Chat log from the meeting on 2018-08-14

From OpenSimulator

Jump to: navigation, search

- OS 0.9.x shows a Land Impact on uploading meshes to assist in creating better meshes with less load on server and clients
- some meshes are rendered incorrectly on the map
- the tree module is suspended due to wildfire breakouts^^

[11:00] Sheera Khan: another question: might it be feasible or desireable to display some "cost" associated with a mesh upload?
[11:00] Sheera Khan: So the user knows how bad or good his/her meshes are
[11:00] Ubit Umarov: Sheera.Khan do uploads on 0.9x then ask again :p
[11:00] Sheera Khan: ahhh :-)
[11:01] Ubit Umarov: there is a pseudo LI
[11:01] Andrew.Hellershanks Sheera, I've wondered about what hooks may be needed (or may already exist) to do a mesh upload cost calculation.
[11:02] Sheera Khan: Ubit seems to have found them ...
[11:02] Ubit Umarov: uff another one not using 0.9x
[11:04] Ubit Umarov: if you have a .dae just test here
[11:04] Ubit Umarov: just don't do the final upload
[11:04] Ubit Umarov: no need to polute osg dbs :)
[11:05] Ubit Umarov: are you testing a mesh upload here ?
[11:05] Sheera Khan: doing it right now...
[11:06] Ubit Umarov: click calculate .. fees and see the answer
[11:06] Sheera Khan: just waiting for the result to appear...
[11:06] Ubit Umarov: ( better a small mesh :) )
[11:06] Sheera Khan: hööööiiii, no fun in uploading something small ^^
[11:06] Ubit Umarov: read the bottom line
[11:07] Sheera Khan: sure, Ubit, I know where those figures are ^^
[11:07] Ubit Umarov: i was telling andrew
[11:07] Sheera Khan: ah ;-)
[11:08] Andrew.Hellershanks Ubit, Last time I did that it didn't show anything re: cost, AFAICR.
[11:08] Ubit Umarov: should be obvious that for this test you should use a simple mesh that does really work :p
[11:08] Sheera Khan: ok, no cost, that's fine:-)
[11:08] Ubit Umarov: and the other numbers?
[11:09] Ubit Umarov: ( of course no cost.. its free here )
[11:09] Sheera Khan: and 137 in LI and Download, 3,7 in Physics and 0,5 in server ...
[11:09] Ubit Umarov: so there you have your "costs"
[11:10] Andrew.Hellershanks :)
[11:10] Sheera Khan: yes, thank you very much :-)
[11:10] Ubit Umarov: not as SL ones..
[11:10] Ubit Umarov: different simpler "math"
[11:10] Sheera Khan: I think that may help a bit generating awareness for the associated loads ...
[11:11] Ubit Umarov: well we have it on opensim since 2015 :p
[11:11] Sheera Khan: I don't think it matters that much how it is actually calculated as long as it is reasonably connected to the load a mesh generates
[11:11] Ubit Umarov: hmm the merge with avn was 2015 right ?
[11:12] Andrew.Hellershanks Was it that long ago?
[11:12] Andrew.Hellershanks time flies
[11:12] Ubit Umarov: yes that was my decision back then... 2012 or 2013 when i made that code
[11:12] Ubit Umarov: just get some simple metrics
[11:13] Bill Blight: yes that long ago
[11:13] Sheera Khan: well, the numbers are reasonably well comparable to what SL told me about that mesh ...
[11:13] Ubit Umarov: sl LI is confusing
[11:13] Ubit Umarov: makes sense.. but confusing
[11:13] Bill Blight: one of the reasons I was so upset that people complain about "Sudden Changes" they have had 3 years to get on board
[11:13] Ubit Umarov: like same mesh on a smaller prim gets larger LI
[11:13] Sheera Khan: hehehe
[11:14] Ubit Umarov: o know why.. makes sense.. but damm confusing
[11:14] Ubit Umarov: and the reason never went to actually enforce LI on parcels control
[11:14] Ubit Umarov: some code for it is there..
[11:15] Ubit Umarov: so.. after 3 years of code on opensim, you seen that feature :p
[11:16] Sheera Khan: right :-) I didn't upload meshes on sims except on Metro...
[11:16] Sheera Khan: and those are on 0.8.x still :-/
[11:16] Ubit Umarov: yeah.. we understand.. 0.82x is the best opensim version ever.. no bugs.. just perfect
[11:16] Sheera Khan: nope...
[11:17] Sheera Khan: just a lot of work to be done to roll a new version out in the grid
[11:17] Ubit Umarov: i was talking to andrew again :p
[11:17] Sheera Khan: ;-)
[11:18] Andrew.Hellershanks Ubit, you know someone might just take that out of context from the meeting log. :)
[11:20] Sheera Khan: there are less and less people reading the log ...
[11:20] Andrew.Hellershanks wonders why Ubit would say something to me about 0.8
[11:20] Ubit Umarov: some of those will take anything out of context, not matter what
[11:20] Andrew.Hellershanks true
[11:20] Sheera Khan: it used to be over 100 ... now it's down to barely 50ies
[11:21] Sheera Khan: Kay is back from his kayak trip in Canada ...
[11:22] Ubit Umarov: ohh still 50? not bad
[11:22] Sheera Khan: he got a bit surprised by the sudden demise of InWorldz ...
[11:22] Ubit Umarov: hi log readers, hope all fine with you...
[11:22] Sheera Khan waves at the reader
[11:22] Andrew.Hellershanks hehe
[11:23] Ubit Umarov: guess we all got surprised by that
[11:24] Sheera Khan: well yes, but only few of us had a commercial commitment to that grid
[11:24] Ubit Umarov: worse.. their own users got surprised
[11:24] Ubit Umarov: but well.. next..
[11:24] Sheera Khan: agrree
[11:25] Ubit Umarov: ( this thing you call beer and put in front of us is really odd Bill )
[11:25] Ubit Umarov: ( can't even read its PH )
[11:26] Bill Blight: HAHA
[11:26] Bill Blight: You sure it is beer
[11:26] Ubit Umarov: no
[11:26] Bill Blight: could be lemonade
[11:26] Bill Blight: LOL
[11:27] Andrew.Hellershanks You don't know how long its been sitting here.
[11:27] Ubit Umarov: so any issues?
[11:27] Ubit Umarov: coding from me is beeing slow.. nothing last weekend :(
[11:27] Sheera Khan: Just a mantis about the tree module ...
[11:28] Ubit Umarov: hm guess i missed that
[11:28] Andrew.Hellershanks I've run across a problem with decoding of mesh while generating the map tile. The error message is useless as it doesn't provide any information about what object has the problem.
[11:28] Ubit Umarov: but trees where working
[11:28] Ubit Umarov: that is a know issue with some mesh decoding Andrew
[11:29] Sheera Khan:
[11:29] Sheera Khan: (about the trees)
[11:29] Ubit Umarov: well that mantis will have to wait.. we are on forest fires season... so .. no trees..
[11:29] Sheera Khan: *giggles*
[11:30] Andrew.Hellershanks Ubit, The objects appear ok inworld so I'm not sure what is the problem with the mesh objects.
[11:30] Ubit Umarov: well 0.9 trees module is dif from 0,8
[11:30] Andrew.Hellershanks It would be nice if the error indication trickled up to where the location of the object could be reported.
[11:30] Ubit Umarov: don't remember the details anymore :)
[11:31] Ubit Umarov: its the decoder andrew
[11:32] Ubit Umarov: we use a external one
[11:32] Ubit Umarov: that apparently never got a fix we have on our own decoders on (ub)Meshmerizer
[11:32] Sheera Khan: OOpsies - RL...
[11:33] Bill Blight: the tree module thing is a no-fly for me, as they seem to want the module to run unchecked and not stop, this has, "Crash my region" written all over it ..
[11:33] Ubit Umarov: yeap
[11:33] Andrew.Hellershanks Ubit, Yes, the problem is in the Meshmerizer of libOMV
[11:34] Ubit Umarov: i may try to fix on the version we do use
[11:34] Ubit Umarov: for now its even my own little fork
[11:35] Ubit Umarov: guess ill just put it back on opensim-libs
[11:35] Ubit Umarov: since by now many use forks
[11:37] Ubit Umarov: but all the meshes i have on test regions work fine on warp3d so.. well a bit hard to check a fix :)
[11:38] Bill Blight: I find that on regions that all the mesh and textures were uploaded new, I don't get those errors andrew
[11:39] Bill Blight: but for Older questionable mesh or HG things that may be missing assets I see those errors on regions that contain those things
[11:39] Andrew.Hellershanks The problem I noticed is with some third party mesh trees.
[11:40] Andrew Hellershanks: I'm not sure where they came from. Possibly from kitely.
[11:41] Bill Blight: sooo, 8x mesh
[11:41] Andrew Hellershanks: Could be. I've asked where they came from.
[11:43] Bill Blight: to me it is related to the same error in regular startup when you see that an item is unable to mesh correctly ... there is no reason to think the map is better at meshing than the physics engine
[11:43] Andrew Hellershanks: I've been told the mesh trees may have come from a public OAR file.
[11:43] Andrew Hellershanks: That probably says something about them right there. :)
[11:45] Andrew Hellershanks: Any other topics for todays meeting?

Personal tools
About This Wiki