Chat log from the meeting on 2022-09-13

 [11:09 AM PT] Ubit Umarov: welcome to another weekly meeting :) [11:09 AM PT]  Ubit Umarov: During last week there where code changes commited [11:10 AM PT]  Ubit Umarov: i did spent sometime playing with dotnet 6.0 migration issues [11:10 AM PT]  Ubit Umarov: in fact got my first win11 BSOD in the process [11:11 AM PT]  Andrew Hellershanks: Bah. Was deep in to some coding of a threaded interpreter language. Wasn't looking at the clock.:P [11:12 AM PT]  Ubit Umarov: but with several hacks i did manage to compile and run a region in windows compiled and mostly using .net6.0 runtime [11:12 AM PT]  Andrew Hellershanks: Hello, everyone. [11:12 AM PT]  Ubit Umarov: in the process i had to remove Xengine, that does depend on features no longer present in dotnet [11:14 AM PT]  Ubit Umarov: also some of out threading contol lost the ability to stop threads, something dotnet did remove, because they decided it is bad [11:14 AM PT] Ubit Umarov: so there is no way to stop a  thread stucked anywhere and try to recover from it [11:15 AM PT]  Ubit Umarov: in a automatic way.. [11:16 AM PT] Ubit Umarov: well can understand why they removed, seems they actually never did it right, if that is even possible.. [11:16 AM PT] Ubit Umarov: hopefully we may live without it.. but it is another lose.. [11:17 AM PT] Ubit Umarov: i did compile out main code in .net6.0,  warp3d and libomv [11:18 AM PT] Ubit Umarov: adding 2 extra dlls.. one to extend the support for image/bitmap to .net6, other to have a component needed for log4net, that they also removed, but still possible to add back [11:19 AM PT] Ubit Umarov: so as i said i did run a basic standalone on sqlite, even did load a oar and tested a car under yengine and ubode [11:19 AM PT] Ubit Umarov: the tested on WSL, the windows linux "emulator" [11:20 AM PT] Ubit Umarov: first time i fired opensim i got my first win11 BSOD [11:21 AM PT] Andrew Hellershanks: :) [11:21 AM PT]  Ubit Umarov: thing mostly a WSL issue... i was out of ram, and WSL did blowup instead of normal use of swap space etc.. [11:21 AM PT]  Ubit Umarov: think.. [11:22 AM PT]  Ubit Umarov: othen ofc opensim just failed to load sqlite low level dlls [11:22 AM PT]  Ubit Umarov: showing that we need to rewairte he entire way how dlls are loads [11:22 AM PT]  Ubit Umarov: dotnet, fails to do what mono does fien [11:22 AM PT]  Ubit Umarov: fine [11:23 AM PT]  Ubit Umarov: so im stil far to check if the image code does still work in linux  on dotnet6.0 [11:24 AM PT]  Ubit Umarov: they did remove it, but they added a flag to allow it again. if one adds the extra dlls [11:24 AM PT]  Ubit Umarov: that are ofc diferent from the windows ones... [11:25 AM PT]  Ubit Umarov: making cross platform even more fun [11:25 AM PT]  Andrew Hellershanks: :) [11:26 AM PT] Ubit Umarov: so.. lots of work to get less features and getting into a lot more unreliable software base [11:27 AM PT] Ubit Umarov: ofc this is also only try to compile and run the software [11:27 AM PT] Ubit Umarov: hard to tell and fix all the hidden breaking changes they made all over the code base [11:28 AM PT] Ubit Umarov: soll well that was some code work during last week... a very frustrating experience [11:29 AM PT] Ubit Umarov: meanwire there was a new FS release [11:30 AM PT] Ubit Umarov: seems working, just profiles images fail to load in full resolution a lot [11:30 AM PT] Ubit Umarov: told Beq abotu it, she may look [11:30 AM PT] Ubit Umarov: meanwhile the had some work on profiles for the next release [11:31 AM PT] Andrew Hellershanks: I downloaded the latest FS for SL and OS this past week. I haven't used either of them much. What I'm noticing right now is that my CPU load seems lower than before. [11:31 AM PT] Ubit Umarov: FS did merge new LL code for profiles that did kill all profiles for opensim [11:32 AM PT] Ubit Umarov: replacing all the lludp based protocol by new ( BAD)  http caps based [11:33 AM PT] Ubit Umarov: Beq was busy putting back support for opensim, and last time we talked, 2 days ago? she had it mostly done [11:33 AM PT] Ubit Umarov: meanwhile i did look to the new profiles protocol,, and it is just really BAD [11:34 AM PT] Ubit Umarov: not easy to jsut add it, without changes from region code up to grid databases apis [11:35 AM PT] Ubit Umarov: since Beq did recover the lludp,  guess i will not try to add that new nonsense ll protocol [11:36 AM PT] Ubit Umarov: see.. change from lludp is change, means use of TCP that is a lot slower and heavy [11:36 AM PT] Ubit Umarov: means changign from a very compact information serielization, to LLSD based one, and XML based oen [11:37 AM PT] Ubit Umarov: that means that 90% of the data payload is XML support only.. [11:37 AM PT] Ubit Umarov: but in this case they made it worse [11:37 AM PT] Ubit Umarov: in profiles, a packet did include several changes [11:38 AM PT] Ubit Umarov: like abotu text, UUIDs if pictures etc etc [11:38 AM PT] Ubit Umarov: so most of the edit on he profile was local on viewer, and only commited to region at end [11:39 AM PT] Ubit Umarov: and sent all in a nice packet [11:39 AM PT] Ubit Umarov: now the new procole makes a CAP call per item changed... [11:39 AM PT] Ubit Umarov: as we do the changess [11:40 AM PT] Ubit Umarov: ir as soon we change about test, it sends it to server.. [11:40 AM PT] Ubit Umarov: the uuid of a pocture.. [11:40 AM PT] Ubit Umarov: etc [11:40 AM PT] Ubit Umarov: picture [11:41 AM PT] Ubit Umarov: that is a extra overhead on comms [11:41 AM PT] Ubit Umarov: and ofc all out code from region to grids only handles that packet with all information [11:42 AM PT] Ubit Umarov: so we would need to add APIs to all the tine changes all over the code base [11:42 AM PT] Ubit Umarov: (ine == tiny) [11:43 AM PT] Ubit Umarov: this does not include things like pics, classified etc [11:43 AM PT] Ubit Umarov: they forgot abotu those and they are still lludp [11:44 AM PT] Ubit Umarov: well.. so it is possible to support this changes, btu a lot of work for absulutly no gain [11:45 AM PT] Ubit Umarov: since Bq did restored old API; other viewers can do it also, so i will not try to add those changes, for now at least [11:46 AM PT] Ubit Umarov: some more as expected as LL find more things usin lludp and moves to http waste [11:46 AM PT] Ubit Umarov: well this are last week code news [11:47 AM PT] Ubit Umarov: what abotu you.. what news do you have? [11:47 AM PT] Andrew Hellershanks: Thank you for that summary of the past week, Ubit. Was a busy week. [11:48 AM PT] Andrew Hellershanks: Nothing new on my end as I was still working on code needed for an industry trade show that started yesterday. [11:50 AM PT] Andrew Hellershanks: We have a few less people here today compared to our usual complement. Does any one have any questions or comments for today? [11:52 AM PT] Andrew Hellershanks: Are we going to finish early for a change? [11:53 AM PT] Jamie.Jordan @grid.kitely.com:8002: looks like [11:54 AM PT] Andrew Hellershanks: I should have some time to do some more testing on FS this coming week. I want to check where the latest FS really has lowered the CPU load. [11:55 AM PT] Andrew Hellershanks: My frame rate also seems a little higher than it was last week. [11:55 AM PT] Ubit Umarov: yes las fs release does include some performance changes ll did [11:55 AM PT] Ubit Umarov: on the rendering etc [11:56 AM PT] Andrew Hellershanks: Yes, I read that in the release notes. It does appear to have made a noticeable difference on my machine. [11:56 AM PT] Ubit Umarov: yeah [11:56 AM PT] Ubit Umarov: sad teh profiles images [11:56 AM PT] Jamie.Jordan @grid.kitely.com:8002: nine too [11:56 AM PT] Ubit Umarov: sad.. [11:57 AM PT] Ubit Umarov: hopefully Beq or other will find why and fix, [11:57 AM PT] Andrew Hellershanks: My profile image came in right away and looks ok. [11:57 AM PT] Ubit Umarov: yeah it is more on load conditions [11:57 AM PT] Ubit Umarov: also at SL [11:58 AM PT]  Ubit Umarov: maybe changes on fetch priorities and timeouts [11:58 AM PT] Andrew Hellershanks: We can provide further reports about the latest FS next week. [11:59 AM PT] Ubit Umarov: only looked to one or 2 of the new features and they seem to work here also [12:00 PM PT] Andrew Hellershanks: We are at the top of the hour. Time to call this meeting to a close. My cat will be happy. She is meowing at me. :) [12:00 PM PT] Andrew Hellershanks: Thank you all for coming. See you again next week.