Chat log from the meeting on 2015-04-14

From OpenSimulator

Jump to: navigation, search
[11:03]  Nebs Metal Bar Stool v1.5 (w/sit & launch): Hello Nebadon Izumi, enjoy your sit..
[11:03]  Nebadon Izumi: hello :)
[11:03]  Alicia.Raven hi nebadon
[11:03]  Sheera.Khan well, alternating ...
[11:03]  Shez Oyen: Hi Neb :)
[11:03]  Jak Daniels: hi Neb
[11:03]  Richardus.Raymaker hi neb
[11:03]  Sheera.Khan sometimes yes, sometimes no ...
[11:03]  Lani Global: Greetings, Gridlings!
[11:03]  Sheera.Khan Hi Neb
[11:03]  Richardus.Raymaker hi lani
[11:03]  Aine.Caoimhe hu UsuerUMMAU42
[11:03]  Lexie Bushy: hello everyone
[11:03]  Richardus.Raymaker see, the ansere is 42
[11:04]  Alicia.Raven user UMMAU42?? see
[11:04]  Alicia.Raven lol
[11:04]  Nebadon Izumi: ya i saw that the other day myself
[11:04]  Lani Global: it is rampant
[11:04]  Robert Adams: hey all!
[11:04]  Richardus.Raymaker hi robert
[11:04]  Alicia.Raven that name is a placeholder for the user managment cache that freaky added last month
[11:04]  Aine.Caoimhe hi Robert
[11:04]  Lexie Bushy: anyone awake
[11:05]  Alicia.Raven hi robert
[11:05]  Richardus.Raymaker i wish diva where around, i guess she knows the answere
[11:05]  Shez Oyen: Hi Lexie :)
[11:05]  Nebadon Izumi: answer for what Richardus?
[11:05]  Lexie Bushy: hi shez
[11:05]  Richardus.Raymaker My offline V2 IM HG dont get send to the avatar. it's stored in the offlineIM table.
[11:05]  Nebadon Izumi: oh
[11:06]  Nebadon Izumi: this is in grid mode?
[11:06]  Shez Oyen: Hi Allen :)
[11:06]  Lexie Bushy: anyone have any skin's ext that i can have please
[11:06]  Aine.Caoimhe applauds Robert's work this week on OSSL ini settings
[11:06]  Simulator Version v0.5 shouts: OpenSim Dev        d96d31b: 2015-04-12 19:32:06 -0700 (Unix/Mono)
[11:06]  Richardus.Raymaker Right now redoing that system, but the json everything get stored. but how do you send it to the avatar when you login ? i expect that need to be done automatic with V2 without extra's
[11:06]  Nebadon Izumi: oh yea Robert about that
[11:06]  Richardus.Raymaker yes robust mode nebadon
[11:07]  Nebadon Izumi: when i updated and restarted i got this >
[11:07]  Andrew Hellershanks: The OSSL ini would be like OpenSimDefaults.ini and is not meant to be changed so you would use another file to customize the defaults?
[11:07]  Lani Global: good work, Robert!
[11:07]  Richardus.Raymaker hi andrew.
[11:07]  Andrew Hellershanks: hey, Richardus
[11:07]  Andrew Hellershanks: Hey, everyone
[11:07]  Alicia.Raven hi andrew
[11:08]  Aine.Caoimhe no, it's more like a gridcommon.ini type of file as I understand it
[11:08]  Richardus.Raymaker spammy neb
[11:08]  Aine.Caoimhe you'd edit it directly
[11:08]  Nebadon Izumi: Richardus, make sure you have the URL in Robust set correctly
[11:08]  Nebadon Izumi: SRV_IMServerURI
[11:08]  Andrew Hellershanks: As soon as you upgrade your OS the file would get clobbered and you would lose all your hanges
[11:08]  Andrew Hellershanks: s/hanges/changes/
[11:08]  Robert Adams: LOL Nebadon... one for each regions!!!
[11:08]  Richardus.Raymaker Nebadon, but nobody seems to know why offlineIM V2 dont work
[11:08]  Richardus.Raymaker lol rpobert.
[11:09]  Nebadon Izumi: no its a single region Robert
[11:09]  Dahlia.Trimble hi
[11:09]  Robert Adams: people were concerned about region owners knowing :)
[11:09]  Nebadon Izumi: a single 256x256 region
[11:09]  Nebadon Izumi: it happened when I loaded an OAR and also everytime i start the region
[11:09]  Shez Oyen: Hi Dahlia :)
[11:09]  Aine.Caoimhe with some messages it would be really handy to know *which* region it is
[11:09]  Richardus.Raymaker hi dahlia
[11:09]  Alicia.Raven hi dahlia
[11:10]  Aine.Caoimhe like "unable to find texture xxx for object yyyy.....if you don't happen to know what region object YYY is in it can be a PITA
[11:10]  Lani Global: Would osGetPhysicsEngineType() and osGetPhysicsEngineName() be allowed everywhere by default for everyone?
[11:10]  Robert Adams: oh... that could be bad.... might be one for each script
[11:10]  Robert Adams: yes Lani
[11:10]  Lani Global: wonderful
[11:10]  Nebadon Izumi: yea it seemd a bit excessive :)
[11:12]  vegaslon plutonian: ya limit to only the first 30
[11:12]  Robert Adams: I'll comment it out for the moment.... there should be a better place to put the warning
[11:13]  Nebadon Izumi: ok cool thanks
[11:13]  Robert Adams: at the moment all the OSSL stuff is embedded in XEngine but really it should be LSL centric (no matter which script engine)
[11:13]  Andrew Hellershanks: what is the difference between the get type and get name? One is number and the other a string?
[11:15]  Robert Adams: no, type is the name specified in the 'physics=' parameter in the [Startup] INI area
[11:15]  Lexie Bushy: 01
[11:15]  Alicia.Raven if osGetPhysicsEngineType is available then why is the new osCheckODE needed?
[11:15]  Lani Global: it isnt
[11:15]  Andrew Hellershanks: Alicia, I was wondering that same thing
[11:16]  Robert Adams: name is the name from the DLL selected.... for BUlletSim, there were different versions planned for for different hardware, etc
[11:16]  Robert Adams: thechnically not Alicia
[11:16]  Alicia.Raven what melanie said about all physics engines should respond the same, just seems wrong to me, they all have differences
[11:16]  Lani Global: phys engine version number may become important at some point
[11:17]  Andrew Hellershanks: Is "physics" just specifying a type of engine? I thought it was really the name of it.
[11:17]  Robert Adams: but the idea is that scripts should be like mentioned in the email -- not BUlletSim specific but SL compatibale with some checks for legacy ODE operation
[11:17]  Dahlia.Trimble I dont agree with it either Alicia, SL couldnt even make different Havok versions act the same
[11:17]  Lani Global: What if some particular aspect of a physics simulation engine is not exactly the same as current SL Havok?
[11:17]  Alicia.Raven if bullet is supposed to work the same as SL then it needs alot more work
[11:17]  Lani Global: Shall OpenSim Physics Engines be allowed to have new additional beneficial features that are unique, not canon?
[11:18]  Andrew Hellershanks: I could see the need for a get version of the physics engine
[11:18]  Nebadon Izumi: Bullet does need a lot more work
[11:18]  Nebadon Izumi: heh
[11:18]  Nebadon Izumi: I dont really forsee another engine anytime soon either
[11:18]  Alicia.Raven but i still dont think it can be made to match SL exactly as melanie seems to expect
[11:18]  Andrew Hellershanks: Perhaps you have some script that only works if a given version of a particular engine would let the script work properly
[11:18]  Lani Global: Would all OpenSim Physics Engines be forced into lock-step with the most recent SL implementation of Havok?
[11:18]  Nebadon Izumi: MOSES is working on physx but its not gauranteed they will actually go the whole way or that it will end up in opensim
[11:18]  Dahlia.Trimble theres some things about bulletsim that I like better than SL havok... like a torus chain :)
[11:18]  Robert Adams: if BulletSim doesn't do it "right", it is supposed to get fixed
[11:18]  OtakuMegane Desu: Opensim already doesn't use the full potential of physics engines. What real point would there be to switching again.
[11:19]  Robert Adams: I am currently just giving tools to the scriptors to do what they need to do
[11:19]  Nebadon Izumi: ya I am not convinced physx is a better solution either honestly
[11:19]  Lani Global: we like those choices, Robert!
[11:19]  Dahlia.Trimble I also don't agree that perfectly emulating SL is "right"
[11:19]  Lani Global: thanks for the tools!
[11:19]  Andrew Hellershanks: DOesn't PhysX need support from the graphics card?
[11:19]  Robert Adams: if they make scripts that only act in one place for one script engine, that is their choice
[11:19]  Nebadon Izumi: well we dont have to emulate SL exactly
[11:20]  Nebadon Izumi: but LSL is a specific language
[11:20]  Nebadon Izumi: we should not be extending
[11:20]  Robert Adams: but now they have enough rope to hang themselves with compatibility problems :)
[11:20]  Nebadon Izumi: our scripting language is OSSL
[11:20]  Dahlia.Trimble we dont extend LSL, thats what os functions are for
[11:20]  Nebadon Izumi: right
[11:20]  Nebadon Izumi: I think the argument here is to extend a LSL function
[11:20]  Nebadon Izumi: to return physics engine
[11:20]  Nebadon Izumi: not going to happen
[11:21]  Robert Adams: it is easy to add modules to OpenSim to add new scripting functions
[11:21]  Lani Global: Physics engines will always have nuances and differences. Some new engine may have a new feature.
[11:21]  Robert Adams: I agree Lani
[11:21]  Nebadon Izumi: we just need to make sure everyone can run that OSSL function is all
[11:21]  Lani Global: OpenSim is modular.
[11:21]  Nebadon Izumi: either way people are going to have to update
[11:21]  Alicia.Raven not what i was saying, i just mean that osCheckODE isnt needed because osGetPhysicsEngineType does the job better
[11:21]  Nebadon Izumi: yea
[11:22]  Allen Kerensky: simply put - extending llGetEnv to return opensim-specifics is EASY and 100% SL forward compatible - so its purely religion and politics if you want to go another route
[11:22]  Dahlia.Trimble there may also be some implicit way to determine physics engine
[11:22]  Allen Kerensky: the osCheckODE is one way
[11:22]  Nebadon Izumi: there was still debate at the time as to turning on OSSL for everyone
[11:22]  Robert Adams: true, Alicia... you now have both
[11:22]  Nebadon Izumi: now we are past that
[11:22]  Dahlia.Trimble such as some unique physical behavior
[11:22]  Allen Kerensky: or opening the osGetPhysics another
[11:22]  Lani Global: osCheckODE, osCheckBullet, osCheckHavok, osCheckNewPhys, osCheckNinja, osCheckUndocumented...
[11:22]  Nebadon Izumi: the ode check was to resolve not turning on OSSL for everyone
[11:22]  Allen Kerensky: but the OSSL requires an enable step that is a significant change and people are throwing more religion and politics
[11:22]  Nebadon Izumi: not anymore
[11:23]  Nebadon Izumi: in core master code anyway
[11:23]  Nebadon Izumi: no releases yet, but that will come
[11:23]  Nebadon Izumi: either way it would require updating
[11:23]  Allen Kerensky: and it uncovered a new mandatory requirement no one knew in advance, all physics must be at least SL/Bullet compatible from a scripting viewpoint
[11:23]  Dahlia.Trimble there are also forks which wont have any of those checks
[11:23]  Dahlia.Trimble like inworldz
[11:23]  Lani Global: the LSL llGetEnv proposal was quick and easy. but it has snowballed into a massive OSSL thing.
[11:23]  OtakuMegane Desu: I dunno why you really need more than one (maybe 2) OS functions for determining physics. One to determine the engine, one perhaps to determine features, etc.
[11:23]  Andrew Hellershanks: I don't have a problem with the ossl config changes. Just that I'm concerned I'll lose any customizations when I update to the next version of grid code since I'll be modifying a file that is shipped with OS
[11:23]  Dahlia.Trimble and I think IW uses physx anyway
[11:24]  Nebadon Izumi: yea IW is very different
[11:24]  Allen Kerensky: and with osCheckODE - bullet scripts will still fail on IW or other physx platforms
[11:24]  Nebadon Izumi: Andrew the old methods still work
[11:24]  Allen Kerensky: so - now, if anything ever gets merged to core from physx, it will need cleanup
[11:24]  Robert Adams: llGetEnv would be easy to program but it really was the same arguements that were needed to get ossl fuctions on in general
[11:24]  Nebadon Izumi: anything you do in OpenSim.ini over rides the defaults
[11:24]  Allen Kerensky: and any physx scripts will need to change
[11:24]  Lani Global: I'm faced with designing products for the entire metaverse.... ODE, Bullet, etc
[11:24]  Robert Adams: getting ossl turned on was a bigger solution to the problem
[11:24]  Allen Kerensky: and they won't - jus tlike ODE - nothing solved
[11:25]  Andrew Hellershanks: nebadon, yes but OS.ini is shipped as os.ini.example so no issue of an override
[11:25]  Dahlia.Trimble would be nice to have a way that didnt require a function that may not exist
[11:25]  Andrew Hellershanks: err.. no issue of it getting clobbered during an update
[11:25]  Nebadon Izumi: yea even with all of these changes
[11:25]  Dahlia.Trimble so your script doesnt crash when you call a nonexistent function
[11:25]  Nebadon Izumi: a huge swath of the metaverse runs old code
[11:25]  Allen Kerensky: and Melanie's position ignore people writing scripts that might get reused in Moses or other places
[11:25]  Nebadon Izumi: getting people to update is far more difficult than fixing this problme
[11:26]  Allen Kerensky: so - we're still Catch-22 it seems
[11:26]  Richardus.Raymaker i dont like ODE too
[11:26]  Nebadon Izumi: ODE is fine, its not being developed
[11:26]  Nebadon Izumi: ODE itself is pretty dead as a project
[11:26]  Robert Adams: there are a lot of people running old versions of OS... that will always be a problem
[11:26]  Dahlia.Trimble Im not against some creative tweaks to llGetEnv
[11:26]  Lani Global: the main reason i proposed the use of LSL extension for llGetEnv was that it would not fail, even in legacy grids and sims
[11:26]  OtakuMegane Desu: Some of the reason for old code hanging around is so much ends up in the air or changing . Once you have something more solid at least some of that I'm sure will be updated.
[11:26]  Nebadon Izumi: and none of the opensim devs or volunteers are really extending it
[11:26]  Allen Kerensky: the last commit to the ODE upstream was over a year ago - anyone updating opensim past that is running their own risk
[11:27]  Robert Adams: the function osGetPhysicsEngineType() does not throw and exception if ossl functions are not enabled... it just returns a zero length string
[11:27]  Lani Global: so, there is no risk in calling that function?
[11:27]  Dahlia.Trimble Robert, on SL?
[11:27]  Richardus.Raymaker nebadon. just other question. do you know if there's betetr option for offlineIM V2 debugging. not found anything
[11:27]  Dahlia.Trimble wont even compile on SL I bet
[11:28]  Robert Adams: true Dahlia... wouldn't work on SL
[11:28]  Andrew Hellershanks: Any script using that function on an older grid (that doesn't have the function) will find that the script won't compile
[11:28]  Lani Global: as i understood it, IF call ossl function without ossl enabled in region, script fail. is that not the case?
[11:28]  Nebadon Izumi: just that one specific function
[11:28]  Nebadon Izumi: others will
[11:29]  Richardus.Raymaker Only the error message could be improved and intercepted instead of spamming a bunch of .net error code lines
[11:29]  Dahlia.Trimble pretty sure the console gets spammed on some ossl functions that arent enabled
[11:29]  Robert Adams: that was my first attempt at adding some script testing
[11:29]  Lani Global: so, i guess i won't close the mantis for llGetEnv extension yet :)
[11:29]  Primitive: Script running
[11:29]  Script saved
[11:29]  Primitive: 
[11:30]  Dahlia.Trimble or maybe its mod functions that spam the console
[11:30]  Lani Global gave you llGetEnv  for OpenSim v1.4.
[11:30]  llGetEnv  for OpenSim v1.4: Rezzed!
[11:30]  llGetEnv  for OpenSim v1.4: SL description of this function:
[11:30]  llGetEnv  for OpenSim v1.4: in Local Chat channel 0, Upon touch, returns a string with the requested data about the region.
[11:30]  llGetEnv: dynamic_pathfinding = 101
[11:30]  llGetEnv: estate_id = 0
[11:30]  llGetEnv: frame_number = 733876
[11:30]  llGetEnv: region_idle = 0
[11:30]  llGetEnv: sim_channel = OpenSim
[11:30]  llGetEnv: sim_version = OpenSim Dev        d96d31b: 2015-04-12 19:32:06 -0700 (Unix/Mono)
[11:30]  llGetEnv: opensim_physics = 
[11:30]  Lani Global: tester
[11:31]  Lani Global: no fail problem in opensim or sl
[11:31]  Dahlia.Trimble sim_channel could say opensim-bullet
[11:31]  Primitive: 
[11:32]  Script saved
[11:32]  Andrew Hellershanks: sim_channel? Why "channel"? Makes me think of some port number would be there.
[11:32]  Dahlia.Trimble LL did that
[11:32]  Andrew Hellershanks: not complaining. Just ciurious
[11:32]  OtakuMegane Desu: lol
[11:32]  Andrew Hellershanks: oh.
[11:32]  Allen Kerensky: that's for the SL server development channel names - magnum, letigre, etc if I remember right
[11:32]  Andrew Hellershanks: of course they would. :P
[11:33]  Andrew Hellershanks: oh,we would call it a grid. They call it channel. ok
[11:33]  Robert Adams: not crazy about changing llGetEnv... anything done there would need the same permission tests as the ossl funtions (extensions to base LSL)
[11:33]  Allen Kerensky: well, the run 1/4th of the main grid on 1 of 4 server software channels - release, or one of 3 dev versions
[11:33]  Robert Adams: and if ossl functionality is available, there are the other functions
[11:33]  Allen Kerensky: so they can rotate regions like farm patches, testing 3 different batches of code in production
[11:33]  Dahlia.Trimble I dont see why it would need permission tests
[11:34]  Lani Global: llGetEnv test unit script is online
[11:34]  Robert Adams: people who don't want the ossl functions, don't want additions to the LSL language.... they want SL compatibility
[11:34]  Lani Global: there is one "greater issue" that we may want to explore, on the discovery of physics engine issue.
[11:35]  Allen Kerensky: well, I like having OpenSim specifics in OSSL calls - but something fundamental like physics engine type needs to always be available for scripters unless a 100% compatibility guarantee is offered by core devs
[11:35]  Lani Global: does the Viewer client need that Physics Engine info?
[11:35]  Allen Kerensky: API stability, or API flexibility through adaptive scripting
[11:35]  Robert Adams: and I don't know why you want llGetEnv with all the other ways of now doing the same thing
[11:35]  Dahlia.Trimble I dont consider it an adition, its more a correction :)
[11:35]  OtakuMegane Desu: At this point, SL compatibility shouldn't be a primary concern, unless it causes a major viewer issue. SL is basically EOL'd in its current form, so why have it hold things back?
[11:35]  Lani Global: would it benefit the Viewer client to know Physics Engine? without some kind of OSSL bridge?
[11:36]  Richardus.Raymaker add new functionality to lsl is asking for problems and confusion. because you get document conflicts. not to mentoin what happens if soemone try it n SL :O
[11:36]  Robert Adams: not that I know of Lani... but a viewer person would have to answer that
[11:36]  Nebadon Izumi: I dont see why the viewer needs to know about physics
[11:36]  Nebadon Izumi: viewer is just rendering data its sent
[11:36]  Allen Kerensky: well Lani was going with a suggestion to try to get something on the wall and that was the easiest way at the time
[11:36]  Nebadon Izumi: it has no concept of physics really
[11:36]  Robert Adams: there are few ways of sending that info... could do it in the login packet (like region size) if it was usable by a viewer
[11:37]  Allen Kerensky: the SL viewer offloads some stuff to a havok lib
[11:37]  Allen Kerensky: but that got paralleled for opensim with the hacd library I believe
[11:37]  Nebadon Izumi: ya thats for uploading only though
[11:38]  Nebadon Izumi: not rendering physics
[11:38]  OtakuMegane Desu: But we don't and won't have Havok so that seems like a non-issue for this anyway.
[11:38]  Allen Kerensky: navmesh stuff in SL may use the havoklib as well, I dunno
[11:38]  Nebadon Izumi: yes it does
[11:38]  Allen Kerensky: yeah exactly Otaku
[11:38]  Nebadon Izumi: the only viewer phsysics is jiggly boobs
[11:38]  Nebadon Izumi: hehe
[11:38]  Dahlia.Trimble it uses it to display the navmesh
[11:38]  Robert Adams: there is all the path following stuff... but I don't know about how to hook into any of that
[11:38]  Dahlia.Trimble but I know a way to display a navmesh without it
[11:39]  Allen Kerensky: OpenSim isn't SL compatible simply for that fact - but Robert has made it possible to get out of the box SL vehicle scripts to work well
[11:39]  Allen Kerensky: so Melanie's point was "if Bullet can - then everyone else should too"
[11:39]  Robert Adams: there is a lot of work needed
[11:39]  Lani Global: impressive how well bullet vehicles perform now.
[11:39]  Allen Kerensky: but no one loves ODE from a coding perspective so its the neglected stepchild here
[11:39]  Andrew Hellershanks: Certain types of vehicles would benefit greatly from being able to use path following. Will be nice when that is implemented.
[11:39]  OtakuMegane Desu: Honestly the viewer shouldn't have to deal with anything physics, that really is best server-side.
[11:39]  Robert Adams: but I want OpenSim to get out of its comfort zone and blaze new trails
[11:40]  OtakuMegane Desu: Agreed, Robert :)
[11:40]  Lani Global: OpenSim is all about Freedom.
[11:40]  Allen Kerensky: so maybe physics in OSSL is wrong - and it should be in a Bullet-specific scripting setup
[11:40]  Robert Adams: I have to leave early... I have to run
[11:40]  Allen Kerensky: and the same for ODE
[11:40]  Andrew Hellershanks: LL won't be doing a lot more innovating on SL soon as they will be focusing on SL 2
[11:40]  Robert Adams: I will check in changes for that spam message soon, Nebadon
[11:40]  Dahlia.Trimble displaying a navmesh in an unmodified viewer without the havok lib
[11:41]  Robert Adams: bye all
[11:41]  OtakuMegane Desu: Yeah. LL has their shiny new project and given their history, they're not going to do more than bare minimum with the current SL now.
[11:41]  Alicia.Raven bye robert
[11:41]  Richardus.Raymaker bye robert
[11:41]  Dahlia.Trimble bye robert
[11:41]  OtakuMegane Desu: Bye
[11:41]  Andrew Hellershanks: See you next week, Robert
[11:41]  Nebadon Izumi: ok thanks Robert
[11:41]  vegaslon plutonian: ah ya just make the mesh in the server and send it out
[11:42]  Dahlia.Trimble vegaslan, ya thats what I did :)
[11:42]  vegaslon plutonian: always would work to make some custom terrains
[11:42]  Allen Kerensky: oh well, I am not sure anything is really resolved for ODE addicts - they still have to change scripts to see if ODE is there somehow and I think that was the biggest objection
[11:42]  Dahlia.Trimble it does that too :)
[11:43]  Dahlia.Trimble I think kitely uses ODE
[11:43]  Dahlia.Trimble maybe avination too
[11:43]  Lani Global: yes, quite a lot of grids still use ode
[11:43]  OtakuMegane Desu: The transition from ODE is going to be a while. Possibly a long while.
[11:43]  Richardus.Raymaker Allen, change scripts so the run with bullet, thats the real solution
[11:43]  Kayaker Magic: Kitely gives you a choice, ODE or Buletsim
[11:43]  Lani Global: the HG market for phys vehicles is about half ODE and half Bullet now.
[11:44]  OtakuMegane Desu: And some grids may never change. Others may use PhysX or even another engine.
[11:44]  Allen Kerensky: yeah - well - bullet is default and compatible with SL/bullet is the requirement for entry into core
[11:44]  Allen Kerensky: so - anyone running something that doesn't meet that - is out of spec and on their own
[11:44]  Dahlia.Trimble requirement?
[11:44]  Allen Kerensky: it is now
[11:44]  Nebadon Izumi: its not a requirement
[11:45]  Nebadon Izumi: but it should be the goal really
[11:45]  Allen Kerensky: that was the practical gist of what Melanie said
[11:45]  Nebadon Izumi: for LSL
[11:45]  Dahlia.Trimble Im not aware of any requirement
[11:45]  Allen Kerensky: if you want a new physics in OpenSim core - then it needs to seem the same as SL and Bullet from a script perspective
[11:45]  Dahlia.Trimble since when?
[11:45]  Allen Kerensky: about 12 hours ago
[11:45]  OtakuMegane Desu: XD
[11:45]  Dahlia.Trimble I'm a core developer and Ive not heard that
[11:46]  Nebadon Izumi: yea thats not really what she said
[11:46]  Dahlia.Trimble maybe someone has that requirement but I dont
[11:46]  Nebadon Izumi: she said the goal for LSL should remain compatible
[11:46]  Lani Global: Would all OpenSim Physics Engines be forced into lock-step with the most recent SL implementation of Havok?
[11:46]  Allen Kerensky: so - all llVehicle* calls made by scripts shoudl EXPECT SL/Bullet-like behavior
[11:46]  Nebadon Izumi: so LSL functions work the same
[11:46]  Nebadon Izumi: outside of LSL you can do whatever you want
[11:46]  Dahlia.Trimble Melanie can impose any requirement on herself that she wants, but dont expect others to abide
[11:46]  Jak Daniels: I don't think it's practically achievable either
[11:47]  Nebadon Izumi: she never said it was a requriement just that was what she recommended
[11:47]  Aine.Caoimhe but then Avination can't keep their leg up on the scripting side by violating that "rule" :p
[11:47]  Allen Kerensky: she said, and I will quote: MUST.
[11:47]  Nebadon Izumi: that doesnt mean it is
[11:47]  Nebadon Izumi: that just her feelings
[11:47]  Allen Kerensky: when you use the word must, in LARGE LETTERS, that implies more than a recommendation
[11:47]  Dahlia.Trimble Allen, she is not the supreme ruler
[11:47]  Nebadon Izumi: you shouldnt take anything any one person says as absolute fact
[11:47]  Nebadon Izumi: thats just how melanie talks, she is very straight forward and literal
[11:48]  Allen Kerensky: not going to argue religion or politics here - just reporting what I read
[11:48]  Lani Global: One can't expect any new physics engine, or any opensim module for that matter, to arise fully formed and bug free, canon compliant.
[11:48]  OtakuMegane Desu: ^
[11:48]  Nebadon Izumi: of course not
[11:48]  Nebadon Izumi: but we really shouldnt be scripting around bugs
[11:48]  Lani Global: lol
[11:48]  Dahlia.Trimble Allen, well you misinterpreted :)
[11:48]  Nebadon Izumi: and then expect developers to nurture bug work arounds
[11:48]  Nebadon Izumi: I am completely against that
[11:48]  Allen Kerensky: pfff scripting around bugs has been an OPenSim requirement for 6 years
[11:48]  Lani Global: we've been scripting around bugs for years!
[11:48]  Aine.Caoimhe the entire "make it just like SL" thing is one of the biggest anchors of the entire development process from what I've seen
[11:48]  Richardus.Raymaker Did you told that to LL in the past to nebadon ? :P
[11:48]  OtakuMegane Desu: BulletSim for how long it's been available and the amount of manpower behind it is progressing wonderfully I think.
[11:48]  Nebadon Izumi: right but we shouldnt be locking those bugs in
[11:49]  Nebadon Izumi: because if something changes and your work arounds break
[11:49]  Nebadon Izumi: then what
[11:49]  Nebadon Izumi: you have to change the scripts anyway
[11:49]  Allen Kerensky: then I have to work around again, speaking from past experience
[11:49]  Nebadon Izumi: so whats the point
[11:49]  Nebadon Izumi: we just keep adding layers of duct tape
[11:49]  Nebadon Izumi: its not good
[11:49]  Richardus.Raymaker points to SL
[11:49]  OtakuMegane Desu: Needs more bubblegum and paperclips.
[11:49]  Andrew Hellershanks: nebadon, that is where adaptive scripts come in if they can check which version of which engine is being used.
[11:49]  Jak Daniels: we will always be scripting against a moving platform
[11:50]  Allen Kerensky: so it was a simple request: give scripters a way to see the damn engine - and everyone turned it into a socio-politico-religious battleground and someone gets stuck out regardless
[11:50]  Nebadon Izumi: we dont have any physics engine versioning
[11:50]  Nebadon Izumi: so no
[11:50]  Nebadon Izumi: that doesnt work either
[11:50]  Lani Global: often we can't wait for the enthusiastic dev to get to a particular bug fix... sometimes it takes several years.
[11:50]  Richardus.Raymaker Allen really ?
[11:50]  BlueWall Slade: is the solution to just pin each subsystem down to one choice and focus the available resources on it?
[11:50]  Allen Kerensky: yeah ODE people don't want to rescript their stuff
[11:50]  Allen Kerensky: that's why they stick on ODE when Bullet has been default for how long?
[11:50]  Richardus.Raymaker Pffff. well, then it's easy. it just stops working sokemday
[11:50]  Richardus.Raymaker and the need to change
[11:51]  Nebadon Izumi: even if we add all of this stuff, there are still going to be changes
[11:51]  Allen Kerensky: yeah I don't have a dog in the race
[11:51]  Nebadon Izumi: that cause skew
[11:51]  Allen Kerensky: so do whatever you want
[11:51]  Nebadon Izumi: so i just dont see the point
[11:51]  Nebadon Izumi: we will never be able to keep parity
[11:51]  Nebadon Izumi: and people are going to have to constantly change scripts
[11:51]  Nebadon Izumi: there is no getting around that
[11:51]  Andrew Hellershanks: If not a physics engine version you could use the build date for it or last commit date
[11:51]  OtakuMegane Desu: Well, if they want to stick with ODE, that's their choice. Someone will probably run an ODE-only grid or two or something. But they don't need to complain when everything else moves on.
[11:51]  Nebadon Izumi: and if people dont update then there is not much we can do
[11:51]  Lani Global: OpenSim is MODULAR.
[11:51]  Nebadon Izumi: a lot of the people who are staying on ODE are also not updating
[11:52]  Allen Kerensky: but its not equally supported modules
[11:52]  Nebadon Izumi: alot of people think if you update
[11:52]  Nebadon Izumi: you cant even use ODE
[11:52]  Nebadon Izumi: seriously
[11:52]  Allen Kerensky: yeah and that clogs things with other problems like TP issues
[11:52]  BlueWall Slade: I think a lot of people just don't change becaus eof fear, etc.
[11:52]  Nebadon Izumi: people refuse to use OSGrid
[11:52]  Nebadon Izumi: because they think thy have to use Bullet
[11:52]  Nebadon Izumi: like seriously?
[11:52]  OtakuMegane Desu: lol
[11:52]  BlueWall Slade: even though they are riding a decaying horse
[11:53]  Dahlia.Trimble nobody updates ODE because noone using it is a contributor
[11:53]  OtakuMegane Desu: I only started using Bullet a few months ago when I got my new server because CentOS 5 puked on Bullet for some silly reason.
[11:53]  Richardus.Raymaker people with old versions on HG onmly cvreate problems. the think ODE use less memory. well mabey it does. because it cannot do much
[11:53]  Nebadon Izumi: yea and the core ODE project doesnt really offer much in terms of updates either
[11:53]  Nebadon Izumi: so anything new done to ODE would be by opensim devs pretty much
[11:53]  Dahlia.Trimble centos 5? isnt that like.... ancient?
[11:53]  Richardus.Raymaker or cpu usage. uhh
[11:53]  Lani Global: people like having a choice.
[11:53]  Lani Global: freedom.
[11:54]  OtakuMegane Desu: Not in terms of server OS, no. It's getting old but hell 4 is still used in numerous places
[11:54]  Andrew Hellershanks: You need someone versed in physics stuff to get in to working on the engine. I've contributed some code but Physics isn't an area of code I'd want to tackle.
[11:54]  Allen Kerensky: well, Lani the bad news is people who want ODE are free to pick up 13.x and start updating it - and then updating the OpenSim module depending on it
[11:54]  Nebadon Izumi: I will still running opensuse 11.4 on the OSCC server up until a few months ago
[11:54]  Nebadon Izumi: lol
[11:54]  Dahlia.Trimble choice is fine as long as people choose to support those choices they want
[11:54]  Nebadon Izumi: it was using Kernel 2.6
[11:54]  Dahlia.Trimble if they expect others to support it on their own time,... not going to happen
[11:54]  Andrew Hellershanks: I didn'trealize ODE is something that existed outside of OS
[11:55]  OtakuMegane Desu: I don't have need for cutting-edge Linux for my server. I need something stable that I don't have to worry about.
[11:55]  Nebadon Izumi: ODE is aincent Andrew
[11:55]  Nebadon Izumi: its mostly for Scientific type work
[11:55]  Nebadon Izumi: not games
[11:55]  Dahlia.Trimble ODE is in a lot of games
[11:55]  Lani Global: there isn't any cry or whine for more ODE support.
[11:55]  Nebadon Izumi: but ya there are some games
[11:55]  Dahlia.Trimble lots of driving games too
[11:55]  Aine.Caoimhe I stopped making any attempt not to use OSSL a couple years ago juest because I was tired of hamstringing myself by using only LSL or by having to double-write huge inefficient portions of things...if people want to stick to old systems that's fine but don't stunt development/progress/new capabilities in an effort to support them
[11:55]  Nebadon Izumi: yea its big with track stuff like tanks
[11:55]  BlueWall Slade: Bullet is probably the best thing goin gright now for game physics.
[11:55]  Jak Daniels: OtakuMegane: It *is* possible to use bullet in centos 5 ;)
[11:55]  Andrew Hellershanks: games?
[11:55]  Allen Kerensky: yes Lani there is - when Dahlia's patch somehow impacted ODE scripting, you were the first person to say "enough"
[11:55]  Lani Global: there is only the need for knowing whether ODE or some other engine is in use.
[11:55]  Nebadon Izumi: but alot of it is use for simulation work
[11:55]  Andrew Hellershanks doens't have a lot of time for games
[11:55]  Andrew Hellershanks: :)
[11:56]  Lani Global: lol
[11:56]  BlueWall Slade: Unles you have deep pockets and can stuf your data center with GPU
[11:56]  Sarah Kline: Theres always a whole load of people trying to hold us back
[11:56]  Dahlia.Trimble Allen, my patch?
[11:56]  OtakuMegane Desu: It works yes, but something about how the module was compiled CentOS 5 didn't like.
[11:56]  Nebadon Izumi: Bullet is a good choice, even though its not perfec
[11:56]  BlueWall Slade: ++ Sarah, lol
[11:56]  Nebadon Izumi: its still being maintained, its open source
[11:56]  Allen Kerensky: something about llLookAt changes made ODE vehicles not work or something, not sure Dahlia
[11:56]  OtakuMegane Desu: It involved SysV and junk,
[11:56]  Nebadon Izumi: its used in a lot of big games and hollywood movies
[11:56]  Jak Daniels: yes, byte 8 in the .so needs changing ;)
[11:57]  Nebadon Izumi: we are not running the latest bullet in opensim yet
[11:57]  Dahlia.Trimble according to the spec
[11:57]  Nebadon Izumi: and even with the version we run we are not taking full advantage of it yet
[11:57]  Lani Global: Dahlia, there was a vehicle ODE banking workaround that borked when llLookAtRot went physical.... but that has been sorted out
[11:57]  OtakuMegane Desu: I compiled it myself for a while. NOw I don't need to at all, since 6 works fine with it.
[11:57]  Dahlia.Trimble llLookAt does *NOT* bank, according to the spec
[11:58]  Dahlia.Trimble if you want banking, llLookAt is the wrong choice
[11:58]  Lani Global: yeah, well, when ODE didn't have Vertical Attractor, a banking method was derived from other functions.
[11:58]  Richardus.Raymaker Anybody here klnow offlineIM v2 debug way. or how to get the messgae from daytabase send to your avatar with V2 ? (it'sin the databse stored)
[11:58]  Dahlia.Trimble well llLookAt is NOT supposed to bank
[11:58]  Richardus.Raymaker teah.
[11:58]  Lani Global: true.
[11:59]  Allen Kerensky: RiRa if its in the DB then the offlineim v2 isn't noticing the login to get it back out?
[11:59]  Lani Global: but, it enabled physical vehicles in OpenSim to have banking for several years in ODE, until BulletSim came around.
[11:59]  Allen Kerensky: that at least narrows down where to look if the messages are in the DB
[11:59]  Richardus.Raymaker Ok. but hopw to fix it ? mabey bug in opensim ?
[11:59]  Andrew Hellershanks: Richardus, I'm still using the older way of handling offline IMs
[11:59]  Aine.Caoimhe is there a Mantis on it?
[11:59]  Richardus.Raymaker i think it's a long standing problem possible. i use 0.8.1
[12:00]  Allen Kerensky: two options RiRa - dig in code or do git biset to identify the breaking commit
[12:00]  Richardus.Raymaker No, because i dont see any error etc. not sure. i can mamke one. but it dont have much info
[12:00]  Allen Kerensky: git bisect that is
[12:00]  Dahlia.Trimble Lani, then those vehicles are scripted improperly, and to make them act as you want it would make *properly* scripted things *not* work properly
[12:00]  Nebadon Izumi: I havent had any trouble with the new Offline IM
[12:00]  Aine.Caoimhe no trouble = can't send IM at all?
[12:00]  Alicia.Raven my offline IM work fine also
[12:00]  Andrew Hellershanks: Richardus, what is offline IMs doing (or not)? Not sending offline IMs to users, not letting someone retrieve them?
[12:00]  Richardus.Raymaker nebadon. it IM send from Hypergrid
[12:00]  Nebadon Izumi: oh
[12:01]  Dahlia.Trimble if you want banking, use other functions
[12:01]  Nebadon Izumi: so local ones work?
[12:01]  Lani Global: Dahlia, yes, you are correct in theory, but in practice that wasn't possible.
[12:01]  Richardus.Raymaker local -> local works
[12:01]  Nebadon Izumi: ok interesting
[12:01]  BlueWall Slade: maybe the database is old and has the bug that was discovered a bout a year ago?
[12:01]  Nebadon Izumi: i probably havent tested that much
[12:01]  Dahlia.Trimble Lani, then use bullet
[12:01]  Richardus.Raymaker but HG -> local ->avatar = fails except it got stored on the grid
[12:01]  Lani Global: Dahlia, as I said, 2 years ago Bullet wasn't an option.
[12:01]  Richardus.Raymaker Hi bluewall, impossible it where fresh database created by opensim 0.8.1
[12:01]  BlueWall Slade: ok
[12:02]  Andrew Hellershanks: Richardus, check the way the offline IM from HG is saved. Perhaps it is saved with the full avatar name plus @ and grid name. Retreiving that locally mayfail as it won't be expecting the @ etc, part when doing a lookup for offline IM's. That's my guess at it.
[12:02]  Richardus.Raymaker the console says, retrieing offline messages for uuid. but more debug code i dont get
[12:02]  Dahlia.Trimble Lani, should I break other people's *properly scripted* objects to fix your *improperly scripted* vehicles?
[12:02]  Richardus.Raymaker then i need to check that with new system is up.
[12:03]  Nebadon Izumi: yea that was my point earlier, one way or the other, someone is going to be broken
[12:03]  Nebadon Izumi: heh
[12:03]  Lani Global: Dahlia, you should feel free to work however you like to make it better.
[12:03]  BlueWall Slade: th atis like the old sit target position bug
[12:03]  Richardus.Raymaker but then it would be bug in OS.. anyway where are local offline IM 's stoired ?
[12:03]  BlueWall Slade: it worked ok, then was changed, then was broken and people made scripts with the wrong position
[12:03]  Nebadon Izumi: yea then everyone fixed their chairs
[12:03]  BlueWall Slade: then when it was fixed, it brok all sit scripts
[12:03]  Nebadon Izumi: and we broke it again
[12:03]  Andrew Hellershanks: There is a separate table for them, Richardus (AFAIK).
[12:04]  OtakuMegane Desu: Never alter function to "support" bugs or flaws unless it's something benificial and desired enough to be of benefit to all.
[12:04]  Nebadon Izumi blames Nebadon for that
[12:04]  BlueWall Slade: and weeping, wailing and gnashing of teeth was heard for weeks
[12:04]  Aine.Caoimhe months....if not years
[12:04]  Lani Global: many creators gave up making chairs for several years.
[12:04]  BlueWall Slade: there is a bug that has been out a while for non-standard avatars.
[12:04]  Nebadon Izumi: I was walking through a mall in sweden voicing with justin convincing him to break it a second time
[12:04]  OtakuMegane Desu: Don't need no chairs. :P
[12:05]  Nebadon Izumi: lol
[12:05]  Sarah Kline: we were chairless
[12:05]  Andrew Hellershanks: hehe
[12:05]  Sarah Kline: lol
[12:05]  Lani Global: funny!
[12:05]  OtakuMegane Desu: Chairs are for pro wrestling.
[12:05]  Aine.Caoimhe I do have one thing I would ask somenone else to test please: it seems under current robust disallow direct login for HG isn't working any more
[12:05]  Nebadon Izumi: ultimiately it was worth it
[12:05]  Nebadon Izumi: even though i had to fix literally 1000s of chairs
[12:05]  Nebadon Izumi: lol
Personal tools
About This Wiki