Chat log from the meeting on 2012-10-09

From OpenSimulator

Revision as of 15:48, 24 October 2015 by Fritigern (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
[11:02] VivK Lowlag: besides less clouds?
[11:02] logger sewell: lol
[11:02] Andrew Hellershanks: 7.5 is still in development?
[11:02] Marcus Llewellyn: There were some large changes to HG in 0.7.5
[11:03] logger sewell: Hello Justin
[11:03] Justin Clark-Casey: hi logger, folks
[11:03] Tiffany Magic: Oh wait... *grabs more coffee*.... the next version under development.
[11:03] Andrew Hellershanks: Hey, justing.
[11:03] Sarah Kline: hi all
[11:03] Andrew Hellershanks: -g
[11:03] Richardus Raymaker: hi sarah, justin
[11:04] Justin Clark-Casey: hi sarah
[11:04] VivK Lowlag: hi Sarah
[11:04] Richardus Raymaker: aha tiffany is compiling the coffee
[11:04] Andrew Hellershanks: I'm mildly surprised that there was no other follow up to the messages re: group code on the mailing list.
[11:04] Justin Clark-Casey: I want to write something but I keep running out of time
[11:04] Andrew Hellershanks: I found another way to improe the code a little.
[11:05] Justin Clark-Casey: I haven't really had a chance to look at the changes yet
[11:05] Andrew Hellershanks: There is a foreach on group chat sending where it makes a new object each iteration but the only thing that changed was the receipient.
[11:05] Andrew Hellershanks: SEeems very inefficient
[11:05] Justin Clark-Casey: but are those being launched on separate thrads?
[11:05] Nebadon Izumi: back, sorry was afk, phone rang
[11:06] Andrew Hellershanks: Not that I see. Its just a foreach followed by copying most of the same data again and again (other than the "to") and fires off the message.
[11:06] Justin Clark-Casey: File and line number?
[11:07] Richardus Raymaker: how many threads can opensim use ?
[11:07] Justin Clark-Casey: a lot
[11:07] Richardus Raymaker: ok. max i see on windows is 100 now.
[11:07] Andrew Hellershanks: That is in SendMessageToGroup() (line 208) in OpenSim/Region/OptionalModules/Avatar/XmlRpcGroups/GroupsMessagingModule.cs
[11:08] Andrew Hellershanks: I haven't run a test to make sure it didn't break but I moved most of the "Copy Message" block to before the foreach
[11:08] Andrew Hellershanks: That was my next thing to do with my standalone.
[11:08] Nebadon Izumi: ya I have been meaning to respond to that email
[11:08] Nebadon Izumi: I just not have had the time, I will try to take a look at it soon
[11:09] Andrew Hellershanks: I wonder how we can set up a test bed to simulate a group with a large number of online users
[11:09] Justin Clark-Casey: andrew: yes - in that case I don't see anything downstream which is currently multi-threaded, so you could just reuse a single object
[11:09] Justin Clark-Casey: though the memory saving is pretty small adn the cpu saving practically non-existant
[11:09] Andrew Hellershanks: justin, that's what I thought. Saves a lot of pointless object creation and data copying.
[11:10] Andrew Hellershanks: justin, you usually say that when I find some way to "improve" the code. It just offends my sensibilities somewhat to see things like that.
[11:10] Justin Clark-Casey: however, i can also imagine a situation where messages rae sent on different threads
[11:10] Justin Clark-Casey: at which point one would have to switch back
[11:10] Andrew Hellershanks: yes, I think that is something we would be heading to.
[11:10] Justin Clark-Casey: andrew: basically because I'm always after the big fish
[11:10] Andrew Hellershanks: would we?
[11:11] Andrew Hellershanks: I don't know the code base enough to go for big fish. I can only deal with the small ones.
[11:11] Nebadon Izumi: hello Dahlia
[11:11] Dahlia Trimble: hi :)
[11:11] Justin Clark-Casey: andrew: maybe - it could, for instance, stop any problems where slow delivery to one recipient holds up message dliveries to eveyone else
[11:11] Justin Clark-Casey: andrew: fair enough. Sorry, I don't mean to be discouraging
[11:12] Andrew Hellershanks: justin, no its not discouraging unless that meant it wouldn't be accepted to the master code base in which I would be wasting time on it.
[11:12] Justin Clark-Casey: but I suspect issues with message delivery probably lie elsehwere
[11:12] Andrew Hellershanks: yes, the big issues were in the email
[11:12] Justin Clark-Casey: oh no, I woul dbe happy to accept a patch which changed that
[11:12] Justin Clark-Casey: as long as we were certain that there wasn't a multi-threaded scenario later on
[11:12] Justin Clark-Casey: e-mail?
[11:12] Richardus Raymaker: mailinglist ?
[11:12] Nebadon Izumi: ya someone made some major fixes to groups
[11:12] Andrew Hellershanks: ok. I'll test it first. I'll add a comment before commiting the change that the code might need to be put inside the foreach if the routine becomes threaded.
[11:12] Justin Clark-Casey: oh that e-mail. Yes, I keep wanting to read that
[11:13] Justin Clark-Casey: keep running out of time
[11:13] Andrew Hellershanks: nebadon, Michelle Argus, IIRC
[11:13] Justin Clark-Casey: because if the changes are good I'd like to ask Michelle to submit a pull req to the Flotasm XMLRPC groups github repo
[11:13] Nebadon Izumi:
[11:13] Andrew Hellershanks: I commented on a few additional changes re: that code
[11:14] Nebadon Izumi: we actually had some issues with groups here at the Town Hall
[11:14] Nebadon Izumi: we had like 25 users
[11:14] Andrew Hellershanks: justin, any idea how we can do load testing on groups?
[11:14] Nebadon Izumi: and the sim froze twice
[11:14] Justin Clark-Casey: I wonder why the group chat causes so much lag./...
[11:14] Nebadon Izumi: due to massive groups spew on the console
[11:14] Nebadon Izumi: it tries to contact everyone
[11:14] Justin Clark-Casey: because it should all be happening on threads which are launched separeate from the packet processing loups
[11:14] Nebadon Izumi: even the offline people Justin
[11:14] Andrew Hellershanks: Don't know. The send request should be in a thread to avoid hanging the rest of what is going on
[11:14] Nebadon Izumi: instead of just the people are are online
[11:15] Nebadon Izumi: it gets hung up apparantly on the offline folks
[11:15] Justin Clark-Casey: nebadon: ah okay, michelle also has changes to opensim itself
[11:15] Andrew Hellershanks: The PHP code looks very ineffient
[11:15] Justin Clark-Casey: yes, I can imagine that
[11:15] Richardus Raymaker: but maptilkes are spamming the console to
[11:15] Richardus Raymaker: maptiles
[11:15] Nebadon Izumi: ya thats what it looked like I need to read it again myself
[11:15] Justin Clark-Casey: don't messages still need to be sent for offline saving?
[11:15] Nebadon Izumi: I dont think so
[11:15] Nebadon Izumi: not group
[11:15] Nebadon Izumi: only personal IMs get stored in SL
[11:15] Justin Clark-Casey: oh ok, so group IM is not saved?
[11:15] Justin Clark-Casey: ok
[11:15] Andrew Hellershanks: You might be confusing saving of offline IM's vs. group chat
[11:15] Nebadon Izumi: correct
[11:15] Justin Clark-Casey: that makes sense now I thikm about it
[11:16] Richardus Raymaker: hmm, let me check but i think i have seen option for offline groups
[11:16] Andrew Hellershanks: I remember there is a switch somewhere to send forward group chat but I forgotten where that is
[11:16] Nebadon Izumi: ya there might be an option for it
[11:16] Nebadon Izumi: but in Second Life, you never get offline group messages
[11:16] Nebadon Izumi: you get notices
[11:16] Nebadon Izumi: but not messages
[11:16] Andrew Hellershanks: But that would only be for certain avatars
[11:16] Dahlia Trimble: I added it and I cant remember either :)
[11:16] Justin Clark-Casey: saving offline group messages would be a lot of data
[11:16] Nebadon Izumi: ya
[11:16] Justin Clark-Casey: considering how much chat there is in certain groups
[11:16] Andrew Hellershanks: yea
[11:17] Richardus Raymaker: SOrry, thats on the robust side:
[11:17] Marcus Llewellyn: I've seen grids attempt to save grid chat as offline messages. It's not pleasant to log in after a few days and get hit with a ton of them.
[11:17] Andrew Hellershanks: Then the persons account would get a bit stick on login retrieving hundreds to thousands of saved messages during login
[11:17] Richardus Raymaker: orwardOfflineGroupMessages
[11:17] Tiffany Magic: I don't see that as a good thing. But, I sure would like to be able to attach NC's or LM's to group nitices.
[11:17] Marcus Llewellyn: group chat, I mean, not grid chat
[11:17] Andrew Hellershanks: attaching to notices is something else I've wanted to look at.
[11:18] Nebadon Izumi: well I would certainly love to get groups updated
[11:18] Nebadon Izumi: but from what i do remember of that email
[11:18] Nebadon Izumi: it sounded like she wanted us to make sure most of the grid got the sim updates
[11:18] Nebadon Izumi: then eventually switch over to the back end updates later
[11:18] Andrew Hellershanks: The SQL queries to retrive data for sending group chat should also get the minimal amount of data. Right now a common routine is used which I think is retrieving way more data than needed which is all getting passed back to the group module.
[11:19] Justin Clark-Casey: ok, looks like there are quite a few changes involved
[11:19] Nebadon Izumi: ya, thats what I thought to Justin
[11:19] Nebadon Izumi: I really havent had time to dig into it
[11:19] Richardus Raymaker: groups is big works to make it better
[11:19] Andrew Hellershanks: Michelle got me aware that there seems to be lots of areas where the code can be improved.
[11:19] Justin Clark-Casey: It would be good if someone could test
[11:19] Nebadon Izumi: I should probably atleast respond to the email and let her know I or we (OpenSim) is interested in the work
[11:19] Justin Clark-Casey: if only to make sure it doesn't break any existing aspect of groups
[11:19] Justin Clark-Casey: nebadon: yes
[11:20] Nebadon Izumi: I dont think i have groups setup anywhere else to test
[11:20] Nebadon Izumi: i may on my local HG grid
[11:20] Nebadon Izumi: i cant remember
[11:20] Nebadon Izumi: if i do i'll try to test it there
[11:20] Andrew Hellershanks: I've done that but would be good if a core dev also said yes
[11:20] Nebadon Izumi: its been quite some time since i started that grid though
[11:20] Nebadon Izumi: will take me a bit to remember everything
[11:21] Andrew Hellershanks: I should test the changes on my standalone
[11:21] Andrew Hellershanks: hm... the fire in the fireplace seems to be burning faster today
[11:21] Justin Clark-Casey: I'll try to make a sensible reply to Michelle end of today
[11:21] Richardus Raymaker: its getting colder outside andrew..
[11:21] Andrew Hellershanks: Richardus, I know. :-P
[11:22] Nebadon Izumi: thanks Justin
[11:22] Andrew Hellershanks: Justin, I'll test my other change to the group module code and if it doesn't appear to break anything I'll commit and make a patch and bug report.
[11:22] Justin Clark-Casey: thanks Andrew
[11:23] Justin Clark-Casey: groups doesn't receive a lot of love - I only really took over the PHP because the Flotsam guy moved on
[11:23] Justin Clark-Casey: whose names temporarily escapes me
[11:23] Nebadon Izumi: Michael Cortez
[11:23] Justin Clark-Casey: ah yes, mcortez
[11:23] Richardus Raymaker: i used the basic group code to.
[11:23] Andrew Hellershanks: justin, I don't like that 4 way JOIN in the SQL code just to get the list of who the chat message needs to go to.
[11:24] Andrew Hellershanks: What bothers me most is the query to get who messages need to go to is done for *every* group chat message.
[11:24] Richardus Raymaker: you want to but the query in a list andrew ?
[11:24] Andrew Hellershanks: It should be cached
[11:24] Simulator Version v0.5 shouts: OpenSim 0.7.5 Dev          1fb53e7: 2012-10-05 04:31:43 +0100 (Unix/Mono)
[11:25] Justin Clark-Casey: The issue with caching is that you need to pick up changes to the group made anywhere on the grid
[11:25] Andrew Hellershanks: Richardus, I want to see if the query can be simplified and return less data. I'd make a new routine to do that.
[11:25] Justin Clark-Casey: If you cache then there will at least be some delay until those are picked up
[11:25] Justin Clark-Casey: thuogh that might just be the price one has to pay to get more efficiency
[11:25] Andrew Hellershanks: Justin, yes, there are issues there but doing a complex query and all that data movement on every group chat message seems a major waste
[11:25] Richardus Raymaker: hi tx
[11:26] Arielle Popstar: is wright configured differently that group chats affect it more then lbbsa or the sandbox?
[11:26] Andrew Hellershanks: I considered a SQLlite database to store who is online in a given chat group and it only needs updating during login/logoff.
[11:26] tx Oh: Eh Oh
[11:26] VivK Lowlag: oh eh tx
[11:26] Justin Clark-Casey: Caching in mmeory would be much simpler
[11:27] Richardus Raymaker: And faster
[11:27] Andrew Hellershanks: True but you can have a lot of groups with active conversations. Wouldn't that use a lot of memory?
[11:27] Justin Clark-Casey: I don't think it would be a major factor
[11:27] Andrew Hellershanks: ok.
[11:28] Andrew Hellershanks: The big thing is being able to do load tests on group chat
[11:28] Richardus Raymaker: the biggest amount of memory you need for the avatar uuid's that need to get the messgae
[11:28] Nebadon Izumi: the only place were going to be able to realistically do large group chat tests is on OSgrid
[11:29] Nebadon Izumi: so step one is really make sure it works without any major breaking
[11:29] Justin Clark-Casey: Well, one either does it as an nunit test or you actually have to gather a large group of people
[11:29] Justin Clark-Casey: The second is far more complicated
[11:29] Justin Clark-Casey: but more realistic
[11:29] Nebadon Izumi: ya thats true
[11:29] Andrew Hellershanks: I have to look more closely at the code but I would think the only thing you need is a list of UUIDs only as to who should receive a message. All the other stuff that is currently retrieved from the DB on a chat message is not needed.
[11:29] Nebadon Izumi: i guess as long as its passing the nunit test
[11:29] Nebadon Izumi: its likely safe for this grid
[11:29] Nebadon Izumi: then it really gets tested
[11:29] Justin Clark-Casey: andrew: essentially, one starts caching group membership
[11:29] Nebadon Izumi: honestly i cant imagine its going to perform much worse than it already does
[11:29] Nebadon Izumi: lol
[11:29] Justin Clark-Casey: well, there are no unit tests that really cover groups atm
[11:29] Nebadon Izumi: we just need to make sure it doesnt break small groups
[11:30] Justin Clark-Casey: mainly because there is no core groups implementation
[11:30] Andrew Hellershanks nods to justin
[11:30] Andrew Hellershanks: well... there is a core groups implementation but it relies on external PHP code and a webserver.
[11:30] Nebadon Izumi: ya
[11:30] Nebadon Izumi: ya its really just hooks right?
[11:30] Nebadon Izumi: not really an implementation
[11:30] tx Oh: i guess you need a list of uuid's for each group, no?
[11:30] Richardus Raymaker: can groups not get some filter to clean the group db from all the 0 user test groups ?
[11:30] Justin Clark-Casey: I would call that an external interface
[11:30] Justin Clark-Casey: yes
[11:30] Andrew Hellershanks: yes
[11:30] Justin Clark-Casey: one could create a core implementation by filling in the blanks
[11:31] Justin Clark-Casey: indeed, it's on my long list but not going to get done by me anytime soon
[11:31] Nebadon Izumi: hehe
[11:31] Andrew Hellershanks: If someone can point me to how to make SQL queries in the OS code base that are independant of the underlying database being used, I would be very happy.
[11:31] Andrew Hellershanks: I'd like to eliminate external PHP/webserver for groups, search, and profile.
[11:32] Justin Clark-Casey: It's all in OpenSim.Data.*
[11:32] Arielle Popstar: wouldnt it be worthwhile to do a group load test on another plaza to see if you get the same result as we did the other day?
[11:32] Justin Clark-Casey: I don't thikn there is much doubt that there are efficiency problems with group messaging
[11:32] Richardus Raymaker: nebadon, where groupIM disabled on town hall ?
[11:32] tx Oh: what result?
[11:33] Andrew Hellershanks: Who knows, we may wind up doing groups better than in SL. :-)
[11:33] Dahlia Trimble: wouldnt be hard lol
[11:33] Richardus Raymaker: SL. there you soemtimes dont get contact with the group. its not gettinbg betetr there
[11:33] Nebadon Izumi: Richardus, I dont know, I will check
[11:34] Richardus Raymaker: normally group IM is the lag maker. i have it disabled because that
[11:34] Nebadon Izumi: Group Messaging is disabled here
[11:34] Andrew Hellershanks: I've been working on my own C# program so I've already dealt with DB queries so I know how to build them to avoid SQL injection issues. I'll take a closer look at OS's DB layer and see what I can use there.
[11:34] Richardus Raymaker: nebadon, then the problems are strange..
[11:35] Nebadon Izumi: but its likely sending out a notice to a large group might have similar results
[11:35] Marcus Llewellyn: I have the impression a lot of s have just gotten used to not using group chat.
[11:35] Arielle Popstar: if it is disaled then how would it make wright crash?
[11:35] Nebadon Izumi: it should theoretically not be sending notices to people who are not online, it should store them
[11:35] Andrew Hellershanks: I still use group chat
[11:35] Nebadon Izumi: i assume in the current setup it just tries everyone
[11:35] Nebadon Izumi: regardless of their status, since we do not store that info anywhere
[11:36] Richardus Raymaker: group notices works most time fien, also for offline
[11:36] Nebadon Izumi: I am sure group size makes a difference
[11:36] Dahlia Trimble: it uses (or used to) use the same mechanism as regular IM
[11:36] logger sewell: group chat works fine in VH it's just it has little lag
[11:36] Andrew Hellershanks: For attachements to group notices it is a matter of making sure that is supported in OpenMetaverse
[11:36] Richardus Raymaker: logger ? small group ?
[11:37] Richardus Raymaker: only to people maby on smae sim ?
[11:37] Tiffany Magic: Group notices work fine... and go to email.... but, nothing can be attached to them. No NC's or LM's.
[11:37] Andrew Hellershanks: without a large group to test it with it is hard to know what impact any changes make.
[11:37] logger sewell: we have had 8 or 9 in chat at the same time
[11:37] Nebadon Izumi: I am talking about groups with 500+ avatars
[11:37] Nebadon Izumi: hehe
[11:37] Justin Clark-Casey: mm, there is a hell of a lot of lookups associated with group chat
[11:37] Andrew Hellershanks: Justin, exactly.
[11:37] Dahlia Trimble: neb are you talking notices or char?
[11:37] Nebadon Izumi: either
[11:37] Dahlia Trimble: *chat
[11:38] Sarah Kline: we will never get that many concurrent users here )
[11:38] Justin Clark-Casey: even group changes won't fully alleviated that - since every non-local agent needs their region resolved
[11:38] Nebadon Izumi: I am just wondering if Notices have a similar effect
[11:38] Andrew Hellershanks: Justin, it does the call with the JOIN's to get the list of online avatars then uses that list in second query to get some group chat permissions
[11:38] Nebadon Izumi: in trying to contact everyone instead of just the people online
[11:38] Dahlia Trimble: notices should be stored, but storing chat is usually just confusing for users
[11:38] Justin Clark-Casey: I suspect that vast amount of delay is looking up user information and then trying to contact regions tbh
[11:38] Richardus Raymaker: hmm, nebadon i think send group notices give lag to. best to ask eryn
[11:38] Justin Clark-Casey: The cost of joins in a db is trivial by comparison
[11:38] Andrew Hellershanks: Hm... I was about to say sort by region UUID but don't think that's possible
[11:38] tx Oh: when will we get 500 concurent users in osgrid?
[11:39] tx Oh: (as average)
[11:39] Dahlia Trimble: I thought it sent messages to a central server if the recipient wasnt in the same sim?
[11:39] Justin Clark-Casey: dahlia: no, simulator itself looks up info then sends to each separate simulator
[11:39] Dahlia Trimble: cant remember now
[11:39] Andrew Hellershanks: It does handle local and remote sending differently
[11:40] Richardus Raymaker: and thats a disaster for the simulator if you have 1000 group memebers
[11:40] Justin Clark-Casey: I also love how methods are labelled Async but are nothing of the sort
[11:40] Andrew Hellershanks: :-)
[11:41] Justin Clark-Casey: ah no I take that back - probably misreading the code
[11:41] Andrew Hellershanks: You will see in the email I sent that I proposed group chat as a project for the metaversity students
[11:41] Arielle Popstar: is there any stats for how many members in the larger groups?
[11:41] Justin Clark-Casey: who are the metaversity students?
[11:41] Justin Clark-Casey: oh, the thing Hiro was organizing?
[11:41] Andrew Hellershanks: yea, that
[11:41] Nebadon Izumi: there were a few groups near 1000 avatars
[11:42] Andrew Hellershanks: There are about 6 in the IRC channel
[11:42] Justin Clark-Casey: This is a very compiclated problem - not a good student one
[11:42] Andrew Hellershanks: Justin, I suggested it as it doesn't require extensive knowledge of the overall code base.
[11:42] Marcus Llewellyn: I hang out in the #mateversity channel. I can't say I'm really active with it, though.
[11:43] Andrew Hellershanks: It is a reasonably isolated set of issues.
[11:43] Lani Global floats into the meeting and takes her seat casually
[11:43] Nebadon Izumi: I didnt even know there was a IRC channel
[11:43] Justin Clark-Casey: but it still requires an extensive knowledge of parts of opensim and an appreicate of the issues of distributed computing
[11:43] Andrew Hellershanks: uh oh. Crash time?
[11:43] Nebadon Izumi: lag..!..!..!
[11:43] Justin Clark-Casey: bung
[11:43] Justin Clark-Casey: also, requires people to actually measure and identify the problem first
[11:43] Andrew Hellershanks: oh. Major chat lag
[11:43] Nebadon Izumi: heh
[11:43] Nebadon Izumi: someone is playing with groups obviously
[11:43] Nebadon Izumi: lol
[11:43] Andrew Hellershanks: hehe
[11:43] Richardus Raymaker: hmm the biggest i now find is 204
[11:43] Sarah Kline: i just had group window open getting numbers of group sorry
[11:43] Nebadon Izumi: it was Richardus!
[11:43] Nebadon Izumi: lol
[11:43] Arielle Popstar: maybe a complicated issue with an overlooked simple solution
[11:43] Sarah Kline: was me lol
[11:43] Sarah Kline: closed
[11:43] Richardus Raymaker: i know the one from doro where hugh
[11:43] tx Oh: chatlags happen when new people rez
[11:44] Nebadon Izumi: this time I am sure it was groups
[11:44] Nebadon Izumi: hehe
[11:44] Nebadon Izumi: it lagged way to long for a new user
[11:44] Richardus Raymaker: sorry... i searched group
[11:44] Sarah Kline: L&P is 1.2K
[11:44] Nebadon Izumi: but that just highlights some of the problems with it
[11:44] Nebadon Izumi: hehe
[11:44] Nebadon Izumi: ya love and Peace
[11:44] Nebadon Izumi: is the biggest one
[11:44] Justin Clark-Casey: group searching will be a diferent issue
[11:45] Arielle Popstar: so disaling groups has no effect on a region?
[11:45] Nebadon Izumi: i actually have a cron task, that automatically disables group chat and stuff on that group
[11:45] Justin Clark-Casey: I would halfway suspect that's another request that isn't being handled asynchronously or something
[11:45] Andrew Hellershanks: That involves a sql query that could take longer to get a response
[11:45] Nebadon Izumi: no matter what the owners change it
[11:45] Richardus Raymaker: if tehre where lag , its my fault. because i used group -> search and also opened the My groups box
[11:45] tx Oh: the whole search topic need to be discussed
[11:45] Nebadon Izumi: it will instantly get reverted
[11:45] Nebadon Izumi: there are 3 or 4 groups i block
[11:45] Dahlia Trimble: block?
[11:45] Richardus Raymaker: i think the search for groups create lag to
[11:45] Nebadon Izumi: ya i have a crontask that changes the group perms
[11:45] Nebadon Izumi: on those large groups
[11:46] Nebadon Izumi: to disable group chat
[11:46] Andrew Hellershanks: tx: I have too many projects to spend a lot of time just digging through the OS code base and getting to know it all so I look for smaller things to handle. search, profile, groups are my areas of interest.
[11:46] Nebadon Izumi: cause its soo nasty
[11:46] Nebadon Izumi: it basically just crashes a sim
[11:46] Marcus Llewellyn: I've seen groups block regions on another grid very recently. They important a ton of data with invalid group UUIDs. As a result, the sims keep requesting data for non-existant groups... over and over and over... abiut every ten sceonds, and the regions become so slow as to be unusable.
[11:46] Richardus Raymaker: Dahlia. image that the simulator need to check and send to 1000 avatars.
[11:46] Arielle Popstar: you do the same for Lbsa Neb?
[11:46] Marcus Llewellyn: important == imported
[11:46] Nebadon Izumi: its unfortunate, but I really did not know what else to do
[11:46] Lani Global joins meeting
[11:47] Nebadon Izumi: this is not at the sims Arielle
[11:47] Nebadon Izumi: i actually modify the groups themselves
[11:47] Justin Clark-Casey: ho hum - clearly groups needs work
[11:47] Nebadon Izumi: so it cant be done on any sim
[11:47] Andrew Hellershanks: Marcus, how could you wind up with invalid group UUIDs?
[11:47] tx Oh: brave andrew hellershanks.
[11:47] Teravus Ousley: Hi rAdams
[11:47] Nebadon Izumi: OARs
[11:47] Marcus Llewellyn: They important Aurora OARs into an OpenSim grid. Land and objects in those OARs all has the aurora group UUIDs.
[11:47] Andrew Hellershanks: tx: :_)
[11:48] Marcus Llewellyn: imported. Argh
[11:48] Nebadon Izumi: if you have group owned objects in OARs
[11:48] Justin Clark-Casey: Hey Teravus
[11:48] Teravus Ousley: Hello Justin
[11:48] Justin Clark-Casey: hi Robert
[11:48] Nebadon Izumi: on my important OARs, like Universal Campus
[11:48] Richardus Raymaker: hello teravus
[11:48] Nebadon Izumi: before i make the oar, i edit the database
[11:48] Richardus Raymaker: hi robert
[11:48] tx Oh: Eh Oh
[11:48] Andrew Hellershanks: tx: I'm somewhat familiar with profile and search as I fixed things relating to them for 0.7
[11:48] Nebadon Izumi: and replace all Group data with Zero UUID
[11:48] Dahlia Trimble: are there ever groups with 1000 people online?
[11:48] Justin Clark-Casey: looks like robert got timedout
[11:48] Nebadon Izumi: this way my oars load cleanly outside of OSgrid
[11:48] Marcus Llewellyn: That's what I'm telling them to do, Neb... clear out all those nonexistant group UUIDs.
[11:48] Dahlia Trimble: what if the sim wasnt aware of users who were not online?
[11:48] Richardus Raymaker: still want to thanks for the undo move fix. its running now for a while but memory is not omeing above the 50% owhat the system have.
[11:49] Andrew Hellershanks: Dahlia, not sure about that in OS based grids. I have seen something close to that in SL.
[11:49] Lani Global: Yesterday at the OSGrid Town Hall Meeting, I brought up discussion about PhysicalPrimMax=10. I would like to talk about it today in the OpenSim meeting also, for those who have missed it.
[11:49] Nebadon Izumi: i would say most of the time here, Love and Peace Group
[11:49] Nebadon Izumi: has 1200 Users
[11:49] Nebadon Izumi: 99% of them are offline
[11:49] Nebadon Izumi: hehe
[11:49] Justin Clark-Casey: so much for Love and peace
[11:49] Marcus Llewellyn: It still seems that opensim could handle nonexistant groups more gracefully, though.
[11:49] Teravus Ousley thinks that the biggest issue with huge physical prim is mass
[11:49] Nebadon Izumi: hey Teravus
[11:49] Lani Global: I would like to propose a change to the default OpenSim dev simulator settings distributed file: Regions.ini there is a line of the file: ; PhysicalPrimMax = 10. This default should be changed to ; PhysicalPrimMax = 64 to support the widespread use and distribution of physical vehicles content in the OpenSim Community.
[11:49] Nebadon Izumi: that was also my Concern Teravus
[11:50] Richardus Raymaker: Dahlia, can you try rebake
[11:50] Nebadon Izumi: i actually planned after this meeting to test that more
[11:50] Nebadon Izumi: and try to find the best #
[11:50] Teravus Ousley: If we could limit the mass above 10, then the only other issue with huge physical prim is collision count
[11:50] Nebadon Izumi: 64 might just sink into the ground and keep on going
[11:50] Nebadon Izumi: to the point you cant find the prims
[11:50] Nebadon Izumi: ah, so even if its 64m
[11:50] Nebadon Izumi: act as if its mass is 10m?
[11:50] Richardus Raymaker: Teravus, lower the mass slowly if the prim get bigger ?
[11:50] Teravus Ousley: huge physical prim will collide with a lot more things
[11:50] Lani Global: a number of us have been testing Physical Vehicles with 64m prim size, and it works fine.
[11:51] Nebadon Izumi: Lani ground or air vehicles?
[11:51] Dahlia Trimble quickly does a vanity rebake...
[11:51] Lani Global: Air
[11:51] Andrew Hellershanks: I've seen physical objects partially in the ground cause major lag in a region. Not sure if that still happens.
[11:51] Teravus Ousley: And, there's another side issue.. that may not be that big of an issue in the short term...
[11:51] Nebadon Izumi: ya Air is not a problem
[11:51] Nebadon Izumi: try ground vehicles
[11:51] Nebadon Izumi: i suspect they will fail spectacularly
[11:51] Lani Global: but these air vehicles are limited by the default
[11:51] Richardus Raymaker: i have plaing with physical train on the list again.. not sure when.
[11:51] Dahlia Trimble: bigger physical prims probably generate more collision tests
[11:52] Teravus Ousley: the other issue.. is the ode physics plugin uses spaces to optimize collision detection. the spaces are a set size of 32m or so.
[11:52] Lani Global: We recognize that it is possible for each simulator owner/operator to set their own Maximum Physical Prim Size to 64 metres. We have changed this setting in some regions and are happy to report that everything works fine with larger physical prim vehicles. However, many people who rent regions in OSGrid or who run their own regions, do not have the skills to change this default.
[11:52] Nebadon Izumi: so maybe 32m should be the max?
[11:52] Richardus Raymaker: what phyical prim size is sl useing now ? i would say 64 meters
[11:52] Nebadon Izumi: all I remember from the past is that larger than 10m prims would sink into the ground quite a bit
[11:52] Teravus Ousley: The idea is that things in one 32m block shouldn't try and collide with things in another 32 m block
[11:53] Andrew Hellershanks: Lani, have you checked what happens to a region if a physical object is partially buried in the ground?
[11:53] Nebadon Izumi: sometimes they would sink below the ground completely and get lost
[11:53] Lani Global: While lone builders and owners of regions can change their own regions' Maximum Physical Prim Size to 64m, it is a different story for Content Creators. Vehicle designers who create free products and put them in the stores for the larger market of the OpenSim Community) must build to whatever is the Default limit
[11:53] tx Oh: who needs such large physical objects anyway?
[11:53] Sarah Kline: heard all this before somewhere ^^
[11:53] Lani Global: Currently, the Content Creators can not deliver all their wonderful designs for vehicles, and users can't modify or convert their existing vehicles, if even one prim in the vehicle is larger than 10 metres.
[11:53] Marcus Llewellyn: All prims in SL are limited to 64m. Physical weight in SL is done by linkset size... 32 prims max. Linkset linkability rules may also have a role in things.
[11:53] Lani Global: the SL maximum prim size is 64m
[11:53] Nebadon Izumi: well I understand that, but if the large prims get lost, and crash sims, then its not worth it
[11:54] Andrew Hellershanks: Hm... might need larger than 10m physical limit for a yacht
[11:54] Nebadon Izumi: so we just need to make sure its ok
[11:54] VivK Lowlag: this was in the townhall minutes
[11:54] Richardus Raymaker: Reminds me at some nightmare's you needed to be very quick to catch a sinking train in the past
[11:54] Andrew Hellershanks: Lani, what is their limit on physical prims?
[11:54] Nebadon Izumi: well just because SL does something, does not mean we can, their Physics engine is much more advanced
[11:54] Nebadon Izumi: there is no Limit in SL
[11:54] Marcus Llewellyn:
[11:54] Nebadon Izumi: any prim can be made physical
[11:54] Nebadon Izumi: even hacked in mega prims
[11:54] Andrew Hellershanks: ok
[11:54] Lani Global: well, these limits are preventing content in opensim.
[11:55] Dahlia Trimble: what kind of content?
[11:55] Nebadon Izumi: this is interesting
[11:55] Lani Global: there are literally THOUSANDS of vehicles in SL with larger than 10m prims
[11:55] Nebadon Izumi: on that page that Marcus just pasted
[11:55] Andrew Hellershanks: Is there some code issue that set the default limit to 10?
[11:55] Nebadon Izumi: Max. physics weight for an object to be set physical - 32.0
[11:55] Richardus Raymaker: cant remeber sl did have much >10m phyical content. it started since SL increased size
[11:55] Nebadon Izumi: so even the super gigantic mega prims their weight is 32
[11:55] Arielle Popstar: the user that doesnt know how to adjust the limit also wont know hopw to fix the issues that might come as a result
[11:55] Nebadon Izumi: according to that page
[11:55] Nebadon Izumi: so there is a cap
[11:56] Lani Global: in my store alone, there are many vehicles with larger than 10m prims
[11:56] Lani Global: in my store in OpenSim
[11:56] Marcus Llewellyn: Looks to me link you simply can't have a physcial linkset of more than 32 prims.
[11:56] Nebadon Izumi: oh
[11:56] Dahlia Trimble: Ive seen lots of large "vehicles" in SL but Ive not seen large ones that actuall move
[11:56] Nebadon Izumi: let me look again
[11:56] tx Oh: i even didn't knew that they allow 64m prims now :-)
[11:56] Lani Global: I currently have a physical vehicle with 60prims that works fine in OpenSim.
[11:56] Andrew Hellershanks: Ah. SL limits prims that can be physical based on their weight, not their dimensions.
[11:56] Nebadon Izumi: oh ya thats worded horribly
[11:57] Richardus Raymaker: Oh tx..
[11:57] Nebadon Izumi: i didnt read the whole thing lol
[11:57] Marcus Llewellyn: It does say mass... so yeah. It's confusing.
[11:57] Lani Global: I has introduced a new MULTI-PASSENGER PHYSICAL FLIGHT VEHICLE and an open source Physical Flight Script Engine developed for OpenSim.
[11:57] Dahlia Trimble: nice
[11:57] tx Oh: yepp, nice
[11:57] Andrew Hellershanks: I take it there is some formula to calculate object mass, or a script call.
[11:58] Nebadon Izumi: there is but I am not sure that really works here
[11:58] tx Oh: i had a flying house once in sl
[11:58] Nebadon Izumi: i tried making a helicopter and using the LSL to calculate and set mass
[11:58] Nebadon Izumi: and I dont think it actually does anything
[11:58] Richardus Raymaker: Does anybody know how things going with llSetKeyframedMotion ?
[11:58] Richardus Raymaker: Someone tried to make it, but its silent now.
[11:58] Nebadon Izumi: also I think we kind of present the avatar mass incorrectly
[11:58] Lani Global: The thing in OpenSim that is now holding back physical vehicle Content Creators is: currently there is a DEFAULT in the OpenSim dev simulator (ini) setup file of "PhysicalPrimMax = 10metres". This Default prim size is a carryover from the old SL 10m prim size standard. It is now obsolete.
[11:58] Andrew Hellershanks: Lani, is that yet another physics engine for OS?
[11:58] Nebadon Izumi: i remember making a helicopter
[11:58] Nebadon Izumi: and calculating the mass
[11:59] Nebadon Izumi: of just the vehicle
[11:59] Nebadon Izumi: it was like 5.0 or something
[11:59] Nebadon Izumi: then i sat avatar and it was like 85
[11:59] Nebadon Izumi: those arent exact numbers but it was something absurd like that
[11:59] Andrew Hellershanks: That's not much. SL has a weight limit of 32.0
[11:59] Lani Global: it is a "LSL Vehicle Script Engine" basically a prim with scripts and other things in it that people use as a root prim for their vehicles.
[11:59] tx Oh: oh, what's that? llSetKeyframeMotion?? is my pupeteer script useless now?
[11:59] Marcus Llewellyn: 32 lindograms?
[11:59] Nebadon Izumi: heh
[11:59] Nebadon Izumi: its been quite some time since i really hacked on vehicles though
[12:00] Nebadon Izumi: my memory needs major refreshing
[12:00] Richardus Raymaker: how did you calculate the mass nebadon. thats very intressting
[12:00] Arielle Popstar: you have a working puppeteer script/
[12:00] Sarah Kline: I go try making a physical megaprim in SL
[12:00] Andrew Hellershanks: I have one, Arielle
[12:00] Arielle Popstar: oh cool
[12:00] tx Oh: not in opensim, the link sets are a bit strange here
[12:00] Lani Global: I've added a proposal to my OpenSim Wiki regarding the proposed change to 64m PhysicalPrimMax default
[12:00] Helicopter v5.9: Vehicle Mass : 0.000000
[12:01] Helicopter v5.9: Vehicle Mass : 82.110878
[12:01] Justin Clark-Casey: A wild helicopter appears
[12:01] Andrew Hellershanks: If weight is calculated the same as in SL, that helicopter would be too heavy to be physical
[12:01] VivK Lowlag: it's a fun copter to ride
[12:01] Marcus Llewellyn: Well... it's no flying dead cat, but it'll do.
[12:01] Andrew Hellershanks: hehe
[12:01] Nebadon Izumi: llGetMass Richardus
[12:01] Lani Global: I'm happy to Demo my Physical Vehicle today for all OpenSim devs.
[12:02] Andrew Hellershanks now wonders when someone will show off a flying dead cat
[12:02] Arielle Popstar: lol
[12:02] Nebadon Izumi: lol
[12:02] Richardus Raymaker: aha ok neb. new command
[12:02] VivK Lowlag: stick a rocket on the cat
[12:02] Teravus Ousley: So here's a potential situation with a huge megaprim
[12:02] Richardus Raymaker: someone need to make a parachute
[12:03] Sarah Kline: yup you can have 64x64 physical prims in SL
[12:03] Andrew Hellershanks: oh. I forgot to check on the guy that is making the parachute jump from 120,000 feet
[12:03] Lani Global: you can have 1024x1024x1024 m Physical Prims in SL.
[12:03] Marcus Llewellyn: What happens if you link a few 64x64's together as physical, Sarah?
[12:03] Master Dubrovna: His jump was cancelled due to high winds
[12:03] Nebadon Izumi: well like lkalif said earlier in IRC
[12:04] Sarah Kline: I didnt stay long enought to try that lol
[12:04] Nebadon Izumi: people in SL have used Physical 256x256x256 prim
[12:04] Teravus Ousley: ... For a long time a plaza exists without problems... then someone rezzes a huge prim and sits it physically in a situation that embeds it within other objects perpetually colliding with things. As a result, it reduces the number of collisions that are available in the stack for proper uses... and people start falling through prim fixtures and objects
[12:04] Nebadon Izumi: they drop it in the center of the sim
[12:04] Nebadon Izumi: and record avatar contact
[12:04] Nebadon Izumi: as a form of radar
[12:04] Nebadon Izumi: lol
[12:04] Nebadon Izumi: such a terrible method
[12:05] Lani Global: Currently if you have a Physical Vehicle in OpenSim, all the prims in the link set must be less than 10m in any dimension. If you go over that dimension, even with one prim, the vehicle fails. This is with the OpenSim Regions.ini default settings.
[12:05] tx Oh: where did they end up?
[12:05] Teravus Ousley: there are only so many collisions available because the collision joints (the things that apply proper forces to keep objects apart) are stored in the stack and there's only so much of that before a stack overflow occurs
[12:05] tx Oh: ZERO_VECTOR?
[12:05] Nebadon Izumi: I will do some hard core testing after the meeting
[12:05] Richardus Raymaker: 64m physical sounds so big..
[12:05] Nebadon Izumi: and see how many 64m prims i can rez before things go bonkers
[12:05] Nebadon Izumi: and how the sim performs
[12:06] Arielle Popstar: so the potential for griefing is much greater
[12:06] Dahlia Trimble waves bye :)
[12:06] Nebadon Izumi: see you Dahlia
[12:06] Richardus Raymaker: Cant be worser then 800.000 cubes nebadon :)
[12:06] Marcus Llewellyn: That's kinda my thing. I picture a 64x64x64 cube... that's not a trivial object to have bouncing around my regions.
[12:06] Sarah Kline: well they need less prims to grief with yes
[12:06] Lani Global: the potential for griefing is not any worse with larger physical prims.
[12:06] Teravus Ousley: well, the stack won't overflow because the number of collision joints is limited.. but people will start falling through prim objects for sure
[12:06] Nebadon Izumi: ya having a 500 normal cubes is pretty bad too
[12:06] Nebadon Izumi: hehe
[12:06] Sarah Kline: i guess if somebodys going to grief they will do
[12:07] Arielle Popstar: nod
[12:07] Lani Global: besides, why should all creativity and features of the entire OpenSim community be stuck at the old 10m obsolete prim size?
[12:07] Richardus Raymaker: Sarah, as long the cant rezz on the sim...
[12:07] Andrew Hellershanks: with a 256m limit on prim dimensions having a physical limit of 10 seems a bit restrictive
[12:07] Teravus Ousley: because the sim will not be able to make enough contact joints to keep up with the requirements of the scene.
[12:07] Nebadon Izumi: Lani if it makes peoples sims run bad
[12:07] Nebadon Izumi: then it will be limited
[12:07] Arielle Popstar: many do know how to change the default Lani
[12:07] Nebadon Izumi: end of story
[12:07] Lani Global: that's my point. Most people don't know how to change the default 10m
[12:08] Nebadon Izumi: we just need to be sure it doesnt make peoples sims run bad
[12:08] Marcus Llewellyn: I guess for me, Lani, it's a case of being able to use that argument for any OSSL feature, for example. Why should I not be able to use osMakeNotecard in all of my scripts? Why can't I force everyone to have that on?
[12:08] Teravus Ousley: In a box stack, for example, this causes prim to fall into other prim and fall through the ground until there are enough contact joints to go around.
[12:08] Nebadon Izumi: ya
[12:08] Lani Global: So, ALL CONTENT made by Content Creators must conform to this obsolete 10m size physical prim.
[12:08] Justin Clark-Casey: it's not obseolete if a larger number would cause the per4formance problems Teravus is talking a bout
[12:08] Nebadon Izumi: only if they want it to run everywhere without instructions
[12:08] Teravus Ousley: once there are enough contact joints to go around, you'll start to see prim springing from other prim.
[12:08] Marcus Llewellyn: Nope. You can do what I do. I include a notevard with my creations telling people they need certain stuff enabled.
[12:08] Nebadon Izumi: right, we are not Second Life Lani
[12:08] Arielle Popstar: no, just include a NC to say to change default on home sim
[12:08] Andrew Hellershanks: nebadon, I've wondered why that function is restricted but can't remember the reason.
[12:08] Richardus Raymaker: Whats wrong with makeing a object from more 10m prims. sl did that for years.
[12:09] Nebadon Izumi: so you really do need to take that out of the argument
[12:09] Richardus Raymaker: uhhmm. wait . thats right what i say ?
[12:09] Nebadon Izumi: we can not just arbitrarily make changes because SL does
[12:09] Nebadon Izumi: we need to make sure if we change it, it doesnt make OpenSimualtor run bad
[12:09] Sarah Kline: how far are people going to travel in these huge spaceships
[12:09] Richardus Raymaker: can you join in sl 2 10 meters prims phyical in the past ?
[12:09] Lani Global: Especially with Mesh and Sculpty vehicles, the larger size is needed for ships, yachts, spacecraft, built to scale.
[12:09] Nebadon Izumi: because if you are distributing things that end up crashing everyones sim
[12:09] Nebadon Izumi: how does that benefit the project?
[12:09] Lani Global: you simply can't have a crew or a party inside a small ship made of 10m prims.
[12:10] Lani Global: Mesh spaceships, airplanes, and ships require the larger size prims.
[12:10] Marcus Llewellyn: You can make ships as big as you like. I'm not sure you should *everyone* to support your ship on every region.
[12:10] Richardus Raymaker: There enough simulators that run low on power. and i know how hard things can get down myself with phyical prims. escpialy if the start to go nuts
[12:10] Nebadon Izumi: ya we understand that, but we are not going to make a change that could potentially cause issues, thats all
[12:10] Andrew Hellershanks: Lani, a small crew perhaps, but not a big party
[12:10] Nebadon Izumi: it needs more testing
[12:11] Nebadon Izumi: were not saying no, were just saying we need to test it more
[12:11] Nebadon Izumi: and make sure we dont end up pissing everyone off
[12:11] Andrew Hellershanks: Your missions is to run tests and report back.
[12:11] Andrew Hellershanks: :-)
[12:11] Lani Global: The limits that are applied to default in OpenSim should be large enough to allow common creative content. Presently this arbitrary limit of 10m stifles content.
[12:11] Nebadon Izumi: ok well you dont seem to be listening
[12:11] Richardus Raymaker: it can get tricky for peple to recover from a crashed phyical sim.. many dont have daily backups etc.
[12:11] Nebadon Izumi: so I am going to change the topic now
[12:11] Nebadon Izumi: hehe
[12:11] Lani Global: You can't even make a yacht or a pirate ship.
[12:12] Nebadon Izumi: we'll let you know how testing goes
[12:12] Lani Global: you can't make a crewed space ship.
[12:12] Arielle Popstar: what happens when one of these vehicles hits the sim boundary>?
[12:12] Lani Global: you can't make a normal to-scale airplane.
[12:12] Nebadon Izumi: repeating yourself 1000 times isnt going to make me change my mind about it I understand what your saying
[12:12] Nebadon Izumi: you do not seem to be grasping the actual problem though
[12:12] Andrew Hellershanks: Lani, so run tests with the limits > 10 and tell us how it goes.
[12:12] Lani Global: i'm making the case today here to emphasize the need for it.
[12:13] Nebadon Izumi: ya I understand the need
[12:13] Sarah Kline: you can make your ships out of mesh
[12:13] Nebadon Izumi: there is no need to keep stressing that part
[12:13] Lani Global: Sarah, I have ships of mesh, but the limit of 10m is much too small to make them fly or float or drive.
[12:13] Richardus Raymaker: people that need it, can change the setting in the opensim.ini. then everybody happy. and if things then go wrong you know why
[12:13] Nebadon Izumi: you need to understand that is not a good reason to just arbitraly make the change without wide scale testing that is all I am trying to say
[12:14] Master Dubrovna: Need to remember some regions here are running on limited resourses. How would that change affect them?
[12:14] Richardus Raymaker: That goes bad master. and hi
[12:14] Marcus Llewellyn: I"m not against a 64m limite, per se. I would like to know my mesh terrain isn't gonna fail if I do it, though.
[12:14] Nebadon Izumi: if all goes well though I am all for the change
[12:14] Andrew Hellershanks: hehe... I see some land on the map that has a bit that looks like a knotted pretzel
[12:14] Lani Global: If you don't care about having more Content Creators provide physical vehicles for the OpenSim Community, then, fine leave the default at 10m.
[12:14] Nebadon Izumi: infact most of the core developers are for it also
[12:14] Arielle Popstar: i have other products that require me to change defaults and osfunctions f0or proper operation
[12:15] Nebadon Izumi: as long as it doesnt cause problems for everyone
[12:15] Richardus Raymaker: sounds intressting to see marcus. i wamnt to test some freefalls on your mesh terrain.
[12:15] Sarah Kline: Lani Neb said he will test
[12:15] Sarah Kline: and if ok then fine
[12:15] Arielle Popstar: those creators didnt stop from creating ecause of current dsefaults
[12:15] Sarah Kline: give him a chance lol
[12:15] Nebadon Izumi: ya well you seem to just be ignoring everyone now Lani
[12:15] Justin Clark-Casey: one would have to test the bulle tplugin as well
[12:15] Marcus Llewellyn: You can find a huge mesh cave and mountain at my region Bear, Rich.
[12:15] Nebadon Izumi: so time to move on to something else
[12:15] Nebadon Izumi: hehe
[12:15] Lani Global: Content Creators, those who make products for the whole community must build their products to whatever is the default. yes, i'm responding to Richardus.
[12:15] Justin Clark-Casey: since the current plan is that bullet will become the default phys eng at some stage
[12:15] Nebadon Izumi: ya things might be completely different with Bulletsim
[12:16] Nebadon Izumi: but thats not quite ready for everyone yet either
[12:16] Richardus Raymaker: Ok. marcus.
[12:16] Nebadon Izumi: eventually it will be far superior to ODE
[12:16] Lani Global: I'm making my case here today, mainly for those who might not have been to the OSGrid Town Hall Meeting, where we discussed this.
[12:17] Richardus Raymaker: I see we lost robert ?, justw anted to know how far he's with bullet.
[12:17] Nebadon Izumi: ok I understand, It will be considered, I dont think anyone else needs to be convinced, now we just need to all test is as best we can
[12:17] Nebadon Izumi: and report on IRC or mantis anything you find unusual
[12:17] Lani Global: I've volunteered to work with OSGrid gods/devs and now I'm volunteering to work with OpenSim devs on this.
[12:18] Nebadon Izumi: Lani i suggest you do more ground based vehicle testing
[12:18] Nebadon Izumi: because I think thats where if any problems will be, it will be on the ground
[12:18] Nebadon Izumi: not with air vehicles
[12:18] Teravus Ousley notes that he tried some SL vehicles in OpenSimulator and they're still behaving badly
[12:18] Tiffany Magic: How do those huge ships and vehicles operate if they can't cross simlines?
[12:18] Nebadon Izumi: ya I would be surprised if any SL vehicles just worked here
[12:18] Nebadon Izumi: without some major mods
[12:18] Justin Clark-Casey: teravus: yeah, I don't think anyone has worked on the vehicle code for a long time
[12:18] Lani Global: ground vehicles physics should not be the preventer of all other types of physical vehicles.
[12:19] Nebadon Izumi: Lani, if everything dosnt work right
[12:19] Nebadon Izumi: it will be the limiting factor
[12:19] Nebadon Izumi: if you dont understand that I don't know what to say to you
[12:19] Lani Global: currently, i have a physical flight vehicle system that crosses sim borders.
[12:19] Nebadon Izumi: it cant just suit your narrow needs
[12:19] Lani Global: i do understand the ground vehicle problem.
[12:20] Nebadon Izumi: if we do find that it is a problem
[12:20] Nebadon Izumi: and we are stuck on 10m
[12:20] Nebadon Izumi: the only thing I can suggest is you include a notecard in the vehciles teaching people how to make the change to OpenSim.ini
[12:20] Nebadon Izumi: and restart their simulator
[12:20] Lani Global: its not 'narrow' because i've already solved it for my own sims. but i'm a representative of a group of content creators that provide content to the opensim community.
[12:20] Arielle Popstar: sounds like the size of the vehicle will dictate a mega anyway for starters
[12:20] Nebadon Izumi: and I represent the entire project not just a group of creators
[12:20] Marcus Llewellyn: It may be there's a compromise size as well. perhaps 20m, or 32m.
[12:21] Lani Global: the 10m prim size was arbitrary.
[12:21] Nebadon Izumi: I do understand your argument, and I hope you can understand ours
[12:21] Richardus Raymaker: think32m is the max. if i hear ode have that as limit for some things
[12:21] Nebadon Izumi: Lani, no its not
[12:22] Nebadon Izumi: its just at the time we decided on 10m
[12:22] Nebadon Izumi: its what worked the best
[12:22] Nebadon Izumi: many of us have done testing on larger prims in the past
[12:22] Nebadon Izumi: and it did not go well
[12:22] Teravus Ousley: at 32m, you start to loose some of the 32m space optimizations
[12:22] Lani Global: things have changed.
[12:22] Nebadon Izumi: beleive me, I have done substantial vehicle testing
[12:22] Nebadon Izumi: yes, things have changed, and thats why we are willing to do more testing
[12:22] Teravus Ousley: it contacts with two of the spaces at once joining the islands essentially.
[12:22] Nebadon Izumi: i agree, we probably can increase things
[12:22] Richardus Raymaker: do you know if kitto tested >10m phyical ?
[12:22] Teravus Ousley: maybe more.. depends...
[12:22] Nebadon Izumi: Richardus, I doubt it
[12:23] Nebadon Izumi: his OpenSim testing was pretty limited
[12:23] Andrew Hellershanks: Time for me to get going
[12:23] Richardus Raymaker: ok so to stay safe 30m is the limit for ode
[12:23] Nebadon Izumi: k Andrew, see you back on IRC im sure :)
[12:23] Justin Clark-Casey: ok, I need to bounce as well
[12:23] Richardus Raymaker: bye andrew
[12:23] Justin Clark-Casey: see you folks later
[12:23] Nebadon Izumi: k Justin, thanks for coming
[12:23] Teravus Ousley: well, if you think about it.. 32m spaces is 3 times the max primsize
[12:23] Marcus Llewellyn: Laters Andrew and Justin :0
[12:23] Richardus Raymaker: bye justin
[12:23] Marcus Llewellyn: :)
[12:23] Nebadon Izumi: ya wow I just realized the time
[12:23] Nebadon Izumi: hehe
[12:23] Teravus Ousley: +2
[12:24] Nebadon Izumi: didnt realize we had gone so far past the hour
[12:24] logger sewell: tc
[12:24] Andrew Hellershanks: Nebadon, oh, definitely. As if I leave when my computer is on.
[12:24] Andrew Hellershanks: ok, take care everyone. See you next week.
[12:24] Sarah Kline: bye justin
[12:24] Andrew Hellershanks waves
[12:24] Nebadon Izumi: see ya
[12:24] Justin Clark-Casey waves
[12:24] Simulator Version v0.5 shouts: OpenSim 0.7.5 Dev          1fb53e7: 2012-10-05 04:31:43 +0100 (Unix/Mono)
Personal tools
About This Wiki