Chat log from the meeting on 2021-04-27

From OpenSimulator

Jump to: navigation, search
[11:00] Pius Noel: hello everybody
[11:01] Ubit Umarov: hi
[11:02] Chat Logger: Meeting chat logging has been enabled.
[11:02] Chat Logger: You can view the log at:
[11:03] Andrew Hellershanks: Hello, everyone.
[11:03] Andrew Hellershanks: Hello, Selby.
[11:03] Selby.Evans hi everyone
[11:03] Pius Noel: Hi Selby
[11:06] Andrew Hellershanks: hm... there was a commit with the text "missing file" that wasn't actually about adding a file.
[11:07] Ubit Umarov: missing file (change)
[11:08] Pius Noel: vivox no longer provides new free accounts for opensim, are there any alternatives that will work integrated with current viewers, like FS?
[11:10] Andrew Hellershanks: the "(change)" part is missing from the commit message if that was meant to be there.
[11:10] Andrew Hellershanks: Pius, There is FreeSwitch if you can get that to work.
[11:11] Andrew Hellershanks: I have found it to be a pain to get configured and working. That said, I haven't tried using it in quite a long time.
[11:11] Pius Noel: ok, I'll take a deeper look on it.
[11:12] Andrew Hellershanks: There is Mumble but I think that also requires a change to something in the viewer.
[11:13] Pius Noel: yes, as far as I know that's the problem.
[11:13] Pius Noel: there was some discussion in the OSGrid forum.
[11:13] Andrew Hellershanks: The people behind SceneGate are working on a replacement but I have no idea how far along they are in the development of a replacement for Vivox.
[11:13] Selby.Evans SceneGate is working on using a mumble derivative, but that is not finished.
[11:14] Andrew Hellershanks: Right. I thought they still had a ways to go on it.
[11:14] Pius Noel: yes, I already looked at SceneGaate
[11:14] Selby.Evans Lisa Laxton would know.
[11:15] Ubit Umarov: ( i did tons of commits where missing file also justs means a change on a file..  duhhh :p )
[11:15] Ubit Umarov: ( and for me that is a missing file copy also :) )
[11:17] Andrew Hellershanks: Ubit, I know the only way to really know what you changed is to look at the contents of the commit. The log entries are mostly useless in that regard.
[11:17] Andrew Hellershanks: but not always.
[11:18] Ubit Umarov: well that was part of last week changes
[11:19] Ubit Umarov: where i did add just some ossl helper funtions
[11:19] Andrew Hellershanks: Yes, I see them. BTW, firts?
[11:20] Andrew Hellershanks: Is that a typo of first?
[11:20] Ubit Umarov:
[11:21] Ubit Umarov: also fixed mantis 8250, making Yengine tick script events per second stat
[11:21] Andrew Hellershanks: ok. More a typo in the log entry.
[11:21] Ubit Umarov: on that, removed script line per second
[11:21] Andrew Hellershanks: I saw the reference to changes in the stats.
[11:21] Ubit Umarov: no longer present on viewers.. and obsolete
[11:22] Andrew Hellershanks: It used to be available via the WebStats.
[11:22] Ubit Umarov: and in fact our code didn't do line per second, was more lsl ll functions calls per seconds
[11:23] Andrew Hellershanks: Yes, that's right. It did provide at least some rough metric of script activity.
[11:23] Ubit Umarov: i know ppl that do scripts on a single line
[11:23] Andrew Hellershanks: I hope those are short scripts.
[11:23] Ubit Umarov: well events per second is a metric that was not working
[11:24] Ubit Umarov: i mena on Yengine.. it was on X
[11:24] Andrew Hellershanks: Does the change in stats affect both XEngine and YEngine?
[11:24] Ubit Umarov: lps, both ofc
[11:24] Andrew Hellershanks nods
[11:26] Andrew Hellershanks: I never paid a lot of attention to the LPS number unless their was lag in a region.
[11:27] Pius Noel: me too
[11:28] Ubit Umarov: script time, is not working on yengine either
[11:28] Andrew Hellershanks: We are fewer in number today than usual. Does anyone have any questions/comments for today or will we wind up wrapping up early?
[11:28] Ubit Umarov: have it showing things
[11:28] Ubit Umarov: but on Y it does show a lot more time
[11:29] Ubit Umarov: not totally clear why
[11:29] Pius Noel: not me, I have to leave early anyway.
[11:29] Andrew Hellershanks: Script runtime numbers in YEngine are very different from XEngine so I couldn't tell if they were right or wrong.
[11:29] Ubit Umarov: one dif to X has to do with operating system interrupts, taking cpu out of a there
[11:29] Ubit Umarov: of a thread
[11:30] Ubit Umarov: on X, the way it works, there is less change of that happening during script event execution
[11:30] Ubit Umarov: on Y it happens a lot more..  that means at least moew 20ms added to script time
[11:30] Ubit Umarov: that is not exactly cpu time on it
[11:31] Andrew Hellershanks: X didn't properly track all the execution time of scripts.
[11:31] Ubit Umarov: but no what ot actually see it
[11:31] Ubit Umarov: in fact it is jsut impossible with normal code
[11:32] Ubit Umarov: by definition of multitask, the code does not see when the os takes cpu from it
[11:32] Ubit Umarov: and the clocks just keep running..
[11:32] Ubit Umarov: well i have a script that on X shows 0,1 ms exec time consistently and on Y shows 13ms
[11:32] Ubit Umarov: grrr
[11:33] Andrew Hellershanks: Yes, I've seen the runtime numbers in Y being an order of magnitude (or so) greater than in X.
[11:33] Ubit Umarov: thats the reason why i didn't commit the code to let Y also tick the script time
[11:34] Ubit Umarov: but this also impacts scripts top etc
[11:35] Andrew Hellershanks nods
[11:37] Andrew Hellershanks: I was trying to find the bubble gum scripts I have to compare some numbers between X and Y. I have the original and modified version of the script. The original had high runtime. The modified one was much better with runtime less than 1 under X.
[11:38] Andrew Hellershanks: Once I can track down those scripts I can do some preliminary comparison between the engines re: good vs bad runtime numbers.
[11:39] Ubit Umarov: well those numbers where never good
[11:39] Ubit Umarov: just 0.1 to 13 is a bit strange
[11:39] Ubit Umarov: when Y is doing it faster
[11:39] Ubit Umarov: wlel not that script.. it is a long event
[11:40] Andrew Hellershanks: In the past X attempted to track time spent executing functions but didn't catch all the time required.
[11:40] Ubit Umarov: it may just be the same thing..  the cpu time slice just ends always in middle of it
[11:41] Ubit Umarov: no the time the thread is waiting for a time slice is counted
[11:41] Ubit Umarov: so..
[11:42] Pius Noel: I've seen some situations where scripts stopped running after sim border crossing on Y.
[11:42] Pius Noel: Y on both sides.
[11:42] Ubit Umarov: no idea
[11:43] Ubit Umarov: can't debug with just statements like that ;)
[11:43] Pius Noel: don't worry, I'll do more testing on that and report on mantis if I know more
[11:43] Ubit Umarov: ty
[11:44] Ubit Umarov: can be a fun timming related thing..  uff
[11:44] Pius Noel nods
[11:44] Andrew Hellershanks nods
[11:45] Pius Noel: I've been looking into the scripts today, but cant see anything wrong.
[11:45] Ubit Umarov: ahh
[11:45] Ubit Umarov: also made the timers and other events sampling time to by configurable
[11:45] Andrew Hellershanks: Pius, where they from SL?  ;)
[11:45] Ubit Umarov: ini setting AsyncLLCommandLoopms ( default 100, 50 a more reasonable value that just eats a bit more cpu looping around
[11:46] Pius Noel: no, I don't think so. the scripts are not written by myself.
[11:46] Ubit Umarov: also mantis 8881: change llHTTPRequest allowed headers control
[11:46] Andrew Hellershanks: Pius, np. I was mostly joking about that.
[11:46] Pius Noel: ;)
[11:47] Ubit Umarov: oh did a bit more coding this last week :)
[11:47] Pius Noel: well I've seen such scripts and actually last week someone asked why a script no longer works on Y and ubODE.
[11:47] Andrew Hellershanks: Ubit, yes you did.
[11:47] Ubit Umarov: well and wasted time to find the 13ms to 0.1ms thing
[11:48] Pius Noel: I know now the reason, it containes mono code.
[11:48] Ubit Umarov: Y only does LSL and its extensions..
[11:48] Pius Noel: yes, but I first had to see the code :)
[11:49] Andrew Hellershanks: Ubit, How does a grid/region owner change the allowed headers for llHTTPRequest?
[11:49] Ubit Umarov: :)
[11:49] Ubit Umarov: does not.. i did :p
[11:49] Ubit Umarov: hardcoded
[11:49] Andrew Hellershanks: Ubit, ok. Possibly changeable via an ini later?
[11:49] Ubit Umarov: nahh
[11:50] Andrew Hellershanks: ok
[11:50] Ubit Umarov: some are just by standard
[11:50] Ubit Umarov: others a bit more questionable
[11:50] Ubit Umarov: i also added the option to stop the script on one or just ignore
[11:51] Ubit Umarov: older code was just ignoring i think
[11:51] Andrew Hellershanks: Pius, It is the compiled code that may or may not have mono. It wouldn't explain why a script stops working on Y that worked in X.
[11:51] Ubit Umarov: when he said mono i did read c# andrew
[11:52] Pius Noel: right... sorry I was not clear enough
[11:52] Andrew Hellershanks: If that is the case that would be a problem. Still basically the same thing I'm getting at is that there is something in the original source that is the problem.
[11:52] Ubit Umarov: ( bc "mono on a script" makes no sense ;) )
[11:52] Andrew Hellershanks: mono vs non-mono doesn't make sense in OS. That is an SL thing.
[11:53] Andrew Hellershanks: Well, for a meeting that I thought was going to be short this week we have managed to fill the time. Almost the top of the hour. Any other topics for today?
[11:54] Pius Noel: no, thank you :)
[11:54] Andrew Hellershanks: Pius, ok.
[11:54] Pius Noel: Hi kayake
[11:55] Pius Noel: +r
[11:55] Andrew Hellershanks: Hey, Kayaker. We are just about to wrap up for today. Did you have anything you wanted to ask before we end todays meeting?
[11:55] Kayaker Magic: Sorry I'm late, up visiting GF and off my schedule.
[11:55] Andrew Hellershanks: np
[11:55] Kayaker Magic: No questions, but I have an ataboy:
[11:55] Andrew Hellershanks: go ahead
[11:55] Ubit Umarov: a what»
[11:55] Ubit Umarov: ataboy?
[11:55] Andrew Hellershanks: :)
[11:56] Andrew Hellershanks: He wants to recognize/thank someone for something they did
[11:56] Kayaker Magic whispers: I converted a bunch of regions from XEngine to Yengine and expected to have some script fixes to do. This was made easier by very good error messages in the log when YEngine found syntax errors!
[11:57] Kayaker Magic: So ataboy Ubit for doing great error messages there!
[11:57] Andrew Hellershanks: :)
[11:57] Ubit Umarov: well some are still.. a mistery :)
[11:57] Selby.Evans :)
[11:57] Kayaker Magic: (You know the exchange rate on ataboys? One ohshit cancels 10 ataboys)
[11:58] Andrew Hellershanks: hehe
[11:58] Ubit Umarov: good i only do oops not ohshit
[11:58] Andrew Hellershanks: What's the exchange rate on them?
[11:58] Kayaker Magic: We have to make up the exhange rate on that....
[11:59] Ubit Umarov: :)
[11:59] Andrew Hellershanks: Hello, Gavin.
[11:59] Gavin.Hird Evening
[11:59] Kayaker Magic: ?You just got here also Gavin?
[11:59] Andrew Hellershanks: Just about to wrap up todays meeting. Anything you want to say or ask?
[11:59] Pius Noel: hi Gavin
[11:59] Gavin.Hird you saw it happening :-)
[11:59] Ubit Umarov: early i talked about letting script time stat also move onY
[12:00] Ubit Umarov: but at least one script shows 13ms on Y and 0.1 on X :p
[12:00] Ubit Umarov: and that is far from real cpu
[12:00] Ubit Umarov: think is that we jsut can't read real cpu on code execution
[12:00] Ubit Umarov: thing..
[12:01] Ubit Umarov: that also impacts scripts top etc
[12:04] Ubit Umarov: yeack
[12:05] Ubit Umarov: region did crash?
[12:05] Pius Noel: wb Kayaker
[12:05] Andrew Hellershanks: Kayaker and Ubit both disappeared for a moment. Um... what happened to you two?
[12:05] Kayaker Magic: Not just me? I got thrown out.
[12:05] Ubit Umarov: oh i also lost comms
[12:05] Andrew Hellershanks: Gavin as well. We didn't crash.
[12:05] Gavin.Hird I crashed
[12:05] Ubit Umarov: guess network things
[12:05] Andrew Hellershanks: Selby, Jagga, and myself stayed put.
[12:06] Gavin.Hird or rather lost connection
[12:06] Pius Noel: me too
[12:06] Gavin.Hird so everyone in Europe got disconnected?
[12:06] Pius Noel: no, not me
[12:06] Ubit Umarov: kay is eu ?
[12:06] Gavin.Hird oh, ok - just a question
[12:07] Pius Noel: I'm from Switzerland... not a member of EU *lol*
[12:07] Ubit Umarov: well i had type something here.. now forgot
[12:07] Ubit Umarov: lol
[12:07] Andrew Hellershanks grins at Pius
[12:07] Kayaker Magic: California
[12:07] Andrew Hellershanks: Go ahead, Gavin.
[12:07] Ubit Umarov: well was strange
[12:07] Ubit Umarov: but at least not a region crash :)
[12:08] Gavin.Hird sorry, the question was if everyone disconnected was in Europe
[12:08] Kayaker Magic: Ubig: "IT'S NOT MY FAULT!"
[12:08] Ubit Umarov: Ubig :)
[12:08] Gavin.Hird did Ubit do something?
[12:08] Ubit Umarov: AHh i remember
[12:09] Gavin.Hird as his region CPU load spiked while I got disconnected
[12:09] Ubit Umarov: gimisa did report me that lm descriptions no longer have region information
[12:09] Ubit Umarov: i do see that with FS and dayturn now
[12:09] Ubit Umarov: guess some LL broken code
[12:10] Gavin.Hird there was a change for that
[12:10] Ubit Umarov: but the comments seem to mean fixing..
[12:10] Ubit Umarov: and just vanished :)
[12:10] Ubit Umarov: i just checked a older fs version and it sets all
[12:11] Ubit Umarov: like  Your Parceldt, ubittestDt (386, 492, 22) - General @
[12:11] Gavin.Hird but I mean what exactly is it that you don't see?
[12:11] Ubit Umarov: now its only  @
[12:11] Gavin.Hird [12:10] Ubit.Umarov and just vanished :)
[12:11] Ubit Umarov: bc the @.. is added by region
[12:12] Gavin.Hird nothing in IM; but in in-world chat
[12:12] Ubit Umarov: im talking about the new lms description :)
[12:12] Gavin.Hird do you need the the region info in IM?
[12:12] Gavin.Hird does it even matter?
[12:13] Ubit Umarov: some ppl have scripts looking to that
[12:13] Ubit Umarov: and was there
[12:13] Gavin.Hird probalby why they took it away
[12:13] Gavin.Hird another stalking vector
[12:13] Ubit Umarov: all code is there
[12:13] Ubit Umarov: and their comments on commits where abotu fixing it
[12:13] Ubit Umarov: guess a bad turn ??
[12:13] Ubit Umarov: no idea
[12:14] Gavin.Hird who's comment - LL?
[12:14] Ubit Umarov: yeah
[12:14] Ubit Umarov: ( but i didn't look to all commits.. )
[12:14] Gavin.Hird what does a typical script use the region for?
[12:14] Ubit Umarov: getting the region name is otherwise very expensive
[12:15] Ubit Umarov: bc we only have the id
[12:15] Ubit Umarov: and need ask grid ( or HG ) about it
[12:15] Gavin.Hird yes
[12:15] Ubit Umarov: and it was all there
[12:15] Gavin.Hird which is probalby why the removed it now they are on AWS and pay per second
[12:15] Ubit Umarov: parcel, region, location, even region age access
[12:16] Ubit Umarov: as i said, it costs more without it
[12:16] Ubit Umarov: bc all that is known at lm creation
[12:16] Gavin.Hird ok
[12:17] Ubit Umarov: as i said, without it then you need ask grid using the region id on the lm data
[12:17] Gavin.Hird but what does a script typically use the region info displayed in the IM message for?
[12:17] Kayaker Magic: Hey! If you look out the window to my right, you will see my first completely sucessful animesh build+running animation: The swimming whale!
[12:18] Ubit Umarov: well gimisa foun that bc he has scripts that where using that
[12:18] Gavin.Hird sure, but what did the script do?
[12:18] Gavin.Hird congrats Kayaker
[12:18] Ubit Umarov: that fish?
[12:18] Andrew Hellershanks: Well done, Kayaker.
[12:18] Pius Noel: nice, Kayaker :)
[12:18] Ubit Umarov: i didn't ask :p
[12:19] Ubit Umarov: well i also asked beq to look at fs side
[12:19] Gavin.Hird ok
[12:19] Ubit Umarov: guess went to her /dev/null
[12:19] Ubit Umarov: hihihi
[12:19] Ubit Umarov: or then she is just afk, as always
[12:20] Ubit Umarov: well not in the mood of adding that info region side
[12:20] Ubit Umarov: but if needed.. well.. lets see
[12:20] Kayaker Magic: Not a fish! A mammal!
[12:21] Ubit Umarov: think i can add things in the data without blowing up viewers
[12:21] Andrew Hellershanks: :)
[12:21] Ubit Umarov: ann yes a fish with its
[12:21] Gavin.Hird I don't have any need to satisfy SL requirements
[12:21] Ubit Umarov: tits
[12:21] Gavin.Hird so if it worked before it can be reversed
[12:21] Kayaker Magic: Also tail flipping up and down, not sideways.
[12:21] Gavin.Hird do you have the specific patch ID?
[12:22] Gavin.Hird like SL-xxxxx
[12:22] Gavin.Hird or something
[12:22] Ubit Umarov: ohh its whale :)
[12:23] Ubit Umarov: looked as a eel when i looked :)
[12:23] Ubit Umarov: gmm patch? no
[12:23] Gavin.Hird since you found the change..
[12:24] Ubit Umarov: well found changes
[12:24] Ubit Umarov: not the one
[12:24] Gavin.Hird OK :-)
[12:24] Ubit Umarov: :)
[12:25] Ubit Umarov: well using desc is always bad
[12:25] Ubit Umarov: that is user editable..
[12:25] Ubit Umarov: and limited in size.. on many cases things just don't fit
[12:26] Ubit Umarov: so i may need to add the info to the asset data
[12:26] Ubit Umarov: and add a ossl to get it
[12:26] Ubit Umarov: grrr
[12:27] Gavin.Hird the question is, is it really needed in person to person IM?
[12:27] Ubit Umarov: its LM not IM
[12:27] Gavin.Hird LM
[12:27] Gavin.Hird sorry
[12:27] Gavin.Hird yeah
[12:28] Gavin.Hird LM is a moving target at the moment
[12:28] Gavin.Hird so
[12:28] Ubit Umarov: yeack
[12:28] Gavin.Hird there are changes, and a slew I have not taken yet as LL keeps chaning their target almost daily
[12:28] Ubit Umarov: so the ll viewer cache went back to drawing board and not rc again ?
[12:29] Gavin.Hird so it seems
[12:29] Ubit Umarov: and now..
[12:29] Ubit Umarov: bahh
[12:29] Gavin.Hird at least they have added support for a bunch of new cursors for MOP
[12:29] Gavin.Hird not sure why
[12:29] Ubit Umarov: sure they do know?
[12:29] Ubit Umarov: ;)
[12:29] Gavin.Hird :-)
[12:30] Gavin.Hird and they fixed a group chat issue for when you are in god mode
[12:30] Ubit Umarov: hm and they increased fees again?
[12:31] Gavin.Hird they increased the fees because it came as a bomb to them that they had to pay per cpu second, and they use more cpu than expected
[12:31] Ubit Umarov: well the move to cloud was going to reduce the costs
[12:31] Ubit Umarov: just they forgot they did own the old machines.. so no costs there
[12:31] Ubit Umarov: and now need to pay
[12:31] Ubit Umarov: so for them is actually more expensive now
[12:32] Ubit Umarov: old machines where more than payed
[12:32] Gavin.Hird It was like on the good old Univac - we coiuld use a whole CPU minute per semester for free!
[12:32] Ubit Umarov: of course they would need to get new machines..
[12:33] Gavin.Hird I think the thing is when you spread your processing all over, there is going to be a lot more expensive cross chatter
[12:33] Ubit Umarov: but was fun.. they did told  that fees would go down bc the move to cloud
[12:33] Ubit Umarov: well and cloud, when you use a cpu, it does cost a lot more than if you do own it :)
[12:33] Ubit Umarov: of rent it all the time
[12:33] Gavin.Hird yes
[12:33] Ubit Umarov: or..
[12:34] Ubit Umarov: but does mk sense..  they sould need a lot less cpus on cloud
[12:35] Ubit Umarov: but wel thats sl  :)
[12:35] Ubit Umarov: their thing
[12:35] Ubit Umarov: so what news do you have abotu opensim?
[12:35] Gavin.Hird Apple came with another Xcode update yesterday and the only machine I can run it on right now is the Apple Silicone one
[12:36] Gavin.Hird but on that the viewer project will not configure
[12:36] Gavin.Hird so...
[12:36] Ubit Umarov: fun fun
[12:37] Gavin.Hird the rate of code change in the viewer does not come close to catching up with Apple
[12:38] Andrew Hellershanks: It is now past the hour and a half mark. Good thing we didn't have much to talk about today. ;)   Any last minute items. I need to get going soon.
[12:38] Gavin.Hird and the latest expectation is you can run Mac apps on the latest iPad later this year
[12:38] Gavin.Hird so that will be even more fun
[12:38] Ubit Umarov: :)
[12:39] Gavin.Hird since they plunged their fastest chip into it and gave it 16 GM or working memory
[12:40] Gavin.Hird and thunderblot
[12:40] Gavin.Hird with support for a 6k display
[12:40] Gavin.Hird 16 GB even
[12:41] Andrew Hellershanks: That seems like a good place to leave things for this week. Thank you all for coming. See you again next week.
[12:41] Andrew Hellershanks: You did well, Pius. You thought you would have to leave early but you are still here.
Personal tools
About This Wiki