Chat log from the meeting on 2017-10-31

From OpenSimulator

Jump to: navigation, search

[11:03] Christopher.Colosi What's on the docket today?
[11:03] Andrew Hellershanks: Christopher: Any OpenSim topic that people care to discuss.
[11:04] Bill.Blight Opensim just nees more cowbell ... ;)
[11:04] Andrew Hellershanks: I do have one announcement which i will make a bit later.
[11:04] Andrew Hellershanks: Bill: Cowbell? :)
[11:04] Kayaker Magic: I saw a note that said MySQL on ubuntu defaults to a safe mode that commits on every query and that this slows down rezzes of objects.
[11:05] Kayaker Magic: Anybody know anything about this?
[11:06] Bill.Blight LOL, it is from a Saturday Night Live sketch .with christopher walken , I guess not as pop culture as I thought .. LOL
[11:06] Andrew Hellershanks: I never watched SNL
[11:06] Christopher.Colosi Oh, it is.
[11:07] Kayaker Magic: I got the cowbell reference, but it no longer makes me laugh....
[11:07] Andrew Hellershanks: Kayaker, I haven't heard or read that before. Where did you see that note?
[11:07] Bill.Blight I use MariaDB , converted everything to the Aria storage engine, seeing a great increase in performance ..
[11:07] Christopher.Colosi search cowbell on google and this is the very first result:
[11:08] Jak Daniels: in MySQL 5.7 manual it says this:
[11:08] Jak Daniels: By default, MySQL starts the session for each new connection with autocommit enabled, so MySQL does a commit after each SQL statement if that statement did not return an error
[11:08] Jak Daniels: ubuntu moved to 5.7
[11:08] Jak Daniels: on 16.04lts
[11:08] Andrew Hellershanks: I have a couple of newer machines using MariaDB. The main machines were I'm using OS are still using MySQL.
[11:09] Kayaker Magic: I heard someone talking about MySQL and OpenSim on a private chat... I'm trying to figure out how to turn that commit on all query mode off on my server.
[11:10] Bill.Blight SET AUTOCOMMIT=0;
[11:10] Bill.Blight to turn it off
[11:10] Andrew Hellershanks: Kayaker, you can see the setting in SHOW VARIABLES.
[11:10] Bill.Blight SELECT @@autocommit; to see it's flag
[11:10] Jak Daniels: [mysqld]
init_connect='SET autocommit=0'
[11:11] Jak Daniels: in your my.cnf file
[11:11] Jak Daniels: or autocommit=0
[11:11] Jak Daniels: in same file depending on version
[11:11] Andrew Hellershanks: I don't know why that would affect rezzing since that would be fetching data and not saving any thing.
[11:12] Bill.Blight it has to update the db for object postion and rotation
[11:12] Bill.Blight also may be a new asset if it is a copy
[11:12] Andrew Hellershanks: yea, I just realized it needs to update the table that keeps track of items in the land.
[11:14] Kayaker Magic: If I recall correctly, autocommit limits the number of queries per second to around 15, then rezzing large complicated objects takes seconds to fetch.
[11:14] Bill.Blight well it is easy to test as you can set that flag on the fly
[11:15] Kayaker Magic: If I log into mysql as root and set that flag to 0, does that effect the instances already running?
[11:15] Jak Daniels: no
[11:15] Bill.Blight SET AUTOCOMMIT=0;
[11:15] Bill.Blight yes it affects new connections, but not existing ones until the close and re-open
[11:15] Jak Daniels: only the current session
[11:16] Kayaker Magic: OK, so restart the instance after setting autocommit=0; I'll try that.
[11:16] Jak Daniels: must set it globally in the config file I would think..
[11:16] Andrew Hellershanks: Would it not immediatley affect any OS instances using the database?
[11:16] Bill.Blight so changing it and restarting the region should make it change
[11:16] Bill.Blight no it would not affect open connections, but when they reconnect it should get the new flag
[11:17] Kayaker Magic: I'll try to measure something rezzing before and after autocommit=0 and report back next week.
[11:19] Andrew Hellershanks: If you turn off autocommit I would think you need to do something when dealing with the database to force the data to be flushed to disk or else you risk losing data if you had an unexpected DB or system crash.
[11:19] Jak Daniels: does opensim wrap a whole load of queries into a transaction anyway for stuff like rezzing?
[11:20] Bill.Blight Yeah that is a risk, like driving your Lamborghini at 200 mile an hour into a wall, fun while it lasted, and was performing well up until the sudden stop ..
[11:20] Andrew Hellershanks: I see flush and flush_time variables in MySQL that may be helpful in dealing with that situation.
[11:20] Andrew Hellershanks: Jak, Good question. I haven't looked at that level of the database interactions of OS.
[11:21] Jak Daniels: I'm guessing it doesn't....
[11:21] OtakuMegane Desu: 15 queries per second seems an odd limit. I know I can shove hundreds of queries per second in with PHP and autocommit is on in a defauly MySQL setup.
[11:21] Jak Daniels: yep, I've never seen that limit before
[11:22] OtakuMegane Desu: Sounds like an artificial limitation or performance issue from something else.
[11:23] Jak Daniels: mysql can do query quotas
[11:23] Andrew Hellershanks: 15 queries per second sounds like a very low value to have on a database that would see a lot of activity.
[11:23] Christopher.Colosi Can you paste in your connection string (sans your password)?
[11:23] Jak Daniels: is it a hosted MySQL server?
[11:24] Bill.Blight appears the default flush_time is 1800 seconds
[11:24] Kayaker Magic: I just tried setting autocommit=0 in mysql, exiting and logging back into mysql as root, autocommit was back to 1 again.
[11:24] Kayaker Magic: I guess I have to set the config and restart mysql to make it stick.
[11:24] Jak Daniels: I think you need to set it globally in my.cnf
[11:24] Bill.Blight yeah it will not stick
[11:25] Bill.Blight it is supposed to , but was just reading that it does not
[11:27] Bill.Blight I actually already had it set in my config, wonder what all nighter I did that on, and forgot about it ... LOL
[11:27] Andrew Hellershanks: :)
[11:28] Kayaker Magic: You have it set to 0 Bill?
[11:28] Bill.Blight yes
[11:29] Bill.Blight I'm sure I just did it recently , I moved my assets a couple of weeks ago to a new server, willing to bet I changed it when I was converting over to the Aria engine
[11:29] Christopher.Colosi gotta run. Ciao all
[11:29] Bill.Blight but as that was a 4 day marathon of no sleep and moving assets, it is kind of a blur
[11:32] Andrew Hellershanks: I've known times like that. Lots of stuff being done in a short time frame and it all blurs together.
[11:32] Andrew Hellershanks: I'll make my annoucement now
[11:33] Andrew Hellershanks: On the subject of some intensive bout of activity I had a session like that this past week.
[11:34] Andrew Hellershanks: I've gone through all the OSSL function pages in the wiki and updated the permissions (where they were wrong, and added in the details for the default permissions and script delay values.
[11:34] Andrew Hellershanks: With 169 functions that took me about a full day to do.
[11:34] Selby.Evans Thanks - -Valuable work
[11:34] Arielle Popstar: that all? :)
[11:34] OtakuMegane Desu: Yay documentation update!
[11:35] Arielle Popstar: but oh so greatfull you did
[11:35] Andrew Hellershanks: I found some discrepancies between the OS code and the osslEnable.ini file. Those still need to be dealt with and the appropriate changes made. I started discussing my findings with the other core developers but haven't finished those discussions.
[11:36] Andrew Hellershanks: Otaku:  :)
[11:37] Arielle Popstar: link?
[11:38] Andrew Hellershanks: I'm thinking I might have missed a script delay value here or there. I am thinking of checking the code for calls to the delay routine to make sure I didn't miss documenting any delays imposed by the OSSL functions.
[11:39] Andrew Hellershanks: Arielle:
[11:39] OtakuMegane Desu: That's probably not a bad idea. I've found there tends to be settings and such buried in the code that never get documented.
[11:40] Andrew Hellershanks: Otaku, some of the OSSL routines call other routines to do the work and the delays could be in the called function. I was checking but I could have missed on.
[11:40] Andrew Hellershanks: Um... talk amongst yourself for a few minutes. I'm being called away to do something.
[11:40] Andrew Hellershanks: bbiab.
[11:42] OtakuMegane Desu: If you have the chance it doesn't hurt to dig into the code more. But it sounds like you've been pretty thorough already so I wouldn't lose any sleep over it. :)
[11:43] Arielle Popstar: anyone see differences between opensim.exe and opensim32.exe on a 64 bit machine?
[11:46] Arielle Popstar: andrew was telling us he did some work on the ossl wiki
[11:46] Bill.Blight opensim32 on windows should make it run in 32 bit mode
[11:46] Bill.Blight on linux does not matter
[11:46] Bill.Blight will always run in 64bit if it is a 64 bit system
[11:46] Arielle Popstar: actually it was meant for the 64 bit till they supposedly made it autodetect on opensim.exe
[11:47] Bill.Blight 32 bit mode according to ubit runs faster for smaller regions, but falls on its face for large ones
[11:47] Bill.Blight Ubit updated the opensim32 not long ago to renable it
[11:47] Sheera.Khan I can imagine since memory is limited then
[11:47] OtakuMegane Desu: I'm guessing unless you're doing big stuff or pushing limits it's probably not much of a noticeable difference.
[11:48] Bill.Blight not really , but on windows it does keep a runaway region from eating all your ram
[11:48] Andrew Hellershanks: Sheera, I was about to mention that we are getting in to that time of year where the clocks start changing in many parts of the world. People need to pay attention to that in determining when to show up for the meetings over the next few weeks.
[11:50] Andrew Hellershanks: i haven't used the OpenSim32 executable since the OpenSim.exe file could be used on either 32 bit or 64 bit.
[11:51] Arielle Popstar: LaNani plays with the code too and has been working on making mesh objects rezz faster
[11:51] LaNani Sundara: i decided to try the 32 bits version today, since Ubit kept hinting that it runs faster
[11:51] Bill.Blight right but opensimi32 forces it to run in 32bit mode, on windows, does nothing on linux
[11:51] Andrew Hellershanks: LaNani, did he say how much faster?
[11:52] LaNani Sundara: i use Win server 2012 R2
[11:52] LaNani Sundara: no not that i recall
[11:52] Bill.Blight this is Ubit, have you ever got a straight answer out of him? ... LOL , me either
[11:52] Sheera.Khan I suppose it depends on what happens on the simulator too
[11:52] Andrew Hellershanks: Bill:  :)
[11:53] Sheera.Khan not every function will perform better by the same amount
[11:53] LaNani Sundara: making things 64 bits does not mean they run faster
[11:53] Bill.Blight yes I think that is what he means by a big sim, not really the size but the process, scripts, yada yada yada
[11:54] Arielle Popstar: does mono come into play on that too
[11:54] Andrew Hellershanks: LaNani, right.
[11:54] Bill.Blight no X64 does not mean faster, but generally that happens because of the expanded memory space and added instruction sets, but not always
[11:54] Sheera.Khan 32 Bit processes are limited at 2 GB of RAM... that's quite a handfull for a sim
[11:54] Bill.Blight actually 3 gig
[11:55] Sheera.Khan oh yes, we are on Win7+ ;-)
[11:55] Arielle Popstar: wasnt there some change to make it good up to 2 terabyte?
[11:55] Bill.Blight as I said mono does not obey the forced 32bit
[11:55] Bill.Blight if it is a x64 system with x64 mono they always run x64
[11:55] Andrew Hellershanks: Ubit just said that is just a bit more. That also only works on Windows. For Linux you would need the 32-bit mono.
[11:56] Andrew Hellershanks: correctcion. On normal linux opensim32 will run in 64b also
[11:56] LaNani Sundara: ask him he just logged in
[11:56] LaNani Sundara: :)
[11:56] Arielle Popstar: he is coming to straighten out the answer....
[11:56] Andrew Hellershanks: I already asked him via IRC
[11:57] Ubit Umarov: today i have a excuse to be late... time changed in EU :p
[11:57] Bill.Blight see we said his name too many times, he showed up
[11:57] LaNani Sundara: yes :) and silly Americans wait a week to catch up :)
[11:57] Jak Daniels: meetings are supposed to be at 7pm UTC..... today's started at 6pm UTC
[11:57] Ubit Umarov: oh you look prettier today Bill.Blight
[11:57] Andrew Hellershanks: I was just mentioning about the time changes are about to happen. I didn't know they have started in some parts of the world.
[11:57] Arielle Popstar: did the computer not know it changed?
[11:58] Jak Daniels: but meeting time is always wrong until the U.S. puts it's clocks back ;)
[11:58] Ubit Umarov: yeap US ppl don't understand UTC :p
[11:58] Bill.Blight I wish we did
[11:58] Bill.Blight I mean I do
[11:58] Bill.Blight but ...
[11:58] Jak Daniels: It's a constant... it never changes
[11:59] Ubit Umarov: actually no.. it does change
[11:59] Ubit Umarov: bc it has a component to match sun positions
[11:59] Bill.Blight see , it screwed up Leighton too
[11:59] Ubit Umarov: ( so leap seconds etc )
[11:59] LaNani Sundara: Ubit i recall you suggesting .sh scripts on Linux to start opensim and they pointed to opensim32
[11:59] Andrew Hellershanks: UTC does have a leap second every so many years.
[11:59] Bill.Blight someone else who forgot about the time change
[11:59] Jak Daniels: technicalities ubit :)
[11:59] Ubit Umarov: yeap
[12:00] Bill.Blight Better late than never Ubit, Leighton
[12:00] Ubit Umarov: just does not have summer time, and it is the same all around
[12:00] Jak Daniels: well this time of year we get to have two meetings lol
[12:00] Leighton.Marjoram am I late clocks changed in the UK over the weekend
[12:00] Bill.Blight yes you are late
[12:00] George Equus: Hi all. One hour late I guess...
[12:00] Bill.Blight LOL
[12:00] Andrew Hellershanks: C&P the meeting log back for the late comers. :)
[12:00] George Equus: sigh
[12:00] Andrew Hellershanks welcomes the new arrivals.
[12:00] Ubit Umarov: seems there was a question about opensim32.exe
[12:00] Leighton.Marjoram darn
[12:00] Jak Daniels: the meeting should be a 7pm UTC. Then people just need to remember their own local time offset
[12:00] Arielle Popstar: next year we do the time change on Euro time
[12:01] Ubit Umarov: yes it will run a bit faster and use less mem on Win...
[12:01] Jak Daniels: depending on time of year
[12:01] Sheera.Khan Hi Chen
[12:01] Andrew Hellershanks: I often wonder if the time of the server(s) used to run OS grid changes its clock.
[12:01] Arielle Popstar: then US and canada will be early?
[12:01] LaNani Sundara: yes, once in a while when we talk you say that it does that... run faster
[12:01] Ubit Umarov: to have the same on linux, you need to have mono setup to 32b also
[12:01] Ubit Umarov: otherwise opensim32 will run in 64b also
[12:01] Andrew Hellershanks: Arielle, probably true. Canada doesn't change its clocks until this coming weekend.
[12:02] Ubit Umarov: ( there is not bc of that )
[12:02] LaNani Sundara: does it run faster in linux on 32 bit mono?
[12:02] Ubit Umarov: yes
[12:02] Ubit Umarov: less memory use also means faster
[12:02] Arielle Popstar: weekend after isnt it Andrew? 2 weeks after euro change usually
[12:03] Ubit Umarov: not that dramatic.. but noticiable on time stats
[12:03] Andrew Hellershanks: Arielle, First Sunday in November.
[12:03] LaNani Sundara: would you run into issues on big region when using opensim32?
[12:03] Arielle Popstar: ok Andrew
[12:03] Ubit Umarov: Release versus Debug also seems to have a bit more impact that it used 2
[12:03] Andrew Hellershanks: LaNani, you would if they use too much memory.
[12:04] LaNani Sundara: but would you really? i mean, how big are we talking about
[12:04] OtakuMegane Desu: You'd have to be pushing 3GB+ memory usage on the instance though.
[12:04] Andrew Hellershanks: I have had regions use more than 2GB of memory
[12:04] Ubit Umarov: sure keep in mind that 32b is limited in Virtual memory size
[12:05] Ubit Umarov: to about 1.9 GB ???
[12:05] LaNani Sundara: opensim is not keeping the whole thing in memory anyway?
[12:05] LaNani Sundara: it loads assets when needed
[12:05] Ubit Umarov: and virtual memory limit is a HARD limit
[12:05] OtakuMegane Desu: It's certainly possible to use large amounts like that but I doubt it's a common case.
[12:05] Ubit Umarov: ( but on a 64b OS that is per instance, not global )
[12:05] LaNani Sundara: yes per instance
[12:06] Ubit Umarov: i did run a oar of wright plaza in 32b
[12:06] Ubit Umarov: with some 800Mb of VM to spare
[12:06] Ubit Umarov: ( just after load.. that does down in use )
[12:06] Chen.Namamoto So it's my first time here, but a while ago I started writing a Rust client for the current UDP protocol (not finished yet), however in a recent log I read that there are going to be big changes from LL's side (switching to JSON/HTTP for many things)? So I was wondering if OpenSim is still going to use the current UDP protocol in the future?
[12:07] Andrew Hellershanks: Welcome, Chen,.
[12:07] Ubit Umarov: well LL is changing to the new thingy
[12:07] Ubit Umarov: not sure how long SL will go on
[12:07] OtakuMegane Desu: From what I understand LL is basically moving more of the data from UDP to HTTP like was done for textures years back.
[12:08] Ubit Umarov: yeah they did move lots of things from lludp to http already
[12:08] Ubit Umarov: moved
[12:08] Andrew Hellershanks: Changing the protocol is a big step. There aren't any plans to make that sort of change in OpenSim any time soon.
[12:09] Ubit Umarov: and in same cases maybe a bad move
[12:09] Chen.Namamoto I realized that they are switching to HTTP when I struggled to find the relevant information for how textures are transmitted etc
[12:09] OtakuMegane Desu: How the protocols work isn't always explained well.
[12:10] Ubit Umarov: the worse is also that those http changes also depend on services like CDN ??
[12:10] Ubit Umarov: same...
[12:10] Chen.Namamoto I'm interested in maintaining compatibility with OpenSim, that's why I'm asking, but if currently there is no plan to switch the protocol here I guess I'll just keep working on my project like before. (Can't promise anything like a new viewer yet, my first life is a bit busy.)
[12:11] Ubit Umarov: they are not changes just of protocol, but changes to match SL backend infrastructure...
[12:11] LaNani Sundara: i guess they would use HTTPS not HTTP?
[12:11] Ubit Umarov: so some are not that nice for us
[12:11] OtakuMegane Desu: I suspect that's as much LL's own choice of delivery. It may require some adapting in third-party but should be able to make it work for Opensim (eventually).
[12:12] Ubit Umarov: or we just will not change :p
[12:12] OtakuMegane Desu: What other things they're changing, I don't know. Haven't kept up with it in a while
[12:12] Ubit Umarov: like we are not already
[12:12] OtakuMegane Desu: Well that too. OS is slowly separating from SL so it doesn't necessarily have to follow along.
[12:13] Ubit Umarov: we want standalones and home run small grids, as long industry standards allow it ( like loopback issues )
[12:14] Ubit Umarov: we at least some of us ;)
[12:14] Sheera.Khan then IPv6 should get top priority...
[12:14] Ubit Umarov: well llUDP ( that is not just UDP) and http is now that a basic easy issue
[12:14] OtakuMegane Desu: Loopback issues is one of the big reasons I've never tried to run servers from home. So much easier to do proper remote.
[12:15] Sheera.Khan quite a lot of our users get IPv6 addresses from their provider already
[12:15] Ubit Umarov: industry is again tring to kill home services
[12:15] Ubit Umarov: moving to central payed services, like it was the idea in the 70's
[12:15] LaNani Sundara: llUDP is a attempt to build a TCP like protocol but using UDP, i always thought
[12:16] Ubit Umarov: the mobile devices gave them a very good excuse
[12:16] Ubit Umarov: no llUPD is not tcp
[12:16] Ubit Umarov: tcp is a stream oriented protocol
[12:16] LaNani Sundara: thats what i just said
[12:16] LaNani Sundara: its not
[12:16] Ubit Umarov: llUDP is a message oriented protocol
[12:17] LaNani Sundara: k
[12:17] Sheera.Khan but let's be honest: an important aspect of a server is the connection speed for uploads. Servers at home are at a big disadvantage there
[12:17] Ubit Umarov: there are other attempts to create such protocols other than LL llUDP
[12:17] Arielle Popstar: not when the user is behind the same network
[12:17] Arielle Popstar: then it is a huge advantage
[12:18] OtakuMegane Desu: Depends where you are. That was my other big reason but it's becoming less and less of an issue unless you're stuck in underserved rural areas.
[12:18] LaNani Sundara: Western Europeans have fast home connections these days, or can have if they want
[12:18] Ubit Umarov: internet is not a democratic thing..
[12:18] Arielle Popstar: i ran home servers for years on 1 mb down and 600 kb up. Good enough for me and 2 friends....pre mesh
[12:18] Ubit Umarov: and it is getting a lot worse on that
[12:19] Arielle Popstar: not all home users needing 20 friends over
[12:19] LaNani Sundara: mesh avatars.. with, i saw some with 42 layers, no wonder they load slow
[12:19] Ubit Umarov: and companies can make a lot more money forcing us to run application on servers than our own copies on pcs
[12:19] Ubit Umarov: see how games are now
[12:20] Ubit Umarov: you no longer bye a copy of a game
[12:20] Ubit Umarov: you by timed subscriptions.. and then " virtual goods"
[12:20] Chen.Namamoto P2P might be an option for busy home servers, but it would require quite some amount of work, or you can just wait and hope that internet connections will get faster
[12:20] Ubit Umarov: well SL did that also
[12:21] Ubit Umarov: insdustry always tried to kill p2p
[12:21] OtakuMegane Desu: The only games I have that fit that description are web stuff and I don't play much of them.
[12:21] Ubit Umarov: p2p is not a nice business for them
[12:22] OtakuMegane Desu: I don't like the subscription stuff like Adobe does now, though.
[12:23] LaNani Sundara: i wonder... sorry to change topics, is there or could there be a way to prebake avatars, mesh ones
[12:23] George Equus holding his Adobe Acrobat XI in a firm grip
[12:24] LaNani Sundara: am i making sense?
[12:24] OtakuMegane Desu: I'm sure it could be possible. What it would take to implement is another matter and probably nothing easy.
[12:24] LaNani Sundara: would it help rezzing?
[12:24] Arielle Popstar: new mesh avi's coming that are a lot lighter
[12:24] Arielle Popstar: i am mesh except head and my complexity is lesss then a lot of others here
[12:25] Andrew Hellershanks: Its almost half past. I need to get going soon. I'll make my announcement for the benefit of the late comers.
[12:26] Andrew Hellershanks: This past week I went through all the OSSL function pages in the wiki and updated the permissions (where they were wrong, and added in the details for the default permissions and script delay values.
[12:26] Andrew Hellershanks: I found some discrepancies between the OS code and the osslEnable.ini file. Those still need to be dealt with and the appropriate changes made. I started discussing my findings with the other core developers but haven't finished those discussions.
[12:27] Arielle Popstar: @Sheera yurs is light in comparison to some at events that are over 500,000
[12:28] Andrew Hellershanks: I wanted to do that for a while. The main reason was to get the default permissions information in to the wiki. People have been using OSSL functions because they were listed in the wiki without realizing they may not be available for use in all grids.
[12:28] Arielle Popstar: thx Andrew
[12:29] Arielle Popstar: the ossl documentation was confusing
[12:31] Andrew Hellershanks: Arielle, It may still be confusing. I haven't spent any time updating the content of the OSSL wiki pages apart from dealing with threat level, permissions, and delay time.
[12:31] Arielle Popstar: and prebaking an avatar in the sense of melding all the parts into one so the final weight would be = to Chen's avi would be cool
[12:31] Ubit Umarov: (rl phone sorry?
[12:32] LaNani Sundara: thats is what i meant yes Arielle
[12:32] Arielle Popstar: similar to the viewer ability to take a set of linked prims and convert them to one mesh
[12:33] LaNani Sundara: its would improve the way things load/rez
[12:34] Arielle Popstar: question is whether it could be done by the viewer on the fly or only through a third party program like blender
[12:35] George Equus: I seem to be in the lead on complexity in present company... yiks.. but at some venues I am the least complex.
[12:35] Arielle Popstar: Andrew question i have about the permissions whether those functions that for example are listed as Estate manager, Owner can also be set to true so anyone can do them?
[12:36] Ubit Umarov: ( back)
[12:36] Arielle Popstar: or are they hardcoded to only allow the default
[12:37] Andrew Hellershanks: The permissions can be changed in the osslEnable.ini file. I just listed the default setting.
[12:37] Ubit Umarov: think osslEnable.ini does have some coments on that
[12:38] Ubit Umarov: a few funtions did got hard restrictions ( ie do need god level no matter what )
[12:38] Ubit Umarov: ( until you change the source code :p )
[12:38] Arielle Popstar: i was having some trouble understanding the comments hence my question here
[12:38] Andrew Hellershanks: Ubit, yes, there are a couple of those. IIRC, I added a comment in the notes part of the page in those cases.
[12:38] LaNani Sundara: did Ubit write them? ;)
[12:38] Arielle Popstar: ahh ok so i did understand it correctly but just was unwilling to believe it
[12:39] Ubit Umarov: ( oops i did LOL )
[12:39] Ubit Umarov: :p
[12:39] LaNani Sundara: haha
[12:40] Ubit Umarov does not understand why that is funny
[12:40] Ubit Umarov: :p
[12:40] Chen.Namamoto I guess if the mesh complexity reduction should be efficient and accurate you only want to do it once because it's probably going to be expensive to compute, but then again on the server you usually cannot use a GPU.
[12:41] Ubit Umarov: hmm chen ??
[12:41] Arielle Popstar: read up Ubit
[12:41] Ubit Umarov: mesh complexity reduction is more a creator issue
[12:42] Arielle Popstar: not for what we were discussing
[12:42] Arielle Popstar: while you were on phone
[12:42] Andrew Hellershanks: I need to get going. Thanks everyone for coming. You don't need to leave right away but whoever is last remember to turn off the lights before you leave. :)
[12:42] Nebs Metal Bar Stool v1.6 (w/sit & launch): Goodbye..

Personal tools
About This Wiki