Chat log from the meeting on 2022-03-01

From OpenSimulator

Jump to: navigation, search
[11:03] Ubit Umarov: and sadly can get a lot worse
[11:03] Gavin.Hird see, there he is
[11:03] Ubit Umarov: lets see
[11:03] Andrew Hellershanks: Hello, everyone.
[11:03] Gavin.Hird I ran an optimize includes job on the viewer - took 22 hours to finish
[11:03] Gavin.Hird Hi Andrew
[11:03] Ubit Umarov: :)
[11:04] Andrew Hellershanks: The cat almost always wants play time just before the meeting is about to start.
[11:04] Ubit Umarov: ofc he does
[11:04] Ubit Umarov: err or she?
[11:04] Andrew Hellershanks: she
[11:04] Gavin.Hird the cat know he will be ignored for the next hour
[11:04] Ubit Umarov: :)
[11:05] Gavin.Hird did you ask the cat for its pronouns? - it might identify as something else
[11:06] Ubit Umarov: well ..
[11:06] Ubit Umarov: so what is new in opensim?
[11:07] Andrew Hellershanks: The version number has been bumped. We are now at
[11:07] Ubit Umarov: :)
[11:07] Ubit Umarov: yeah release basicly done
[11:07] Gavin.Hird announced too?
[11:07] Ubit Umarov: just didn't sent the emails
[11:07] Gavin.Hird ah
[11:08] Ubit Umarov: site ready, i hope
[11:08] Ubit Umarov: well the release notes may need a little love
[11:08] Ubit Umarov whispers: but not much to tell :(
[11:08] Gavin.Hird what are the highlights?
[11:09] Ubit Umarov: guess none
[11:09] Gavin.Hird there must be one jucy bug fix?
[11:09] Ubit Umarov: just a few bug fixes and the things we spoke abotu here
[11:10] Jamie.Jordan hey everybody
[11:10] Vincent.Sylvester Some changes to messaging that had a funny bug in there... after it was first fixed, is now fixed again. Bunch of code refactoring to bring it more in line with reasonable standards
[11:10] Vincent.Sylvester Some libomv updates here and there
[11:10] Ubit Umarov: a change does make live easier on standalones
[11:10] Ubit Umarov: again no need to also add the region name with region flags
[11:11] Ubit Umarov: that was a pain for standalones
[11:12] Vincent.Sylvester Enforcement of notecard size restriction which is nice
[11:12] Ubit Umarov: and without that, on  pl just can't login
[11:12] Ubit Umarov: well we hope all changes are nice :P
[11:12] Ubit Umarov: just no shinny one
[11:12] Vincent.Sylvester Haven't had to make another "revert bad core patch" commit yet so all good so far
[11:13] Ubit Umarov: well that can happen always
[11:13] Ubit Umarov: :)
[11:13] Ubit Umarov: in fact that was the role of
[11:14] Ubit Umarov: i did plan to release in jan, with any possible "revert bad core patch
[11:14] Vincent.Sylvester A lot of changes going in to save some cpu timing, which is nice to see given cpu load remains a big factor. While memory consumption has been going down steadily, cpu remains high
[11:14] Ubit Umarov: well cpu is needed to do things :)
[11:15] Ubit Umarov: but does do a few less sily things now
[11:15] Vincent.Sylvester I'm still advocating let's waste 100mb of memory and see what cpu we can trade for that
[11:15] Vincent.Sylvester memory slightly easier to get now than cpu power
[11:16] Vincent.Sylvester And I don't mean more caches
[11:16] Ubit Umarov: things like  if(auuid == ) where a waste of cpy
[11:16] Ubit Umarov: cpu
[11:16] Vincent.Sylvester Yep a lot of those were changed up
[11:17] Ubit Umarov: i added uuid.iszero and isnotzero :)
[11:17] Ubit Umarov: string == "" also
[11:17] Ubit Umarov: jezz
[11:18] Ubit Umarov: so some CO2 reduction :)
[11:19] Vincent.Sylvester Wonder how many more of that is left now, can't be that much
[11:20] Ubit Umarov: MB did fix a typo od mine on a ini file
[11:21] Ubit Umarov: so no code changes dusring last week
[11:21] Ubit Umarov: during..
[11:21] Ubit Umarov: hmm i did add some code to avoid a null ref also
[11:21] Ubit Umarov: both on the release
[11:22] Vincent.Sylvester A few of these I sometimes wonder if anyone even ran into them, but I know not to try, cause with my luck I know I'll probably find something even more obscure
[11:23] Andrew Hellershanks: :)
[11:24] Ubit Umarov: :)
[11:25] Vincent.Sylvester Did some work on the wiki fixing up some ancient pages with accurate information and built out the distributed robust section a little bit. Resolving and confirming some mantis stuff to clean that mess up.
[11:26] Ubit Umarov: btw having issues with pgsql gavin?
[11:26] Ubit Umarov: vincent tells there are some now
[11:26] Ubit Umarov: well he went play with recent versions...
[11:26] Vincent.Sylvester a foreign concept to some it seems
[11:27] Vincent.Sylvester There is a mantis on this as well
[11:27] Gavin.Hird huh?
[11:27] Vincent.Sylvester 8959
[11:27] Gavin.Hird issues, why?
[11:28] Gavin.Hird sorry I was way into optimizing includes ...
[11:28] Ubit Umarov: :)
[11:29] Vincent.Sylvester Using ubuntu focal, latest mono and postgres you get a cert error, if you then update npgsql and recompile you get OpenSim to run, but type conversion errors breaking login
[11:29] Gavin.Hird they both built and run
[11:29] Vincent.Sylvester npgsql requiring some hacks to fully work, which are just a result of poor code using unsafe conversions
[11:30] Gavin.Hird I have not tested Postgresql 12
[11:30] Vincent.Sylvester Think the version I tried was 14 something
[11:31] Gavin.Hird I think it will not work because of the ancient npgsql
[11:31] Gavin.Hird it does not have the security mechanism
[11:31] Vincent.Sylvester It can probably be fixed by building an altered version of npgsql to remove some of the hacks in there
[11:31] Gavin.Hird hacks?
[11:32] Andrew Hellershanks: Hello, Michael.
[11:32] Vincent.Sylvester compiler.unsafe something thing
[11:32] Ubit Umarov: you can't
[11:32] Ubit Umarov: that is a deep change
[11:32] Michael.Christopher hello, sorry im late
[11:32] Gavin.Hird I was about to say good luck with that
[11:32] Gavin.Hird Hi Michael
[11:33] Vincent.Sylvester Looking at npgsql source being a big mess, no idea who worked on it, but left behind just hacks in a lot of places to get around postgres changes it seems. Makes it hard to follow the code paths
[11:33] Ubit Umarov: you just need to use things really made for .net framework
[11:33] Ubit Umarov: not confusing and messy .net crap 1.0, crap 3.2--- e tc
[11:33] Gavin.Hird where did you look at npgsql code - on git?
[11:34] Vincent.Sylvester I think it was github
[11:34] Ubit Umarov: at this point a lot of c# software is just BAD
[11:34] Ubit Umarov: ppl lost on the ms mess
[11:34] Vincent.Sylvester cloned the whole thing and went back to the release tag for 3.2.7
[11:35] Vincent.Sylvester Was just going to remove the namespaces that force those two dlls to be required and try to fix whatever error compile throws one at a time
[11:35] Vincent.Sylvester Which is not a fun task
[11:35] Ubit Umarov: you can't fix things like that :p
[11:35] Ubit Umarov: use of the new unsafe hacks is deep low level code :)
[11:35] Vincent.Sylvester Just means I need a bigger shovel
[11:36] Andrew Hellershanks: hehe
[11:36] Michael.Christopher whispers: thats a more polite phrasing than I was thinking :))
[11:36] Gavin.Hird it is probalby easier to fix the postgres code in opensim...
[11:36] Ubit Umarov: those , just need to wait the rest of code is all in .net7.0
[11:36] Ubit Umarov: ot 8.0
[11:36] Ubit Umarov: or..
[11:37] Ubit Umarov: .net5.0 and .net6.0 may stll just be useless trash
[11:37] Vincent.Sylvester There is a reason for all that though. As database software adds new functions and performance enhancements it is looking like it might be time to rework the database interfaces we do have to get some performance back
[11:37] Ubit Umarov: at least does not justify any migration
[11:38] Vincent.Sylvester With mariadb now adding uuid fields storing the same data with less bytes there are savings to be made
[11:38] Ubit Umarov: a lot of things this old dbs add is just "flowers"
[11:38] Michael.Christopher but if more flowers bring in more of the general public to enjoy those flowers, we'll take it, right?
[11:39] Gavin.Hird unless I remember wrong Postgresql 14 add serious chages for security and scalability
[11:39] Vincent.Sylvester There have been a lot of changes to both postgres and mariadb, which, I think at least, warrant a look at as they could provide some performance improvements
[11:40] Vincent.Sylvester As data grows you eventually need every little byte
[11:40] Ubit Umarov: well and mariadb case is a bt workse
[11:40] Michael.Christopher especially to increase graphics quality as equipment capacity rises
[11:41] Ubit Umarov: since it did split from mysql, that may mean we need to add all the stack for it
[11:41] Ubit Umarov: all our data stack for it
[11:41] Ubit Umarov: currently it is jsut the mysql one
[11:41] Vincent.Sylvester I did create migrations for it, haven't cleaned them yet though
[11:41] Gavin.Hird withe legacy password
[11:42] Ubit Umarov: it is migrations and all access code
[11:42] Vincent.Sylvester Also changed everything needed for it to connect, but can't get tests to work proper still
[11:42] Ubit Umarov: if it keeps spliting from mysql
[11:42] Vincent.Sylvester Meanwhile mysql is moving forward at a snails pace
[11:42] Gavin.Hird not if you pay for it
[11:43] Gavin.Hird isn't that the whoe point - drive you into depseration and the paid for version?
[11:43] Gavin.Hird whole*
[11:44] Vincent.Sylvester Technically goes for most of this type of software out there
[11:44] Michael.Christopher least any that ofer paid versions
[11:44] Gavin.Hird yes
[11:45] Gavin.Hird Postgres does not exist as a paid for option at all
[11:45] Michael.Christopher hence why it isnt partially disabled :P
[11:45] Gavin.Hird many offer services, but not the db
[11:45] Gavin.Hird exactly Michael
[11:46] Vincent.Sylvester Well switching to postgres is a bit off the table if the latest version doesn't play well with OpenSim, cause at the very least in terms of security you want to keep your things updated or within lts support
[11:47] Ubit Umarov: who cares about security :p
[11:47] Michael.Christopher wasnt paying enough attention to realize it had an issue with OS
[11:47] Vincent.Sylvester Sticking to the edge of EOL like OpenSim used to do with mono is not fun
[11:47] Gavin.Hird there is basically  a missing cast in libopenmetaverse that keeps us from updating pgnsql
[11:47] Ubit Umarov: those things should be buried inside boxes only visible to opensim or other code
[11:47] Michael.Christopher ubit - only companies and people with privacy concerns
[11:47] Ubit Umarov: who cares abotu tls etc in those cases..
[11:47] Ubit Umarov: bahh
[11:47] Vincent.Sylvester Yes database bound to localhost only, that's a given I would hope
[11:48] Vincent.Sylvester Doesn't stop bugs entirely though
[11:48] Ubit Umarov: nothing stops bugs
[11:48] Ubit Umarov: bugs will rule the world one day
[11:48] Michael.Christopher i thought that step was in the basic setup tutorial (localhost access only setting)
[11:48] Ubit Umarov: well other kind of bugs :p
[11:49] Gavin.Hird nonorobots
[11:49] Gavin.Hird nanorobots
[11:49] Vincent.Sylvester People often elect to open their db, because they want to connect to it from the outside to change things, yet don't understand what ssh tunnels are
[11:49] Gavin.Hird the grey gooey syndrome
[11:49] Michael.Christopher :)) @Gavin - I was just thinking how long will it be till we have actual bugs going after electricity as a food :))
[11:49] Andrew Hellershanks: Vincent, or how to firewall the box to limit access to a given machine(s).
[11:50] Michael.Christopher vincent - or don't know how to set up a tunnel
[11:50] Ubit Umarov: well, unrelated, hopr i don't forget to send the release emails :)
[11:50] Ubit Umarov: ofc no need.. everyone keeps looking to our site
[11:50] Ubit Umarov: ;)
[11:50] Vincent.Sylvester If you firewall OpenSim properly everyone starts yelling they can't get their kitely deliveries
[11:51] Andrew Hellershanks: Ubit, I left you a reminder message.
[11:51] Michael.Christopher hmm.
[11:51] Ubit Umarov: andrew :)
[11:52] Michael.Christopher so how do we firewall properly while still allowing people to bring their purchases with them
[11:52] Andrew Hellershanks: Vincent, they just need a hole for Kitely's delivery service.
[11:52] Michael.Christopher a single delivery service acces spoint or one for every grid they want to bring from?
[11:53] Vincent.Sylvester Used to be that kitely kept changing their plug on that which was a nightmare, think the ip is now fixed
[11:53] Motoko.Karu whitelist kitelys IPs... route approp port
[11:54] Andrew Hellershanks: Vincent, yeah that would be a pain if Kitely kept changing IP. If the firewall could accept a fully qualified domain it would avoid that issue.
[11:54] Andrew Hellershanks: I think they did do something to address that issue.
[11:55] Michael.Christopher the changing IPs or the FQDN?
[11:55] Andrew Hellershanks: It probably was to used a fixed IP. IIRC, they had some notes about their delivery service on their web site.
[11:56] Vincent.Sylvester It would be nice if at some point in future focus would be put on the database interfaces, get them updated and maybe cleaned a bit to improve overall performance and adjust for the changes, that's all I ask
[11:57] Andrew Hellershanks: Sounds reasonable. No idea how much work that would entail.
[11:58] Michael.Christopher thats what I was thinking about andrew
[11:58] Vincent.Sylvester Getting that scoped out would be a start
[11:59] Andrew Hellershanks nods
[12:00] Andrew Hellershanks: We are at the top of the hour already. Any other last minute topics for today?
[12:00] Michael.Christopher I've got to head out, school pickup time.  Have a great week, DM me if you need anything
[12:00] Michael.Christopher nice timing andrew :))
[12:00] Andrew Hellershanks: :)
[12:00] Michael.Christopher see you all later
[12:01] Ubit Umarov: :)
[12:01] Andrew Hellershanks: ok, Michael. See you another time.
[12:01] Ubit Umarov: enjoy crashes
[12:01] Ubit Umarov: err i mean features
[12:01] Andrew Hellershanks: Not .2?  :)
[12:02] Motoko.Karu lol
[12:02] Ubit Umarov: release is
[12:02] Ubit Umarov: is current dev
[12:03] Andrew Hellershanks: Right. It does have a dev tag.
[12:03] Andrew Hellershanks: If there is nothing more for today I'll call this meeting to a close.
[12:03] Andrew Hellershanks: Thank you all for coming. See you again next week.
[12:04] Motoko.Karu Thanks everyone..
[12:04] Gavin.Hird I am still making progess of viewer version 3.0
[12:04] Selby.Evans Bye All
[12:04] Ubit Umarov: :)
[12:04] Motoko.Karu laters
[12:04] Jamie.Jordan thanks yall
[12:04] Gavin.Hird but not much specific to add
[12:04] Ubit Umarov: cya ppl
[12:04] Andrew Hellershanks: version 3.0? What is the current version number?
[12:05] Gavin.Hird 2.8.x for macOS and 1.9.8 on Windows
[12:05] Andrew Hellershanks: ok. Sounds like you have some big changes in the works.
[12:05] Gavin.Hird yes
[12:05] Andrew Hellershanks: :)
Personal tools
About This Wiki