Chat log from the meeting on 2023-11-28

From OpenSimulator

Jump to: navigation, search
[12:09 PM PT]  Ubit Umarov: Welcome :)
[12:09 PM PT]  Ubit Umarov: about last week code changes,  finally we moved  out git master to dotnet 6.0
[12:10 PM PT]  Ubit Umarov: merging into it the branch dotnet6, that now is "dead"
[12:11 PM PT]  Ubit Umarov: the previus master with code for .net framework 4.x and Mono is stored on a branch called Mono-Net4X
[12:13 PM PT]  Clifford.Hanger ok
[12:14 PM PT]  Lyr Lobo: Thank you, Ubit
[12:14 PM PT]  Lyr Lobo: What does this mean for future development?
[12:14 PM PT]  Lyr Lobo: [My apologies, was stepping out of a class]
[12:15 PM PT]  Ubit Umarov: Well Mono and .net framework are now dead, replaced by dontnet
[12:15 PM PT]  Ubit Umarov: so it means we do have a path into future :)
[12:16 PM PT]  Vincent.Sylvester I made quite a lot of progress on a migration check routine last week, unfortunately in testing for that I noticed that newer versions of Postgres are completely buggered. The migrations themselves fail. So either the migrations need changing, which might break older Postgres versions or need to do version detection and run different migrations.
[12:17 PM PT]  Clifford.Hanger sounds messy
[12:17 PM PT]  Ubit Umarov: :(
[12:17 PM PT]  Ubit Umarov: ofc current dotnet version is 8.0, but 6.0 is a LTS so will have support for at least another year
[12:18 PM PT]  Ubit Umarov: and we will not rush into last versions, ofc
[12:20 PM PT]  Vincent.Sylvester Let others work out the bugs first heh
[12:22 PM PT]  Vincent.Sylvester Postgres blowing up did make me look at alternatives again, especially in regards to data structure, but whether having a dual DB structure with mySQL and CouchDB is going to improve performance any I'll have to see. Just seems to me like data that is identified to belong to a user doesn't need to share a large table
[12:23 PM PT]  Clifford.Hanger i can see that
[12:25 PM PT]  Clifford.Hanger silly question?
[12:25 PM PT]  Orbert.Tatham CouchDB? Are we really interested in adding the whole Erlang baggage to this?
[12:25 PM PT]  Clifford.Hanger is there any where i can look to find out which .cs file has what purpose?
[12:26 PM PT]  Vincent.Sylvester Well it's an option. I did originally look at Mongo, but Couch is backed by Apache so we know that won't just go poof from one day to the other
[12:26 PM PT]  Clifford.Hanger some are descriptive
[12:26 PM PT]  Vincent.Sylvester You'll have to read the code basically, we don't document to that depth
[12:27 PM PT]  Clifford.Hanger a lot of reading lol!
[12:28 PM PT]  Vincent.Sylvester I don't think anyone knows the entire codebase, when changes are necessary you just read what's there and fix things
[12:28 PM PT]  Clifford.Hanger its knowing where to look is my problem
[12:28 PM PT]  Ubit Umarov: savings in anti-insomnia drugs.!
[12:28 PM PT]  Lyr Lobo: hehe
[12:29 PM PT]  Vincent.Sylvester You can search the codebase for keywords or look into the specific module, what are you looking for?
[12:30 PM PT]  Clifford.Hanger copying of objects and checking god mode
[12:31 PM PT]  Clifford.Hanger found some bits
[12:31 PM PT]  Vincent.Sylvester That's probably somewhere in Permissions
[12:31 PM PT]  Clifford.Hanger but not all
[12:34 PM PT]  Clifford.Hanger hi andrew
[12:34 PM PT]  Orbert.Tatham He lives
[12:34 PM PT]  Andrew Hellershanks: Hello, everyone.
[12:35 PM PT]  Andrew Hellershanks: I had to help the PSW who is helping my mother. She arrived at the time the meeting started.
[12:35 PM PT]  Orbert.Tatham Timing is everything
[12:35 PM PT]  Andrew Hellershanks: Yes, it is.
[12:36 PM PT]  Andrew Hellershanks: What have I missed?
[12:36 PM PT]  Vincent.Sylvester Speaking of. Change in LSL spec moving llCreateLink delay from 1000ms to 100ms. I did some tests with large linksets and it adds quite a bit of cpu. Any lower intervals, like 10ms, just crash the threads directly
[12:38 PM PT]  Clifford.Hanger i have to ask
[12:38 PM PT]  Ubit Umarov: our code stays in 1000ms
[12:38 PM PT]  Clifford.Hanger any development on voice?
[12:39 PM PT]  Ubit Umarov: the default. tha tis
[12:39 PM PT]  Clifford.Hanger hides
[12:39 PM PT]  Vincent.Sylvester I don't see the point of changing that either, the cpu increase really isn't worth it
[12:40 PM PT]  Vincent.Sylvester 500ms I could justify perhaps, but a tenfold change is a bit much
[12:42 PM PT]  Andrew Hellershanks: My thought is to wonder why it requires so much CPU.
[12:42 PM PT]  Ubit Umarov: and fun when some objects can have 1k prims or more...
[12:42 PM PT]  Vincent.Sylvester Linking is a heavy operation, it has to create a new object and copy all the parts
[12:44 PM PT]  Ubit Umarov: content that can be changed at runtime is great..  but ll is just suicidal
[12:44 PM PT]  Andrew Hellershanks: Why does it need to create a new object. You have two SOP/SOG items. It should not that much more complicated than updating one to link in to the other.
[12:44 PM PT]  Andrew Hellershanks: It should be not that much more...
[12:44 PM PT]  Ubit Umarov: seems they do not understand that is main reason why viewers at so bad relative to game engines
[12:45 PM PT]  Ubit Umarov: and one thing is content modified by humans, other it by scripts..
[12:45 PM PT]  Joe Magarac: Linking viewer side is cheap. It's just reparenting.
[12:46 PM PT]  Ubit Umarov: not really
[12:47 PM PT]  Ubit Umarov: but details...
[12:47 PM PT]  Joe Magarac: I've had to code it from scratch. It doesn't do much. Doesn't even cost an object creation.
[12:53 PM PT]  Andrew Hellershanks: The code handling linking of objects in the OS code used to have a lot of unneeded overhead but it was a lot better the last time I looked at it.
[12:53 PM PT]  Clifford.Hanger I take it, its a no?
[12:53 PM PT]  Andrew Hellershanks: Clifford, you would have better luck asking the people working on voice solutions for OS.
[12:53 PM PT]  Ubit Umarov: you call it unneeded :p
[12:54 PM PT]  Clifford.Hanger who would that be?
[12:54 PM PT]  Vincent.Sylvester Still have to figure out how to fix that unlink bug
[12:55 PM PT]  Andrew Hellershanks: Ubit, It used to do a lot of work unlinking nodes from one list and then linking them to the other list, one at a time. With two linked lists the tail of one can point at the head of the next and you're done. The only tricky part is changing the root prim of the second object in to an ordinary prim then doing some renumbering.
[12:57 PM PT]  Ubit Umarov: wel i did change that link code yes..
[12:57 PM PT]  Andrew Hellershanks: Clifford, The first set of people that comes to mind are the ones behind EchoVoice.
[12:57 PM PT]  Andrew Hellershanks: Ubit, yes. It was changed and is a lot better.
[12:58 PM PT]  Vincent.Sylvester Voice is going to require hundreds of hours of work most likely in just figuring out how it works and how it can be replaced. Unless it starts to rain money from the sky that'll take months of spare time
[12:59 PM PT]  Clifford.Hanger understood
[12:59 PM PT]  Clifford.Hanger but i thought it had been decided that echovoice was a non starter
[1:00 PM PT]  Vincent.Sylvester Anything IMA is a non starter
[1:00 PM PT]  Andrew Hellershanks: Clifford, keep an eye on the OSCC schedule. There might be a panel about voice support. That would be a good time to find out what is going on.
[1:01 PM PT]  Joe Magarac: Have to go, next meeting.
[1:01 PM PT]  Clifford.Hanger isnt that an IMA gig?
[1:01 PM PT]  Andrew Hellershanks: ok, Joe. Thanks for dropping by.
[1:01 PM PT]  Clifford.Hanger yes Joe sorry me too
[1:01 PM PT]  Clifford.Hanger
[1:01 PM PT]  Andrew Hellershanks: IIRC, EchoVoice is an IMA thing. I don't remember off hand what other projects there are underway related to voice.
[1:02 PM PT]  Clifford.Hanger if you want to point me in a better direction
[1:02 PM PT]  Lyr Lobo:
[1:03 PM PT]  Lyr Lobo: Here is a compact version of the OSCC schedule
[1:03 PM PT]  Ubit Umarov: ty :)
[1:03 PM PT]  Lyr Lobo: more events are adding this week for the zoom sessions following the conference and community events
[1:03 PM PT]  Lyr Lobo: if you wish to host one, please add it to the Get Involved link
[1:03 PM PT]  Lyr Lobo:
[1:03 PM PT]  Lyr Lobo: Many thanks *smiles*
[1:04 PM PT]  Andrew Hellershanks: Thank you, Lyr.
[1:04 PM PT]  Lyr Lobo grins
[1:05 PM PT]  Andrew Hellershanks: Any other questions or comments for today before we wrap things up?
[1:06 PM PT]  Andrew Hellershanks: I don't see anyone typing so I will take that as us being done for this week.
[1:06 PM PT]  Ubit Umarov: :)
[1:07 PM PT]  Andrew Hellershanks: The start time for this meeting coincides with the time of a person coming to the house to help my mother so I will often be stepping away during the meeting of may be late depending on when the PSW arrives on the day.
[1:07 PM PT]  Lyr Lobo: Thank you!
[1:07 PM PT]  Andrew Hellershanks: Thank you all for coming. See you again next week.
[1:07 PM PT]  Lyr Lobo: Most understandable
[1:08 PM PT]  Ubit Umarov: ok rl calls..  cya later :)
[1:08 PM PT]  Andrew Hellershanks: s/of may/or may/
[1:08 PM PT]  Lyr Lobo: Have a great week
[1:08 PM PT]  Lyr Lobo: off to make breakfast and chase a class
[1:08 PM PT]  Lyr Lobo: *waves*
[1:08 PM PT]  Andrew Hellershanks: Vincent, I am hazarding a guess that you will be posting the meeting log any moment now as I wasn't here for the first part of it?
[1:09 PM PT]  Andrew Hellershanks waves to Lyr.
[1:09 PM PT]  Vincent.Sylvester heh
Personal tools
About This Wiki