Chat log from the meeting on 2022-01-25

From OpenSimulator

Jump to: navigation, search
[11:02] Selby.Evans Hi Michael
[11:02] Vincent.Sylvester There he is
[11:02] Michael.Christopher how has everyone been
[11:02] Selby.Evans Hi Andrew
[11:02] Gavin.Hird Hi Andrew
[11:02] Andrew Hellershanks: Hello, everyone.
[11:02] Andrew Hellershanks: Michael, Keeping busy. :)
[11:03] Michael.Christopher wish I could be less busy some days lol
[11:04] Andrew Hellershanks: I know what you mean, Michael.
[11:05] Vincent.Sylvester Gavin how is the object cache thing now, viewers fixed up nicely?
[11:06] Gavin.Hird The macOS version is fixed, but the latest Windows 10 update clobbered my build environment
[11:06] Gavin.Hird so I am not even sure if I sensibly can build a Windows version any more
[11:06] Gavin.Hird that is I can build it but not test it
[11:07] Michael.Christopher .....gotta love windows.
[11:07] Vincent.Sylvester As is tradition with Windows well I wish you the best of luck with that
[11:08] Gavin.Hird it is an ufortunate combination of Windows progress, vmware drivers lagging because of the move to Apple Silicon
[11:08] Vincent.Sylvester If all goes well I can give that a test this week, knock on wood nothing else gets in the way
[11:09] Gavin.Hird I was already working with a developer version of VMware they have not released because they broke OpenGL earlier this fall
[11:09] Michael.Christopher gavin - virtualizing your windows test environment?
[11:09] Gavin.Hird I don't think I will invest the €3000 needed to get a sensible Windows environment for building and testing the viewer
[11:10] Gavin.Hird it is being built in vmware on a Mac Pro
[11:10] Michael.Christopher nice
[11:10] Gavin.Hird I don't own any Windows native machines
[11:10] Ubit Umarov: ohh guess you can do  it with $300
[11:11] Michael.Christopher ok
[11:11] Gavin.Hird yeah, try get a  GPU for $300
[11:11] Ubit Umarov: with a tiny atom like cpu..  tiny gpu..
[11:11] Ubit Umarov: oh cpu internal one
[11:11] Michael.Christopher lol
[11:12] Vincent.Sylvester I got a spare machine you can use with enough horsepower, just gotta get around to installing windows on it
[11:12] Ubit Umarov: just compiles will take a bit lonnnnggg and viewers will do fpm not fps
[11:12] Gavin.Hird lol Ubit
[11:12] Ubit Umarov: ( frames per minute
[11:12] Michael.Christopher rofl@ubit
[11:12] Gavin.Hird I know
[11:12] Gavin.Hird the current build machine has a 12 core Xeon and a viewer rebuild takes 1 1/2 hours
[11:13] Ubit Umarov: 11 cores plaing poker, one working
[11:13] Andrew Hellershanks: ouch
[11:13] Michael.Christopher hey, i've got a windows xp era system in the other room that can still run firestorm or opensimulator...
[11:13] Vincent.Sylvester Debug or release config?
[11:13] Gavin.Hird 12 cores 24 threads
[11:13] Gavin.Hird release
[11:13] Vincent.Sylvester Yikes
[11:14] Vincent.Sylvester Can't say I am surprised though given the mess viewers are in a lot of ways
[11:14] Gavin.Hird just sorting out the icludes...
[11:14] Gavin.Hird if you open the build project in VS, it sorts out over 230000 includes
[11:14] Ubit Umarov: includes are just a huge thing on current c++
[11:15] Vincent.Sylvester Getting the correct python version and then the environment setup properly took me two hours, it's insane
[11:15] Ubit Umarov: tons and tons and tons of files to open
[11:15] Gavin.Hird right
[11:15] Gavin.Hird it is very fragile the whole thing
[11:15] Kayaker Magic: I like to joke that the makefile for many of these projects has more code in it than the actual source code of the projects.
[11:16] Vincent.Sylvester Really kills development that, no rapid iterations at that level of compile speed
[11:16] Michael.Christopher but IS it a joke kayaker?
[11:16] Gavin.Hird if you are lucky a small change will take 7ish minutes to build
[11:16] Kayaker Magic: The fact that you are not laughing makes it even funnier!
[11:17] Vincent.Sylvester It's kinda insane that level of mess is accepted and nothing seems to be getting done in terms of refactoring that stuff
[11:17] Gavin.Hird I have not made up my mind what to do with it
[11:17] Gavin.Hird 1.5 million lines of code
[11:18] Gavin.Hird excluding the libs
[11:18] Vincent.Sylvester No wonder then it cannot manage 60 fps most days even on the newest hardware huh, but we knew that already
[11:19] Vincent.Sylvester My concern is mainly if there might not be a better way for you to build and test, I'd sponsor the hardware, but shipping computers across europe in these times I wonder if it'll even get there
[11:19] Ubit Umarov: several years, several devs.. ofc it is comples
[11:19] Ubit Umarov: complex
[11:20] Gavin.Hird I have not made up my mind
[11:20] Gavin.Hird LL also seems to be kinda desperate trying to run the macOS viewer in vine
[11:21] Gavin.Hird or rather the Windows viewer in vine on macOS
[11:21] Ubit Umarov: not enough wine?
[11:21] Ubit Umarov: cofff coff ;)
[11:21] Vincent.Sylvester Those jokes write themselves lol
[11:21] Gavin.Hird whine
[11:21] Ubit Umarov: :)
[11:22] Kayaker Magic: Who's the whinemaker now?
[11:22] Gavin.Hird :-)
[11:22] Ubit Umarov: those things devs can't help? guess it is a big bug
[11:22] Vincent.Sylvester On the plus side at least the fix was quick and it's another drop in the bucket for OpenSim teaching LL once again how it's done
[11:22] Vincent.Sylvester So I am happy it got resolved, least of no longer thing I am just seeing things
[11:23] Ubit Umarov: nahh you do see things.. just other things now
[11:23] Vincent.Sylvester Script ghosting is a new thing, which I got an email about on saturday
[11:23] Vincent.Sylvester Apparently down to really slow HDD
[11:23] Andrew Hellershanks: Script ghosting?
[11:24] Ubit Umarov: script what?
[11:24] Vincent.Sylvester So the script has to make changes to a prim and that causes an update into the database, which is too slow to react
[11:24] Vincent.Sylvester Something strange going on with locking and timing somewhere
[11:25] Vincent.Sylvester Script checks if prim has new size and color and all sorts of other things done to it, but still sees old data
[11:25] Michael.Christopher hmm
[11:25] Vincent.Sylvester After moving the whole thing to SSD problems gone
[11:25] Ubit Umarov: most prim changes store are defered in time
[11:25] Vincent.Sylvester I thought so too, I been looking all over trying to find if it somehow uses database over local data, but nope
[11:25] Ubit Umarov: like for a few minutes even
[11:26] Gavin.Hird Isn't there a default dealy of 30+ seconds to flush to db?
[11:26] Ubit Umarov: there is even a rolling delay
[11:26] Gavin.Hird right
[11:26] Vincent.Sylvester Really weird, could be something going bad with the disk too who knows, now that it is moved anyways
[11:26] Ubit Umarov: ( up to a max delay that don't remember ofc )
[11:26] Andrew Hellershanks: The data may not have been persisted to the DBs yet but I would have expected it would know the current state of prims in the local scene.
[11:26] Michael.Christopher so the problem only happened when using an HDD, switching to  an SSD solved?
[11:27] Gavin.Hird I think max delay is around 300 sec
[11:27] Gavin.Hird unless you change it
[11:27] Gavin.Hird I flush every 10 seconds on my simulators
[11:27] Ubit Umarov: long since i touch that
[11:27] Vincent.Sylvester Really strange one that, never seen that in all the years
[11:27] Vincent.Sylvester I suspect there might have been something wrong with the disk too, too late to check now
[11:27] Michael.Christopher oh ok
[11:28] Michael.Christopher still a very interesting glitch to see
[11:28] Ubit Umarov: ofc a thing that does impact disk is logs
[11:28] Ubit Umarov: we have 2 many of those
[11:28] Ubit Umarov: but that case,  guess more and better debug needed
[11:29] Ubit Umarov: well assuming the case itself is already defined
[11:29] Vincent.Sylvester I might have some time to try this limiting iops or something
[11:29] Vincent.Sylvester Though quite a lot of variables I can't simulate
[11:30] Vincent.Sylvester Just been going through scene update code and so on to see if there is anything that doesn't just re-use the local data
[11:30] Gavin.Hird In OpenSim,ini
[11:30] Gavin.Hird ;# {MinimumTimeBeforePersistenceConsidered} {} {Time before un-changed object may be persisted} {} 60
    ;; Objects will be considered for persistance in the next sweep when they
    ;; have not changed for this number of seconds.
     MinimumTimeBeforePersistenceConsidered = 10

    ;# {MaximumTimeBeforePersistenceConsidered} {} {Time before changed objects may be persisted?} {} 600
    ;; Objects will always be considered for persistance in the next sweep
    ;; if the first change occurred this number of seconds ago.
     MaximumTimeBeforePersistenceConsidered = 100
[11:30] Gavin.Hird defaults are 60 and 600 seconds
[11:31] Ubit Umarov: well assuming comments do match code :)
[11:31] Gavin.Hird lol
[11:31] Gavin.Hird I think it does
[11:31] Ubit Umarov: backup does it
[11:31] Kayaker Magic: But I thought the code was the documentation!
[11:31] Ubit Umarov: so does shutdown
[11:32] Ubit Umarov: all objects waiting to store are stored..
[11:32] Vincent.Sylvester Weird thing is that it failed the first "self test", but a few seconds later the script would report all good, so I am suspecting some timing or locking in code, which given the async nature is always a potential
[11:32] Ubit Umarov: but old things
[11:32] Ubit Umarov: guess that is not changed basicly since
[11:32] Vincent.Sylvester In the end just one more reason to use SSD over slow disks to me
[11:33] Vincent.Sylvester The more data you deal with the clearer that becomes just from perspective of latency
[11:33] Ubit Umarov: well possible looking to script may help
[11:33] Gavin.Hird even the slowest disk does not give you a 60 second delay before comitted to platter :-)
[11:34] Ubit Umarov: some do have write cash even :p
[11:34] Gavin.Hird maybe if you have the origian 5 MB RAMAC
[11:34] Vincent.Sylvester Yeah that's the hiccup point for me as well, but if other data is queued up too
[11:35] Gavin.Hird all disks have a write cache
[11:35] Ubit Umarov: but think before jumping down to disk, one needes better debug before
[11:35] Gavin.Hird of 32 Mb+
[11:35] Vincent.Sylvester That's the other idea I had, bad write cache
[11:35] Ubit Umarov: like looking to what the script os doing
[11:35] Ubit Umarov: is..
[11:35] Vincent.Sylvester I did a smart test, I don't know how well it checks for that, didn't do a long test
[11:35] Tom Tomson: Can I get some help setting up and logging into a local grid here?
[11:36] Ubit Umarov: im prety sure a bad disk would showup on a lot more things than just a script
[11:36] Vincent.Sylvester I have seen pigs fly as well I don't trust anything anymore
[11:36] Vincent.Sylvester Well nothing I haven't broken with my own two hands that is
[11:36] Gavin.Hird Hi Tom
[11:36] Tom Tomson: hello Gavin
[11:36] Andrew Hellershanks: Hello, Tom.
[11:36] Gavin.Hird what are the issues you have
[11:37] Ubit Umarov: well this is more about opensim development not use help :(
[11:37] Andrew Hellershanks: Kayaker, the code is the documentation but the comments in it may not always reflect changes that may have been made to the code.
[11:37] Gavin.Hird I am sure we can tackle a few questions
[11:37] Ubit Umarov: and on that.  not mucj last week in new code
[11:38] Tom Tomson: basically: Login failed, region not found. Ubit, I'm new, someone at the welcome center suggested I come here
[11:38] Ubit Umarov: i fact only a commit reverting one i made a few dasy earlier
[11:38] Michael.Christopher what viewer are you using tom?
[11:38] Andrew Hellershanks: Have a seat, Tom.
[11:38] Gavin.Hird operating system isntalled on, what version of opensim?
[11:39] Andrew Hellershanks: Ubit, there was one other small change to the code related to prebuild.
[11:39] Ubit Umarov: i did revert a time penalty i did add to objectMessage and attachments
[11:39] Gavin.Hird SL-16699 Increase default cache size to 4GB
[11:39] Gavin.Hird lol
[11:39] Vincent.Sylvester Yeah ms vs 2022 which apparently still has tons of bugs in it
[11:39] Vincent.Sylvester I couldn't even install it yesterday
[11:39] Tom Tomson: OS is windows 7, opensim
[11:40] Andrew Hellershanks: Tom, what grid are you trying to access? Which viewer are you using?
[11:40] Ubit Umarov: those did have negative impact on some little action games ww have around
[11:40] Gavin.Hird just a standalone Tom?
[11:40] Tom Tomson: I'm using FirestormOS, and trying to access a standalone
[11:40] Ubit Umarov: Ohh that was this week?
[11:40] Ubit Umarov: yes i did add dectection of vs2022 to the runprebuild.bat
[11:40] Gavin.Hird ok, is there a need to open ports in the FW ofr a standalone on Windows?
[11:41] Tom Tomson: What's an FW?
[11:41] Gavin.Hird firewall
[11:41] Andrew Hellershanks: Firewall.
[11:41] Tom Tomson: ok
[11:41] Vincent.Sylvester I just don't see the need to switch right now anyways, buggy mess as usual with vs releases, pain
[11:41] Ubit Umarov: tom
[11:41] Tom Tomson: yes?
[11:41] Andrew Hellershanks: If it is a local standalone the problem is probably more a loopback issue rather than firewall.
[11:41] Ubit Umarov: did you added region name on to region flags also ?
[11:42] Ubit Umarov: like told on release notes?
[11:42] Vincent.Sylvester We'll have to see what comes next and not being able to install it on so many systems is a bit of a problem. Plus the few plugins useful for doing OpenSim debug no idea if they are updated yet either
[11:42] Vincent.Sylvester Not that local nunit tests outside of jenkins ever worked well to begin with
[11:43] Tom Tomson: I added the region name when OpenSim.exe asked me to, I haven't seen the release notes
[11:43] Ubit Umarov: ohh ofc you didn't
[11:43] Ubit Umarov: well go see them :p
[11:43] Andrew Hellershanks: Tom, Is the standalone running on the same computer as the viewer you are using?
[11:43] Vincent.Sylvester I did have some success writing those though in regards to adding compliance testing for both script engines, maybe one day that code will make it in, if Ubit remembers where he stuffed my emails into
[11:43] Gavin.Hird Is there a need to add it elsewhere on a bog standard standalone?
[11:43] Tom Tomson: @Andrew, yes it is
[11:44] Ubit Umarov: yes there is a extra step now
[11:44] Ubit Umarov: on
[11:44] Gavin.Hird ouch
[11:44] Ubit Umarov: and it is on the release note
[11:44] Ubit Umarov: notes
[11:44] Tom Tomson: @Ubit, where can I find the release notes?
[11:44] Gavin.Hird probalby should be on the install instructions too
[11:44] Ubit Umarov:
[11:45] Ubit Umarov: wil have that relaxed on standalones, not robust
[11:45] Andrew Hellershanks: Tom, did you see any red text in the console of the Standalone when it started?
[11:45] Ubit Umarov: "
[11:45] Ubit Umarov: On Standalones make sure you add an entry for your region in section [GridService] of config-include/StandaloneCommon.ini. For example for region "My Region" there should be Region_My_Region = "DefaultRegion, DefaultHGRegion" (i.e. start with Region_ and spaces in the region name replaced by _ )
On Grids make sure you have at least one region with flags DefaultRegion, DefaultHGRegion (not necessarily same regions) on similar entries in section [GridService] of Robust.ini
[11:45] Ubit Umarov: "
[11:45] Ubit Umarov: this
[11:46] Ubit Umarov: there is no red on this
[11:46] Ubit Umarov: jsut region not found
[11:47] Ubit Umarov: what is you region name tom ?
[11:47] Andrew Hellershanks: Ubit, perhaps it should throw red text if it can't find default region information.
[11:47] Gavin.Hird Good news: Nvidia Likely Pulling Out of Arm Acquisition After Failing to Win Regulatory Approval
[11:48] Vincent.Sylvester hehe
[11:48] Vincent.Sylvester Noice
[11:48] Vincent.Sylvester Now we just need Intel to get their collective shit together and release some GPUs and we might get a market once again for consumers and not against them
[11:48] Ubit Umarov: well the ini system, and regions init has several flaws
[11:49] Ubit Umarov: this name thing is one of them
[11:49] Gavin.Hird the GPU market is nuts
[11:50] Andrew Hellershanks: Tom, I would suggest also checking what IP address you are specifying for the standalone. I use the actual IP address assigned to my machine within my local network. I wouldn't use
[11:50] Ubit Umarov: the decision to have nothing to tell  region mode, just change active dlls, ( already loaded ) is another
[11:50] Jamie.Jordan nuts is an understatement
[11:51] Jamie.Jordan i've never seen it like this
[11:51] Ubit Umarov: oh maria did kick hard on the cript currencies
[11:51] Ubit Umarov:
[11:51] Ubit Umarov: on the comments on that
[11:52] Gavin.Hird nft
[11:52] Ubit Umarov: more on why nft are NOT any usefull
[11:52] Gavin.Hird whispers: are not people abandoning nft already
[11:52] Ubit Umarov: yeah
[11:53] Vincent.Sylvester nft and blockchain about as useful as a bullet wound
[11:53] Ubit Umarov: but teh crypto guys spent $$ on gpus
[11:53] Ubit Umarov: need to use them now :p
[11:53] Michael.Christopher i failed to really see the point of nft
[11:53] Gavin.Hird Maria's ability to predict the market has been worse than the UK MetOffice seasonal forecasts
[11:53] Ubit Umarov: :)
[11:54] Vincent.Sylvester That was more fiery than the pepper steak I had ealier
[11:54] Michael.Christopher right?
[11:54] Ubit Umarov: well was about time to see more ppl telling how bad and useless those things are
[11:55] Ubit Umarov: ( got it fixed tom ? )
[11:55] Vincent.Sylvester Crypto market going up and down is so weird given the whole idea was independence so it depending on other markets to define its own value is a conceptual flaw so big it could sink the Titanic twice as fast
[11:56] Gavin.Hird A crypto miner here got a $200000 extra power bill for December, cancellng out all eranings
[11:56] Selby.Evans I wonder if there is an NFT for the Brooklyn Bridge.
[11:56] Andrew Hellershanks: I can't even read that page. Too many tracking things to disable for the site.
[11:56] Ubit Umarov: :)
[11:57] Ubit Umarov: think ppl do not understand that value of a currency at end of day depends on the power of the armed forces of its country
[11:57] Andrew Hellershanks: Gavin, that is one of the big problems with mining for crypto.
[11:57] Vincent.Sylvester crypto and machine learning killing the gpus, at least ML has some uses even if a bunch of people just use it to generate images that look like diarrhea on canvas
[11:57] Gavin.Hird yes
[11:58] Vincent.Sylvester "you can use machine learning for that" - I can also just write good code on my own
[11:58] Gavin.Hird you mean a Hunter Biden piece of art Vincent?
[11:58] Tom Tomson: ok, I've added my region to the StandaloneCommon.ini file
[11:59] Tom Tomson: I also uncommented the line "    MapTileDirectory = "./maptiles""
[12:00] Vincent.Sylvester I tried two of these drawing bots that you can give words to and they create something, it just looks like van Gogh having a stroke and re-acquainting himself with his lunch
[12:00] Ubit Umarov: (then start it and test tom)
[12:00] Vincent.Sylvester Makes me want to yeet myself into the sun lol
[12:01] Michael.Christopher lol
[12:02] Ubit Umarov: but it is not only gpu market that is nuts
[12:02] Ubit Umarov: it is everything..
[12:02] Gavin.Hird just sign it with Hunter Biden and it will sell for $1/2 million
[12:02] Ubit Umarov: well and opensimulator? any more news ?
[12:03] Ubit Umarov: before andrew looks to clock ;)
[12:03] Michael.Christopher nothing else this week?
[12:03] Gavin.Hird not much
[12:03] Michael.Christopher :)) I was thinking the same lol
[12:03] Kayaker Magic: Tom: setting up robust and Opensim from scratch is difficult, you might consider connecting a region to OSGrid first, where the INI files are mostly set up for you.
[12:03] Ubit Umarov: wle i plan to release
[12:03] Ubit Umarov: not much really
[12:04] Motoko.Karu looks like tom left
[12:04] Ubit Umarov: well tom issue will be relaxed in standalones case
[12:05] Ubit Umarov: to set those regions will be needed on robuts, standalones will return again Any region of none found
[12:05] Andrew Hellershanks: Not much in the way of OS news today as we already mentioned. Just a few commits.
[12:05] Ubit Umarov: DIva, RL allowing will release her things for
[12:06] Ubit Umarov: and we will be on
[12:06] Ubit Umarov: ;)
[12:07] Gavin.Hird that is cruel :-))
[12:07] Andrew Hellershanks: If Tom hasn't sorted out his problem I hope he knows to try IRC or the mailing list.
[12:07] Andrew Hellershanks: hehe
[12:07] Gavin.Hird a case of the good old "DOS ain't done till Lotus won't run"?
[12:09] Selby.Evans Must go -- bye all
[12:09] Andrew Hellershanks: We are just past the hour mark now. Any other final comments, questions, or thoughts re: OS for today?
[12:09] Michael.Christopher Looks like i've got to head out, have a great week everyone!
[12:09] Andrew Hellershanks: ok, Selby. tc
[12:09] Ubit Umarov: cya Selby.Evans
[12:09] Andrew Hellershanks: ok, Michael. see you next week.
[12:10] Jamie.Jordan great meeting thanks yall
[12:10] Ubit Umarov: :)
[12:10] Andrew Hellershanks: yw, Jamie.
[12:10] Andrew Hellershanks: Time to wrap it up before we have many more people needing to leave.
[12:11] Andrew Hellershanks: Thank you all for coming. See you again next week.
[12:11] Ubit Umarov: :)
Personal tools
About This Wiki