Chat log from the meeting on 2015-08-11

[10:55] Andrew Hellershanks: How things going, Kayaker?
[10:55] Simulator Version v0.5 ruft: OpenSim 0.8.2.0 Dev        0b5385f: 2015-08-11 11:55:51 +0300 (Unix/Mono)
[10:56] Kayaker Magic: Pretty good.
[10:56] Andrew Hellershanks: Haven't seen you in VH in quite a while.
[10:56] Kayaker Magic: I'm spending a lot of time on Kitely, planning on porting things from there to VH
[10:56] Andrew Hellershanks: oh, ok
[10:57] Andrew Hellershanks: That's not one of the grids I've been to.
[10:57] Kayaker Magic: It's the virtual worlds on demand place, they create your region on Amazon servers only when someone tries to go there.
[10:58] Kayaker Magic: They also can export products from their marketplace to any Hypergrid enabled grid
[10:58] Kayaker Magic: So my store there is doing well.
[10:59] Richardus.Raymaker @hypergrid.org:8002: kayayer, about what you talk ?
[10:59] Kayaker Magic: We were talking about the Kitely grid
[11:00] Andrew Hellershanks: Kayaker, That might be a reason to check out Kitely. My stores don't do well elsewhere. I just tend to avoid HG grids as I don't want people easily able to take my stuff else where and perhaps take things apart and duplicate my work.
[11:00] Kayaker Magic: I was paranoid about that at first, but I have relaxed.
[11:01] Alicia.Raven @grid.spellscape.co.uk: hi everyone
[11:01] Andrew Hellershanks: I'd want to know about the controls in place before I'd allow my stuff to be sold in an HG enabled grid.
[11:01] Kayaker Magic: Yes, someone can steal the stuff, but there are also a lot of people buying that are not stealing.
[11:01] Sarah Kline: I should close down my freebie store here and start selling stuff
[11:02] Kayaker Magic: When the export bit is fully implemented things will get a little better.
[11:02] Dahlia Trimble kommt in Chat-Reichweite (18.61 m).
[11:02] Andrew Hellershanks: I've got my particle generator. I wouldn't want people making their own copies of it for sale. I put a lot of work in to it.
[11:02] Shez Oyen: Kitley's market doesn't like my Paypal.. I use it all over the web but it won't work on Kitely.. I'm flumoxed
[11:02] Dahlia Trimble: hi
[11:03] Jak Daniels: hello all
[11:03] Shez Oyen: Hi Dahlia and everyone :)
[11:03] Andrew Hellershanks: hey, everyone.
[11:03] Aine.Caoimhe @refugegrid.com:8002 waves
[11:03] Andrew Hellershanks wonders if the cat has given up calling for me
[11:03] Richardus.Raymaker @hypergrid.org:8002: hi shez dahlia
[11:04] Shez Oyen: Rich :)
[11:04] Kayaker Magic: Dang, i forgot about lunch on the BBQ and burned the corn on the cob...
[11:04] Kayaker Magic: In RL
[11:04] Richardus.Raymaker @hypergrid.org:8002: Andrew, then the future is a problem for you
[11:05] Richardus.Raymaker @hypergrid.org:8002: Kayaker, burning nobs fine. but corn. Noooo
[11:05] Kayaker Magic: Let's just call this batch "carmelized". I think it is still edible.
[11:08] Andrew Hellershanks: Otherwise known as Cajun style. ;)
[11:10] Andrew Hellershanks: I've just tried pinging nebadon. Not sure if he is around.
[11:11] Alicia.Raven @grid.spellscape.co.uk: gets quiet with out nebadon here lol
[11:11] Kayaker Magic: Well, besides lunch and Hypergrid marketplaces, any OpenSim topics to talk about?
[11:12] Jak Daniels: Does anyone know if its possible to actually visualise (the mesh?) that the physics engine generates on a region for collisions?
[11:12] Andrew Hellershanks: Richardus, why would the future be a problem for me?
[11:12] Jak Daniels: I have a few all mesh terrains that sometimes seem to have invisible walls on or near the edges sometimes
[11:12] Sarah Kline: developer, render metadata physics shapes
[11:13] Dahlia Trimble: Jak, there used to be a way with ODE to visualize the colliders
[11:13] Richardus.Raymaker @hypergrid.org:8002: well, you keep having the same question that others mabye can sell your work.
[11:13] Aine.Caoimhe @refugegrid.com:8002: @Dahlia: I was a little surprised by Melanie's comment here: http://opensimulator.org/mantis/view.php?id=7278#c29133
[11:14] Aine.Caoimhe @refugegrid.com:8002: I was under the impression that having scripts recompile at sim start-up was generally a recommended thing unless there was a strong reason not to
[11:14] Jak Daniels: ah I asked my question too soon as Robert is now here and it is with bullet physics I want to visualise the collision data it creates
[11:14] Aine.Caoimhe @refugegrid.com:8002: but she seems to be saying the exact opposite
[11:14] Jak Daniels: hi robert btw
[11:14] Sarah Kline: go to your devloper menu
[11:14] Alicia.Raven @grid.spellscape.co.uk: deleting scripts at startup does not delete the state files, or i didnt think it did
[11:15] Sarah Kline: render metadata
[11:15] Shez Oyen: Robert :)
[11:15] Sarah Kline: then physics shapes
[11:15] Aine.Caoimhe @refugegrid.com:8002: yeah
[11:15] Aine.Caoimhe @refugegrid.com:8002: I'm confused
[11:15] Jak Daniels: is that advanced in replex?
[11:15] Dahlia Trimble: Aine, Melanie wrote most of XEngine so I guess she knows :)
[11:15] Jak Daniels: is that Advanced in replex?
[11:15] Robert Adams: hello all
[11:15] Alicia.Raven @grid.spellscape.co.uk: hi robert
[11:15] Jak Daniels: oops sorry for the double type
[11:16] Dahlia Trimble: Jak, I'm not aware of a way to visualize the colliders in bullet. If the viewer shows anything it's just what the viewer would thing is there in SL I guess, not actual collision geometry
[11:17] Richardus.Raymaker @hypergrid.org:8002: hi robert
[11:17] Dahlia Trimble: *would think
[11:18] Andrew Hellershanks: Richardus, I'm concerned about people being able to take items to grids lacking in control over exporting items.
[11:18] Sarah Kline: the geometry itself you would see in a preview when you uploaded it
[11:18] Dahlia Trimble: the way I mentioned for ODE used a separate opengl program
[11:18] Jak Daniels: ok thanks Dahlia. I want to visualise what the physics engine sees as colliders
[11:18] Alicia.Raven @grid.spellscape.co.uk: aine, im pretty sure state files are not deleted, if they were it would have cause problems for me at some point
[11:18] Jak Daniels: surely the preview is what the viewer thinks not bullet sim?
[11:18] Dahlia Trimble: yes its just a simulation
[11:18] Robert Adams: The colllsions would have to be made available.... collisions are per object and saved in collection between ticks
[11:19] Robert Adams: one could add an extended physics LSL function to fetch the collisions but there would have to be some accumulator added... what collides is not kept around for more than one tick
[11:19] Dahlia Trimble: I had a lot of difficulty when I wrote most of the collision geometry code. Most shapes were tested by just walking on them. Took me a *long time*
[11:19] Jak Daniels: so it doesn't 'render' then to an internal mesh or anything?
[11:20] Kayaker Magic: Andrew: I think you are worrying about a low probability event, when there is a high probability of many more grids and people out there who would use your products without stealing them.
[11:20] Robert Adams: the Bullet physics simulation reports a collision at a point between two objects
[11:20] Jak Daniels: so a large mesh terrain's collision zone is recalcualted every time or just once on sim start?
[11:20] Andrew Hellershanks: Aine, hm... that comment is a little odd. I would have to look at the code. I'm surprised if it delete state files. I would only expect the compiled code files to be deleted on start up.
[11:21] Robert Adams: BulletSim gets a list of (obj`, obj2, collisionPorint, penetrationDepth)
[11:21] Jak Daniels: excuse my poor understanding of it lol :)
[11:21] Aine.Caoimhe @refugegrid.com:8002: it's also set to TRUE by refulat
[11:21] Aine.Caoimhe @refugegrid.com:8002: which seems odd
[11:21] Robert Adams: there is caching of the collisions inside Bullet so it doesn't start from scratch each tick
[11:22] Aine.Caoimhe @refugegrid.com:8002: why set the default of something to the setting that you high recommend not using?
[11:22] Richardus.Raymaker @hypergrid.org:8002: i notiched that sometimes objects get a solid bounding box. instead of shaped
[11:22] Jak Daniels: so a mesh ground tile 32x32m is one object in the cache?
[11:23] Andrew Hellershanks: Aine, DeleteScriptsOnStartup defaults to false.
[11:23] Robert Adams: I thought the problem was in a grid accepting a pre-compiled script -- none of the sanity or saftey checking done by the compiler would be in place when a grid accepted a binary script
[11:24] Robert Adams: a bounding box is created by default... if BulletSim can mesh the object, the mesh is used (whether prim or actual mesh)
[11:24] Andrew Hellershanks: I tell the mesh import process to use physics shape from object.
[11:24] Robert Adams: if it doesn't mesh, it stays a bounding bo
[11:24] Nebadon Izumi: hello everyone sorry runnig late today
[11:24] Jak Daniels: ok so if it meshes can it be somehow exported and visualized?
[11:25] Shez Oyen: Neb :)
[11:25] Robert Adams: if the physics shape mesh exist, that should be used by BulletSm
[11:25] Robert Adams: hello Nebadon
[11:25] Dahlia Trimble: hi Nebadon
[11:25] Nebadon Izumi: hello :)
[11:26] Alicia.Raven @grid.spellscape.co.uk: hi neb :)
[11:26] Robert Adams: the mesher in OpenSim just creates a shape mesh... it doesn't have all the prim information (faces, ...)
[11:26] Jak Daniels: ok, thanks Robert
[11:26] Dahlia Trimble: the collision mesh is just a single shape, yes
[11:26] Robert Adams: The mesher in libomv (MeshmerizerR) will output a displayable prim mesh
[11:27] Dahlia Trimble: the code that makes collision shapes for prims is the same code that makes prims for display in Radegast
[11:27] Jak Daniels: ah ok, that's where I'll look then :) I'll do some code digging
[11:27] Dahlia Trimble: so if you want to see what a collision shape for a prim looks like, look at the prim in radegast
[11:28] Dahlia Trimble: meshes are handled differently
[11:28] Dahlia Trimble: a mesh assset can have a separate physics collider that may not look at all like the display mesh
[11:28] Dahlia Trimble: that collider could be a mesh or it could be "convex hulls"
[11:29] Dahlia Trimble: if none exists, opensim will attempt to use a display mesh for physics
[11:29] Dahlia Trimble: but may fail, and you get a box
[11:29] Jak Daniels: ok, now I understand it better. thank you
[11:29] Dahlia Trimble: all of that code is in Meshmerizer.cs
[11:30] Richardus.Raymaker @hypergrid.org:8002: Dahlia, in secondlife you need to use many times prim mode too for a good collision shape
[11:30] Richardus.Raymaker @hypergrid.org:8002: conmvex hull is bad
[11:31] Dahlia Trimble: Richardus, I can't say Ive tried much in SL
[11:31] Alicia.Raven @grid.spellscape.co.uk: nice outfit nebadon :)
[11:31] Nebadon Izumi: do we even support convex hull here?
[11:31] Robert Adams: BulletSim has a software convex hull approximator builder...if a mesh is made physical, it is turned into a set of convex hulls
[11:32] Dahlia Trimble: when I was doing the mesh colliers for opensim, I tried to make them better than SL :) Not sure I did but I did try :)
[11:32] Nebadon Izumi: Robert I meant if we select Convex hull as the physics type
[11:32] Nebadon Izumi: does that actually do anything?
[11:32] Richardus.Raymaker @hypergrid.org:8002: Dahlia, sofar i know opensim works with collision shapes better then Secondlife
[11:32] Robert Adams: if a physics mesh is supplied as part of the mesh, that collection of hulls is uesed
[11:33] Robert Adams: an authored physical object usually has a display mesh (high detail) and a physical mesh (lower detail for faster physics computation)
[11:34] Dahlia Trimble: that would be preferred
[11:34] Robert Adams: Nebadon simulates that by creating cars with a phantom mesh exterior and a physical, transparent cube for the body shape
[11:34] Robert Adams: this makes for good looking cars that a quicker at physical simulation
[11:35] stiofain nbmcmedia: 0000
[11:35] Robert Adams: use spheres for wheels (transparent) as they are quick to compute collisions for
[11:35] stiofain nbmcmedia: hi folks
[11:35] Dahlia Trimble: hi
[11:36] Nebadon Izumi: yea set the wheels to Glass too
[11:36] Andrew Hellershanks: Why glass?
[11:37] Aine.Caoimhe @refugegrid.com:8002: they brake better? :p
[11:37] Nebadon Izumi: less friction
[11:37] Richardus.Raymaker @hypergrid.org:8002: WHat aine say, perfect brakes
[11:37] Alicia.Raven @grid.spellscape.co.uk: you can reduce friction more by reducing the restitution value
[11:37] Andrew Hellershanks: The materials options offer different amounts of friction?
[11:37] Nebadon Izumi: you could probably experiment with the different types I dont think there is huge variations
[11:37] Dahlia Trimble: those who drive cars with glass wheels should no drive over stones?
[11:38] Nebadon Izumi: haha
[11:38] Andrew Hellershanks laughs
[11:38] Robert Adams: reducing friction (by changing material or using the script functions) works
[11:39] Robert Adams: also can reduce gravity that applies to an object
[11:39] Nebadon Izumi: yea I have done that in motorcycles
[11:39] Nebadon Izumi: used bouyancy to keep the motorcycle upright
[11:39] Nebadon Izumi: not a great solution
[11:39] Nebadon Izumi: hehe
[11:40] Alicia.Raven @grid.spellscape.co.uk: not if they float away lol
[11:40] Nebadon Izumi: well it wasnt enough to float
[11:40] Nebadon Izumi: but enough to keep it upright
[11:40] Robert Adams: there is a vehicle feature to keep motorcycles upright (with leaning when turning)
[11:40] Nebadon Izumi: be neat if we had some kind of gyroscopic stabilizer function
[11:41] Nebadon Izumi: ah nice
[11:41] Andrew Hellershanks: The modern day equivalent of training wheels on a bike?
[11:41] Nebadon Izumi: does that work Robert?
[11:41] Nebadon Izumi: i did this a long time ago
[11:41] Robert Adams: I remember debugging it :)
[11:41] Robert Adams: if it doesn't work, I'll take another look
[11:43] Andrew Hellershanks: I was talking with the person who does tech support for a grid. They were having trouble with Open Simulator sending out email. They kept getting authorization issues.
[11:44] Nebadon Izumi: are they using SSL?
[11:44] Richardus.Raymaker @hypergrid.org:8002: Possible because opensim and many other porograms dont support pop before smtp
[11:44] Andrew Hellershanks: I did some looking in to it and the problem is that the SMTP mail code used in Open Simulator is ancient. It doesn't support a lot of the modern authentication protocols such as SSL, and TLS.
[11:44] Nebadon Izumi: I had trouble with SSL in newer mono versions
[11:44] Nebadon Izumi: on mono 4.x
[11:45] Andrew Hellershanks: Nebadon, not sure what they have on their mail server.
[11:45] Nebadon Izumi: we do support ssl for SMTP
[11:45] Nebadon Izumi: because thats the only way you can use gmail
[11:45] Nebadon Izumi: but newer mono it doesn work
[11:45] Nebadon Izumi: I had to revert back to mono 3.12
[11:45] Andrew Hellershanks: OS is using DotNetMail (or something like that). Its ancient. My research says it was made when there was no mail support in .NET but that is no longer the case.
[11:46] Andrew Hellershanks: There is a mail class in .NET now that handles some of the newer authentication methods.
[11:47] Alicia.Raven @grid.spellscape.co.uk: when i was messing with the email module i didnt even think to check if .net had any classes, interesting news
[11:47] Nebadon Izumi: would be good to modernize that stuff if we can
[11:47] Andrew Hellershanks: I was planning to file a mantis about this issue. I think the existing mail class needs to be replaced.
[11:47] Nebadon Izumi: the only time i really use mail is with WiFi addon
[11:47] Nebadon Izumi: never setup the llEmail stuff
[11:48] Andrew Hellershanks: The .NET one is listed as available in .NET 3.5 and 4.0 at least so that shouldn't cause any problems for OS as we now require one of those versions.
[11:48] Nebadon Izumi: yea question is does it work in mono
[11:48] Andrew Hellershanks: I have a grid where we support sending forwarding IMs to offline users via email.
[11:49] Robert Adams: maybe we can talk Diva to make the email change since it's used by the wifi installations
[11:49] Andrew Hellershanks: nebadon, Would have to check but the mail class has been in .NET before 3.5 so one would hope mono picked it up already.
[11:49] Alicia.Raven @grid.spellscape.co.uk: we have offline im to email but we run an SMTP server so have full control there for security
[11:50] Andrew Hellershanks nods
[11:51] Andrew Hellershanks: I gave him ideas on how to work around the issue such as having an SMTP server to only receive from their grid servers which could pass mail on to their primary mail server for deliver.
[11:51] Alicia.Raven @grid.spellscape.co.uk: running a mail server is a nightmare, i hope i never have to touch the configs again lol
[11:51] Andrew Hellershanks: Nebadon, which mono is the minimum suggested/required for OS these days?
[11:52] Andrew Hellershanks: Alicia, if it is Sendmail based, I don't blame you.
[11:52] Alicia.Raven @grid.spellscape.co.uk: postfit
[11:52] Jak Daniels: I use exim ;)
[11:52] Alicia.Raven @grid.spellscape.co.uk: postfix*
[11:52] Andrew Hellershanks: For a while I had been using Qmail (mostly on my own machine)>
[11:53] Jak Daniels: mono 2.10 I thought Andrew
[11:53] Andrew Hellershanks: ok. I'll check on the mail support in that version of mono.
[11:54] Jak Daniels: but you'd be better with 3.10/3.12 I think, but 2.10 is still the baseline irc
[11:55] Andrew Hellershanks: yes, I just wanted to make sure support for the newer mail class is in the lowest level of mono used with OS.
[11:55] Dahlia Trimble: I always had good luck with 2.10.8.1
[11:55] Dahlia Trimble: using 3.99 now, is good so far
[11:55] Andrew Hellershanks: I was using that but I'm on 3 something now.
[11:55] Simulator Version v0.5 ruft: OpenSim 0.8.2.0 Dev        0b5385f: 2015-08-11 11:55:51 +0300 (Unix/Mono)
[11:56] Andrew Hellershanks: 3.2.8 is what I have now.
[11:56] Dahlia Trimble: most 3.x I had trouble with, often memory trouble
... to be  continued later ...