Chat log from the meeting on 2022-09-20

From OpenSimulator

Jump to: navigation, search
[11:00] Selby.Evans hi
[11:00] Vincent.Sylvester This beer tastes funny, but it was free
[11:00] Andrew Hellershanks: Hello, everyone.
[11:00] Ubit Umarov: beer? jezzz yr so dead
[11:00] Selby.Evans hi
[11:00] Ubit Umarov: that is so so old..
[11:00] Andrew Hellershanks: All beer tastes funny
[11:01] Vincent.Sylvester True, but all the Rum is actually gone
[11:01] Vincent.Sylvester Let's see this week we had a few minor commits I think Ubit can best describe what they were I haven't really looked
[11:02] Ubit Umarov: oh? there where commits?
[11:02] Andrew Hellershanks: one or two. :)
[11:02] Jamie.Jordan Hi everybody
[11:03] Ubit Umarov: ahh yes
[11:03] Andrew Hellershanks: Hello, Jamie
[11:03] Ubit Umarov: change some code using things that will be gone from .net
[11:03] Ubit Umarov: some in fact obsolete for ages
[11:04] Ubit Umarov: other that i seen that will no work he same way
[11:04] Ubit Umarov: so i changed before i forget :)
[11:05] Andrew Hellershanks: good idea
[11:05] Ubit Umarov: one that would totally stop working is our local console code
[11:06] Ubit Umarov: in fact looking to the ms changes,  one can question how it did work
[11:07] Ubit Umarov: thing is that despite ms claims so far, console code is actually not that thread safe
[11:07] Ubit Umarov: namely using console.readkey
[11:08] Ubit Umarov: i meant using that at same time onther thread reads or changes cursor positions for exampel
[11:09] Ubit Umarov: in later frameworks ms added locks, so readky will block
[11:09] Ubit Umarov: i added a workaround that seems working
[11:10] Vincent.Sylvester Their changes make me wonder what they think the usecase of .net is as I would say what OpenSim does isn't particularly "out there". MS making changes to please "someone" or themselves perhaps not making logical changes, only changes that work for their usecase everyone else be damned huh.
[11:10] Ubit Umarov: sadly adds some lag to console typing, and makes a thread awake up  every 100ms  when in past it was possible just sleeping waiting for a signal
[11:11] Ubit Umarov: well on this, thing is that console is a bit complex thing
[11:11] Ubit Umarov: in fact, at least on linux, the framework does connect to a terminal
[11:12] Ubit Umarov: kinda like old rs232 connections to a vt100 terminal
[11:12] Ubit Umarov: to read a cursor position means  sending a control sequence of chars and getting a answer back
[11:13] Ubit Umarov: on the same virtual serial link...
[11:14] Ubit Umarov: ms found they could not "interleave" that on waiting readkey
[11:14] Ubit Umarov: so they added locks to serielize that..
[11:14] Ubit Umarov: totally breaking code like ours
[11:15] Ubit Umarov: if that is a realy issue ms did found, or just a big ms mistake,, can't tell
[11:15] Vincent.Sylvester Sounds like a sledgehammer through door kinda fix on their end
[11:15] Ubit Umarov: fact is that they changed the code, and refuse any fixes
[11:16] Ubit Umarov: did told they would change consose on .net7.0..  but seen nothing
[11:16] Vincent.Sylvester At this point you can basically expect such "fixes" from MS, seems their modus operandi in terms of .net and most other things
[11:17] Andrew Hellershanks: NIH syndrome?
[11:17] Ubit Umarov: well with this, i do have opensim running compiled to .net6.0 and under its runtime
[11:17] Ubit Umarov: both on windows and linux
[11:17] Vincent.Sylvester Well most of it, not everything
[11:18] Ubit Umarov: sadly still some hand made dlls changes
[11:18] Ubit Umarov: since all we had for mono, will just not work
[11:18] Vincent.Sylvester Then again I think XEngine has served us well over the years, but YEngine just works so much better
[11:18] Ubit Umarov: and ofc course we lost things like Xengine
[11:19] Ubit Umarov: also i have like 570 warnings abotu obsolete code
[11:19] Vincent.Sylvester I did some scripting past week and it's so nice to just refer to spec and things work as intended not having to massage every piece into place like XEngine had a tendency to
[11:19] Ubit Umarov: that will not work on .net7.0
[11:19] Andrew Hellershanks: 570 warnings? Eek. That will take a while to sort out.
[11:19] Ubit Umarov: like the bitmaps, fundamental to us
[11:19] Ubit Umarov: not reay
[11:19] Ubit Umarov: its like 3 or 4 things
[11:20] Ubit Umarov: like All  Image code...
[11:20] Vincent.Sylvester Andrew just disable all the pragma stuff and you get a ton of warnings using .net4.8 already. Only some of them are really hinting at issues
[11:20] Ubit Umarov: and of httprequest they consider dead
[11:20] Ubit Umarov: and a few simllar things
[11:21] Ubit Umarov: but those will work on .net6.0 adding some extra dlls
[11:21] Ubit Umarov: will not on net7 as i said
[11:22] Ubit Umarov: i have zero warnings now after adding orders to ignore some warnings :)
[11:22] Ubit Umarov: <SuppressWarnings>CA1416,SYSLIB0011,SYSLIB0014</SuppressWarnings>
[11:23] Vincent.Sylvester "Jesus take the wheel" type coding xD
[11:23] Andrew Hellershanks: hehe
[11:23] Ubit Umarov: is there any other form of coding with ms ?
[11:23] Andrew Hellershanks: Always have to be careful when you ignore warnings. You need to be sure which ones are safe to ignore.
[11:23] Ubit Umarov: well on opensource
[11:23] Ubit Umarov: remember that mono was a big disaster also
[11:24] Vincent.Sylvester Some of the warnings it spews even on 4.8 are stupid, like it complains about events, but how else are you supposed to handle that stuff, even MS doesn't offer a replacement when you look for the warning id
[11:24] Ubit Umarov: mono 4 was a huge fail for us
[11:24] Ubit Umarov: well those warnings are on another class of warnings
[11:24] Ubit Umarov: the ones im talking abotu are really thigns will will need to fix someday
[11:25] Ubit Umarov: fact this things do work on net6 means thats a few more years
[11:25] Ubit Umarov: hopefuly
[11:25] Andrew Hellershanks nods
[11:26] Vincent.Sylvester Well take everything MS says with a grain of salt these days
[11:26] Vincent.Sylvester "Windows 10 is the last windows"...
[11:26] Ubit Umarov: well meanwhile, some are using a new fs beta
[11:26] Ubit Umarov: that only has the new SL stupid profiles code
[11:26] Ubit Umarov: so profiles are all broken
[11:27] Andrew Hellershanks: Vincent; yeah, we know how that turned out.
[11:27] Ubit Umarov: does not have the fixes Beq made to recover llUDP; like i said last wee
[11:27] Ubit Umarov: possible similar issues with other beta viewers out there
[11:28] Vincent.Sylvester That profile thing I really don't get, it makes sense to only send partial updates sure, but then sending them over http... instead of oh idk caching and waiting till they are done making edits every two seconds and sending a full update then over unblocked udp
[11:28] Ubit Umarov: no it makes no sense
[11:28] Vincent.Sylvester One step forward two steps back
[11:28] Ubit Umarov: their changes are just... well whatever
[11:29] Vincent.Sylvester Right idea, horrible execution, as usual
[11:29] Ubit Umarov: and for us, a lot harder to implement, as i said
[11:29] Ubit Umarov: not even right idea
[11:29] Ubit Umarov: ahh hey removed Interest also
[11:30] Ubit Umarov: that tab is gone
[11:30] Ubit Umarov: "interests"
[11:30] Vincent.Sylvester Well I am not sure how many did use that in earnest to be fair, but it is nice to have and doesn't hurt anything
[11:30] Ubit Umarov: at least on that Beta fs
[11:30] Vincent.Sylvester Far as I remember it's just some bit field thing in the db anyways, so not like it takes up that much space
[11:31] Andrew Hellershanks: I know I set that flags in my profile but I rarely if ever looked at those entries in the profile of someone else.
[11:31] Ubit Umarov: well LL had delete all this profiles
[11:31] Ubit Umarov: using only web ones
[11:32] Ubit Umarov: now decided to put them back, using firestorm code..  but changing he protocol
[11:32] Ubit Umarov: mas made a stupid protocol
[11:32] Ubit Umarov: sorry the "stupid"  but that is a exact description :p
[11:33] Vincent.Sylvester Trying to do everything, succeeding at nothing, that's what it feels like lately
[11:33] Ubit Umarov: well it is ll, not our thing, except that we get the effects
[11:34] Ubit Umarov: lets hope Beq can add her fixes to it
[11:34] Vincent.Sylvester There is not quality control internally I feel, nothing that says "hey let's think for a minute"
[11:34] Ubit Umarov: nah hthe have control
[11:35] Ubit Umarov: they break things on a very controled way!!
[11:35] Vincent.Sylvester lol
[11:36] Andrew Hellershanks: That's another reason we need OS specific viewers that don't depend on LL for their main code base.
[11:36] Ubit Umarov: possible this where the server side guys...
[11:36] Ubit Umarov: whatever...
[11:36] Ubit Umarov: i will not add the caps
[11:36] Ubit Umarov: viewers for opensim MUST use current api
[11:37] Ubit Umarov: well at least in near times
[11:37] Ubit Umarov: i mean will not add in near time
[11:38] Vincent.Sylvester Always treating their changes as long term tests, often enough they reverted or changed to something more logical later on
[11:38] Ubit Umarov: as i told would need to add a very twisted caches collectorm to make a entire new API from the viewer cap to all the databases
[11:39] Ubit Umarov: because our protocol only suports changes on a block of data they now did split
[11:39] Ubit Umarov: protocols even
[11:39] Ubit Umarov: ofc and there are profiles modules
[11:39] Ubit Umarov: that wil be dead also
[11:40] Andrew Hellershanks: The change affects the third party profile modules?
[11:40] Ubit Umarov: yeah
[11:40] Andrew Hellershanks: :P
[11:40] Vincent.Sylvester Ultimately it's what, same behavior just 30% more traffic used for some minor profile update
[11:41] Ubit Umarov: less behavior
[11:41] Ubit Umarov: no interests
[11:42] Ubit Umarov: really LL think http is a instantaneos crap
[11:42] Ubit Umarov: current coders could even add a cap to send a char by char on local chat
[11:42] Ubit Umarov: :p
[11:43] Ubit Umarov: well whatever..
[11:43] Ubit Umarov: sad we get the effects of their things like this..
[11:43] Andrew Hellershanks nods
[11:44] Ubit Umarov: as i said Beq already spent time putting our API back...
[11:44] Ubit Umarov: lets hope it gets on fs for opensim viewers soon
[11:44] Ubit Umarov: and others keep it also
[11:45] Andrew Hellershanks: We can but hope. :)
[11:45] Andrew Hellershanks: It is now quarter to the hour. Cuga has an announcement.
[11:45] Andrew Hellershanks: Go ahead, Cuga.
[11:46] Andrew Hellershanks: Hm... have we lost Cuga?
[11:47] Vincent.Sylvester Let me get my poking stick
[11:47] Andrew Hellershanks: While we wait for Cuga, did anyone else have a question/comment for today?
[11:49] Andrew Hellershanks: I can mention one minor news items. This past weekend Kitely grid celebrated their 14th anniversary.
[11:49] Ubit Umarov: already 14?  time flies
[11:50] Andrew Hellershanks: I know. I didn't realize they had been running that long. One of the last events was a live music performance done via a media stream. The video had the audio and video of the person performing their music live.
[11:50] Cuga.Rajal sorry
[11:50] Cuga.Rajal some was at door
[11:51] Cuga.Rajal back
[11:51] Andrew Hellershanks: np, Cuga. You may go ahead with your announcement.
[11:51] Cuga.Rajal I was just going to mention that I re-wrote the HG authentication module that I was asking about 3 weeks ago
[11:51] Cuga.Rajal I put it on a github repo
[11:51] Cuga.Rajal
[11:52] Cuga.Rajal it's for inbound HG TPs, like the form they use at OSCC party
[11:53] Cuga.Rajal forces a form to be signed before allowing inbound HG
[11:53] Cuga.Rajal hopefully useful to others
[11:54] Cuga.Rajal thats all I had
[11:54] Andrew Hellershanks: I'm sure otherse will find it useful. You can let us know how much of a response you get to the project.
[11:55] Andrew Hellershanks: ty, Cuga.
[11:55] Cuga.Rajal I got one star already LOL
[11:55] Andrew Hellershanks: :)
[11:55] Andrew Hellershanks: That's a start.
[11:56] Andrew Hellershanks: Five minues left. Any other comments/questions/announcements for today?
[11:57] Jamie.Jordan great meeting yall
[11:57] Andrew Hellershanks: Anyone seen Kayaker recently? It seems a little unusual for him to be absent several weeks in a row.
[11:57] Jagga Meredith: *snores quiet;y*
[11:57] Andrew Hellershanks: ok, Jamie. tc. Thanks for coming.
[11:57] Andrew Hellershanks: It there is nothing more I will wrap up todays gathering.
[11:58] Jamie.Jordan think he's traveling
[11:58] Andrew Hellershanks: ok. That would explain it.
[11:58] Andrew Hellershanks: That will do it for this week. Thank you all for coming. See you again next week.
Personal tools
About This Wiki