Chat log from the meeting on 2023-10-17

From OpenSimulator

Jump to: navigation, search
[11:14 AM PT]  Ubit Umarov: Welcome :)
[11:14 AM PT]  Web Rain: i found i small bug ^^ llJsonGetValue return something that is not empty string when the key is not found :( eg llJsonGetValue("{}",["-"]) will return something
[11:14 AM PT]  Clifford.Hanger any voice news?
[11:14 AM PT]  Ubit Umarov: lets start by telling that the code had not changes during last week
[11:15 AM PT]  Gavin.Hird I looked at the mumble thingy, but I need a dedicated machine to test it on because it seems flaky
[11:15 AM PT]  Vincent.Sylvester Web, if you have a reproducible bug please write a mantis ticket
[11:16 AM PT]  Web Rain: oki :)
[11:16 AM PT]  Gavin.Hird so the plan is to free up a Mac running a VAR and move it to teh PI4 I got working with the latest dotnet6
[11:16 AM PT]  Clifford.Hanger ok
[11:17 AM PT]  Vincent.Sylvester If you need hardware why not say so, I got stuff sitting around collecting dust
[11:17 AM PT]  Gavin.Hird then I can fiddle around with it without disturbing anything else
[11:17 AM PT]  Gavin.Hird I jsut told you I have hardware :-)
[11:17 AM PT]  Object: Script running
[11:18 AM PT]  Object: Hello, Avatar!﷐
[11:18 AM PT]  Vincent.Sylvester I had some more fun with xunit over the last couple days, preparing a setup for prime time so to speak
[11:18 AM PT]  Ubit Umarov: yeah llJsonGetValue("{}",["-"]) does return something.. a empty list
[11:18 AM PT]  Ubit Umarov: :)
[11:18 AM PT]  Ubit Umarov: same as at sl
[11:18 AM PT]  Clifford.Hanger hey Joe
[11:18 AM PT]  Gavin.Hird what is xunit?
[11:19 AM PT]  Web Rain: ha ?!? strange
[11:19 AM PT]  Vincent.Sylvester The primary issue of how to add it to the solution I solved with a bit of a hack, but better than poking around in the Prebuild
[11:19 AM PT]  Joe Magarac: Hello.
[11:19 AM PT]  Vincent.Sylvester xunit is a testing framework like nunit, just a bit more modern and maintained
[11:19 AM PT]  Orbert.Tatham Hey, Joe
[11:19 AM PT]  Gavin.Hird ok, thx
[11:19 AM PT]  Vincent.Sylvester It's one of the recommended ones for dotnet testing
[11:20 AM PT]  Vincent.Sylvester Given we need new tests and nunit had syntax changes I felt it better to upgrade to something since re-write was bound to be needed anyhow
[11:20 AM PT]  Ubit Umarov: hmm there should abe a json_invalid somewhere on that answer
[11:20 AM PT]  Object: Hello, Avatar﷐
[11:20 AM PT]  Ubit Umarov: actually it is there... displayed as that box
[11:20 AM PT]  Vincent.Sylvester xunit is quite a bit faster it seems, so that's a plus. I also restructured the tests for now to not have them littered around
[11:21 AM PT]  Gavin.Hird sounds goo Vincent
[11:21 AM PT]  Gavin.Hird good*
[11:21 AM PT]  Vincent.Sylvester Putting them all in one place under one project so they aren't built all the time along the other stuff, making them more of an optional thing for those that want to use them
[11:21 AM PT]  Gavin.Hird not goo, haha
[11:22 AM PT]  Ubit Umarov: well one day i may look into those unit tests on core :)
[11:22 AM PT]  Vincent.Sylvester I still have to go through the old tests and make plans on what to test and how, so it'll be a while before I can push anything out the door
[11:22 AM PT]  Ubit Umarov: for now other things on todo ( one day )
[11:22 AM PT]  Ubit Umarov: like moving git master to dotnet6 :)
[11:22 AM PT]  Joe Magarac: Oh, good.
[11:23 AM PT]  Clifford.Hanger if anyone missed it -  i can report that the kcozems searc module runs fine on dotnet
[11:24 AM PT]  Web Rain: what kind of tests do you need for the .net6 ?
[11:24 AM PT]  Gavin.Hird I started decoding if there is anything - like in anything opensim can uses in the 420 commit AIS update they made to the viewer code
[11:24 AM PT]  Clifford.Hanger but classifides section in query needs to be pointed at the grid db
[11:24 AM PT]  Clifford.Hanger not search
[11:25 AM PT]  Vincent.Sylvester Primarily I plan to test basic functions, login, inventory, data serializations and simulator behavior. Those are usually the things that break first or have in the past. The most critical operations really.
[11:26 AM PT]  Vincent.Sylvester More specific stuff like script functionality is fairly easy to just spin up a dummy script engine and run a script to check output. Harder are things with physics calls and complex behavior that require dummy regions to be spun up
[11:26 AM PT]  Vincent.Sylvester Stuff like crossings and teleports are quite the things to test for. The old tests on that are basically praying for things to work
[11:27 AM PT]  Gavin.Hird it it only was the test...
[11:27 AM PT]  Vincent.Sylvester There are a few systems like scene helpers etc. that provide ways to create tests without the need to run a full environment
[11:28 AM PT]  Vincent.Sylvester Have to see how they work and integrate, but I think that's worth the effort. Knowing if a change breaks basic functions is quite vital as we move to replacing larger chunks of code to newer dotnet standards
[11:29 AM PT]  Gavin.Hird to be fair region crossings and teleports works pretty great
[11:29 AM PT]  Web Rain: yep is a major change
[11:29 AM PT]  Clifford.Hanger hi jamie
[11:30 AM PT]  Vincent.Sylvester I recall backporting some dotnet change to mono and a test blew up so definitely not a bad idea to test things even if just the basics
[11:30 AM PT]  Vincent.Sylvester Thankfully still have a few people actively following development and writing tickets when stuff breaks
[11:31 AM PT]  Jamie.Jordan hi everybody sorry i'm late
[11:32 AM PT]  Web Rain: i'm working on an Opensim plugin for BTCPay server that allow cryptocurrencies payments in-world so , i can test the HTTP I/O and Json things...
[11:32 AM PT]  Web Rain: Hi Jamie :)
[11:33 AM PT]  Lyr Lobo smiles at Jamie
[11:34 AM PT]  Gavin.Hird maybe a Tilia plugin, but I bet it cost an arm and a leg to use the service
[11:35 AM PT]  Joe Magarac: There is a Tilia plug-in for Unity. Nobody uses it. But look into that.
[11:35 AM PT]  Gavin.Hird if nothing else, as long as SL is around, Tilia should be there too
[11:36 AM PT]  Kayaker Magic: And the reverse.
[11:36 AM PT]  Kayaker Magic: Didn't Tilia buy LL?
[11:36 AM PT]  Joe Magarac: No, no.
[11:36 AM PT]  Joe Magarac: Tilia is partly owned by JP Morgan Chase Payments.
[11:37 AM PT]  Gavin.Hird I don't think so
[11:37 AM PT]  Joe Magarac: But they haven't done much with it,.
[11:38 AM PT]  Gavin.Hird The thing is of course for anyone in Europe to use a US payment service, it is acomplete can of worms both tax vise and consumer legilation
[11:38 AM PT]  Web Rain: this is a proof of work video... is easy as 123 and the most important is a self-hosted solution and not a service !
[11:38 AM PT]  Web Rain:
[11:39 AM PT]  Web Rain: my modest work is here
[11:39 AM PT]  Web Rain:
[11:39 AM PT]  Web Rain: in wait of a BTCPay team code review and audit
[11:40 AM PT]  Vincent.Sylvester In regards to commerce the overall feedback I tend to get is that most just want something that is stable, volatility of crypto doesn't seem to fit that well, each to their own I guess
[11:41 AM PT]  Gavin.Hird right
[11:41 AM PT]  Clifford.Hanger agreed
[11:41 AM PT]  Web Rain: i agree but not all coins are good for the opensim usage
[11:41 AM PT]  Gavin.Hird paynets in BTC is more or less useless with the wild fluxtuations in valuation for the micro transactions done in opensim
[11:41 AM PT]  Gavin.Hird payments*
[11:42 AM PT]  Web Rain: and you don't care about that because what you display to your end-user is fiat prices, BTCPay calculate in the fly the amounts in cryptos before the invoice part
[11:43 AM PT]  Web Rain: is a pretty nice software and secure because its don't use sensive data like the private keys
[11:44 AM PT]  Joe Magarac: Yes, that's a good feature.
[11:44 AM PT]  Web Rain: the payment gose directlly to your own wallet and BTCPay helps to manage the things
[11:44 AM PT]  Joe Magarac: But if the are not charging for transactions, they must have some other revenue plan.
[11:45 AM PT]  Joe Magarac: In crypto, that's usually either investing the customer's money, or just stealing it.
[11:45 AM PT]  Web Rain: ehhh? is a software that you install at home or a VPS like opensim ^^
[11:45 AM PT]  Vincent.Sylvester Anything even remotely more complex than PayPal is likely to fly over half the populations heads
[11:45 AM PT]  Joe Magarac:
[11:46 AM PT]  Joe Magarac: Yes.
[11:46 AM PT]  Clifford.Hanger agreed again
[11:46 AM PT]  Gavin.Hird yeah but people don't hold BTC, so it must (for the very majority) be purchased soemwhere
[11:46 AM PT]  Ubit Umarov: Core will stay out of money modules..
[11:46 AM PT]  Joe Magarac: Has anyone found a bank that supports FedNow?
[11:46 AM PT]  Joe Magarac: FedNow costs US$0.045 per transaction.
[11:46 AM PT]  Ubit Umarov: "money" is a a thing a lot more complex than the supporting software..
[11:47 AM PT]  Joe Magarac: The US is finally getting a Government-run payment system.
[11:47 AM PT]  Gavin.Hird sounds like a US payment processor. useless for European use
[11:47 AM PT]  Ubit Umarov: and crytoblabla?  yeah.. nahh :p
[11:47 AM PT]  Clifford.Hanger that figures
[11:47 AM PT]  Joe Magarac: Right. It works like SEPA, but not internationally.
[11:47 AM PT]  Web Rain: agree the current opensim and hypergrid citizens are not familiar with cryptos in general but adopting cryptos in the hypergrid open the thing to a new category of users... from the crypto world
[11:47 AM PT]  Ubit Umarov: ofc we do provide the normal hooks for external modules to link to..
[11:48 AM PT]  Joe Magarac: The cryto world is much smaller than it used to be. NFTs are down 98%.
[11:48 AM PT]  Gavin.Hird lol
[11:48 AM PT]  Web Rain: NFTs are scam lol
[11:48 AM PT]  Vincent.Sylvester Single handedly raising the ceo population in federal prisons
[11:49 AM PT]  Clifford.Hanger any new thoughts on voice Ubit?
[11:49 AM PT]  Joe Magarac: Check out the Bankman-Fried trial news. It's not going well for him.
[11:49 AM PT]  Ubit Umarov: looking to same chat before the meeting.. the testing of opensim on raspPI was done with master/mono or dotnet6 ?
[11:50 AM PT]  Gavin.Hird dotnet6
[11:50 AM PT]  Ubit Umarov: no Clifford.. no changes on voice issue
[11:50 AM PT]  Clifford.Hanger Vincent?
[11:50 AM PT]  Gavin.Hird standalone does not work with SQlite, but it probalby doesn't anyway
[11:50 AM PT]  Gavin.Hird so you need to connect to a MySQL instance
[11:51 AM PT]  Vincent.Sylvester Will be a while before we see anything happening with voice most likely, Rome wasn't built in a day
[11:51 AM PT]  Ubit Umarov: as i said last meeting, Vivox is not a option, because they told me so
[11:51 AM PT]  Joe Magarac: Right.
[11:51 AM PT]  Vincent.Sylvester Replacements require quite a bit of documentation to form a plan of action
[11:51 AM PT]  Ubit Umarov: i mean vivox5...  and that is above technical issues..
[11:51 AM PT]  Joe Magarac: Someone said Mumble is no good. What's the problem there?
[11:52 AM PT]  Ubit Umarov: no point of having code to use a service that is not there to use
[11:52 AM PT]  Clifford.Hanger i agree
[11:52 AM PT]  Gavin.Hird IDK _ I am going to isntall the macOS version and see what it does, but it seems flaky in that you need to disable system intergrity protection to run it
[11:53 AM PT]  Vincent.Sylvester Document the process, I'll give it a go on linux when I get some time next week or so :)
[11:53 AM PT]  Ubit Umarov: about mumble..  we have no client side code.. the old slvoice for it, sure no longer works
[11:53 AM PT]  Web Rain: freeswitch is fully dead ?
[11:53 AM PT]  Gavin.Hird and it only works in Intel Mac
[11:53 AM PT]  Gavin.Hird but I'll give it a wirr
[11:54 AM PT]  Vincent.Sylvester Freeswitch should still work, but is kinda limited, though it has some other interesting features
[11:54 AM PT]  Vincent.Sylvester Freeswitch is a full sip backend so those are fun to play with
[11:54 AM PT]  Clifford.Hanger ty Vincent
[11:54 AM PT]  Web Rain: i will give it a try too this week
[11:54 AM PT]  Ubit Umarov: as i said,  IMA was working on their mumble solution ( EchoVOice?)  but the crowdfunding the asked did not got the amount that asked.. not even close
[11:54 AM PT]  Joe Magarac: Ah.
[11:55 AM PT]  Gavin.Hird 2%-ish?
[11:55 AM PT]  Vincent.Sylvester I think not even
[11:55 AM PT]  Ubit Umarov: well vivox started as just SIP variant also
[11:55 AM PT]  Clifford.Hanger i dont see that as having much hope, sorry
[11:55 AM PT]  Orbert.Tatham Someone wanted to retire on it, I think
[11:55 AM PT]  Clifford.Hanger :)
[11:55 AM PT]  Gavin.Hird :-)
[11:55 AM PT]  Clifford.Hanger hi Dennis
[11:56 AM PT]  Ubit Umarov: freeswithc does depend on vivox viwers side
[11:56 AM PT]  Joe Magarac: Oh.
[11:56 AM PT]  Ubit Umarov: and only allows it to work with the basic worse codec
[11:56 AM PT]  Ubit Umarov: so sound quality is bad... no 3d etc etc
[11:56 AM PT]  Vincent.Sylvester I think I wrote in the mantis on this that a more realistic funding goal would have been somewhere between 10 and 20 thousand to get some form of prototype, if a developer with a discounted rate could be sourced
[11:56 AM PT]  Gavin.Hird 2.5 Khz voice channel
[11:57 AM PT]  Vincent.Sylvester OpenSim community coming up with 75k was never going to happen
[11:57 AM PT]  Gavin.Hird or something like that
[11:57 AM PT]  Gavin.Hird kHz
[11:57 AM PT]  Ubit Umarov: but last time i looked ( 3 yrs ago) it was working
[11:57 AM PT]  Clifford.Hanger you did Vincent
[11:58 AM PT]  Joe Magarac: Have to go to another meeting. Thanks, all.
[11:58 AM PT]  Vincent.Sylvester The positive development we have in regards to voice is that quite a few platforms are building those capabilities into their software so the renewed interest is spawning new projects and codecs
[11:58 AM PT]  Clifford.Hanger and it seems our best/only hope
[11:58 AM PT]  Ubit Umarov: well 75K is nothing much..  less than a year salary for a junior code in usa
[11:58 AM PT]  Ubit Umarov: ( and uk, etc )
[11:58 AM PT]  Ubit Umarov: junior coder..
[11:58 AM PT]  Ubit Umarov: like Beq told on that mantis
[11:59 AM PT]  Ubit Umarov: ( sure a lot less in pakistan )
[11:59 AM PT]  Gavin.Hird soon to be replaced by an AI coder, that runs for a fraction of the cost
[11:59 AM PT]  Gavin.Hird yek
[11:59 AM PT]  Vincent.Sylvester I did find some interesting codecs with open licenses, but I am no audio tech so what the specifics mean I got no clue
[11:59 AM PT]  Vincent.Sylvester Remains to be tested what the capabilities of those are
[11:59 AM PT]  Ubit Umarov: codec is just a simple part
[12:00 PM PT]  Ubit Umarov: old speex (now opus)  is free...
[12:00 PM PT]  Vincent.Sylvester The more open source there is the easier it is to build something from scratch
[12:00 PM PT]  Ubit Umarov: issue is protocols.. etc
[12:00 PM PT]  Ubit Umarov: in fact major issue is cost of servers infrasture
[12:01 PM PT]  Ubit Umarov: voice is not that tolerant to latency variations...
[12:01 PM PT]  Ubit Umarov: you will not do it with a VM tiny server
[12:01 PM PT]  Gavin.Hird setting up servers is probalby beyond the scope for most small grids
[12:01 PM PT]  Clifford.Hanger sorry i need to go
[12:01 PM PT]  Clifford.Hanger event starting
[12:01 PM PT]  Clifford.Hanger thanks everyone
[12:01 PM PT]  Gavin.Hird so there must most likely be a shared service of sorts
[12:02 PM PT]  Ubit Umarov: for mumble there are still servers that are possible to rent
[12:02 PM PT]  Vincent.Sylvester The server stuff can be solved, what strikes me as the biggest issue is the viewer end. Creating a clean-room reverse engineered slvoice would be one way to solve that, but doing that is easier said than done
[12:02 PM PT]  Gavin.Hird what's the cost?
[12:02 PM PT]  Ubit Umarov: like 1200 USD for 2 years with 400 slots last time i looked .)
[12:02 PM PT]  Kayaker Magic: If you don't need to chat between regions, there should be a solution that runs on the region server, like another module.
[12:02 PM PT]  Ubit Umarov: in fact same cost as teamspeak servers on same provider :)
[12:03 PM PT]  Vincent.Sylvester Voice as region module is not going to solve the latency issue, might make it even worse
[12:03 PM PT]  Ubit Umarov: well and ppl do expect also voice IMs, grid wide, etc
[12:04 PM PT]  Vincent.Sylvester Ideal would be a servers in multiple geo locations that negotiate which one is best to use for a given conversation or region
[12:04 PM PT]  Gavin.Hird server side need QoS
[12:04 PM PT]  Vincent.Sylvester Potentially even using inter-server connections to reduce latency between peers
[12:04 PM PT]  Vincent.Sylvester At least for those things technology exists already
[12:04 PM PT]  Ubit Umarov: well for a reason ll went for a external solution, when they where doing all things in house )
[12:05 PM PT]  Orbert.Tatham If Vivox didn't want a left body part for it, they do use that technology, but....
[12:05 PM PT]  Ubit Umarov: ok rl calls
[12:05 PM PT]  Ubit Umarov: any other opensim issue? :)
[12:05 PM PT]  Gavin.Hird not this week for me
[12:05 PM PT]  Web Rain: no for now :) ty Ubit ^^
[12:06 PM PT]  Ubit Umarov: ok then see tou all next week :)
Personal tools
About This Wiki