clark-Casey: marcus: I heard that their github repo builds on Ubuntu
[12:05]  Justin Clark-Casey: I don't get the snese they are relying on Mono
[12:05]  Nebadon Izumi: ya we could try teleporting to the pi
[12:05]  Richardus.Raymaker Oh justin, do you expect RC2 this week ? or just delay is a bit ?
[12:05]  Marcus.Llewellyn Woot! Good to know.
[12:05]  Nebadon Izumi: see if it explodes or melts
[12:05]  Nebadon Izumi: lol
[12:05]  Nebadon Izumi: OKC World 05
[12:05]  Justin Clark-Casey: rich: I can't say for sure. Having to work on some other bits atm, and those other bugs to address of course
[12:05]  Marcus.Llewellyn Next question then is how much of the framework will be available under that CLR.
[12:06]  Seth.Nygard I would expect the RCs to have he asset transfer issue with scripts
[12:06]  Nebadon Izumi: ya it doesnt even seem like mono is doing anything obvious from the microsoft code yet either
[12:06]  Billy.Bradshaw ned, if the mono passes there jenkins, it runs. We have see no issues with running Opensim
[12:06]  Nebadon Izumi: which is kind of surprising
[12:06]  Dahlia.Trimble tries to crash test the pi. bye all :)
[12:06]  Billy.Bradshaw their**
[12:06]  Nebadon Izumi: want to jump over?
[12:06]  Marcus.Llewellyn Pi load test? ;)
[12:06]  BlueWall.Slade ok, off to the pi
[12:06]  Aine.Caoimhe I'll give it a whirl
[12:06]  Nebadon Izumi: sure be interesting
[12:06]  Justin Clark-Casey: lol
[12:06]  Nebadon Izumi: see how well itholds up
[12:06]  Justin Clark-Casey: it's hg enabled?
[12:06]  Marcus.Llewellyn Woohoo
[12:06]  Nebadon Izumi: yes
[12:06]  Arielle.Popstar map search world 05
[12:06]  Nebadon Izumi: jump on over lets see
[12:06]  Justin Clark-Casey: alright, this should be itneresting
[12:06]  Seth.Nygard Have a great evening/night everyone, take care
Personal tools
About This Wiki