[11:04 AM] Andrew Hellershanks: Hello, everyone.
[11:04 AM] Andrew Hellershanks: Wow. Small group today. People must be on vacation or back at school.
[11:04 AM] Andrew Hellershanks: Hello, Lry.
[11:04 AM] Lyr.Lobo @cc.opensimulator.org:8002: Hello *grins*
[11:04 AM] Lyr.Lobo @cc.opensimulator.org:8002: Great to see you!
[11:05 AM] Lyr.Lobo @cc.opensimulator.org:8002: Yes school started today
[11:06 AM] Andrew Hellershanks: Yes. The CNE ended yesterday so that was always the indicator that summer is over and it is back to school.
[11:06 AM] Lyr.Lobo @cc.opensimulator.org:8002: And we picked a date for the OpenSimulator Community Conference - the first weekend in December, if that is good for you. December 6-7, 2025
[11:06 AM] Lyr.Lobo @cc.opensimulator.org:8002: I've been out sick, but we'll start working on the conference site soon
[11:06 AM] Andrew Hellershanks: I'll put that on my calendar
[11:07 AM] Lyr.Lobo @cc.opensimulator.org:8002: Thanks!
[11:07 AM] Andrew Hellershanks: The OSCC date, not you being out sick. :)
[11:07 AM] Ubit Umarov: :)
[11:07 AM] Lyr.Lobo @cc.opensimulator.org:8002: /me chuckles
[11:09 AM] Andrew Hellershanks: Speaking of vacations, someone appears to have been on one. I don't see any changes to the source tree being made this past week.
[11:10 AM] Ubit Umarov: oh? jezz lazy devs
[11:10 AM] Lyr.Lobo @cc.opensimulator.org:8002: /me grins
[11:11 AM] Vincent.Sylvester @hg.zetaworlds.com: I started work on writing migration and schema files for master, after a few months of testing it appears the system is solid enough. I have a large migration coming up which will give that another test though
[11:11 AM] Vincent.Sylvester @hg.zetaworlds.com: Convincing my brain to work on boring things isn't easy
[11:11 AM] Lyr.Lobo @cc.opensimulator.org:8002: /me steps away for a moment
[11:12 AM] Ubit Umarov: ( wonde how is weather in antarctica now...)
[11:12 AM] Andrew Hellershanks: Sounds good, Vincent. Yes, we do have to deal with the boring stuff every now and then so we can get to the fun stuff.
[11:13 AM] Vincent.Sylvester @hg.zetaworlds.com: None of it was really fun given the utter insanity that is the non-standard of sql
[11:14 AM] Vincent.Sylvester @hg.zetaworlds.com: Multiple rewrites of queries and routines just to deal with the weirdness of sqlite for example
[11:14 AM] Vincent.Sylvester @hg.zetaworlds.com: Along with a pesky bug in the map rendering which I have been unable to track down it's been a fun week
[11:15 AM] Vincent.Sylvester @hg.zetaworlds.com: Something about additive zoom level generation causes weird artifacts over time
[11:15 AM] Ubit Umarov: you mean, on your fork right?
[11:15 AM] Andrew Hellershanks: What sort of bug in map rendering?
[11:15 AM] Vincent.Sylvester @hg.zetaworlds.com: Don't you worry, you'll get the pleasure of integrating that at some point :P
[11:15 AM] Andrew Hellershanks: :)
[11:16 AM] Ubit Umarov: (hope iglo rent is cheap this time og year)
[11:16 AM] Vincent.Sylvester @hg.zetaworlds.com: I kept the original routine of adding new tiles to existing ones to reduce overhead, but somehow they end up rendering the water wrong, you can see spots on the map
[11:16 AM] Ubit Umarov: ( err hmm are there iglos in antarctica? not sure )
[11:17 AM] Vincent.Sylvester @hg.zetaworlds.com: It's strange as if it doesn't read the tile and assumes water color instead, but the tiles exist and doing a full refresh works perfectly
[11:17 AM] Vincent.Sylvester @hg.zetaworlds.com: I did add some file locking so I suspect it's either that or the routine to wait until all tiles are present isn't waiting long enough
[11:18 AM] Jagga.Meridith @hg.zetaworlds.com: (no too far north. the only thing up there is a research station)
[11:18 AM] Vincent.Sylvester @hg.zetaworlds.com: When the full refresh of 17000 tiles only takes 30 seconds I'm not even sure if I should worry about it or just fire that once in a while instead
[11:18 AM] Andrew Hellershanks: Is there no signal to let you know when all the tiles are present?
[11:19 AM] Vincent.Sylvester @hg.zetaworlds.com: I check the file modification time and if one is found that is newer than 5 minutes old I wait
[11:19 AM] Ubit Umarov: ( you mean other that the nazi base where hitler is living ?)
[11:19 AM] Andrew Hellershanks: /me nods
[11:19 AM] Andrew Hellershanks: I've done that type of file checking before.
[11:19 AM] Vincent.Sylvester @hg.zetaworlds.com: Course that needs refreshing as .net doesn't subscribe to inode or whatever to notice changes
[11:20 AM] Vincent.Sylvester @hg.zetaworlds.com: Putting maptile refresh on an hourly cron basically masks the problem now since that seems to work fine, but ideally the additive routine should work too since it's less work for the system
[11:22 AM] Vincent.Sylvester @hg.zetaworlds.com: We had some more trouble with libgdiplus as well
[11:23 AM] Vincent.Sylvester @hg.zetaworlds.com: After discovering the font rendering issue I rolled back to the previous builds that were public on apt repos from Canonical, but those builds have a massive memory leak
[11:23 AM] Vincent.Sylvester @hg.zetaworlds.com: So instead I went for the build published by mono, which is just slightly above that, but with the leak fixed and working font rendering
[11:23 AM] Vincent.Sylvester @hg.zetaworlds.com: Adjusted the wiki to note how one might need to install that now
[11:24 AM] Vincent.Sylvester @hg.zetaworlds.com: Canonical has yet to respond to the issue ticket
[11:24 AM] Vincent.Sylvester @hg.zetaworlds.com: Frankly not making a great case for their subscription stuff if they can't even fix such bugs within 6 months
[11:25 AM] Vincent.Sylvester @hg.zetaworlds.com: The fixed version does compile on ubuntu so there is always that as an option. I have yet to try the version winehq is maintaining
[11:25 AM] Andrew Hellershanks: libgdiplus is an ongoing source of pain.
[11:26 AM] Vincent.Sylvester @hg.zetaworlds.com: It just drives me mad when such obvious issues that are documented and even have suspect lines of code pointed just sit there to detriment of not just us, but apparently other software is broken because of this as well
[11:26 AM] Andrew Hellershanks: /me nods
[11:26 AM] Vincent.Sylvester @hg.zetaworlds.com: Canonical is not known for being all that active in fixing these things, prolly the third time in recent history some of their stuff was so old that it had bugs I ran into
[11:27 AM] Andrew Hellershanks: I feel your pain. It has affected me with another project.
[11:27 AM] Vincent.Sylvester @hg.zetaworlds.com: I tried the debian version, that's broken too and the dependencies it has are strange since they have no actual record in the build binary yet required them by package definition
[11:28 AM] Vincent.Sylvester @hg.zetaworlds.com: Getting libtiff5 and libgdiplus via wget and dpkg is easy enough for now to fix things
[11:28 AM] Vincent.Sylvester @hg.zetaworlds.com: If the winehq version works with font rendering there is always the option to make a .deb for that one or straight include the dll in the shipped binary
[11:29 AM] Vincent.Sylvester @hg.zetaworlds.com: Have not tested if this also breaks on mac os, but it might be broken there too
[11:31 AM] Vincent.Sylvester @hg.zetaworlds.com: Hearing from FS that they had reports of textures failing to load on the latest release, but so far apparently no reports from OS side of things on that. Latest I have is that there is some suspect code being narrowed down so might see a fix for that
[11:31 AM] Vincent.Sylvester @hg.zetaworlds.com: I tried reproducing that a couple days ago, but no luck
[11:32 AM] Vincent.Sylvester @hg.zetaworlds.com: Seems rather random and only happening when a lot of regions are loaded as well, something we no longer really see
[11:32 AM] Vincent.Sylvester @hg.zetaworlds.com: Maybe that's just my lower view distance setting, but I rarely see more than 5 regions loaded at once these days even if they just regular size
[11:34 AM] Ubit Umarov: now all are improving their regions using 2k textures everywhere, including nano prims
[11:34 AM] Ubit Umarov: :)
[11:34 AM] Vincent.Sylvester @hg.zetaworlds.com: It's apparently not overload of vram or whatever. I suspect more cache or network things, but that's just my gut feeling
[11:35 AM] Vincent.Sylvester @hg.zetaworlds.com: After the object cache thing I am distrustful of viewer caches xD
[11:35 AM] Ubit Umarov: well degradated cpus did error telling nor enough vram in same games
[11:36 AM] Ubit Umarov: but no idea "not load"can mean a lot of things
[11:36 AM] Vincent.Sylvester @hg.zetaworlds.com: Supposedly is hardware independent error, but yeah crossed my mind too
[11:36 AM] Vincent.Sylvester @hg.zetaworlds.com: Well they remain gray
[11:36 AM] Vincent.Sylvester @hg.zetaworlds.com: Some load, but most are gray
[11:37 AM] Ubit Umarov: and sl only opensim only or both ?
[11:37 AM] Vincent.Sylvester @hg.zetaworlds.com: From what I heard no reports on our end, so that makes me think it's network since we have slight differences there
[11:38 AM] Vincent.Sylvester @hg.zetaworlds.com: Terse updates and such
[11:38 AM] Ubit Umarov: what does "our end"mean in that context ?
[11:38 AM] Vincent.Sylvester @hg.zetaworlds.com: Supposedly reverting a lot of code from fixes seems to help so there seems to be a clue as to what it is
[11:39 AM] Ubit Umarov: and reverting code from what fixes?
[11:39 AM] Vincent.Sylvester @hg.zetaworlds.com: Well no reports related to OpenSim, only a problem at SL
[11:39 AM] Vincent.Sylvester @hg.zetaworlds.com: Viewer side code related to changes from last release
[11:39 AM] Ubit Umarov: ok ty
[11:41 AM] Vincent.Sylvester @hg.zetaworlds.com: I recall a similar problem years and years ago where textures would randomly stop loading, but that was around the time Hippo was a thing so don't remember what the problem was. Don't think it's related
[11:42 AM] Ubit Umarov: users may just became color blind
[11:43 AM] Vincent.Sylvester @hg.zetaworlds.com: Reject textures, embrace basic colors!
[11:45 AM] Lyr.Lobo @cc.opensimulator.org:8002: /me smiles
[11:48 AM] Ubit Umarov: ( no snoring here please)
[11:50 AM] Ubit Umarov: i was hoping Cuga could come to this meeting
[11:51 AM] Ubit Umarov: we have code contributions for Bullet on mantis, without apple silicon support
[11:51 AM] Ubit Umarov: maybe next week :)
[11:51 AM] Ubit Umarov: gavin also
[11:51 AM] Vincent.Sylvester @hg.zetaworlds.com: I did talk to him about that and he wants to poke mb to see if the changes to that are useful
[11:52 AM] Vincent.Sylvester @hg.zetaworlds.com: No point in building mac os version when it doesn't offer anything useful or breaks stuff in the process
[11:52 AM] Ubit Umarov: last seems added more threads to its use
[11:52 AM] Ubit Umarov: guess coder does not know that some run 100 regions per PI core
[11:53 AM] Ubit Umarov: well per box ( pi or not )
[11:53 AM] Vincent.Sylvester @hg.zetaworlds.com: Throwing all good sysadmin practices out the window isn't really something you can adjust for
[11:53 AM] Ubit Umarov: opensim is not a desktop game
[11:54 AM] Ubit Umarov: well all apps need proper balance of thread use
[11:54 AM] Vincent.Sylvester @hg.zetaworlds.com: True
[11:54 AM] Ubit Umarov: something dotnet only gets worse, if allowed
[11:55 AM] Vincent.Sylvester @hg.zetaworlds.com: Throwing more at it doesn't necessarily mean more performance when in the end it all has to sync up to build a frame
[11:55 AM] Ubit Umarov: wome forks did remove SmartThreadpool
[11:55 AM] Ubit Umarov: ignoreing that it does allow for better threading control
[11:56 AM] Ubit Umarov: and does not use ms thread managing p[arameters
[11:56 AM] Vincent.Sylvester @hg.zetaworlds.com: You can rip out and replace all sorts of things, most of that doesn't really offer any better experience given it then still has to run other slow code
[11:57 AM] Ubit Umarov: like only asking for a real new thread 2 seconds later
[11:57 AM] Ubit Umarov: hmm or 1 or something that high
[11:58 AM] Ubit Umarov: well that bullet will add its own manager from cpp libs
[11:58 AM] Ubit Umarov: so possible bullet will run on tyhe nice E-cores :)
[11:58 AM] Vincent.Sylvester @hg.zetaworlds.com: And you can hardly force these things either. I tried doing manual gc after map tile render to free that memory up quicker, but no amount of force gc stuff I added did anything, it just held onto that memory for 5 minutes and that's that
[11:59 AM] Ubit Umarov: you have no real way to tell gc to release anuthing
[11:59 AM] Vincent.Sylvester @hg.zetaworlds.com: That's become pretty clear now yeah
[12:00 PM] Ubit Umarov: force gc is only like ask 'can you please check for mem to release?'
[12:00 PM] Ubit Umarov: many times gc just ignores it
[12:00 PM] Vincent.Sylvester @hg.zetaworlds.com: Throwing everything but the kitchen sink at it and it appears more stubborn than me so I capitulate and just left a delay in there
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: Many thanks, dear friends, for a great meeting. We are thankful for your ideas and for keeping our dream alive. *hugs all around*
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: I must dash away to work and dream of space exploration. *winks*
[12:01 PM] Ubit Umarov: :)
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: Wishing you a lovely day! Pardon my interruption.
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: /me waves
[12:01 PM] Ubit Umarov: yeah rl also calls me...
[12:02 PM] Ubit Umarov: have a corvette stuck on a planet somewhere
[12:02 PM] Ubit Umarov: oops nm
[12:02 PM] Lyr.Lobo @cc.opensimulator.org:8002: Teaching classes and building corvettes...
[12:02 PM] Lyr.Lobo @cc.opensimulator.org:8002: /me chuckles
[12:02 PM] Lyr.Lobo @cc.opensimulator.org:8002: bon voyage!
[12:02 PM] Ubit Umarov: :)
[12:02 PM] Ubit Umarov: so any another for todya ?
[12:03 PM] Ubit Umarov: other issue i mean..
[12:04 PM] Ubit Umarov: ok cya next week if not before :)
[12:11 PM] Andrew Hellershanks: Time to wrap up todays meeting as we are no past the hour mark.
[12:11 PM] Andrew Hellershanks: now past
[12:11 PM] Andrew Hellershanks: Thank you all for coming. See you again next week.
