Chat log from the meeting on 2021-09-21

From OpenSimulator

Jump to: navigation, search
[11:01] Gavin.Hird Hi Vincent, Andrew, Jamie and Selby
[11:01] Andrew Hellershanks: Hello, everyone.
[11:01] Ubit Umarov: hi all
[11:01] Jamie.Jordan Hi everybody
[11:01] Selby.Evans hi everyone
[11:02] Gavin.Hird Hi Sefanie
[11:02] Stefanie.Ferdinand Hello everyone
[11:03] Selby.Evans hi Stefanie
[11:03] Andrew Hellershanks: Hello, Stefanie.
[11:06] Andrew Hellershanks: We might as well get started.
[11:06] Ubit Umarov: so what news from opensim things?
[11:06] Andrew Hellershanks: Someone has been busier making changes to the code this past week. :)
[11:06] Vincent.Sylvester So I have some positive news in that I have not heard any complaints regarding the changes to beginInvoke negatively impacting crossing performance or stability so that's good
[11:07] Vincent.Sylvester Remains the question of said method still being in various other places and dependencies shipped with OpenSim
[11:07] Ubit Umarov: begininvoke is jsut to use other threading scheme
[11:08] Ubit Umarov: many say it was removed on newer .net versions
[11:08] Ubit Umarov: so removed here
[11:08] Gavin.Hird For a second I thought you wrote beingwoke
[11:08] Ubit Umarov: similar :p
[11:09] Vincent.Sylvester I did a binary search and found mentions in various dll like the sql connectors and even more basic stuff
[11:10] Ubit Umarov: well that was the "fun" part i seen for example last mysql connector for using it
[11:11] Ubit Umarov: while at "web" all telling it was removed from .net core and .net5
[11:11] Ubit Umarov: well it is one of the possible forms to ask something to be done by another thread
[11:12] Ubit Umarov: guess we had all forms of doing it.. so one less
[11:12] Vincent.Sylvester Question is with .net you can, at least at the moment, still install older versions for backwards compatibility, but how does that translate to mono or, well, if mono even survives and isn't axed in favor of a .net nix version
[11:12] Grid: Items successfully shared.
[11:12] Ubit Umarov: now it is just using a thread pool workjob like elsewhere
[11:13] Ubit Umarov: not really
[11:13] Ubit Umarov: .net 4.8 to the newer ones is a big break
[11:13] Ubit Umarov: on linux, mono will only do .net 4.x
[11:14] Ubit Umarov: wnything else is ms .dotnet something
[11:14] Ubit Umarov: oyu need to install .net5 sdk etc etc
[11:14] Eldovar.Lamilton Hello everyone
[11:14] Ubit Umarov: 2 diferent things
[11:14] Andrew Hellershanks: Hello, Eldovar.
[11:14] Gavin.Hird Hi Eldovar
[11:14] Ubit Umarov: in fact seems ms is just doing a big mess breaking things apart
[11:15] Selby.Evans Hi Eldovar
[11:15] Stefanie.Ferdinand Hi Eldovar
[11:16] Ubit Umarov: migration of opensim to .net5 will mean it would not work on several platforms
[11:16] Ubit Umarov: i mean using .net unique new things like mathf
[11:17] Ubit Umarov: ofc having opensim to actually "compile" on .net5 will be a pain
[11:17] Gavin.Hird so in reality Win only?
[11:17] Ubit Umarov: XEngine will just not work...
[11:17] Andrew Hellershanks: Ubit, that is one way to force a migration to YEngine. ;)
[11:17] Ubit Umarov: dynamic textures and map will requeire the use to external library
[11:18] Ubit Umarov: bc bitmap and things are only supported on .net5-windows
[11:18] Vincent.Sylvester It worries me that I have been unable to find clear information as to the future of mono as of late, it seems the .net Enterprise thing backfiring now means radio silence
[11:19] Ubit Umarov: well don't see .net5 or 6 a thing for us on the next 2 or 3 years
[11:19] Ubit Umarov: acording to ms ppl it is dead
[11:19] Ubit Umarov: it is only .net framework 4.8
[11:19] Ubit Umarov: so will only get fiexes
[11:20] Ubit Umarov: ofc mono is still owned by a foundation with more than ms
[11:20] Vincent.Sylvester That's my hope actually, because if it was to be completely abandoned without receiving updates and security that could be very bad
[11:20] Ubit Umarov: well ms did a big push on it
[11:20] Ubit Umarov: mono 6 is just the best mono ever was
[11:21] Ubit Umarov: but ms ppl said so many things till now
[11:21] Ubit Umarov: guess we need to wait even more :)
[11:22] Vincent.Sylvester Also it is open after all so can be community maintained, question there is only how far that goes which is based on how many "need" mono to stay in the end
[11:22] Ubit Umarov: on .net core 1 and 2 the things they said mean that even current win apps  would not run without deep rewrite
[11:22] Ubit Umarov: they changed a lot of that adding a ton of 4.8 to net5
[11:23] Ubit Umarov: (even bitmap on windows )
[11:23] Ubit Umarov: but appdomains ( that xengine wants) etc, gone for good
[11:24] Vincent.Sylvester From what I gather not many really notice the change to Y from X, the other way round can be quite annoying though
[11:25] Ubit Umarov: if you use full Y features you can't go back to X
[11:25] Ubit Umarov: even if you do proper lsl in some cases
[11:25] Vincent.Sylvester llsay I think not accepting ints in X is one thing
[11:25] Ubit Umarov: like the basic normal use of llSleep
[11:26] Vincent.Sylvester That too
[11:26] Eldovar.Lamilton isn't llSleep considered harmful anyways?
[11:26] Ubit Umarov: only on Xengine
[11:27] Vincent.Sylvester Timing in Y can be a bit odd when using it, but that's mostly by design and just hints at bad script design
[11:28] Ubit Umarov: timing is never exact on scritps
[11:28] Vincent.Sylvester Far less issues getting complex things to work and time right in Y when done to LSL spec, which is the whole idea behind it
[11:28] Ubit Umarov: well timing is never exact on windows or linux
[11:29] Ubit Umarov: for correct timing there are special operating systems RTOS
[11:29] Ubit Umarov: real time operating systems
[11:29] Ubit Umarov: linux can be compiled with options ot improve that
[11:29] Gavin.Hird macOS has some very precise and high resolution timers, which is why it was/is so good for midi applications
[11:30] Ubit Umarov whispers: but a LOT worse on Y or X
[11:30] Ubit Umarov: things will happen.. one day..
[11:30] Vincent.Sylvester The fun doesn't begin until your clocksource cannot raise tsc anymore and everything gets real wonky
[11:31] Ubit Umarov: raise tsc ?
[11:31] Vincent.Sylvester kernel tries to dial in clocksource, falls back to idk what when it cannot get tsc to be stable, but without tsc mono and to that extend OpenSim just doesn't behave
[11:32] Vincent.Sylvester Had that once, was fun
[11:32] Ubit Umarov: tsc is a cpu register
[11:32] Ubit Umarov: in older ones had a HUGE flaw..
[11:32] Ubit Umarov: each core had one..
[11:32] Ubit Umarov: its frequency was per core .. etc..
[11:33] Ubit Umarov: totally useless
[11:33] Vincent.Sylvester by default hardware clocksource from cpu, if that don't work it's software clock I think, but that ends up messing up a lot, causing apps to fail
[11:33] Ubit Umarov: that got improved
[11:33] Ubit Umarov: but also now all MBs have another clock
[11:33] Ubit Umarov: around 10Mhz
[11:33] Ubit Umarov: that can be used for timing
[11:34] Ubit Umarov: that is what StopWatch uses
[11:34] Ubit Umarov: the ibm pc like many others was not made for real time apps
[11:35] Ubit Umarov: time precision was around 10ms
[11:35] Ubit Umarov: 15.6 on windows
[11:35] Ubit Umarov: arounf 1ms on most linux
[11:36] Andrew Hellershanks: I seem to remember something like 55 ticks per second or ~18mS.
[11:36] Ubit Umarov: windows can also be made to be 1ms precision
[11:36] Ubit Umarov: but will burn a lot more cpu
[11:36] Ubit Umarov: it is a deep hidden setting
[11:36] Vincent.Sylvester As with everything useful in windows
[11:37] Ubit Umarov: well you all know that windows has more configuration options than linux right?
[11:37] Ubit Umarov: :)
[11:37] Ubit Umarov: just all well hidden
[11:37] Ubit Umarov: well on last week code changes
[11:38] Ubit Umarov: made a few changes around llTeleport and osTelepor
[11:38] Ubit Umarov: like not try to teleport child agents.. duhhh
[11:39] Ubit Umarov: the osteleport(agent, pos, llokat ) only did work on same region
[11:39] Ubit Umarov: it now can work anywhere
[11:40] Ubit Umarov: like pos = <-10,10,20> will tp you to the region to west of curretn
[11:40] Vincent.Sylvester Welcome changes that made teleports more stable, still glitches happen due to god knows what's going on, in a few hundred attempts I could find no patterns, but much less with said changes which is nice :)
[11:40] Ubit Umarov: :)
[11:41] Ubit Umarov: that means that, that strange dataserver read of lms can be used directly with this last variant
[11:41] Ubit Umarov: well after checking the string is really a vector
[11:42] Eldovar.Lamilton Teleporting to the west of current means, only for adjacent regions, right?
[11:42] Ubit Umarov: it can be very far
[11:42] Eldovar.Lamilton oh wow
[11:42] Vincent.Sylvester If there is no region at your destination then it won't teleport you
[11:42] Ubit Umarov: but pos must be a position in local region coords
[11:43] Ubit Umarov: ie its <0.0.0> origint is current region reference corner
[11:44] Ubit Umarov: ie left lower corner on map
[11:45] Ubit Umarov: also did some fixes on llTeleport to it should do HG teleports from landmarks
[11:45] Ubit Umarov: or do them better
[11:46] Ubit Umarov: follow ll rules, it will not work on temporary attachments
[11:46] Ubit Umarov: following..
[11:46] Ubit Umarov: guess that was a security issue
[11:47] Vincent.Sylvester I took some time this week looking at the wiki, the grid list page lists about 10 or so grids that no longer exist, domains expired and everything. Spoke to Melanie about it and the consensus was that should domains expire and grids be offline with certainty to move them into the inactive listing as well as removing all hyperlinks from the listings themselves to prevent the wiki being blacklisted for potentially linking to websites engaging in phishing when those expired domains get registered again. Would be nice if someone could give that list a look over at some point, I don't have a wiki account myself :)
[11:48] Ubit Umarov: not easy to keep a active grids list
[11:48] Ubit Umarov: hypergridbusiness does keep one i think
[11:48] Ubit Umarov: and few other projects
[11:48] Vincent.Sylvester I keep one based on just basic ping requests to the IPs, but that alone isn't a guarantee of course
[11:49] Ubit Umarov: yeah
[11:49] Ubit Umarov: some grids did died and returned on a totally dif dns
[11:50] Vincent.Sylvester I'm just mentioning it since I found it :)
[11:50] Eldovar.Lamilton congrats, Vincent :-)
[11:50] Ubit Umarov: yeah guess we have it somewhere on wiki
[11:50] Ubit Umarov: grid owners where suposed to keep that updated
[11:51] Ubit Umarov: on the open spirit of wiki
[11:51] Ubit Umarov: that did not consider free porn pub uses etc
[11:52] Vincent.Sylvester Well on expired domains the consensus was to move them to prevent bad links
[11:52] Vincent.Sylvester Google has gone haywire lately blacklisting lately, guess to boost their ad revenue
[11:53] Gavin.Hird friends don't let friends use Google
[11:53] Ubit Umarov: so more opensim news?
[11:53] Andrew Hellershanks grins at Gavin
[11:53] Ubit Umarov: :)
[11:53] Andrew Hellershanks: Ubit, I think you have pretty much covered it for this week regarding the code side of things.
[11:54] Andrew Hellershanks: Anything from you, Gavin, on the viewer side?
[11:54] Gavin.Hird I'll probalby release another viewer version this week
[11:54] Gavin.Hird the LL TPV meeting was interesting to say it the least
[11:54] Gavin.Hird they have hires some new tech / product manager
[11:54] Ubit Umarov: oh i told you about the viewer side bug on requestparcel?
[11:54] Gavin.Hird an ex Microsoft type
[11:54] Gavin.Hird not particularly interested in talking open source at all
[11:55] Gavin.Hird Mojo Linden
[11:55] Andrew Hellershanks thinks of a line from Austin Powers
[11:55] Ubit Umarov: ( i did a fs jira on it.. other viewers may have it )
[11:55] Gavin.Hird they are also adding a viewr type that is non interactive
[11:56] Gavin.Hird you can only log in and observe
[11:56] Ubit Umarov: ..
[11:56] Ubit Umarov: you get the argument posGlobal from:
LLVector3d pos_global = region->getCenterGlobal();
[11:56] Ubit Umarov: this wrong. The center of a region is not a region identifier, its left corner is.

so this must me
LLVector3d pos_global = region->getOriginGlobal();
[11:56] Ubit Umarov: ...
[11:56] Ubit Umarov: on some calls to requestRegionParcelInfo
[11:57] Ubit Umarov: ...
[11:58] Ubit Umarov: no idea where the use of region center did come from
[11:58] Gavin.Hird ok, bookmarked jira
[11:58] Ubit Umarov: wlel viewers...
[11:58] Ubit Umarov: :p
[11:58] Ubit Umarov: center is only usefull like to place the water and things like that
[11:58] Ubit Umarov: not as region reference
[11:59] Vincent.Sylvester I did get a report a few weeks ago of parcel crossing issues perhaps related to that, well I hope anyways, because that would mean is fixed now :)
[11:59] Ubit Umarov: HG and varaible size regions, always "magic"
[12:00] Andrew Hellershanks: We are at the top of the hour. Does anyone have a question/comment before we wrap up this meeting?
[12:00] Ubit Umarov: ( also on opensim code! ofc )
[12:00] Andrew Hellershanks: a final question/comment that is.
[12:00] Ubit Umarov: i actually did add extra code to compare region handle
[12:01] Ubit Umarov: a:handle == b.handle will not do it..
[12:02] Ubit Umarov: so still "vodoo" code abotu large regions, still on opensim. even
[12:02] Vincent.Sylvester Ubit, removing voodoo one bug at a time
[12:02] Ubit Umarov: well we try :)
[12:04] Ubit Umarov: ofc some broken lms outthere...
[12:04] Ubit Umarov: lms are, or where, basicly made by viewers
[12:04] Ubit Umarov: so some are just brokne
[12:04] Ubit Umarov: even broken
[12:04] Gavin.Hird then they must be broken ;-)
[12:04] Eldovar.Lamilton :-)
[12:05] Ubit Umarov: recent fs had jsut stopped putting proper descritption
[12:06] Ubit Umarov: wel the one we had.. not exactly "proper" either
[12:06] Ubit Umarov: but it just did remove the @ gridid
[12:06] Ubit Umarov: oh HG lms i mean
[12:06] Ubit Umarov: on..
[12:07] Andrew Hellershanks is offline.
[12:08] Eldovar.Lamilton oops
[12:08] Andrew Hellershanks is online.
[12:08] Gavin.Hird did Andrew carsh?
[12:08] Ubit Umarov: andrew cat did alt-f4 on his viewer
[12:08] Eldovar.Lamilton lol
[12:08] Eldovar.Lamilton wb Andrew
[12:08] Andrew Hellershanks: My internet connection went down for a moment. :P
[12:08] Ubit Umarov: ( ok cat tried to pull the cable instead )
[12:08] Andrew Hellershanks: What did I miss?
[12:08] Eldovar.Lamilton stretched your leg and hit the power button
[12:08] Gavin.Hird the-sky-is-falling moment
[12:09] Andrew Hellershanks: Can't blame the cat. She isn't anywhere near me.
[12:09] Ubit Umarov: ok cat is on the post eating the fiber connection
[12:09] Eldovar.Lamilton byte by byte
[12:09] Andrew Hellershanks: hehe
[12:09] Ubit Umarov: :)
[12:10] Vincent.Sylvester And then Level3 is down, as is tradition
[12:10] Andrew Hellershanks: Final call for any last minute questions/comments before we wrap up.
[12:10] Gavin.Hird I friend of min's rabbit removed the insulation of a live 220V core over a lenght of 2 metere without getting zapped
[12:10] Gavin.Hird mine''s
[12:11] Ubit Umarov: oh osgrid his having a fundraise week
[12:11] Ubit Umarov: events at well event plaza
[12:11] Andrew Hellershanks nods
[12:11] Vincent.Sylvester Start employing it, clearly it wants to be an electrician lol
[12:11] Gavin.Hird :-)
[12:11] Ubit Umarov: will have auction next week end
[12:11] Jamie.Jordan have a great week yall
[12:11] Ubit Umarov: auctions
[12:12] Andrew Hellershanks: Bye, Jordan.
[12:12] Ubit Umarov: cya jordan
[12:12] Selby.Evans must go -- bye all
[12:12] Andrew Hellershanks: Bye, Selby.
[12:12] Ubit Umarov: cya Selby.Evans
[12:12] Gavin.Hird bye Selby
[12:12] Grid: Eldovar.Lamilton gave you possibly missed conversation 2021-09-21.
[12:13] Ubit Umarov: so more news?
[12:13] Vincent.Sylvester I wrote more ossl functions this week for the fun of it
[12:13] Vincent.Sylvester even documented them o_o
[12:13] Ubit Umarov: Oh dear.. :9
[12:13] Eldovar.Lamilton wow!
[12:14] Ubit Umarov: life on zeta must be fun...
[12:14] Ubit Umarov runs and hides
[12:14] Eldovar.Lamilton where are those good ole times when someone asking for documentation got the answer RTFS ...
[12:14] Ubit Umarov: aks me for doc :p
[12:15] Gavin.Hird ask LL for viewer documentation
[12:15] Vincent.Sylvester lol
[12:15] Eldovar.Lamilton :-)
[12:15] Gavin.Hird pretty much RTFS
[12:15] Ubit Umarov: well i did update out wiki
[12:15] Vincent.Sylvester Yeah between you and phil things stay up to date pretty well
[12:15] Gavin.Hird even aks them for documentation how to build their viewer
[12:16] Ubit Umarov: changed
[12:16] Ubit Umarov: even did add a script.. jezz
[12:17] Vincent.Sylvester Wiki is better than it gets credit for a lot of things
[12:17] Eldovar.Lamilton right
[12:17] Ubit Umarov: ( no one did notice "A teleport to other region has a 5 second delay" so far :p )
[12:17] Andrew Hellershanks: :)
[12:17] Vincent.Sylvester Well that's also not totally consistent as well so maybe that's why
[12:18] Vincent.Sylvester Plus some regions just... hate you and take their time anyways
[12:18] Ubit Umarov: kinda is now
[12:18] Ubit Umarov: and will add it to llTeleport also
[12:19] Ubit Umarov: ofc this delays are not that good on Xengine
[12:19] Ubit Umarov: details :p
[12:19] Gavin.Hird try walking across the border of 3 regions meeting up in a corner
[12:19] Ubit Umarov: didi it several times
[12:19] Ubit Umarov: ages ago
[12:19] Andrew Hellershanks: How about flying across?
[12:20] Gavin.Hird it might work
[12:20] Gavin.Hird but sometimes the avtar gets in a serious mess
[12:20] Ubit Umarov: tip: do not try with bullet
[12:20] Ubit Umarov: :p
[12:20] Vincent.Sylvester It works quite well, but oh boy when it fails it fails hard with a lot of fireworks
[12:20] Ubit Umarov: well opensim crossings and teleports..
[12:20] Ubit Umarov: work fine until they don't.. then well a mess
[12:21] Vincent.Sylvester Wrote a database search to retrieve back the whatever people lost
[12:21] Vincent.Sylvester Don't drive what you cannot lose
[12:22] Vincent.Sylvester take copy is a button not pressed enough
[12:23] Andrew Hellershanks nods
[12:23] Andrew Hellershanks: Time to wrap up todays meeting. Thank you all for coming. See you again next week.
[12:24] Eldovar.Lamilton Thank you for chairing, Andrew
Personal tools
About This Wiki