Chat log from the meeting on 2017-07-11

From OpenSimulator

Jump to: navigation, search

[11:09] Andrew.Hellershanks What is new and exciting since last week?
[11:10] Bill.Blight Ubit can't drive boats to save his life, that is new ... Or was that already known?
[11:10] Arielle.Popstar you mentioned a few weeks back about looking into whats still needed for a .9 official release
[11:10] Bill.Blight :P
[11:11] Andrew.Hellershanks Any thoughts on that, Ubit?
[11:12] Andrew.Hellershanks I haven't gotten any sort of list of what is still to be done before it can be officially released.
[11:12] Gavin.Hird LL has defined a caps for animating rigged objects (meshes) that the viewer will respond to : ObjectAnimation
[11:12] Arielle.Popstar perms are done and that was the thing that was supposedly holding it up
[11:12] Ubit Umarov: no thoughts on release other that we should by now be on 0.9234334
[11:12] Ubit Umarov: or something
[11:12] Andrew.Hellershanks :)
[11:13] Andrew.Hellershanks The perms testing is complete?
[11:14] Ubit Umarov: did you ever started it Andrew ?
[11:15] Arielle.Popstar wasnt diva looking into automating the testing?
[11:15] Ubit Umarov: diva did add some minimal automatic testing
[11:16] Andrew.Hellershanks I haven't been working on the OS code in the last few weeks. I had some paid RL work I was doing.
[11:16] Ubit Umarov: actually testing it all that way is a huge task..
[11:17] Arielle.Popstar any recent mantis reports of anything not working properly then?
[11:17] Arielle.Popstar perms wise?
[11:17] Andrew.Hellershanks One way to generate the tests would be to describe tests to be done in a way that code for the tests can be generated. That is still a big job.
[11:17] Ubit Umarov: to have reports ppl that care about perms need to be using 0.91
[11:18] Andrew.Hellershanks I haven't seen any new perms reports being filed.
[11:18] Ubit Umarov: and the large majority of current users only want to know how to activate GOD mode
[11:18] Arielle.Popstar that fixes broken perms i find :)
[11:18] Ubit Umarov: so.. no reports but one or two im private
[11:18] Bill.Blight I have had no issues to report with perms, other than the group permissions getting funky , which we already talked about before
[11:19] Gavin.Hird are the changes to the permission code production ready?
[11:19] George Equus flüstert: my problem with Blamgates was found between chair and keyboard.. solved
[11:19] Ubit Umarov: if no reports could mean no issues.. yes
[11:19] Leighton.Marjoram lol @George
[11:19] Arielle.Popstar hehe
[11:20] Ubit Umarov: but we are on a loop
[11:20] Ubit Umarov: many will not test 0.9 bc there is no release...
[11:20] Ubit Umarov: release keeps delayed bc there is no testing...
[11:20] Arielle.Popstar well maybe some issues will be discovered when there is an official .9 release that other grids can get onto to test
[11:20] Ubit Umarov: so actually i start not to care about releases
[11:21] Ubit Umarov: anyone runs what they want
[11:21] Bill.Blight I'm with Ubit on that
[11:21] Arielle.Popstar could stop caring about detailed testing
[11:21] Arielle.Popstar and release a .9
[11:21] Bill.Blight does that matter
[11:21] Ubit Umarov: and it seems clear now some just do not want any upgrades
[11:21] Bill.Blight who cares if it is called a release or as it is, same code
[11:21] Ubit Umarov: they are happy with what they have
[11:21] Bill.Blight calling it a release is just a warm and fuzzy, get over that
[11:22] Ubit Umarov: some on 0.7.x versions
[11:22] Ubit Umarov: so.. really why bother ?
[11:22] Sheera Khan: no, that is not totally true...
[11:22] Arielle.Popstar tell the grid owner and managers that
[11:22] Bill.Blight I am a grid owner so I tell me
[11:22] Bill.Blight lol
[11:22] Arielle.Popstar consider your self told :)
[11:22] Sheera Khan: Metropolis would like to have a thing (called release) which is a state we can build on
[11:22] Andrew.Hellershanks Release or no, master has a ton of changes and bug fixes over 0.8
[11:23] Sheera Khan: it is no easy thing to upgrade a whole grid.
[11:23] Andrew.Hellershanks One area that will give some people concern is how the changes may affect scripted objects that use physics.
[11:23] Ubit Umarov: as time goes by.. the harder it will be
[11:23] Arielle.Popstar i really cant understand why people wouldnt want to upgrade
[11:23] Sheera Khan: We would like to have a state where the devs say: that's our base to work on ...
[11:23] Arielle.Popstar other then bullet
[11:23] Bill.Blight If you called it a release this hour, it would not change the status of it an hour ago ..
[11:24] Andrew.Hellershanks Sheera, depends on what you mean by that. It isn't hard to do depending on how one manages the OS instances.
[11:24] OtakuMegane Desu: Mostly because scripts or other things they have somehow rely on factors that have changed.
[11:24] Ubit Umarov: osgrid does releases every week
[11:24] Arielle.Popstar no but it gives that warm fuzzies thing
[11:24] Leighton.Marjoram I like 09 upgraded my grid yesterday.
[11:24] Gavin.Hird A release is a milestone and snapshot in time that everyone can relate to. It is also easier to identify issues when people are on the release version, and not some random commit.
[11:24] Leighton.Marjoram also a release is an endorsement that the code is ok for production.
[11:24] Andrew.Hellershanks There are some people who won't touch the master branch code until someone sticks a release label on it.
[11:25] Arielle.Popstar exactly
[11:25] Leighton.Marjoram nods @ Andrew
[11:25] Sheera Khan: yes, that's the point
[11:25] Ubit Umarov: yes we all know the theory
[11:25] Arielle.Popstar so why not just make an official one and move on\
[11:25] Gavin.Hird it is not theory
[11:25] Leighton.Marjoram its more than a theory its an observation Ubit
[11:25] Selby.Evans Commercial grids have to avoid worrying their customers
[11:25] Sheera Khan: OSGrid is explicitly called a testgrid for a reason
[11:25] Gavin.Hird it is used successfully by all software companies out there that counts
[11:25] Bill.Blight Commercial grids on free alpha software should not be so picky
[11:26] Ubit Umarov: well release is not my dep..
[11:26] Arielle.Popstar whose is it then?
[11:26] Andrew.Hellershanks Not mine either.
[11:26] Ubit Umarov: i did stop changing code for some periods waiting...
[11:26] OtakuMegane Desu: I can understand waiting a short time after a release to be sure there aren't any huge issues but staying years behind, that's a bit much.
[11:26] Ubit Umarov: nothing happened.. i go on
[11:27] Arielle.Popstar whose decision are you waiting for then Ubit?
[11:27] Sheera Khan: explicitely states as the stable release to build a grid upon
[11:27] Ubit Umarov: well Diva is the one making them now
[11:27] Arielle.Popstar is she around?
[11:27] Gavin.Hird Ubit you can go on as much as you like. The Release relates to a spcific (past) commit
[11:27] Ubit Umarov: not for weeks .. posible months
[11:28] Arielle.Popstar then why is she is control of it?
[11:28] Ubit Umarov: lost in ugly rl
[11:28] Arielle.Popstar dont we need devs who are active?
[11:29] Ubit Umarov: well a good idea is not having the most active devs ( adding bugs) actually doing releases
[11:29] Sheera Khan: or simply say - "ok, that's it for the moment - all big stuff is in, you can use it now"
[11:29] Arielle.Popstar nods
[11:30] Arielle.Popstar you could do that too Ubit. Recommend everyone go ahead and update
[11:30] Gavin.Hird use it @commit but not newver cause that is ...untested
[11:30] OtakuMegane Desu: Active dev can mean a couple things. If it's someone who mostly chases bugs and refinements that'd actually be a good input for choosing release.
[11:30] Sheera Khan: I don't expect the development to come to a halt after a relase... (at least I hope so)
[11:31] Sheera Khan: we, as users , don't even know if testing the perms is still in progress or if it is finished
[11:31] Arielle.Popstar nods @Sheera
[11:31] Ubit Umarov: well someone should review commits etc to add to release
[11:31] Andrew.Hellershanks Code development doesn't come to a halt after a release. At some point the code is tagged as a release. A branch is created for that version in case other fixes need to go in to the code marked as released. Code development continues on the master branch.
[11:31] Ubit Umarov: ( specially on postfixes releases.. )
[11:32] Sheera Khan: well, Metropolis i.e. is exactly waiting for this to happen
[11:33] Ubit Umarov: yeap as i said on loop now :p
[11:33] Sheera Khan: bc such a tag would tell us: this is comlete and working for now, dev continues in a new branch and new features go there
[11:33] Ubit Umarov: and hmm there was also a call to boycott 0.9x ..
[11:34] Ubit Umarov: not sure how many are active on that boycott ??
[11:34] OtakuMegane Desu: Seriously?
[11:34] Sheera Khan: I woudn't know of anyone ...
[11:34] Selby.Evans I never heard of boycott
[11:34] Andrew.Hellershanks I've heard some people are wary of 0.9
[11:35] George Equus: Whenever Dan publish an update I grab it and if I see any radical change in behaviour from before I let him know, after a week or so, Don't want to report stuff I managed to mess up myself... like the Blamgate.
[11:35] OtakuMegane Desu: I've been out of the loop for a while so I wouldn't have heard of it but a boycott seems a bit much.
[11:35] Bill.Blight They only boycott I ever heard of was over the perceived mesh issue and the sit target thing .. and that was just fear mongering
[11:35] Ubit Umarov: yeap aine is a name that cames to mind related to that boycott
[11:35] Andrew.Hellershanks I haven't heard of an actual boycott
[11:36] George Equus: Never listen to such rumors,
[11:36] Gavin.Hird not sure of a boycott, but there were discussions in this forum on reverting to 0.8.2 and start a more moderate change process from there
[11:36] George Equus: just try myself...
[11:36] Andrew.Hellershanks Gavin, yes I heard something along that line.
[11:36] Gavin.Hird just go back read the transcripts of this meeting
[11:37] Ubit Umarov: moderate was just not possible
[11:37] Ubit Umarov: ubOde dependend on fixes on framework etc
[11:37] Bill.Blight I love the term moderate change .. Translation (I don't understand it or want to change so no one should)
[11:37] Ubit Umarov: after the merge
[11:37] Andrew.Hellershanks The thinking is that the massive avn code merge has caused various issues in the code and that a more controlled update of the code should have been done to allow the code to be more properly tested as it went.
[11:37] Ubit Umarov: other fixes where needed
[11:37] Arielle.Popstar it was the merge that was a bit of an issue
[11:38] Ubit Umarov: var regions suport was.. well needing fixes etc
[11:38] Sheera Khan: ok, so be it... that's why there are changes in the major version number... smaller stuff (changes) are in the minor numbers
[11:38] Andrew.Hellershanks To go back to 0.8.2 and start applying all the changes from there would waste a lot of time and just further delay any release.
[11:38] Ubit Umarov: well if you look merge is long gone
[11:38] Arielle.Popstar yes agreed
[11:38] Gavin.Hird the merge felt pretty messy for a period, but the result has been very good
[11:38] Ubit Umarov: even ubOde is no longer as it was at avn
[11:38] Ubit Umarov: etc
[11:38] OtakuMegane Desu: Yeah. I should think any issues from the merge have been fixed by now.
[11:38] Arielle.Popstar .9 is stable enough to warrant a release in my opinion
[11:38] Bill.Blight The merge is well past any issues caused directly by the merge,
[11:38] Selby.Evans I have heard of concerns about problems still being in the code, but with more people using .9 the concern should diminish
[11:39] Ubit Umarov: well many do say that 0.9 is very broken
[11:39] Gavin.Hird Ubit has been a star keeping at it
[11:39] Ubit Umarov: when we ask details.. no answer
[11:39] Bill.Blight exactly
[11:39] Sheera Khan: well, in every software product the issues are within the code, aren't they?
[11:39] Arielle.Popstar then they can come here and state the issue
[11:39] Andrew.Hellershanks More people using the code will make it easier to determine if the people concerns about the code are right or not.
[11:40] Andrew.Hellershanks Sheera, usually but there are often still some pebkac issues.
[11:40] Andrew.Hellershanks ;)
[11:40] Arielle.Popstar think the only real issue i have heard is the physics collision thing
[11:40] Sheera Khan: nobody ecpects a perfect and flawless software
[11:40] Ubit Umarov: the early reports of 0.9 issues at Maria's where just a joke
[11:40] Selby.Evans I have seen no problems with .9, but I am not really challenging the capabilities.
[11:40] OtakuMegane Desu: There's always going to be some problems in the code of some sort. Long as it's nothing particularly disruptive or damaging I'm usually ok with a given state at this point of development.
[11:40] Bill.Blight Biggest issue with .9x has nothing to do with the code, it is the Nay Sayers out there bashing it at every turn , because it does not do EXACTLY what THEY want EXACTLY the way THEY want it ..
[11:40] Ubit Umarov: ubOde will break the uploads of all avatar meshs ?? REALLY ???
[11:41] George Equus: whining is easy... doing something about it not that common for the average person
[11:41] Sheera Khan: just a point where one can say: "I've got it all together"
[11:41] Ubit Umarov: ( or the Ubit Umarov is just a alt of Melanie.. )
[11:41] Andrew.Hellershanks I haven't updated a grid to to latest 0.9 for a while. It is still running 0.9 from a few months ago. I've got a few people sensitive to changes in physics affecting scripts.
[11:42] Andrew.Hellershanks Bill, those naysayers will never be happy even if you could give them exactly what they say they want.
[11:42] Gavin.Hird the fact (and this is a fect) that converting existing mesh uploaded on Bullet to uBode breaks some of the physics is a valid one
[11:42] Sheera Khan: @Andrew: pebkac issues permitted though ,-)
[11:43] Ubit Umarov: ubOde is a option
[11:43] Selby.Evans I hear good things about ubODE.
[11:43] Gavin.Hird it is but it is not pushed as that
[11:43] Ubit Umarov: you can run bullet
[11:43] Arielle.Popstar but if bullet is updated like Misterblue intends to do then those who have such builds can use one region for that and ubode for the rest
[11:43] Bill.Blight I am considering changing my name to Bill ubODE ... LOL
[11:43] Gavin.Hird there general understanding is 0.9 = ubode
[11:43] Bill.Blight Bullet is still the default in the config right?
[11:43] Sheera Khan: It would be a good thing if one could remove the old and broken meshes from te asset server... ours is at 4 TByte now ...
[11:44] Ubit Umarov: you will still have issues.. bc 082 handling of physics shape type. is.. lets say odd
[11:44] Bill.Blight Odd is a very nice way to put that ubit
[11:44] Ubit Umarov: Sheera we can't do that
[11:44] Sheera Khan: and there will be a lot of reuploads then ...
[11:44] Gavin.Hird granted it may be odd, but there is a heck of a lot of mesh nobody can reupload for all kinds of reasons
[11:44] Ubit Umarov: bc some are not really broken..
[11:44] George Equus: you can't keep all happy and be backwards compatible for ever, at some point you have to break away the old
[11:44] Andrew.Hellershanks What issues will people have with previously uploaded mesh with the current code?
[11:44] Sheera Khan: that's why I said: It _would be_ a good thing ...
[11:45] Selby.Evans The physics engine may increase demand for .9 as people get more experience with it.
[11:45] Gavin.Hird backward compatible from one version to the next is not forever Geroge
[11:45] Bill.Blight There comes a point where you have to move forward and burn the bridge behind you , so the bad does not follow
[11:45] Ubit Umarov: to upload them as they did is a valid option.. just not for the use they gave them inworld after
[11:45] Arielle.Popstar what you mean Ubit?
[11:46] Ubit Umarov: it is valid to upload a mesh without physics high res collsions
[11:46] Andrew.Hellershanks The main issue with previously uploaded mesh would be based on whether a person remembered to include physics when it was uploaded?
[11:46] Bill.Blight It means if it is a mesh you need to walk in, on or collide with , it needs to be correct, if it is just decoration it does not matter
[11:46] Ubit Umarov: yeap Bill.Blight
[11:47] Arielle.Popstar or walk through
[11:47] Sheera Khan: so clothing would be unaffected ...
[11:47] Ubit Umarov: it is a builder option at upload time... the 82 and bullet just ignored
[11:47] Andrew.Hellershanks To walk through a mesh you can always mark it phantom.
[11:47] Arielle.Popstar but that has to be done for the whole thing then
[11:48] Gavin.Hird people were told by all kinds of o"tutorial" they did not have to make collision models
[11:48] Ubit Umarov: well i never used .82 but seems on it NONE physics type still collides???
[11:48] Gavin.Hird because bullet would (mostly) take care of it
[11:48] Gavin.Hird so there is a large number of such meshes in use out there
[11:48] Gavin.Hird give, sold, spread all over
[11:48] Ubit Umarov: and CONVEX means full resolution collisions on 82
[11:48] Ubit Umarov: bahh
[11:48] Arielle.Popstar the longer we wait with a release though, the more bad mesh builds will be made
[11:49] Bill.Blight bad mesh is not the fault of opensimulator
[11:49] Ubit Umarov: if it was like that, even using bullet somethings are now diff on .91
[11:49] James.atLLOUD I missed the answer, bullet is default physics?
[11:49] Bill.Blight bad mesh is bad mesh
[11:49] Ubit Umarov: type NONE is not sent to physics at all
[11:49] Gavin.Hird it was marketed like that by opensimulator at the time
[11:49] Andrew Hellershanks: James, yes
[11:49] Ubit Umarov: yes bullet is the default
[11:49] James atLLOUD: OK ty .
[11:49] Gavin.Hird bad or not has nothing to do with the issue
[11:49] Gavin.Hird preserving backward compatibility has everything to do with it
[11:49] Ubit Umarov: ( also to easy this issues on ppl that do not know them )
[11:50] Gavin.Hird otherwise people will simply give up and leave
[11:50] Bill.Blight At some point you can't propagate a bug or flaw, regardless of backward compatibility ..
[11:50] Ubit Umarov: well this mesh issue is irritating now
[11:50] Arielle Popstar: is there a way to give UBode the toggable option?
[11:50] Gavin.Hird try use the same logic with Windows
[11:50] Ubit Umarov: SL meshs had always been as ubOde does them ( or tries )
[11:51] Ubit Umarov: toggable option ?
[11:51] Bill.Blight At that level of backwards compatibility I should be able to use Model T tires on a 2017 car
[11:52] Gavin.Hird thankfully I have made collision models everywhere it counts, so nothing sold on Kitely by me will stop working
[11:52] Gavin.Hird but there are tons of items sold out there that will break
[11:52] Ubit Umarov: ppl that did learn a bit about meshs have no issues
[11:52] Bill.Blight LOL
[11:53] Gavin.Hird if models sold on Kitely don't have proper collision models, they will break on an ubode region
[11:53] Ubit Umarov: the main diff relative to SL is that they recommend STEP 2 and you should not used it in OS still
[11:53] Andrew Hellershanks: We can put a page together about how to properly upload mesh so people know what they should or shouldn't do based on their intended use of mesh but we can't make them read it.
[11:54] Andrew Hellershanks: Some people will just continue doing what they have been doing as it has worked in the past.
[11:54] Ubit Umarov: ( because opensim viewers do not use the havoc mesh tools and use worse ones)
[11:54] George Equus: "from one version to the next is not forever" Just a general observation Gavin.. say e.g. physics.. at some point you will have to settle for one or the other, whatever is chosen should by the time it is declared the official code be working and bug free for all and scripts available for those that must make adjusts. If other versions can be available without interfering with the rest of the code, (backwards comp) sure.
[11:54] Gavin.Hird virtually every tutorial on the subject addressing OpenSim can be wiped
[11:54] Sheera.Khan Those we can redirect to the documentation and say: Read the fantastic manual"
[11:55] OtakuMegane Desu: Manuals are good.
[11:55] Gavin.Hird RTFM
[11:55] Andrew Hellershanks: yup
[11:55] Selby.Evans I could run an article about that problem --Andrew if you make a page, send me the link .
[11:55] Andrew Hellershanks: If we have docs out there and people continue to do "the wrong thing" we can point them to the doc.
[11:55] Sheera.Khan I certainly will post a tutorial on that in our forum...
[11:56] Sheera.Khan if there is a final decision, how it's done correctly
[11:56] Ubit Umarov: there are several "fixed ones" around for opensim
[11:56] Ubit Umarov: but also a few wrong ones sure
[11:56] Gavin.Hird Actually the viewers need to disable all the Analyze buttons when logged to an OpenSim grid
[11:56] Gavin.Hird or not show them at all
[11:56] Ubit Umarov: a few days ago i was talking with a kitely builder
[11:57] Ubit Umarov: he had no idea of what CONVEX means
[11:57] Ubit Umarov: because he never saw it working there
[11:57] Andrew Hellershanks: I can't say I have noticed that Analyze does anything other than take some time before you can click other buttons.
[11:57] Ubit Umarov: i had to invite him here and show
[11:57] Ubit Umarov: ( i could had invited to SL also :p )
[11:58] Selby.Evans Anybody that posts tutorial on this, please send me link.
[11:58] Gavin.Hird Analyze is meant to inveract with Havok and since Havok is not present on OS, it leaves the physics in an undefined state
[11:58] Bill.Blight Ubit keeps referring to this being like SL, personally I think it is less about being like SL and more about just doing it correctly ...
[11:59] Ubit Umarov: think aaack may have a few
[11:59] Ubit Umarov: at least i had chat with him about meshs etc
[12:00] Ubit Umarov: actually some here are made by him ( i don't like some )
[12:00] Andrew Hellershanks: There is a wiki page about Collada Mesh
[12:00] Ubit Umarov: ( i did not check each video.. so.. )
[12:01] Ubit Umarov: forget collada
[12:01] Ubit Umarov: that is a complex thing
[12:01] Ubit Umarov: and it is only a file protocol for opensim
[12:01] Andrew Hellershanks: What do you mean by forget collada? How are we supposed to upload mesh if not in Collada format?
[12:01] Gavin.Hird it must be explained in a very simple How-To list
[12:01] Ubit Umarov: builders need to know about opensim on blender or Maya 3ds etc
[12:01] Gavin.Hird the main thing is that making a collision mesh is extra work
[12:01] Sheera.Khan Since I'm not a gifted mesher and I don't have much experience in that I can only write a tutorial if I have some documentation known to be good and correct. Otherwise I'm prone to spread errors and don't even realize them^^
[12:02] Ubit Umarov: collada is just the file format to give the result to the viewer
[12:02] Ubit Umarov: and it is a complex format..
[12:02] Gavin.Hird people are lazy by nature and want to avoid that extra step
[12:02] Ubit Umarov: not by far a good reference for our meshs uses
[12:03] Gavin.Hird particularly if you just copied the mesh from "somewhere" and don't have any skills in even making a collision mesh
[12:03] Andrew Hellershanks: Sure it is a file format. It is the only format we can use to upload mesh, afaik.
[12:03] Andrew Hellershanks:
[12:04] Andrew Hellershanks: If there is information on that page that is no longer valid it should be updated.
[12:04] Arielle Popstar: how hard is it to add a collision mesh to a build?
[12:04] Gavin.Hird not particularly hard, but it requires you can use a modeller
[12:05] Arielle Popstar: modeler like blender?
[12:05] Gavin.Hird and it often needs to be tested quite a bit for more complex builds
[12:05] Gavin.Hird yes like Blender (shrugs)
[12:05] Ubit Umarov: well it is not the purpose of this meetings to teach about meshs
[12:06] Arielle Popstar: so builder who created for kitely market can update with relative ease?
[12:06] Gavin.Hird possibly Arielle
[12:06] Bill.Blight a true builder should have no issues updating their mesh
[12:06] Ubit Umarov: really to talk about meshs issues, ppl should have minimal understanding on how they work
[12:06] Gavin.Hird I agree Bill
[12:06] Gavin.Hird but extra work
[12:06] Ubit Umarov: or we keep wasting time on the same issue
[12:07] Bill.Blight I don't see it as extra work .. I see it as , should have done it right the first time ..
[12:07] Andrew Hellershanks: That is why it would be very useful for those that know how they work to write it down for others to read.
[12:07] Arielle Popstar: it is about whether it will delay adoption to a .9 release Ubit
[12:07] Arielle Popstar: thats why it is important
[12:07] Ubit Umarov: no
[12:07] Ubit Umarov: meshs are not a release issue
[12:07] Andrew Hellershanks: Or document what impact people may expect to see.
[12:07] Bill.Blight people can still use bullet
[12:07] Gavin.Hird For me, when making a model that is not supposed to be phantom, the physics model is an integral part of the workflow
[12:07] Ubit Umarov: Meshs are a closed issue since the merge
[12:07] Bill.Blight no one is twisting their arm to switch to ubODE
[12:08] Selby.Evans So sellers could update -- if they knew the problem.
[12:08] Ubit Umarov: in code terms i mean
[12:09] Bill.Blight If you are a mesh seller, you should do them correctly , regardless of the destination, cutting corners to eliminate "work" is not the correct answer ...
[12:09] Gavin.Hird Selby, they basically have to test all their models on an ubode region, and fix those that don't work as expected. The replace the mesh on the market and potentially send updates to very customer
[12:09] Ubit Umarov: we can't break fundamental building tools because opensim devs and users did not know about meshs in past
[12:10] Bill.Blight if the mesh was right to begin with it will work on bullet or ubode
[12:10] Ubit Umarov: physics shape type is a fundamental thing
[12:10] Arielle Popstar: a lot of this is user generated, not commercial professionals
[12:11] Gavin.Hird You only have to read G+ to see a number of people selling stuff on Kitely have proclaimed physics models are not for them
[12:11] Ubit Umarov: any user that takes so 30mins more learning about meshs have no issues on this
[12:11] Bill.Blight cutting corners to eliminate faces, or physics to make the model "lighter" because one particular upload program will "fake" it for you is not good practice ..
[12:11] Ubit Umarov: there are other problems on blender tutorials
[12:11] Andrew Hellershanks: Where does a person go to spend that 30 minutes learning about meshes for use with OpenSim?
[12:11] Ubit Umarov: bc most of then are not about modeling for real time simulation
[12:12] Selby.Evans ty Gavin
[12:12] Ubit Umarov: for only for static rendering
[12:12] Ubit Umarov: that means 100k tris meshs etc
[12:12] Gavin.Hird exactly Ubit
[12:12] Gavin.Hird very few tutorials for how to model for OpenSim
[12:12] Sheera.Khan I would even be willing to spend a full hour, if I knew where to find the info...
[12:13] Bill.Blight any tutorial for any low poly 3d game modeling with proper physics should work for OS as well
[12:13] Gavin.Hird most models are super high density compared to what they should be
[12:13] Ubit Umarov: i know aaack did fic some tutorials after talking with me
[12:13] Arielle Popstar: if opensim works like s/l for mesh stuff on ubode, then the tuts for s/l should be good for opensim?
[12:13] Andrew Hellershanks: I want to create some musical instruments as mesh objects. I'm not sure if I would be creating them in a way that is "the right way" or not.
[12:13] Ubit Umarov: it is a good base yes arielle
[12:13] Andrew Hellershanks: I just use the model file itself as the physics model.
[12:14] Gavin.Hird Most SL stuff has the same issue with too much topology
[12:14] Ubit Umarov: and like bill said.. a mesh for ubode will work on both engines
[12:14] Bill.Blight if you are creating a mesh that does not require collisions then it does not really need a physics model at all
[12:14] Andrew Hellershanks: I have a problem with triangle counts because the modeller I use doesn't deal in triangles.
[12:14] Gavin.Hird the most popular package for designing clothes used in SL is made for static renders
[12:14] Ubit Umarov: see this example
[12:14] Bill.Blight if you don't have to go , IN it, ON it , or run into it, then physics are irrelevant ..
[12:15] Gavin.Hird the SL avatar is about 6000 vertices
[12:15] Ubit Umarov: this mesh should not have been uploaded with physics at all
[12:15] Gavin.Hird a typical mesh jeans in SL is in the order of 50000 -70000
[12:15] Sheera.Khan Aaack Aardvark: Latest Youtube video: one year ago
[12:15] Andrew Hellershanks: why not if you need to get in to it to drive it?
[12:15] Ubit Umarov: in a car it must have physics type NONE
[12:16] Bill.Blight unless you like to live dangerously like me
[12:16] Bill.Blight LOL
[12:16] Andrew Hellershanks: hehe
[12:16] Ubit Umarov: see the wood prim?
[12:16] Andrew Hellershanks: yes
[12:16] Ubit Umarov: that is the only prim physics knows about
[12:17] Ubit Umarov: this is how things need to be build
[12:17] Ubit Umarov: one thing is the VISUAL aspect
[12:17] Bill.Blight of course I ignore that
[12:17] Bill.Blight lol
[12:17] Ubit Umarov: another thing is the PHYSICS representation
[12:18] Ubit Umarov: this is just a extrem example
[12:18] Ubit Umarov: this is ugly sure
[12:19] Andrew Hellershanks: That is information that seems to be known to a select few and needs to be documented for the rest of the people who may want to make the occasional mesh object.
[12:19] Bill.Blight a Low poly physics model works quite well as a physics prim shape, but not a 100k tris car
[12:19] Ubit Umarov: andrew this is BASIC building information
[12:19] Bill.Blight but IF anybody that is charging money for mesh builds should already know these things
[12:19] Bill.Blight I hate to agree with Ubit
[12:19] Ubit Umarov: it was already like that with sculpts
[12:19] Andrew Hellershanks: I don't consider mesh "basic".
[12:19] Gavin.Hird they don't Bill
[12:19] Bill.Blight and this is Opensim's fault?
[12:19] Gavin.Hird both because in SL you can get away with the Analyze
[12:19] Ubit Umarov: about this car example.. we should add a few more boxes to give it better collsion volume
[12:20] Gavin.Hird that often generates a reasonable model for you
[12:20] Gavin.Hird so Sl developers thinks it works like that here too
[12:20] Andrew Hellershanks: I'll be happy to spend 30 minutes or an hour learning about mesh to make sure I understand them and their use properly if someone can point me to where I'm supposed to go.
[12:20] Gavin.Hird hence all the erroneous tutorials
[12:20] dj phil: more box or more spheres Ubit ?
[12:21] Ubit Umarov: in car case think boxes
[12:21] dj phil: sphere no better collision ?
[12:21] Ubit Umarov: but depends on the model
[12:21] Gavin.Hird + people sell and distribute copybotted items + items uploaded from 3D graphcis sites
[12:21] Gavin.Hird and they don't have a clue how to make the physics
[12:21] Ubit Umarov: well you want a cruce car shape
[12:21] Bill.Blight spheres will give you more responsiveness to terrain, but that is not always wanted
[12:21] Ubit Umarov: and a car is a box :)
[12:22] Ubit Umarov: spheres is wheels
[12:22] Andrew Hellershanks: I've heard that sphere or box is the best to use for the physics object of a build.
[12:22] dj phil: sphere don't reduce the calculation ?
[12:22] Bill.Blight spheres reduce what one vector?
[12:23] Gavin.Hird sphere is useless for a building
[12:24] Arielle Popstar: so can we come to any agenda
[12:24] Ubit Umarov: another basic physics car
[12:24] James atLLOUD: lol cammed in on the car then tried to turn on my flashlight from Fallout. lol
[12:24] Ubit Umarov: you can think as chassis
[12:24] Arielle Popstar: about how to get a release
[12:26] Ubit Umarov: this ugly car uses this chassis
[12:27] Ubit Umarov: well i do have a more complex example
[12:27] Ubit Umarov: as usual don't find it :)
[12:28] Andrew Hellershanks: Arielle, first step is check if there are any open bug reports, or work in progress, related to show stopper or serious bugs.
[12:28] Ubit Umarov: well so this cars are examples that you can and should NOT upload with mesh for physics
[12:29] Ubit Umarov: bc it easy and better to add the physics using a few prims
[12:29] Ubit Umarov: in fact in SL if you try to use a mesh with physics on a vehice it will violate LI limits
[12:30] dj phil: Hmmmm i like use custom physic files
[12:30] Arielle Popstar: so what is a show stopper?
[12:30] Ubit Umarov: what show stopper ?
[12:30] Sheera.Khan Just to show that I don't understand it still: I would mesh the car in say Blender - ok. What about the physics model? Do I do that inWorld with real Prims and link it to the car?
[12:31] Arielle Popstar: a few months ago it was the perms only and that has now mostly been resolved
[12:31] Gavin.Hird Arielle, you can make a prim physics model, set it to invisible and link with the mesh
[12:32] Ubit Umarov: you don't remember old sculpts stairs?
[12:32] Gavin.Hird or you can model the physics model as a very low poly mesh and upload with the model
[12:34] dj phil: why no upload with physic if we made custom physic file ?
[12:34] Ubit Umarov: well mesh optimization can be complex..
[12:34] dj phil: Custom physic is cool no ?
[12:34] Ubit Umarov: that is why there are Pros
[12:34] Andrew Hellershanks: I don't think there are any "show stopper" but some people might consider changes to the physics engine they use that break scripted objects as a show stopper.
[12:34] Ubit Umarov: and only those will be able to build on the new SL think and Space it seems
[12:35] Ubit Umarov: we do want all to build here..
[12:36] Ubit Umarov: but some learning is needed ..
[12:36] Andrew Hellershanks: nebadon has told me about using a simple object for physics.
[12:36] Arielle Popstar: so while misterblue is giving some love to bullet, it should be delayed?
[12:36] Andrew Hellershanks: If misterblue is working on Bullet and has some thing he thinks needs to be fixed before a release that might be a reason for a delay.
[12:36] Ubit Umarov: MB last words on that where that he is finishing a current project
[12:36] Ubit Umarov: and will look to bullet after
[12:37] Bill.Blight Improvise , overcome and adapt ...
[12:37] Bill.Blight ;)
[12:37] Ubit Umarov: :)
[12:37] Andrew Hellershanks: If he can't deal with it for some time it may just have to wait until after a release and the changes can be applied to the post fix branch for the release.
[12:37] Ubit Umarov: well and another meeting about meshs :p
[12:37] Andrew Hellershanks: hehe
[12:37] dj phil: meeting or wiki updates :)
[12:38] Ubit Umarov: ( and yes DJ you can use custom meshs.. but is a more advanced topic, and some times a simple in world prim is a better option )
[12:38] Arielle Popstar: alright so then what about contacting Diva about doing a .9 release then?
[12:38] dj phil: and/or
[12:38] Andrew Hellershanks: Who ever knows of a tutorial about the proper way to create/upload/use mesh, drop the URL here now so people reading this weeks meeting notes can look at the information.
[12:38] Sheera.Khan I would soooo love to see a meeting about how meshes work and what one has to do to make them work correctly
[12:39] Ubit Umarov: some one on kitely is giving blender classes
[12:39] Arielle Popstar: Mike Lorrey
[12:39] Ubit Umarov: but.. still needs some talk about meshs.. because he had never seen CONVEX working there it seems
[12:39] Andrew Hellershanks: I know some people in SL that used to give Blender classes.
[12:39] Gavin.Hird Sheera, for OpenSim the basic message on mesh is that every mesh that is not supposed to be phantom needs a physics model uploaded with the mesh, or made from prims in world
[12:39] Ubit Umarov: yeap Mike Lorrey
[12:40] Selby.Evans I will ask Mike about correct blender tutorials
[12:40] Gavin.Hird it also needs to be double sided in most cases
[12:40] Ubit Umarov: if you do a house where ppl needs to walk in, you need physics mesh
[12:40] Ubit Umarov: or if you want ppl to sit there using physics sits
[12:41] Gavin.Hird bullet accepts planes, but once moved into an ubode region, you will collide with one side of the plane, and walk through it like butter from the other side
[12:41] Ubit Umarov: or want to use in war games and things must stop bullets
[12:41] Ubit Umarov: etc
[[12:42] Ubit Umarov: good thing for us in opensim is that prims don't cost much
[12:42] Andrew Hellershanks: If you want to upload physics with a model you can make a very simplified model to encase the object and use it for physics part of a mesh.
[12:42] Ubit Umarov: ohh this was the OLD sculpts example (rezzes a staircase)
[12:43] dj phil: Still a staircase :o lol
[12:43] Andrew Hellershanks: Part of the problem with sculpt or mesh stairs is the rise on them. A step of more than about 0.25m will pretty much act like a wall.
[12:43] Gavin.Hird in some cases using the lowest LOD can work for physics, but needs to be tested
[12:44] Ubit Umarov: with ubOde it does collide and we can climb
[12:44] Andrew Hellershanks: How high can each step be before you can't climb it properly?
[12:44] Ubit Umarov: 0.6m +- on ubode
[12:44] Gavin.Hird not more than 30 cm on bullet
[12:45] Andrew Hellershanks: ok. That's quite a change. With regular box prims I couldn't enter a building that had much more than 0.25m step.
[12:45] Ubit Umarov: but even in ubOde this is not nice
[12:45] Bill.Blight It is something like .25 on ubode ...
[12:45] Andrew Hellershanks: Hm... my avatar keeps acting like it wants to do a table dance. :)
[12:46] Andrew Hellershanks: The use of invisible prim ramp is how stairs were made climbable in the past.
[12:46] Bill.Blight might be a bit more, stairs seem to work much better .. invisible ramp on straight stairs still a better option
[12:46] Ubit Umarov: this is a better option
[12:46] dj phil: Nice build Ubit :)
[12:47] Ubit Umarov: added a ramp, linked making ramp root
[12:47] Ubit Umarov: and the sculpt is now type NONE
[12:47] Andrew Hellershanks nods
[12:47] Ubit Umarov: so stairs are just the ramp for physics
[12:48] Ubit Umarov: this is a 10years or more SL technic
[12:48] Andrew Hellershanks: The ramp would keep the physics simple instead of having to worry about the step height.
[12:48] Andrew Hellershanks: invisible ramps for stairs has been used in SL for many years.
[12:48] Ubit Umarov: yes
[12:48] Gavin.Hird you upload a physics model like the ramp with the mesh stair
[12:48] Ubit Umarov: even with ubOde that does the steps
[12:49] Ubit Umarov: on bullet you still need to do this
[12:49] Gavin.Hird walks fine and the user don't have to worry about reshaping an invisi prim
[12:50] Ubit Umarov: ( not even sure is bullet does sculpts collisions )
[12:50] Gavin.Hird sculpts needs to die
[12:50] Gavin.Hird they are horrible meshes
[12:50] Gavin.Hird viewer killers # 1
[12:50] Ubit Umarov: depends on use Gavin.Hird
[12:51] Gavin.Hird still horrible meshes
[12:51] Ubit Umarov: they can be smaller
[12:51] Bill.Blight if you are really sold on having mesh stairs, you can upload the no mesh stairs, create a ramp, export as a collada and reimport with physics, instant mesh stairs with physics ramp
[12:51] Andrew Hellershanks: I can more easily make a mesh object than I can make some sculpties.
[12:51] Gavin.Hird not if made right
[12:51] dj phil: Here is a old image i made for staircase
[12:51] Ubit Umarov: there are nice sculpts around
[12:51] Andrew Hellershanks: Gavin, not if what is made right?
[12:52] dj phil: and the "Analyse" button
[12:52] Gavin.Hird the mesh
[12:52] Gavin.Hird import a sculpt into a modeller and look what it actually looks like
[12:52] Andrew Hellershanks: I think the analyze button is partly to calculate how much L$ one will have to spend to upload the mesh.
[12:53] Ubit Umarov: gavin i did import sculpts into blender testing the physics :)
[12:53] dj phil: you don't need to use it Andrew ... It breaks your mesh
[12:53] Ubit Umarov: no andrew
[12:53] Ubit Umarov: that is done on calculate
[12:53] Andrew Hellershanks: djp, I haven't seen it break a mesh yet. What does it actually do?
[12:54] Andrew Hellershanks: Ubit, oh, right. I did say its been a while since I uploaded one.
[12:54] Bill.Blight it may close openings if it is a mesh you have to move through
[12:54] Bill.Blight or just close them completely with a face
[12:54] dj phil: It breaks everything and causes alert in console andrew
[12:54] Gavin.Hird Bill it only does that on SL
[12:54] Ubit Umarov: analyse step2 is also a mesh simplification
[12:54] Ubit Umarov: and actually is more that that.. is a different format
[12:54] dj phil: Hmmm analyse work now Ubit ??? :o
[12:55] Gavin.Hird on OpenSim analyze leaves a mesh in an uninitialized state
[12:55] Bill.Blight I tested i here a mesh I could walk in without analyze could not with it .. (not supposed to use it anyway in opensim)
[12:55] Andrew Hellershanks: I haven't had a problem using it with the mesh objects I've uploaded. I had a mesh with some holes and it didn't close those holes. I wish it had.
[12:55] Ubit Umarov: ubode will do meshs with step 2
[12:55] Bill.Blight why I said "may" close holes
[12:56] Ubit Umarov: just you do need to test them fully
[12:56] Ubit Umarov: bc as we said viewers may had changed it , breaking it
[12:56] Andrew Hellershanks: More details for the page about proper ways to deal with mesh. That and the maximum tri count are all useful information.
[12:56] Ubit Umarov: seen very ugly cases..
[12:56] Andrew Hellershanks: yes, it is another long meeting. Almost two hours again today.
[12:57] Andrew Hellershanks: Any last thoughts before I say "class dismissed"?
[12:57] Ubit Umarov: and still mostly on meshs :(

Personal tools
About This Wiki