Chat log from the meeting on 2023-04-25

From OpenSimulator

Jump to: navigation, search
[10:58 AM PT]  Gavin.Hird Hi everyone
[10:59 AM PT]  Andrew Hellershanks: Hello, everyone.
[11:00 AM PT]  Lyr Lobo: Hello *grins*
[11:02 AM PT]  Andrew Hellershanks: Hm... I don't see Ubit online. Not sure if/when he will join us today.
[11:02 AM PT]  Vincent.Sylvester Let's see we had some changes to parcel details to go along with last weeks changes, so the code is a bit more readable now. Also finally updated the syntax file for new functions and constants
[11:03 AM PT]  Vincent.Sylvester LL meanwhile been pushing some strided lists stuff that will need looking at when anyone feels like subjecting themselves to that nonsense
[11:04 AM PT]  Selby.Evans Hi everyone
[11:04 AM PT]  Andrew Hellershanks: Vincent, Where is that information listed?
[11:04 AM PT]  Ubit Umarov: hi
[11:04 AM PT]  Gavin.Hird 'ello
[11:04 AM PT]  Andrew Hellershanks: Hello, Selby.
[11:04 AM PT]  Vincent.Sylvester It's on the LSL wiki now, it was part of the changes a couple days ago
[11:04 AM PT]  Vincent.Sylvester They only now actually listed them on the functions list page
[11:05 AM PT]  Kayaker Magic: What LSL function are we talking about?
[11:05 AM PT]  Andrew Hellershanks: Vincent, Ah. I see them flagged as new.
[11:05 AM PT]  Vincent.Sylvester I never used strided lists, but apparently some do for those new functions to have use
[11:06 AM PT]  Andrew Hellershanks: llList2ListSlice, llListFindStrided, and llListSortStrided.
[11:06 AM PT]  Vincent.Sylvester Though I think that also has something to do with the linkset data thing they added
[11:06 AM PT]  Andrew Hellershanks: I've used strided lists but not very often.
[11:07 AM PT]  Andrew Hellershanks: Hello, Ubit. I didn't see you in IRC so I wasn't sure if/when you were going to make it here.
[11:07 AM PT]  Ubit Umarov: was a bit delayed at rl, but here now
[11:08 AM PT]  Andrew Hellershanks: LL keeps those "new" flags around a long time. The page about one of the "new" functions was last changed in September of last year.
[11:09 AM PT]  Vincent.Sylvester Making it seem like they are doing more than they actually are sneaky buggers
[11:10 AM PT]  Cuga.Rajal Been a while since I looked over the LSL implementation list
[11:11 AM PT]  Gavin.Hird They are making plenty enough of other changes...
[11:12 AM PT]  Vincent.Sylvester Annoyingly
[11:12 AM PT]  Orbert.Tatham They seem to be quite efficient at creating new problems
[11:13 AM PT]  Vincent.Sylvester Feels more and more like a handful of people writing code actively ignoring each other, only explanation I have for some of those decisions being made
[11:13 AM PT]  Gavin.Hird lol
[11:13 AM PT]  Andrew Hellershanks: :)
[11:13 AM PT]  Vincent.Sylvester But what else is new that's their modus operandi for years
[11:14 AM PT]  Gavin.Hird such as writing a bunch of code for Sansar only to dump it completely
[11:14 AM PT]  Orbert.Tatham I think it has gotten worse in the last few since they dumped the few Lindens whose job was staying in touch with the users
[11:15 AM PT]  Gavin.Hird they got new investors with a diffrent focus
[11:15 AM PT]  Ubit Umarov: i will look to those functions..  but im not that fan of them...
[11:15 AM PT]  Gavin.Hird plus their head tech guy is a Microsoftie, so gotta do it the MS way
[11:15 AM PT]  Orbert.Tatham That would explain why they constantly screw up
[11:16 AM PT]  Vincent.Sylvester I don't think many are a fan of those functions, was talking to Jeff on IRC about implementing them to make your life easier, but he didn't seem to care much for them either
[11:16 AM PT]  Vincent.Sylvester It's all a bit silly
[11:16 AM PT]  Vincent.Sylvester Then again if LSL made sense half the functions would need to be renamed and changed
[11:17 AM PT]  Orbert.Tatham I do wonder if more syntactic sugar is useful anyway
[11:17 AM PT]  Jagga Meredith: I took strided lists back at University when dinosaurs roamed the earth.  It may be an attempt at speed but I hesitate to suspect any meaningful improvement
[11:17 AM PT]  Vincent.Sylvester Underneath it operates the same way it just a different name and slightly different output
[11:18 AM PT]  Vincent.Sylvester One is basically same as a function already there just operating on a different part
[11:18 AM PT]  Vincent.Sylvester Should probably wait a bit anyways in case they decide to add more caveats or bugs... ahem features... later
[11:19 AM PT]  Gavin.Hird right now they are keeping themselves busy going full on AIS
[11:22 AM PT]  Ubit Umarov: ais is still something i do not plan to work on
[11:23 AM PT]  Gavin.Hird good
[11:23 AM PT]  Jagga Meredith: "ais"?
[11:23 AM PT]  Ubit Umarov: so we will not have those new tooltips things  images(??) they are adding to inventory
[11:23 AM PT]  Gavin.Hird asset interface something?
[11:23 AM PT]  Jagga Meredith: ah
[11:23 AM PT]  Ubit Umarov:
[11:24 AM PT]  Gavin.Hird they will remove all UDP
[11:24 AM PT]  Ubit Umarov: inventory
[11:24 AM PT]  Vincent.Sylvester I looked at the docs for that and couldn't help but think that someone worked really hard on that and it's just too bad it's a pile of shit
[11:24 AM PT]  Gavin.Hird because, of course Azure don't like that
[11:24 AM PT]  Gavin.Hird sorry Amazon
[11:25 AM PT]  Ubit Umarov: yeah so lets use low and heavy http and LLSDxml
[11:26 AM PT]  Ubit Umarov: well or worse..
[11:26 AM PT]  Ubit Umarov: low meant slow
[11:26 AM PT]  Gavin.Hird but that will of course mean a break between viewer code for OS and SL
[11:26 AM PT]  Ubit Umarov: another one yes
[11:26 AM PT]  Gavin.Hird yes
[11:26 AM PT]  Vincent.Sylvester Feeling the need to mess with things that work just fine while ignoring the actual issues everyone's been crying about for a decade
[11:27 AM PT]  Vincent.Sylvester Did they get some new execs from Microsoft by chance lol
[11:27 AM PT]  Gavin.Hird they did
[11:27 AM PT]  Orbert.Tatham Another shovel full, incoming
[11:27 AM PT]  Vincent.Sylvester *head desk*
[11:30 AM PT]  Orbert.Tatham All cynicism aside, I have to ask if anyone has figured out how long Firestorm is going to deal with the semi-schizophrenic separation between OpenSim and SL before they drop us on our heads?
[11:30 AM PT]  Jagga Meredith: who has the bigger user base?
[11:31 AM PT]  Andrew Hellershanks: Orbert, That would be a question for the devs who work on FS.
[11:31 AM PT]  Orbert.Tatham My understanding is SL by order(s) of magnitude
[11:31 AM PT]  Jagga Meredith: ah
[11:31 AM PT]  Gavin.Hird I am more wondering how long they actually can be bothered to make a viewer for SL, because LL is implementing just about every functionality in FS in their own viewer   some
[11:32 AM PT]  Vincent.Sylvester Stealing other peoples things badly, another thing they do poorly
[11:32 AM PT]  Orbert.Tatham Their viewer is as bad as everything else, that is what has kept fs alive
[11:32 AM PT]  Gavin.Hird FS has donated code and made pull requests, which LL "enhance" after
[11:32 AM PT]  Vincent.Sylvester Thems some big quotation marks there lol
[11:33 AM PT]  Vincent.Sylvester Thankfully most of it is open source, albeit nearly impossible to read if you aren't used to it
[11:34 AM PT]  Gavin.Hird plus the half million branches
[11:34 AM PT]  Orbert.Tatham Andrew, I ask it here because I have no access to FS devs, but we in large part here survive on the fact that someone makes a reasonably competent viewer
[11:34 AM PT]  Vincent.Sylvester And massive pulls of LL code that take forever to load in git
[11:34 AM PT]  Gavin.Hird which come and go faster than the daylight changes
[11:34 AM PT]  Ubit Umarov: we only have one fs dev that does look to our needs now and them
[11:34 AM PT]  Ubit Umarov: most others jsut gave up long ago
[11:35 AM PT]  Vincent.Sylvester We try to help them as much as possible with that as well down to providing lines of code to change
[11:35 AM PT]  Ubit Umarov: some still need to look because the code base is same, splited by in code condicionals, not full diferent files like it was for  short period
[11:36 AM PT]  Vincent.Sylvester I do detect some hostility whenever someone from OpenSim inputs any tickets with them, but that's been the case for a long time now
[11:36 AM PT]  Ubit Umarov: if we where on a splited fork, we wouild possible be on a 3 years old fs or worse
[11:37 AM PT]  Ubit Umarov: alchemy also told me that is dropping opensim
[11:37 AM PT]  Jagga Meredith: OK, I was wondering about that
[11:37 AM PT]  Vincent.Sylvester Like many times in the past yet comes crawling back each time
[11:37 AM PT]  Ubit Umarov: so we have dayturn, scenegate.. and no idea about others
[11:38 AM PT]  Gavin.Hird Rye Mutt seems to work for LL now
[11:38 AM PT]  Vincent.Sylvester Radegast still somewhat works too
[11:38 AM PT]  Andrew Hellershanks: How many of them are available on all three of the main operating systems (Windows, Mac, and Linux)?
[11:39 AM PT]  Gavin.Hird mono dependent?
[11:39 AM PT]  Gavin.Hird Radegast I mean
[11:39 AM PT]  Gavin.Hird mine will be win   macOS
[11:39 AM PT]  Lyr Lobo smiles and nods
[11:40 AM PT]  Ubit Umarov: cinder is keeping it..  not sure if fully dotnet now
[11:40 AM PT]  Vincent.Sylvester With mac linux ain't that far, though still a bit of work that
[11:41 AM PT]  Gavin.Hird what makes you say that?
[11:42 AM PT]  Cuga.Rajal limux being like mac is prob more true for server than client
[11:42 AM PT]  Vincent.Sylvester For the most part dependencies are or were at least similar. That said working and working well are two different things. Getting a macOS thing working on linux can be relatively simple, but it likely won't work well
[11:42 AM PT]  Gavin.Hird yes
[11:42 AM PT]  Vincent.Sylvester Biggest issue being graphics
[11:42 AM PT]  Gavin.Hird biggest linux issue are the libs
[11:42 AM PT]  Cuga.Rajal mac dropped opengl
[11:42 AM PT]  Vincent.Sylvester Yeah that
[11:42 AM PT]  Gavin.Hird to get them built for all the distros
[11:43 AM PT]  Andrew Hellershanks: libs dependencies is a big issue for cross platform work.
[11:43 AM PT]  Gavin.Hird we'll see at WWDC 2023 if Apple puts the last nail in the OpenGL coffin
[11:44 AM PT]  Vincent.Sylvester Which reminds me
[11:44 AM PT]  Vincent.Sylvester We still have that libgdiplus dependency at the moment for some graphics stuff
[11:44 AM PT]  Vincent.Sylvester Is there anything equivalent in dotnet6 that can be used for those things?
[11:45 AM PT]  Ubit Umarov: totally diferent graphic libs
[11:45 AM PT]  Ubit Umarov: plus future ms MAUI ??
[11:45 AM PT]  Orbert.Tatham dotnet is trying to get rid of everything win32
[11:46 AM PT]  Ubit Umarov: well Image things for windows are supported
[11:46 AM PT]  Vincent.Sylvester Well is there anything that can potentially produce the same output at least? Requiring a rewrite of the graphics wrapper is a given, but is there at least something that will work the same way if integrated?
[11:46 AM PT]  Gavin.Hird But MS has time to put code for more ads everywhere in their clients
[11:46 AM PT]  Ubit Umarov: with a full rewrite
[11:47 AM PT]  Orbert.Tatham I am with Ubit - i have not seen any thing in common
[11:47 AM PT]  Ubit Umarov: from dynamic textyres, jp2000 decode ( for lludp and scult maps), warp3d, etc
[11:48 AM PT]  Ubit Umarov: dynamic textures.. not even sure if all primitives we use are supported ( like drawy elipses etc )
[11:50 AM PT]  Vincent.Sylvester Well I was just going to take those functions and throw them at that fancy new full text thing some claim to be intelligence and ask it to rewrite them using x instead of libgdiplus so just wondering what that would be if any
[11:50 AM PT]  Orbert.Tatham Good luck with that
[11:51 AM PT]  Gavin.Hird :-))
[11:51 AM PT]  Vincent.Sylvester It kinda worked for the map tiles... after a couple attempts
[11:51 AM PT]  Cuga.Rajal should train chatgpt to write metal code lol
[11:51 AM PT]  Gavin.Hird that is reserved for Siri
[11:51 AM PT]  Cuga.Rajal tru :)
[11:52 AM PT]  Gavin.Hird The biggest issue with Metal is it is tile based
[11:52 AM PT]  Gavin.Hird so a completely different way of rendering
[11:53 AM PT]  Vincent.Sylvester For the maptiles the changes I made to that have been proving themselves for over a month now in production. The generation of the zoom levels works flawlessly without the weird patches used to seeing when grid registrations overlapped
[11:53 AM PT]  Gavin.Hird but again WWDC2023 could bring us Metal 4 with new fuctionality
[11:53 AM PT]  Vincent.Sylvester Haven't seen it mess up yet and it's a lot quicker at doing them too
[11:54 AM PT]  Vincent.Sylvester So we need something that has jpeg2000 support to generate images, which not much has I take it?
[11:55 AM PT]  Gavin.Hird KDU
[11:55 AM PT]  Ubit Umarov: we have 2 libs
[11:55 AM PT]  Ubit Umarov: one in c#, other openjpeg C# wrapper
[11:56 AM PT]  Ubit Umarov: thing is that those depend on .net image
[11:56 AM PT]  Ubit Umarov: read to bitmap.. etc etc
[11:57 AM PT]  Ubit Umarov: fully written to interface with .net framework, ofc
[11:57 AM PT]  Ubit Umarov: so all that needs rewrite..
[11:57 AM PT]  Ubit Umarov: not fun
[11:57 AM PT]  Andrew Hellershanks nods
[11:58 AM PT]  Ubit Umarov: and then, we are moving from a "standard" by big ms, to something that may be dead in 2 years...
[11:58 AM PT]  Andrew Hellershanks: A lot can happen in 2 years.
[11:58 AM PT]  Ubit Umarov: yeah
[11:59 AM PT]  Cuga.Rajal never dead, just not upgradeable
[11:59 AM PT]  Cuga.Rajal dead for developers
[11:59 AM PT]  Ubit Umarov: well..  for now  libgdi   and win gdi   does work
[11:59 AM PT]  Orbert.Tatham dead in the marketplace
[11:59 AM PT]  Ubit Umarov: so withing dotnet 6.0 we should be ok
[11:59 AM PT]  Ubit Umarov: lets see dotnet 8.0 ;)
[12:00 PM PT]  Ubit Umarov: btw our http code is also now "obsolete" by ms
[12:00 PM PT]  Ubit Umarov: may also stop working on future dotnet
[12:01 PM PT]  Ubit Umarov: they replaced it by httpclient
[12:01 PM PT]  Ubit Umarov: that is a LOT worse.. since assumes a app only opens a few web pages at most
[12:01 PM PT]  Ubit Umarov: while opensim needs to more than a ton :)
[12:02 PM PT]  Ubit Umarov: to open
[12:02 PM PT]  Ubit Umarov: that is another problem
[12:02 PM PT]  Gavin.Hird a ton for what purpose?
[12:02 PM PT]  Ubit Umarov: the ton of services we have..
[12:02 PM PT]  Ubit Umarov: comms to any other region...
[12:02 PM PT]  Ubit Umarov: teleports.. etc etc
[12:02 PM PT]  Gavin.Hird oh, you mean all the reuests to asset server?
[12:03 PM PT]  Ubit Umarov: several tons :)
[12:03 PM PT]  Ubit Umarov: those also
[12:03 PM PT]  Ubit Umarov: all services
[12:03 PM PT]  Vincent.Sylvester Without http you can't even login
[12:03 PM PT]  Ubit Umarov: well not that http
[12:04 PM PT]  Ubit Umarov: we do have a http server that they say is obsolete,  but diferent reason
[12:04 PM PT]  Orbert.Tatham That's another service
[12:04 PM PT]  Ubit Umarov: i mead the parts where the code is a http client
[12:04 PM PT]  Ubit Umarov: the httpwebrequest things
[12:05 PM PT]  Ubit Umarov:
[12:05 PM PT]  Ubit Umarov: still there for dotnet 8.0
[12:05 PM PT]  Vincent.Sylvester httprequest from scripts, region intercom, hg, etc
[12:05 PM PT]  Ubit Umarov: if oyu look there
[12:06 PM PT]  Ubit Umarov: you see
[12:06 PM PT]  Ubit Umarov: "
[12:06 PM PT]  Ubit Umarov: Remarks

We don't recommend that you use HttpWebRequest for new development. Instead, use the System.Net.Http.HttpClient class.
[12:06 PM PT]  Ubit Umarov: "
[12:06 PM PT]  Ubit Umarov: as i said httpClient is WORSE
[12:06 PM PT]  Ubit Umarov: because assumes a app only uses a few pages/sites
[12:06 PM PT]  Ubit Umarov: not the massive need we have
[12:07 PM PT]  Gavin.Hird give them developer feedbak
[12:07 PM PT]  Vincent.Sylvester xD
[12:07 PM PT]  Ubit Umarov: but future problem.. as you see still there on dotnet 8
[12:07 PM PT]  Selby.Evans bye all
[12:07 PM PT]  Cuga.Rajal so they are not dropping HttpWebRequest, they are just recpommending a diff class
[12:07 PM PT]  Gavin.Hird yeah, the earlier they gets a heads up about issues the better I suppose
[12:07 PM PT]  Orbert.Tatham See you Selby
[12:08 PM PT]  Ubit Umarov: Cuga that is VERY diferent :P
[12:08 PM PT]  Ubit Umarov: in fact another massive rewrite
[12:08 PM PT]  Ubit Umarov: because we do use that all over..
[12:08 PM PT]  Cuga.Rajal si why not just ignore their recommendation and keep using HttpWebRequest
[12:08 PM PT]  Ubit Umarov: but the problem is not only the rewrite
[12:08 PM PT]  Ubit Umarov: is that the new is WORSE
[12:09 PM PT]  Ubit Umarov: tunned for a simple app use
[12:09 PM PT]  Ubit Umarov: as i said.. a few pages open now and then
[12:09 PM PT]  Gavin.Hird rewrite everything in serverside Swift
[12:09 PM PT]  Gavin.Hird :-)
[12:10 PM PT]  Orbert.Tatham I say Rust
[12:10 PM PT]  Kayaker Magic: RL calling, I need to run
[12:10 PM PT]  Vincent.Sylvester It's a kinda one-sided thing as the name somewhat implies that it isn't really meant for moving tons of data just making requests for single things here and there
[12:10 PM PT]  Vincent.Sylvester Doesn't like to keep them open either
[12:10 PM PT]  Ubit Umarov: well that is like 1000 warnings i did mute on dotnet6
[12:10 PM PT]  Ubit Umarov: all teling "httpweb.. is obsolete.. " :)
[12:12 PM PT]  Cuga.Rajal oh the message does say it is depreciated.. hmm ok
[12:12 PM PT]  Vincent.Sylvester Typical MS though, deprecate something and then point to something inferior or completely different and call it a day
[12:12 PM PT]  Andrew Hellershanks: :P
[12:13 PM PT]  Orbert.Tatham Welcome to the Micro$haft treadmill
[12:13 PM PT]  Cuga.Rajal thats different than "use class y instead of x cause we think its better"
[12:14 PM PT]  Cuga.Rajal which is sort of what the messages on the dotnet webpate were saying
[12:14 PM PT]  Vincent.Sylvester Maybe a library exists that can handle this that we can use instead... let's check nuget...
[12:14 PM PT]  Orbert.Tatham LOL
[12:14 PM PT]  Andrew Hellershanks: :)
[12:15 PM PT]  Vincent.Sylvester I would not be opposed to doing that in principle given that the current http requests have a weird limit on them in some form that I never managed to track down
[12:15 PM PT]  Vincent.Sylvester Perhaps a purpose-built lib that isn't bloated from years of MS development on it might even yield some better performance
[12:16 PM PT]  Vincent.Sylvester A plan B of sorts
[12:16 PM PT]  Cuga.Rajal http and libgdi  ... that will keep developers busy
[12:17 PM PT]  Andrew Hellershanks: Just need to find someone to build a lib for the purpose.
[12:17 PM PT]  Vincent.Sylvester How hard can it be, it's just http /s
[12:18 PM PT]  Vincent.Sylvester Ideally something like mysqlclient which is mostly a drop-in replacement over the current connect should the day ever come we can't use the one we have
[12:18 PM PT]  Gavin.Hird rl call....
[12:18 PM PT]  Cuga.Rajal that woud be ideal
[12:18 PM PT]  Vincent.Sylvester One can dream right :)
[12:18 PM PT]  Andrew Hellershanks: ok, Gavin. Thanks for being here.
[12:18 PM PT]  Orbert.Tatham Peace, Gavin
[12:19 PM PT]  Lyr Lobo: bye *waves*
[12:19 PM PT]  Cuga.Rajal tc Gavin
[12:20 PM PT]  Andrew Hellershanks: It is almost 20 past the hour. Any other questions or comments before we wrap up for today?
[12:20 PM PT]  Cuga.Rajal I have just a quick question about
[12:21 PM PT]  Cuga.Rajal see the comment for llSetBuoyancy
[12:21 PM PT]  Cuga.Rajal should same comment be there for llSetForce()?
[12:21 PM PT]  Vincent.Sylvester That page has not been kept up to date really, it needs a checkup really
[12:21 PM PT]  Cuga.Rajal ok
[12:21 PM PT]  Vincent.Sylvester Feel free to retest implementation and update where needed, it's been neglected quite a bit
[12:21 PM PT]  Cuga.Rajal Im assuming nothing new with llSetForce
[12:22 PM PT]  Cuga.Rajal ok that answers my question :)
[12:22 PM PT]  Andrew Hellershanks: :)
[12:22 PM PT]  Vincent.Sylvester Until we can get rid of the avatar motors those functions can't really work
[12:23 PM PT]  Cuga.Rajal thats what I thought, just saw llSetForce in all green so wanted to check :)
[12:23 PM PT]  Cuga.Rajal I think it should have same color/comment as llSetBuoyancy
[12:23 PM PT]  Andrew Hellershanks: It has been quite a while since I last tested buoyancy. I had a balloon script hat used it but it didn't work back then. OpenSim has seen a lot of changes since then,.
[12:24 PM PT]  Cuga.Rajal I'm mostly interested in avatar attachments affecting avatar fall up/down
[12:24 PM PT]  Cuga.Rajal parachutes, etc
[12:24 PM PT]  Vincent.Sylvester I did make a parachute that somewhat works to stop you going splat, but it's not exactly smooth
[12:25 PM PT]  Cuga.Rajal I can use llApplyImpulse with a timer
[12:25 PM PT]  Andrew Hellershanks: That wiki page has not been updated since September 20 of last year.
[12:25 PM PT]  Cuga.Rajal but not smooth
[12:25 PM PT]  Vincent.Sylvester Yeah basically that
[12:25 PM PT]  Vincent.Sylvester You can smooth that a bit more using while and sleep
[12:26 PM PT]  Cuga.Rajal still too bumpy... I'll wait :)
[12:27 PM PT]  Andrew Hellershanks: Cuga, just avoid jumping off any tall buildings for now. ;)
[12:27 PM PT]  Cuga.Rajal lol
[12:28 PM PT]  Andrew Hellershanks: FInal call for any questions/comments for today?
[12:29 PM PT]  Andrew Hellershanks: I'm not seeing signs of anyone typing so I will say we are done for today.
[12:29 PM PT]  Andrew Hellershanks: Thank you all for coming. See you all again next week.
Personal tools
About This Wiki