Chat log from the meeting on 2018-03-27

From OpenSimulator

Jump to: navigation, search

[11:07] Sheera Khan: Ubit, how's your httptest going?
[11:07] Ubit Umarov: did not mess much with it
[11:08] Bill.Blight I use the test branch everyday ... Well a flavor of it ...
[11:08] Ubit Umarov: last work/pending.. Yengine .. needs more cleanup, and decision abotu features we really want, from the xmrones
[11:09] Bill.Blight Yengine is pretty smooth, I think people will complain about it being a tad bit more strict to syntax though
[11:09] LaNani Sundara: is the Yengine the main difference with that dev master?
[11:09] Ubit Umarov: bf that i was changing some serialization on inventory fetch, still more to do
[11:09] Ubit Umarov: Yengine ( from xmr) should be closer to SL on syntax
[11:10] Bill.Blight httptests has a LOT of differences to master, increased .NET version is one of them
[11:10] Ubit Umarov: well its error reporting is just BAD
[11:10] LaNani Sundara: hmm ok
[11:10] Ubit Umarov: most cases gives useless line numbers, and totally useless stack dump
[11:10] Ubit Umarov: scripters do not care about internal c# stack
[11:11] Ubit Umarov: well X used to do that also..
[11:11] Bill.Blight there are some really nice C# functions, that it has that would like to use , such as arrays
[11:11] Ubit Umarov: but Y error reporting is as now a lot worse than X
[11:11] Bill.Blight Was going to say, "What error reporting?" concerning Yengine
[11:12] Ubit Umarov: well thing about those extentions is that they look like c# but are not
[11:12] LaNani Sundara: remind me again, besides you enjoying it, why you are making a new engine?
[11:12] Ubit Umarov: some are non standard.. so.. not sure we should keep them as they are
[11:12] TG.Lucan Is the commit to master ok, dos'nt seem to do much?
[11:12] Bill.Blight The new engine is quite a bit faster,
[11:12] LaNani Sundara: faster sounds good
[11:12] Bill.Blight Yengine is not in Master
[11:12] LaNani Sundara: uses less CPU? less RAM?
[11:13] Ubit Umarov: well faster is not WHY lol
[11:13] Kayaker Magic: Hopefully the new engine has a fix for the llSleep bug.
[11:13] Ubit Umarov: it does solve the llSleep issue
[11:13] TG.Lucan did not say it was, there was a comit to naster today
[11:13] Bill.Blight NO faster is not WHY but is a result
[11:13] Ubit Umarov: it does "micro multithreading"
[11:13] LaNani Sundara: hmm
[11:14] Ubit Umarov: llSleep and similar do actually release the thread, that can do somethign else
[11:14] Sheera Khan: I really would like to see real arrays with multiple dimensions ;-)
[11:14] Ubit Umarov: and, not tested, deleted scripts should release memory
[11:14] Bill.Blight State changes are faster on Yengine and do not seem to stutter
[11:14] Ubit Umarov: and the parsing is a lot more like SL one..
[11:15] Ubit Umarov: it does parse somethings X does fail to
[11:15] LaNani Sundara: *looking at the commit...
[11:16] Ubit Umarov: of course it is a LSL( OSSL) only engine.. no options fo use c# VB etc that X can co
[11:16] Ubit Umarov: it does its own IL code generation
[11:16] Kayaker Magic: Does anyone actually use those VB, etc options?
[11:16] Ubit Umarov: a few use C#
[11:17] TG.Lucan I use c#
[11:17] Ubit Umarov: but that can only be used on a very closed region
[11:17] LaNani Sundara: VB? i doubt it...
[11:17] Ubit Umarov: another thing about Y is scripts passing
[11:17] Ubit Umarov: on crossings and tps
[11:18] LaNani Sundara: what if you cross from a region with xengine to one with yengine and teh other way round
[11:18] Ubit Umarov: it does not send binary dll, but a previus step data ( AST ) that compiles to IL faster and hopefully safer
[11:18] LaNani Sundara: different sims ofc
[11:18] Bill.Blight Scripts will reset and if they did not exist they will need to recompile
[11:18] Ubit Umarov: so "trustbinary ) is more safe.. well it actually ignores that option  :)
[11:19] Ubit Umarov: yes not that compatible with X
[11:19] Bill.Blight So regardless of how good Yengine may be, we will see the , "OMG IT IS DIFFERENT , KILL IT!!" cries ... HAHA
[11:19] Ubit Umarov: that is a issue.. some scripts that depend on script state lose it
[11:19] LaNani Sundara: if it doesnt have obvious benefits or improvements, yes probably Bill
[11:20] Ubit Umarov: ie those things from sl that are sold with paramaters on scripts variables, will fail
[11:20] LaNani Sundara: same reason people stay on older opensim versions
[11:20] Bill.Blight Benefits don't have to be a flashing neon sign to be relevant, and unfortunately that is what it seems people look for ..
[11:20] LaNani Sundara: *nods
[11:21] Ubit Umarov: opensim is a open source project, people use what they want
[11:21] LaNani Sundara: as you know i do Ubit ;)
[11:22] Ubit Umarov: the reverse of that, is that we can ignore people and move on :p
[11:22] LaNani Sundara: haha :p
[11:22] Ubit Umarov: ( well within reason .p )
[11:22] LaNani Sundara: i do stay uptodate with the master, even if i merge it with my code
[11:22] Ubit Umarov: well master is outdated
[11:23] Sheera Khan: leaving Windows XP behind is reasonable imho ;-)
[11:23] Ubit Umarov: current master is httptests
[11:23] Ubit Umarov: we are having a organizational issue on that..
[11:23] Bill.Blight httptests is so far ahead of master it is not even really the same version any more
[11:23] LaNani Sundara: o idk... :) its ahead or the offcial release still
[11:24] Ubit Umarov: basicly our machine is now outdated, we can't upgrade it to a new kernel etc
[11:24] Ubit Umarov: so we can't use a recent mono..
[11:24] Ubit Umarov: as you know httptests now targets -net4.6
[11:24] LaNani Sundara: why not?i have build on the latest monos
[11:24] Ubit Umarov: that means a mono 5.x
[11:24] Ubit Umarov: and win vista at least
[11:25] LaNani Sundara: i always use last mono and update mono
[11:25] Ubit Umarov: so, at least for me, current master is 0.9 bugs fixes
[11:25] Bill.Blight The Core Dev system that runs their jenkins and build environment cannot be upgraded so it is tough
[11:25] TG.Lucan I run on LInux 4.5 and windows 4.6.1 with a few small adjustments
[11:26] LaNani Sundara: ah ok
[11:26] Ubit Umarov: its a ubuntu 10 thing
[11:26] Ubit Umarov: if we try to upgrade it, we may lose it :p
[11:26] Bill.Blight Mono 5.x really has better memory managment and is compatible to a higher level of .NET
[11:26] LaNani Sundara: ok i understand
[11:26] LaNani Sundara: ts just that one machine thats the issue
[11:27] Ubit Umarov: but well im not a linux sysadm.. so can't do much on that
[11:27] Bill.Blight I think there is also some resistance to moving to a newer .NET codebase, as some people who are very LOUD would also need to upgrade
[11:27] Sheera Khan: well, do it anyways...
[11:27] LaNani Sundara: upgrade from win 7
[11:27] LaNani Sundara: :D
[11:28] Ubit Umarov: " Project opensim » mono- build #5869: SUCCESS in 8 min 59 sec:"
[11:28] Bill.Blight I cringe everytime I see that Ubit
[11:28] Ubit Umarov: well mono 2.10... not really good anymore
[11:28] TG.Lucan Maybe a 12 months notice would soften the blow :)
[11:29] LaNani Sundara: seems mostly just a funding issue?
[11:29] LaNani Sundara: to get a better test machine
[11:29] Ubit Umarov: well httptests has 4.6 as target.. may change to 4.6.2 ( losing vista suport )
[11:29] Andrew.Hellershanks Could the machine run 3.2.8 at least?
[11:29] Ubit Umarov: mono 3 is also useless now
[11:29] Sheera Khan: Anyone using Vista around?
[11:30] LaNani Sundara: Vista... i hope not
[11:30] Ubit Umarov: most 4,x suports .net4.5 only. and most are bad
[11:30] Andrew.Hellershanks Perhaps but it would still be newer than 2.10
[11:30] TG.Lucan Mono master at 5.13
[11:31] LaNani Sundara: master builds and runs fine on uptodate mono
[11:31] Ubit Umarov: what?? was 5.10 a few days ago
[11:31] Ubit Umarov: lol
[11:31] Sheera Khan: Support for MS Vista ended on 11th of april last year
[11:31] Ubit Umarov: well you need to fix prebuid to use msbuild
[11:31] TG.Lucan odd number are pre release
[11:31] Kayaker Magic: msbuild does not like our makefiles.
[11:32] Sheera Khan: anyone using it still will have a lot of security issues by now and most probably a locked down computer due to some ransomware^^
[11:32] Ubit Umarov: httptests has a hacked prebuild that does msbuild projects fine ( hacked bc was not a proper fix :p )
[11:32] LaNani Sundara: i use monodevelop
[11:32] LaNani Sundara: GUI
[11:32] Ubit Umarov: well what made me move on from 4.0 was a security issue on .net
[11:32] TG.Lucan you can manually preset prebuild.xml to 4.5 and things should work
[11:32] Ubit Umarov: i mean .net4.0
[11:32] Kayaker Magic: The documentation still says you should use nant for testing. But it does not work.
[11:33] Ubit Umarov: docs are for mono 2.10 :p
[11:33] TG.Lucan you nay need to clean up old assemblies, manually though
[11:34] Kayaker Magic: My observation on projects like this are that the makefiles are a write-only language that seems to have more code in them than the project itself.
[11:34] Ubit Umarov: master does not even compile on some mono installs if you do not tell tools to use .net4.5
[11:34] Ubit Umarov: bc .net4.0 is no longer suported
[11:35] Ubit Umarov: but that was a issue on some installs.. that did only install things for .net4.5 comp
[11:35] Ubit Umarov: guess master does compile on mono5x using xbuild
[11:36] TG.Lucan Win 10 complains for 4.6, likes 4.6.1
[11:36] Kayaker Magic: Yes, but it gets ~150 warning errors. I HATE THAT!
[11:36] Sheera Khan: it would be nice to have a working description of the newer build environment
[11:36] Sheera Khan: not one about mono 2.10^^
[11:36] Ubit Umarov: hmm mine does not complain about 4.6
[11:36] LaNani Sundara: i use VS 2017 on win 10
[11:36] Ubit Umarov: but win10 does like 4.7.x now :)
[11:37] Bill.Blight get 3 warnings on proper mono version and .NET version on linux ..
[11:37] Bill.Blight on httptests
[11:38] Kayaker Magic: Most of the warnings are about linking multiple copies of object files, i.e. bugs in the makefile.
[11:38] Misterblue Waves: we're probably supposed to change everything to '.netstandard2.0'
[11:38] Ubit Umarov: yeah i did clean up some warnings on httptests
[11:38] Kayaker Magic: The rest of the warnings are about variables set but never used.
[11:39] LaNani Sundara: master sure has plenty warnings... unused variable here and whatever there
[11:39] Ubit Umarov: hidding warnings is not a good idea.. some are really valid :)
[11:39] TG.Lucan TRue Mr Blue
[11:39] Bill.Blight yep most of those are gone in httptests
[11:39] LaNani Sundara: :) nice clean up then
[11:39] Ubit Umarov: 2 soon to think about .netstandard2.0
[11:39] Ubit Umarov: and that is a subset of 4.7.x
[11:40] Bill.Blight Core needs a upgraded dev box before can cleanly work current let alone future standards
[11:40] Ubit Umarov: and actual code revison to really change a framework is .. a pain
[11:40] Ubit Umarov: there is no backwards comp anymore
[11:41] Ubit Umarov: at least on same things..
[11:41] TG.Lucan Good that people do experiment with current versions, avoids nasty surprises down the road
[11:41] Ubit Umarov: and MS seems to had deleted all code migration informatiion from older versions
[11:41] Bill.Blight There comes a time when in war(coding) you need to burn the bridges to keep the enemy from following you ..
[11:42] Andrew.Hellershanks When master is compiled with mono 5.10 it spits out 52 warnings. A far cry from the days I remember seeing around 300 warnings.
[11:42] LaNani Sundara: its 2018... time to stop worrying about win and linux and mono version from the previous century
[11:42] Bill.Blight ^^^^^
[11:43] Sheera Khan: Bill - the Klingon code warrior ;-)
[11:43] Ubit Umarov: well i do like machines and code from last century.. but there is a point when we must change :(
[11:43] Bill.Blight IMHO, a lot of the same people that scream that opensim is going nowhere are the same ones holding it back ..
[11:43] TG.Lucan I see 138 warnings compiling the current master on mono 5.x, and the roslyn compilers
[11:43] Ubit Umarov: well for any new users, what hold us if the LOOK
[11:43] Ubit Umarov: this viewers rendering is getting outdated :(
[11:44] Bill.Blight Yep
[11:44] LaNani Sundara: i am new, basicly, 0.9.1 was my first opensim version
[11:44] LaNani Sundara: :D
[11:44] Ubit Umarov: ( yes just seen a game i play upgrade to a new rendering engine :) )
[11:44] Ubit Umarov: uff dam good visual now on all games
[11:44] Bill.Blight One of the reasons I forked httptests, was just to give Ubit a stroke, when he sees my changes ...
[11:45] Ubit Umarov: of course my cpu and gpu does smoke :)
[11:45] Ubit Umarov: gpu at 100% all its 2GB internal ram in use :)
[11:46] Ubit Umarov: you forked to have your own funny crashs :p
[11:47] Bill.Blight hey I crash very rarely these days ... I have one sim that has been running for over a month and it is one of my vehicles sims that I abuse ... LOL
[11:47] Ubit Umarov: ( but that new game was not devel by amateurs.. it took them at least one year to upgrade some 20 maps ( == regions ) )
[11:48] Bill.Blight Plus those games have usually a gig of preloaded resources, not everything streamed over the wire
[11:48] Ubit Umarov: lol
[11:48] Ubit Umarov: 30GB this one :)
[11:48] Bill.Blight haha
[11:48] Bill.Blight Point proven
[11:48] Bill.Blight :P
[11:49] Ubit Umarov: sorry 38.8GB now
[11:49] Bill.Blight It is much easier to render "Pretty" if you know what you are going to need to render in advance ..
[11:49] Ubit Umarov: and loading a region, my disk makes it freeze for a few seconds
[11:50] Ubit Umarov: filling the 2GB of the gpu ram
[11:50] Ubit Umarov: ppl now tells that to play games one needs to have ssd
[11:50] Ubit Umarov: but well details unrelated to us..
[11:50] Andrew.Hellershanks oh, I misread the 2GB reference earlier. I thought Ubit was saying his computer only had 2GB of memory.
[11:50] Bill.Blight That is also what a lot of people that talk about Opensimulator performance vs other games, is we stream EVERYTHING over the wire
[11:51] Bill.Blight They just don't grasp that
[11:51] Ubit Umarov: no andrew. its the GPU internal RAM .. it uses it all
[11:51] Andrew.Hellershanks Ubit, yes. I did figure that out after I re-read the earlier comment.
[11:51] Ubit Umarov: and using all the tricks in the book
[11:52] Ubit Umarov: distance dependent terrain tesselation etc etc etc
[11:52] Ubit Umarov: ie closer terrain as more vertices etc
[11:52] Ubit Umarov: so any kid look to those games, will not stay here
[11:53] Ubit Umarov: in fact even the last version was better than this :(
[11:53] Ubit Umarov: since ll now has sansar.. not sure they will improve the rendering engine
[11:53] Andrew.Hellershanks We aren't likely to see a better rendering engine anytime soon. LL would be keeping something for Sansar/
[11:53] Bill.Blight Yeah but that is part of the issue, Opensimulator is not a "Game" it is a simulation, and primarily a social one at that, it is not a "run and gun" game ..
[11:54] Ubit Umarov: does not matter its a game or not
[11:54] Bill.Blight I know
[11:54] Ubit Umarov: look matters :(
[11:55] Bill.Blight we are using 10 year old + opengl rendering in the viewers, not a ton the server side can do about that
[11:55] Ubit Umarov: well and with skyrim you can have girls has you like
[11:55] Ubit Umarov: err never mind :)
[11:55] Ubit Umarov: ( skyrim has allows for mods with more advanced texturing and models than the stock one )
[11:56] Ubit Umarov flüstert: ( so ppl add special looking girls npcs etc to it :) )
[11:56] TG.Lucan Ubit, can you look at the last commit to master, just seems to delete a line,
[11:56] Andrew Hellershanks: Before we start losing people just a quick update about the birds module.
[11:57] Bill.Blight right, but yet again , that is all client side ... with gigs of preloaded data, and a much better rendering engine, not much that can be done server side to fix that here ... Would need a brand new viewer with a new render engine ..
[11:57] Ubit Umarov: errr
[11:57] Andrew Hellershanks: I made some more updates to it this past week. It is getting its settings properly now. Unfortunately, I also found another few issues with it.
[11:57] Ubit Umarov: damm
[11:58] Ubit Umarov: ok will fix it
[11:58] TG.Lucan kk
[11:58] Andrew Hellershanks: I may have to leave it for a while as I have some other projects to work on. I will send Jak a pull request to pick up my latest set of changes.
[11:59] Andrew Hellershanks: I'll also file an issue or two against my fork so to remind myself of one bug along with some other enhancement I want to make to the module.
[12:00] Andrew Hellershanks: Me speak english real good. :P Seem to be having some trouble typing today. :)
[12:01] Andrew Hellershanks: We are at the top of the hour. Any last minute items?
[12:01] Ubit Umarov: ok pushed the right source file this time
[12:01] TG.Lucan k Ubit
[12:02] Kayaker Magic: I'm still warming up to submit a patch for the arc trig functions. Will bug Andrew for more help with that.
[12:02] Andrew Hellershanks: Oh, hey Misterblue. You were hiding behind the chat window I had open in my viewer.
[12:02] Andrew Hellershanks: Kayaker, np
[12:03] Misterblue Waves: been lurking :)
[12:03] Andrew Hellershanks: That's a wrap for this weeks meeting. See you again next week.
[12:04] Ubit Umarov: kay
[12:04] Andrew Hellershanks: I had a machine crash again because Robust chewed up all the system memory.
[12:04] Ubit Umarov: most likely il not accept such a patch
[12:04] Ubit Umarov: those resolution issues should be handled on the scripts where and only where needed
[12:04] Andrew Hellershanks: I'm going to get the machine up to mono 5 and see if the memory usage improves at all.
[12:04] Ubit Umarov: code as is, just follows c#
[12:06] Misterblue Waves: how is it this meeting ended up at Hurliman Plaza?
[12:06] Ubit Umarov: well region had a rename also :)
[12:06] Andrew Hellershanks: Not sure. Ubit didn't like the lag at Wright Plaza so we wound up moving to Projet Secret. After that we wound up here.
[12:06] Ubit Umarov: this is Project Secret renamed i think
[12:07] Andrew Hellershanks: The region completely changed its look from one week to the next if that is the case.
[12:07] Ubit Umarov: still not the best for meetings.. but ok
[12:07] Ubit Umarov: just no sense on having working meetings on a 35K prims region
[12:07] Andrew Hellershanks: Perhaps the only good region for you would be an empty region except for a few chairs. :)
[12:08] Ubit Umarov: yeap almost yes
[12:08] Ubit Umarov: we want ppl to came and not lose time downloading things
[12:08] Bill.Blight and really need a region with build and scripting , so if we need to test or demonstrate anything during the meetings , we have that ability .. .then just restore the region to defaults before the next
[12:09] Ubit Umarov: yeap
[12:09] Andrew Hellershanks: The only items that take ages to rez tend to be these chairs behind me.
[12:09] Ubit Umarov: the lighter the best
[12:09] Andrew Hellershanks: We could look for a sandbox to hold the meeting.
[12:10] Bill.Blight Is it so hard to find someone to keep a region up for the dev meetings?
[12:10] Ubit Umarov: well if not for organization issues, we should to this on several grids also :)
[12:10] Ubit Umarov: but just impossible to schedule
[12:10] Bill.Blight I suggested that a while back ubit, but as you say, nightmare to arrange
[12:11] Ubit Umarov: yeap
[12:11] Andrew Hellershanks: Um... you mean more than one meeting a week each in a different grid?
[12:11] Bill.Blight no
[12:11] Bill.Blight the weekly meeting but on different grids
[12:11] Bill.Blight so it is a floating meeting
[12:11] Andrew Hellershanks: How would that cause a scheduling problem?
[12:11] Andrew Hellershanks: The problem is for people to find out where it is being held on any given week.
[12:12] Ubit Umarov: just summer time changes is a problem, imagine location also :)
[12:12] Bill.Blight people show up late or in the wrong spot now even though it is always in the same place, imagine if it was in different places
[12:12] Andrew Hellershanks: I know. That's what I was thinking.
[12:13] Bill.Blight have a static region here on OSGRID that has nothing but a hypergate to the current meeting location
[12:13] Andrew Hellershanks: We would need a device in a fixed location that points to the location for the meeting but that has its own headaches.
[12:13] Misterblue Waves: it would be like one of those HG tours
[12:13] Ubit Umarov: well if we had another machine, we could run our own region
[12:13] Bill.Blight Imagine that, maybe even call the region "Dev Meeting" how complicated ... LOL
[12:13] Andrew Hellershanks: :)
[12:13] Ubit Umarov: in fact that would be the best..
[12:14] Ubit Umarov: but even the main machine is a issue now :(
[12:14] Bill.Blight I have offered 3 times to put up a perm region here for the dev meetings, not offering again
[12:14] Ubit Umarov: well we are osgrid guests here.. so.. wel..
[12:15] Ubit Umarov: we did ask them for a region.. got this

Personal tools
About This Wiki