[11:03 AM] Andrew Hellershanks: Hello, everyone.
[11:03 AM] Gavin.Hird @grid.xmir.org:8002: Hi Andrew
[11:03 AM] Ubit Umarov: hi
[11:03 AM] Gavin.Hird @grid.xmir.org:8002: maybe he is hinting at the new embedded browser
[11:03 AM] Andrew Hellershanks: I see we finally cobbled together enough pennies to hire a new bartender.
[11:03 AM] Gavin.Hird @grid.xmir.org:8002: which LL does a lot of work on right now, or threaded UDP
[11:05 AM] Vincent.Sylvester @hg.zetaworlds.com: I suspect it's just a word of caution to watch LL changes as they work on the renderer as well, that some fundamental approach to that isn't changing in a way you end up having to redo something. I doubt it's catastrophic
[11:05 AM] Gavin.Hird @grid.xmir.org:8002: two viewer version posted today at https://www.dayturn.com/viewer/index.php?resources/ 3.0.2 has the goodies
[11:05 AM] Gavin.Hird @grid.xmir.org:8002: we can't do anythign with the renderer till all UI is moved out of the GL view
[11:06 AM] Ubit Umarov: :)
[11:07 AM] Vincent.Sylvester @hg.zetaworlds.com: Speaking of viewers. Past week fixed a bunch of bugs in FS with TJ doing most of the work. Newest mac os beta saw multiple icons show in the dock, because the plist specifics were not set correctly and apparently Apple now cares about standards. Just today fixed the bug on world map clicking var regions would not load the region information
[11:08 AM] Vincent.Sylvester @hg.zetaworlds.com: The ??? name thing has seemingly been fixed proper, hasn't shown up all week so that's looking fixed
[11:08 AM] Vincent.Sylvester @hg.zetaworlds.com: Still working on the cache thing all week I did find a sort of fix yesterday, which just raises more questions though
[11:09 AM] Gavin.Hird @grid.xmir.org:8002: great. I added iCloud sync of usernames and passwords per grid to the viewer published today
[11:09 AM] Vincent.Sylvester @hg.zetaworlds.com: It seems as if the way bakes are handled is timing dependent in some way that triggers rebakes while the viewer is still sending baked textures as it seems to query or receive bad textures during the same time
[11:10 AM] Ubit Umarov: well vicent is talking about assets cache
[11:10 AM] Vincent.Sylvester @hg.zetaworlds.com: Why that doesn't happen with Flotsam is confusing the heck outta me
[11:11 AM] Ubit Umarov: he is been working on a version to improve cahce sharing by several regions
[11:11 AM] Ubit Umarov: and having same issues, not visible on main code
[11:12 AM] Gavin.Hird @grid.xmir.org:8002: ok
[11:12 AM] Ubit Umarov: and yes bakes have always been temperamental
[11:12 AM] Vincent.Sylvester @hg.zetaworlds.com: It more and more seems like a case of existing code masking a bug, but I don't yet know where exactly
[11:12 AM] Gavin.Hird @grid.xmir.org:8002: how does it manifest?
[11:12 AM] Ubit Umarov: and for some reason.. well easy to understand.. only use udp texture fecthc
[11:13 AM] Vincent.Sylvester @hg.zetaworlds.com: The cache I implemented initially did all in memory, so as fast as it can be. That still looped like 3 bakes and the actual textures the server side got never worked
[11:13 AM] Ubit Umarov: bet this days viewerr changes did degradate that fetch even more
[11:13 AM] Ubit Umarov: ( they can t use http on that, bc http is CDN
[11:13 AM] Vincent.Sylvester @hg.zetaworlds.com: It's like the viewer sends new bakes while already sending them and it trips over itself looping over and over until you kill 3 threads and then it stops looping, but also never gets correct textures
[11:14 AM] Ubit Umarov: ( and bakes do not 'fit there'
[11:14 AM] Ubit Umarov: specially with their servcer side bakes, etc etc
[11:14 AM] Vincent.Sylvester @hg.zetaworlds.com: I ended up adding a manual bake save in avatar factory to force the internally cached bakes to be stored. That strangely runs in between the caching of bakes according to log output
[11:14 AM] Ubit Umarov: also some content protection... their http is open... you know the uuid you get the content
[11:15 AM] Vincent.Sylvester @hg.zetaworlds.com: So there is some weird timing thing going, which given I made it as fast as possible with memory cache means the bug is not in the cache itself and Flotsam working is because it is slow, somehow
[11:15 AM] Vincent.Sylvester @hg.zetaworlds.com: Just doesn't really add up does it
[11:16 AM] Gavin.Hird @grid.xmir.org:8002: probably the same timeout mechanism we ran in to, identified and fixed
[11:16 AM] Cuga.Rajal @rajal.org:9000: viewer side problem?
[11:17 AM] Gavin.Hird @grid.xmir.org:8002: yes
[11:17 AM] Gavin.Hird @grid.xmir.org:8002: totally
[11:17 AM] Vincent.Sylvester @hg.zetaworlds.com: I'm not yet sure which side it is. I haven't found the part that signals to the viewer "yes I got all bakes thank you"
[11:17 AM] Vincent.Sylvester @hg.zetaworlds.com: It just seems to send and then much too quickly request them, seeing they are wrong and sends new ones overlapping the running send
[11:17 AM] Gavin.Hird @grid.xmir.org:8002: Claude read the opensim server code compared what it did to what the viewer did and waht opensim expected and fixed it
[11:18 AM] Ubit Umarov: actually viewer does ask abot cached bakes
[11:18 AM] Vincent.Sylvester @hg.zetaworlds.com: Why a different region side asset cache would impact that timing is what confuses me, since that's kinda independent from that and does not provide any feedback on when it has finished
[11:18 AM] Lyr.Lobo @cc.opensimulator.org:8002: /me smiles and waves
[11:18 AM] Vincent.Sylvester @hg.zetaworlds.com: All void functions
[11:18 AM] Ubit Umarov: hi Lyr.Lobo
[11:19 AM] Lyr.Lobo @cc.opensimulator.org:8002: Hello! 1st day of classes, pardon me for my distractions
[11:19 AM] Gavin.Hird @grid.xmir.org:8002: Hi Lyr
[11:19 AM] Lyr.Lobo @cc.opensimulator.org:8002: Great to see you *grins*
[11:19 AM] Vincent.Sylvester @hg.zetaworlds.com: So something signals the viewer to resend while it, seemingly, is still sending bakes already and then it never "sees" the correct data or determines something is missing and it needs to send the whole set again. It's utterly confusing
[11:20 AM] Vincent.Sylvester @hg.zetaworlds.com: Like I said the fix was to tell it to store the memory cached bakes
[11:20 AM] Vincent.Sylvester @hg.zetaworlds.com: Essentially runs a store on assets immediately rather than queuing it up
[11:21 AM] Andrew Hellershanks: Hello, Lry. They are starting classes about a month lather than I usually associate with start of classes.
[11:21 AM] Vincent.Sylvester @hg.zetaworlds.com: If I can find what seems to signal the viewer to send again and again I might get somewhere, something triggers it to do this, but I don't see the server sending a "bakes received" signal of any kind
[11:22 AM] Vincent.Sylvester @hg.zetaworlds.com: It's a bit like it sends the bake then immediately requests the asset for it and somehow it gets the wrong one or gets told it is missing, triggering rebake
[11:22 AM] Vincent.Sylvester @hg.zetaworlds.com: As if a foreach has the wrong sort order or something
[11:22 AM] Ubit Umarov: order to rebake is there.. somewhere
[11:23 AM] Ubit Umarov: also a viewer queiry abotu cached bakes
[11:23 AM] Ubit Umarov: a bit confusing process.. and yeah time critical
[11:24 AM] Ubit Umarov: mostly because how fs decides to ask for things withou waiting for current work to end
[11:24 AM] Vincent.Sylvester @hg.zetaworlds.com: I see that "Received texture update" and been trying to go backwards from what triggers that to see where the entry point is. Going forward from that point I see nothing that signals to the viewer that all bakes were received, so that's why I think it is the viewer "checking" if they all present and that somehow runs while it is still sending them over
[11:24 AM] Vincent.Sylvester @hg.zetaworlds.com: Sends the first few bakes, then requests them going in reverse so they are not found, rebake rinse and repeat
[11:25 AM] Vincent.Sylvester @hg.zetaworlds.com: Cause it's like the last 4 or 5 bakes are always labeled missing
[11:25 AM] Ubit Umarov: well spent hours fighting that years ago.. then so did Beq viewer side
[11:25 AM] Ubit Umarov: things are working with core cache.. somehow
[11:25 AM] Ubit Umarov: but i do see me as cloud longer now
[11:26 AM] Vincent.Sylvester @hg.zetaworlds.com: I'll probably dig into that some more with TJ when I get some willpower back. The explicit saving in avatar factory seems to work nicely and shouldn't impact anything negatively
[11:26 AM] Ubit Umarov: yeah and burn more disk
[11:26 AM] Vincent.Sylvester @hg.zetaworlds.com: Doesn't blow up with Flotsam, essentially just bypasses that queue to store and forces it directly
[11:27 AM] Ubit Umarov: flotsam is working.. so issue is most likely on the new code..
[11:27 AM] Vincent.Sylvester @hg.zetaworlds.com: If it is viewer side can fix it in FS, but that doesn't help over viewer versions any
[11:27 AM] Ubit Umarov: that needs to be well check, bf going mess up other part of code, still working
[11:27 AM] Vincent.Sylvester @hg.zetaworlds.com: Well what I haven't done is test Flotsam with memory cache enabled
[11:28 AM] Ubit Umarov: dont thing GC is good enought for that still
[11:28 AM] Ubit Umarov: but sure.. a test
[11:28 AM] Vincent.Sylvester @hg.zetaworlds.com: It's essentially what I implemented to debug, simple dictionary of id and asset data, stuffing the whole asset into memory and it looped still
[11:29 AM] Ubit Umarov: wel flotsam as a even faster hack to remeber assets in memory
[11:29 AM] Vincent.Sylvester @hg.zetaworlds.com: If Flotsam works on that then... yeah. Even so the cache should be independent from bakes, all it is supposed to do is store things, where doesn't really matter so long as it can retrieve them
[11:30 AM] Vincent.Sylvester @hg.zetaworlds.com: Flotsam file cache defers writing to disk, only updating weak reference cache
[11:30 AM] Ubit Umarov: c# has a odd thing that allows to get back memory that lost references, but still in memory waith for gc to free
[11:31 AM] Ubit Umarov: weak references...
[11:31 AM] Cuga.Rajal @rajal.org:9000: I have unrelated question
[11:31 AM] Cuga.Rajal @rajal.org:9000: Dotnet 8 support EOL is about a month away. Should we care?
[11:32 AM] Gavin.Hird @grid.xmir.org:8002: it is lie asking Windows 10 users if they care 😀
[11:32 AM] Ubit Umarov: no we should not care
[11:32 AM] Vincent.Sylvester @hg.zetaworlds.com: Win10 technically went eol a year ago, yet I'm still here. Microslop overestimates how quick things move or how stable their new shit is
[11:33 AM] Cuga.Rajal @rajal.org:9000: k
[11:33 AM] Ubit Umarov: version9 and 10 dont seem to give any relevant improvement and have breaking changes
[11:33 AM] Ubit Umarov: liek floatsam is dead on it
[11:34 AM] Vincent.Sylvester @hg.zetaworlds.com: Nonsensical breaking changes
[11:34 AM] Ubit Umarov: KFM.. and a few more things
[11:34 AM] Cuga.Rajal @rajal.org:9000: no improvements but will continue security fixes.. not sure if we should care about that
[11:34 AM] Ubit Umarov: yeah security fixes...
[11:35 AM] Vincent.Sylvester @hg.zetaworlds.com: Version 20 will just be a hello world written in python, can't have bugs without code
[11:35 AM] Ubit Umarov: great excuse to keep breaking your things
[11:35 AM] Ubit Umarov: ms is the virus this days :p
[11:35 AM] Cuga.Rajal @rajal.org:9000: Well I don;t use Windows... =)
[11:35 AM] Ubit Umarov: ( getting worse than norton possilbe )
[11:35 AM] Vincent.Sylvester @hg.zetaworlds.com: Removing stuff instead of fixing it and then having to re-invent the wheel goes against my compassion to help
[11:35 AM] Gavin.Hird @grid.xmir.org:8002: We used to call DOS a boot sector virus
[11:36 AM] Ubit Umarov: dotnet is ms, is case you missed ti :)
[11:36 AM] Vincent.Sylvester @hg.zetaworlds.com: Oh your knee hurts, let's cut the leg off!
[11:36 AM] Cuga.Rajal @rajal.org:9000: yeah thats as close as I get to them lol
[11:37 AM] Ubit Umarov: hmm didnt look what dotnet11 will break more
[11:37 AM] Cuga.Rajal @rajal.org:9000: so no real security fixes, just "security" removals
[11:37 AM] Gavin.Hird @grid.xmir.org:8002: will probably only run in the cloud
[11:38 AM] Vincent.Sylvester @hg.zetaworlds.com: Dictionaries exhibit different sort orders between windows and unix, so logically instead of changing the sort or allocation, let's just remove them all - Microslop
[11:38 AM] Ubit Umarov: well win11 is been turned into a cloud terminal
[11:38 AM] Gavin.Hird @grid.xmir.org:8002: yup
[11:39 AM] Vincent.Sylvester @hg.zetaworlds.com: Even the gutted version I been using isn't growing on me any
[11:39 AM] Ubit Umarov: hmm ll did add a new texture to PBR assets ?
[11:39 AM] Gavin.Hird @grid.xmir.org:8002: one or more?
[11:40 AM] Ubit Umarov: seen ref to one more
[11:44 AM] Cuga.Rajal @rajal.org:9000: RL is calling... see you all later
[11:44 AM] Ubit Umarov: have fun
[11:44 AM] Lyr.Lobo @cc.opensimulator.org:8002: bye Cuga
[11:44 AM] Andrew Hellershanks: Bye, Cuga
[11:46 AM] Ubit Umarov: ....
[11:46 AM] Ubit Umarov: The .NET 11 runtime includes: Updated minimum hardware requirements for x86/x64 and Arm64 architectures, requiring more modern instruction sets to improve performance and reduce maintenance complexity.
[11:46 AM] Ubit Umarov: ...
[11:46 AM] Ubit Umarov: that is another... all love to kill cpus that are fine
[11:47 AM] Gavin.Hird @grid.xmir.org:8002: of course: you have to buy a new machine. Reminds me about the slogan for Vista: "Ask not what Vista can do for you. Ask what you can buy for Vista"
[11:47 AM] Luisillo Contepomi: ^_^
[11:47 AM] Vincent.Sylvester @hg.zetaworlds.com: One other thing that came up in the bug fixing crusade was the inability to hg teleport to a region with the same coordinates. I'm don't remember if that was a bug on both sides or just a server side thing though. If it's something the viewer side can fix as well that might make it in if it can be found there
[11:48 AM] Vincent.Sylvester @hg.zetaworlds.com: I seem to recall it was something on both sides needing changes to handle this though
[11:48 AM] Vincent.Sylvester @hg.zetaworlds.com: Something with the regionhandle encoding
[11:49 AM] Gavin.Hird @grid.xmir.org:8002: does that collision occure often?
[11:49 AM] Ubit Umarov: that is a HG feature
[11:50 AM] Vincent.Sylvester @hg.zetaworlds.com: Given the default coordinates for regions are sort of hardcoded the collision is bound to happen unless that is randomized, even then if you happen to get lucky it'll fail
[11:50 AM] Ubit Umarov: viewers do user map coords as region identifier
[11:50 AM] Ubit Umarov: ( region handler is basicly map coords )
[11:51 AM] Gavin.Hird @grid.xmir.org:8002: for lack of a gid ID?
[11:51 AM] Gavin.Hird @grid.xmir.org:8002: grid
[11:51 AM] Ubit Umarov: they do not use uuid
[11:51 AM] Ubit Umarov: terhy are make for a single grid... so map coors is a good identifier
[11:52 AM] Ubit Umarov: made..
[11:52 AM] Vincent.Sylvester @hg.zetaworlds.com: That's what I'm wondering too. Is it just viewer not being hg aware in that regard
[11:52 AM] Gavin.Hird @grid.xmir.org:8002: ok, but the grid can send a unique identifier for the grid, no?
[11:52 AM] Ubit Umarov: no
[11:53 AM] Gavin.Hird @grid.xmir.org:8002: ok, but if there is a collision the viewer could ask the destination grid who it is
[11:53 AM] Ubit Umarov: well things do seem to work...
[11:53 AM] Vincent.Sylvester @hg.zetaworlds.com: If I remember correctly last time it came up. Something about needing to move actual grid coordinates to inflict a teleport that can jump to a different address and thus grid, since the regionhandle is the same it drops into teleport within region mode or something, but then the address is wrong so it fails
[11:54 AM] Ubit Umarov: but tps to regions with handlers the viewer has in memory my be fun
[11:54 AM] Ubit Umarov: i mean regions in memory with same handler... ir in view
[11:55 AM] Gavin.Hird @grid.xmir.org:8002: build a destination coordinate which includes the destination grid and teleport to it
[11:55 AM] Ubit Umarov: we do put hg regions on lower part of map.. ( y = 0)
[11:56 AM] Ubit Umarov: but to actully tp.. the real handler on target region needs to be used
[11:56 AM] Ubit Umarov: anywy.. things are working...
[11:56 AM] Ubit Umarov: so such collisions are rare..
[11:56 AM] Gavin.Hird @grid.xmir.org:8002: just swap in what is needed at the right time in the tp sequnece
[11:57 AM] Vincent.Sylvester @hg.zetaworlds.com: There is a fairly recent report on mantis about same coordinate hg teleport failing. I think when that came in I did test it and it failed as they said
[11:57 AM] Ubit Umarov: oops
[11:57 AM] Ubit Umarov: missed that
[11:57 AM] Vincent.Sylvester @hg.zetaworlds.com: Been a few months tho
[11:58 AM] Vincent.Sylvester @hg.zetaworlds.com: http://opensimulator.org/mantis/view.php?id=5210
[11:58 AM] Ubit Umarov: wel no surprise if that are "corner"cases
[11:58 AM] Ubit Umarov: 5210 is many months ago
[11:58 AM] Gavin.Hird @grid.xmir.org:8002: few months 👍
[11:58 AM] Ubit Umarov: like years
[11:58 AM] Vincent.Sylvester @hg.zetaworlds.com: Evidently I did test it, not that I remember it
[11:58 AM] Lyr.Lobo @cc.opensimulator.org:8002: /me grins at Gavin's Thumbs up
[11:59 AM] Lyr.Lobo @cc.opensimulator.org:8002: Great work, Gavin, on the emojis in your Mac viewer
[11:59 AM] Vincent.Sylvester @hg.zetaworlds.com: Trying to clean up mantis of old bugs, sue me ;P
[11:59 AM] Gavin.Hird @grid.xmir.org:8002: apparently Gen Z thinks it is offensive
[11:59 AM] Ubit Umarov: 0.7.1 dev master
[11:59 AM] Ubit Umarov: :)
[11:59 AM] Gavin.Hird @grid.xmir.org:8002: thanks Lyr
[11:59 AM] Ubit Umarov: 2010-11-15 13:36
[11:59 AM] Vincent.Sylvester @hg.zetaworlds.com: Comment is from 2021, but I don't recall there being a specific fix to this since then
[12:00 PM] Ubit Umarov: well like i said.. it may be unfixable
[12:00 PM] Vincent.Sylvester @hg.zetaworlds.com: Like I said, from what I recall how it works not something that is easily fixed either given it relies on regionhandle
[12:00 PM] Luisillo Contepomi: Many thanks to all the developers for their effort and dedication. See you later. Good night.
[12:00 PM] Ubit Umarov: without viewer changes to support hg
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: Reminder: The OpenSimulator Community Conference's Call for Proposals is open - let us know if you have cool tech, designs, research, education, creativity, AI stylings, or artistry that you want to share with the community. https://conference.opensimulator.org/call-for-proposals/
[12:01 PM] Ubit Umarov: have fun
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: Take care, Luisillo
[12:01 PM] Andrew Hellershanks: Night, Luisillo
[12:01 PM] Gavin.Hird @grid.xmir.org:8002: I guess you can call my entire viewer transformation AI syling, haha
[12:01 PM] Gavin.Hird @grid.xmir.org:8002: styling*
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: when the bar closes, the bar closes *watches the barkeep leave*
[12:01 PM] Ubit Umarov: :)
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: yes, gavin
[12:01 PM] Vincent.Sylvester @hg.zetaworlds.com: That's what I'm trying to determine, cause if that's something viewer side can fix I can look at it with TJ and see if we can figure out a way. It may be unfixable, but we can look at least
[12:01 PM] Lyr.Lobo @cc.opensimulator.org:8002: you should share your accomplishments
[12:02 PM] Lyr.Lobo @cc.opensimulator.org:8002: sessions are 20 minutes and fun
[12:02 PM] Vincent.Sylvester @hg.zetaworlds.com: Just trying to determine whether it's viewer side, server side or both
[12:02 PM] Gavin.Hird @grid.xmir.org:8002: I know. Currently shared as code and running viewers
[12:02 PM] Gavin.Hird @grid.xmir.org:8002: actually I could as Claude to write the entire presentation
[12:02 PM] Gavin.Hird @grid.xmir.org:8002: aks
[12:03 PM] Andrew Hellershanks: I told a person for whom I do some work about virtual worlds. He asked we discussed the possibility of doing something that could wind up being worth of a talk.
[12:03 PM] Ubit Umarov: back then there was also a issue with tps 1024 grid coords way, if i remember
[12:03 PM] Ubit Umarov: grid units...
[12:03 PM] Lyr.Lobo @cc.opensimulator.org:8002: /me grins at Gavin
[12:04 PM] Vincent.Sylvester @hg.zetaworlds.com: Wasn't that 4096
[12:04 PM] Ubit Umarov: think it was 1024.. but who knows..
[12:04 PM] Vincent.Sylvester @hg.zetaworlds.com: Somewhere in the depths of the entity transfer module
[12:04 PM] Ubit Umarov: possilbe 4k
[12:05 PM] Ubit Umarov: think viwer thing also
[12:05 PM] Vincent.Sylvester @hg.zetaworlds.com: All delegates and events everywhere almost impossible to figure out what goes where
[12:06 PM] Ubit Umarov: wel rl again..
[12:06 PM] Ubit Umarov: any other issue for today?
[12:06 PM] Lyr.Lobo @cc.opensimulator.org:8002: take care, me too
[12:06 PM] Lyr.Lobo @cc.opensimulator.org:8002: off to life
[12:06 PM] Lyr.Lobo @cc.opensimulator.org:8002: Many thanks!
[12:06 PM] Ubit Umarov: :)
[12:06 PM] Andrew Hellershanks: Ubit: Hey! That's supposed to be my line. :)
[12:07 PM] Ubit Umarov: :)
[12:07 PM] Andrew Hellershanks: Lyr, tc. Thanks for dropping by. Have fun now that school has started again.
[12:07 PM] Andrew Hellershanks: Yes, what Ubit said as we are just past the hour.
[12:08 PM] Gavin.Hird @grid.xmir.org:8002: Just be sure to download the new Mac viewer build people out there and tell me what works and what don't
[12:08 PM] Ubit Umarov: :)
[12:09 PM] Ubit Umarov: ok see you all next week, if not before
[12:09 PM] Gavin.Hird @grid.xmir.org:8002: I have a couple settings that needs admin password to change
[12:09 PM] Gavin.Hird @grid.xmir.org:8002: people will hate it or love it
[12:09 PM] Ubit Umarov: not sure why stand anim also fails now
[12:09 PM] Andrew Hellershanks: That will do it for this week. Thank you all for coming. See you again next week.
