Chat log from the meeting on 2023-05-02

From OpenSimulator

Jump to: navigation, search
[10:59] Andrew Hellershanks: Hello, everyone.
[10:59] Selby.Evans hi everyone
[11:00] Motoko.Karu Hihi everyone
[11:00] Tess JL: Hello
[11:00] Cuga.Rajal Hello all :)
[11:00] Lyr Lobo: Hello *smiles*
[11:00] Jamie.Jordan Hi everybody
[11:01] Joe Magarac shouts: Back shortly. All inventory missing.
[11:02] Vincent.Sylvester Let's see last weeks changes were mostly adding strided list stuff and so on
[11:02] Vincent.Sylvester Minor fixes
[11:02] Andrew Hellershanks: Welcome, Joe and Tess.I don't recall seeing either of you before. Glad to have you.
[11:02] Tess JL: Thank you
[11:05] Andrew Hellershanks: Hello, Gavin.
[11:05] Vincent.Sylvester Something something about srgb colors, which are something viewer side if I remember correctly
[11:05] Gavin.Hird Hi Andrew, everyone
[11:05] Vincent.Sylvester Still nothing all that solid on pbr so far
[11:06] Tess JL: You mean from LL?
[11:06] Gavin.Hird The PBR code is still a moving target
[11:07] Joe Magarac: Sorry about that. Somehow lost outfits folder.
[11:07] Gavin.Hird ouch
[11:08] Andrew Hellershanks: wb, Joe
[11:10] Vincent.Sylvester Ubit did some updates on the lsl implemented functions list on the wiki
[11:11] Ubit Umarov: yeah..  s a bout code changes i did add a few lsl functions
[11:11] Lyr Lobo: Hello *smiles*
[11:11] Ubit Umarov: namely those srgb to linear color and reverse  that ll did add something ago
[11:12] Ubit Umarov: llLinear2sRGB and llsRGB2Linear
[11:12] Ubit Umarov: thins things have to do mostly with Gamma correction..
[11:13] Ubit Umarov: i also did add llListSortStrided, llList2ListSlice and llListFindStrided
[11:14] Ubit Umarov: found that our implementation of llList2ListStrided  had little relation to ll one, so i changed it
[11:14] Ubit Umarov: did kept our old code as ossl osOldList2ListStrided
[11:15] Tess JL: What was the difference?
[11:15] Gavin.Hird what did our implementation do?
[11:15] Ubit Umarov: llListFindList  was also not correct..
[11:15] Ubit Umarov: mostly handly of start and end ranges
[11:15] Ubit Umarov: handling i mean
[11:16] Ubit Umarov: on those, those are simple, have no option to define a exclude range
[11:16] Ubit Umarov: our code did it etc
[11:16] Cuga.Rajal I was using llListFindList a lot recently and noticed a case that produced a thread crash
[11:16] Andrew Hellershanks: I think the additions have all (mostly?) been mentioned. The full list of functions that were added are: llLinear2sRGB, llList2ListSlice, llList2ListStrided, llListFindList, llListFindStrided, llListSortStrided, llsRGB2Linear, osListSortInPlaceStrided.
[11:16] Cuga.Rajal would that be fixed now?
[11:16] Ubit Umarov: llListFindList compare was also not as spec
[11:17] Cuga.Rajal I'll re-test that
[11:17] Vincent.Sylvester Do you have a log of that crash Cuga?
[11:17] Ubit Umarov: bu tthat one i just did replace
[11:17] Cuga.Rajal it wasn't a crash per se, it did cause a server error at the console
[11:17] Ubit Umarov: i did update the release notes :)
[11:18] Cuga.Rajal I donlt have the output now but can try to reproduce
[11:18] Ubit Umarov:
[11:18] Cuga.Rajal I'll send it to you
[11:18] Vincent.Sylvester There was something a while back with, I think, implicit casts or something that caused some issues that kept pointing to list functions
[11:19] Ubit Umarov: i actually found a auto one
[11:20] Ubit Umarov: like list a = aIntegerFuntionReturnValue
[11:20] Cuga.Rajal I will try to reproduce. At the time I was trying just to get something to work, so I found a work-around
[11:20] Ubit Umarov: does set a = [aIntegerFuntionReturnValue]
[11:20] Ubit Umarov: at sl it is a error
[11:20] Ubit Umarov: i left it as is for now..
[11:21] Joe Magarac: A question, while lists are being discussed. Lists are passed to functions by value in SL. Does OS also pass them by value, or by reference? Since lists are immutable, scripts can't tell the difference, Pass by reference would help with script space bloat.
[11:21] Ubit Umarov: well this new code only got basic testing..
[11:22] Ubit Umarov: hm not sure sl does pass them by value
[11:23] Cuga.Rajal If you are updating, can I request updating llSetForce to have same comments as llSetBuoyancy
[11:23] Joe Magarac: I've had scripts blow up because they passed a large list by value, causing a stack/heap collision.
[11:23] Ubit Umarov: internallly at opensim they are passed as reference
[11:24] Ubit Umarov: since they are implemented as a c# class
[11:24] Joe Magarac: Good to know. Will bug the SL devs.
[11:24] Andrew Hellershanks: :)
[11:24] Ubit Umarov: guess they will not change that now
[11:25] Joe Magarac: they might. They just doubled script space from 64KB to 128KB. So that area is getting attention.
[11:25] Ubit Umarov: very unrelated thing :p
[11:25] Andrew Hellershanks: Joe, They did? Hm... my particle generator script may work over there now.
[11:25] Ubit Umarov: by value or ref is language spec..  max memory is just a detail
[11:25] Misterblue Waves: a different question: does anyone know if login checking has changed in the last month or so? My testing environment is getting "Logins currently restricted" errors trying to log into OSGrid but logging in with Firestorm works.
[11:26] Ubit Umarov: osg did change some block by mac address
[11:26] Andrew Hellershanks: Misterblue, What viewer are you using when you aren't using FS?
[11:26] Misterblue Waves: my testing setup is odd in that I'm logging in through an Nginx XMLRPC proxy to get the https request (since OSGrid doesn't use https)
[11:27] Ubit Umarov: well a reason for restricted answer is a forbidden mac address
[11:28] Ubit Umarov: or id0
[11:28] Misterblue Waves: it's my browser viewer for OpenSim... I have to proxy since browsers need https (or, at least, once https, always https)
[11:28] Misterblue Waves: guess that means that DigitalOcean's mac addrs have been restricted
[11:29] Misterblue Waves: guess I need to talk to OSGrid admins
[11:29] Ubit Umarov: no idea
[11:29] Ubit Umarov: ask dan
[11:29] Andrew Hellershanks: I was about to mention Dan. :)
[11:29] Ubit Umarov: and test with him looking to console
[11:29] Vincent.Sylvester Normally for secure logins you can just proxy pass, don't need anything more fancy than that
[11:30] Ubit Umarov: does your viewer send proper mac ?
[11:31] Ubit Umarov: ( the block works on information sent by the viewers.. reason why is so.. useless )
[11:31] Misterblue Waves: I am doing a proxy pass to do a => and it used to work. Stopped working sometime n the last few months so something changed
[11:32] Ubit Umarov: real mac physical mac can't ever be used.. macs are local to a physical connected network
[11:32] Ubit Umarov: see what you are sending as mac
[11:32] Ubit Umarov: .. and change it :P
[11:33] Misterblue Waves: I'll talk to Dan and see if anythings changed
[11:35] Andrew Hellershanks: I was trying to see if Dan is around on IRC but I'm not sure which server/channel he is using.
[11:35] Joe Magarac: I use the MAC address from in my Rust viewer, and  that works for SL, OSGrid, and Kitely at least.
[11:37] Ubit Umarov: about lists immutability i did add a clear violation long ago, and again now
[11:38] Ubit Umarov:
[11:38] Ubit Umarov: does sort the input list in place
[11:38] Ubit Umarov: so saving memeory
[11:38] Ubit Umarov: memory
[11:39] Ubit Umarov: like
[11:40] Object: Script running
[11:41] Object: Script running
[11:42] Object: Script running
[11:44] Ubit Umarov: nap moment?
[11:44] Andrew Hellershanks: I was just checking something.
[11:44] Vincent.Sylvester Trying to think if there is anything else that happened last week
[11:45] Vincent.Sylvester I managed to eject myself from my own region, that was fun, but not really a bug either
[11:45] Gavin.Hird naw, just updating a bunch of SL branches after they promoted another one
[11:45] Andrew Hellershanks: It would have been even funnier if you had banned yourself.
[11:45] Ubit Umarov: for those using fs beta ( not gavin:p) fs did new beta release
[11:46] Ubit Umarov: that as images/tgextures upload broken
[11:46] Ubit Umarov: has
[11:46] Gavin.Hird oh
[11:46] Ubit Umarov: they did fix it on git already
[11:46] Orbert.Tatham Have any of the beta users tried any mesh uploading to see how it works now?
[11:46] Vincent.Sylvester I was working on a system to fix an avatar in place on an object via pushobject and the calculation for the applied force rolled the integer over and moved me across the region in less than a frame time apparently. Found myself outside the region, but still in it
[11:46] Motoko.Karu ^_^
[11:47] Ubit Umarov: the issue im taliking is on image uploads
[11:47] Gavin.Hird teleporting someone to coordinate 0,0,10000 is always fun
[11:47] Ubit Umarov: vincent Vodoo is a topic for other ppl meetings
[11:47] Ubit Umarov: :p
[11:49] Ubit Umarov: and Vincent are you sure it was same region on same universe?
[11:49] Vincent.Sylvester We have no bound checks on movement, there are some, but nothing "solid" so to speak, hence yeah could try and fix that, but not really worth it, edge case... literally
[11:49] Ubit Umarov: duhh ofc we do have checks...
[11:50] Ubit Umarov: how do you think crossings happen?
[11:50] Ubit Umarov: :p
[11:50] Gavin.Hird stumbling across the region border?
[11:51] Vincent.Sylvester Position yes, movement no. Not much tests "does this movement even work", well at least not directly
[11:51] Vincent.Sylvester Move quickly enough it never catches you
[11:51] Vincent.Sylvester Just fun experience
[11:51] Joe Magarac: SL side, region crossings trigger when the object root passes 1M past the border.
[11:52] Vincent.Sylvester No point wasting nanoseconds adding such a check ;)
[11:52] Ubit Umarov: a check you are dreaming about
[11:52] Ubit Umarov: and in fact some of those things do depend on the physics engine in use also
[11:52] Ubit Umarov: at least in part
[11:54] Gavin.Hird so if the main link object in a link set is 1 meter into the next region, doe SL suck the entire object over to that region? (way to ruin a build)
[11:54] Ubit Umarov: obejct position is root prim position
[11:55] Gavin.Hird yes, so if root position is 1 meter in, will an entire linked build follow?
[11:55] Joe Magarac: Yes.
[11:55] Ubit Umarov: yeap
[11:55] Gavin.Hird woops
[11:55] Ubit Umarov: same here, ofc
[11:55] Ubit Umarov: 0.1m even :)
[11:57] Joe Magarac: I know it's 1m in SL. I've tested.
[11:57] Jagga Meredith: I built a rectangular object to surround another parcel that was autoreturn.  No part touched it.  It got sucked in anyway.  Had to build four separate pieces.
[11:58] Vincent.Sylvester If you place it right on the edge that may happen
[11:58] Jagga Meredith: yup
[11:58] Bill's Chair whispers: This is Bill's Chair
[11:58] Vincent.Sylvester Move it a couple centimeters in and it should be fine
[11:59] Gavin.Hird Hi Popstar
[11:59] Arielle Popstar: hi Hird
[11:59] Cuga.Rajal that autoreturn would be based on the center point of the prim
[11:59] Andrew Hellershanks: Hello, Arielle. There is an empty seat to the right of Mr. Blue.
[11:59] Lyr Lobo: Hi Arielle
[12:00] Arielle Popstar: Hi Andrew, Hi Lyr
[12:01] Cuga.Rajal a single prim "surrounding" the parcel would have a center point inside the parcel
[12:01] Andrew Hellershanks: We are already at the top of the hour. Does anyone who may have to leave shortly have a question/comment.
[12:01] Jagga Meredith: ah
[12:01] Ubit Umarov: think here, only the root prim position matters
[12:01] Arielle Popstar:
[12:02] Joe Magarac: I have a new viewer out, and am working through OS/SL differences. So I will be back with obscure questions.
[12:02] Vincent.Sylvester Should be given sim surround things use that inversely
[12:02] Tess JL: I have a quick question if there's time
[12:02] Ubit Umarov: i gave beq a code solution for that..
[12:02] Ubit Umarov: seems she forgot it :)
[12:02] Cuga.Rajal I think thats correct.. make an object to surround the parcel with the root prim center outside the parcel
[12:02] Ubit Umarov: i told about it last week
[12:03] Vincent.Sylvester Beq still exploring rl probably
[12:03] Vincent.Sylvester Weird grid that, heard the admins can be mean
[12:04] Andrew Hellershanks: Tess, Sure. You can ask your question now
[12:04] Jagga Meredith: yeah, big blue room
[12:04] Ubit Umarov: gavin, had a chance to test my code ?
[12:05] Ubit Umarov: dayturn also does not display terrain textures on minimap on such regions
[12:05] Ubit Umarov: with size not a power of 2
[12:05] Tess JL: I know two grids, Dreamworldz and the Gor Grid have a function that will automatically start regions when somebody enter a neighbour region. Is that a function suproted by opensim or have they added it themselves?
[12:05] Ubit Umarov: added tess
[12:05] Gavin.Hird I have not looked at it. there is so much else going on with the code and build issues and a million changes from MS, Apple and LL topping up at the same time
[12:05] Gavin.Hird so first things first
[12:05] Andrew Hellershanks: Tess, it isn't part of OpenSim.
[12:06] Andrew Hellershanks: Kitely Grid does the same thing.
[12:06] Orbert.Tatham Seems like Kitely does that too
[12:06] Vincent.Sylvester It's an extension to the neighbor service that calls a remote point to issue the starting command
[12:06] Tess JL: Yes but Kitely doesn't have adjacent regions
[12:06] Cuga.Rajal Arielle last time you said your debug console was set to "Info".. is there a lower logging setting like logging errors only
[12:06] Vincent.Sylvester Kitely does that on initial teleport request
[12:07] Ubit Umarov: (ariele i did change this region size to 1024 to stop the warnings here )
[12:07] Arielle Popstar: didn't check but still wouldnt solve the significant impact on the FPS ?
[12:08] Cuga.Rajal it might
[12:08] Cuga.Rajal worth trying
[12:08] Andrew Hellershanks: Hm... I can't think of why those messages would have an impact on FPS
[12:08] Ubit Umarov: dreamworld does have that the option to start/stop regions acording to use
[12:09] Andrew Hellershanks: Is that done at the region level or is it handled through Robust?
[12:09] Arielle Popstar: Joe, did you set up a media site for the viewer  chat you were talking about last week? Discord?
[12:09] Tess JL: IF you mean Dreamworld, it's handled through Robust
[12:09] Misterblue Waves: it would be nice to have a standard way to do that in OpenSim core
[12:10] Vincent.Sylvester jsonsimstats root agents = 0 shut it down, would be simplest way
[12:10] Vincent.Sylvester You could build a module that checks on teleport in or out whether anyone is left and issue shutdown
[12:11] Vincent.Sylvester Plenty of ways to accomplish that
[12:11] Tess JL: I have some selfish reasons to ask. I'm working on a continent with more than a thousand 4x4 regions. There's no way I cn afford to put it online if all regions are to be live all the time ;-)
[12:11] Cuga.Rajal if avatars in adjacent region have draw distance that's reaching to neighboring region, that might make sense to keep the region alive
[12:11] Cuga.Rajal that might be what Kitely is doing?
[12:11] Arielle Popstar: So thats where dreamgridz smart start would be good
[12:12] Ubit Umarov: kitely as that feature for ages
[12:12] Ubit Umarov: has also
[12:12] Misterblue Waves: I guess the problem is that start/stop is not a feature of the region simulator but a feature of the management system around running the simulator
[12:12] Cuga.Rajal a lot of grids make enhancements to the code from trunk
[12:12] Ubit Umarov: yes
[12:13] Ubit Umarov: current core opensim robust can't even know what regios are up
[12:13] Andrew Hellershanks: hm... That raises an interesting questions. If you TP in to a region that gets auto started would adjacent regions also get started at the same time? That isn't really an issue in Kitely as most(?) of there regions have water around them.
[12:13] Vincent.Sylvester If the regions are mostly empty they'll only consume some memory. Not a whole lot of cpu, starting and stopping however are high in load, if you do fancy maptiles even more so.
[12:14] Arielle Popstar: how about a dynamic varregion?
[12:14] Vincent.Sylvester Speaking from experience when it comes that, you can put up hundreds of regions if they are empty on relatively little hardware, but starting them all at once without making sure you don't overload will blow up
[12:14] Andrew Hellershanks: Vincent, not to mention the added time to start up regions if they have lots of objects, or scripts in them, and it is doing fancy map tiles.
[12:15] Orbert.Tatham I know that kitely's region startups take on the order of minutes
[12:15] Lyr Lobo: bye Selby
[12:15] Selby.Evans bye all
[12:15] Motoko.Karu cya selby
[12:15] Andrew Hellershanks: One option to reduce the start up time related to maps is to generate the map on region shutdown.
[12:15] Andrew Hellershanks: Bye, Selby.
[12:15] Ubit Umarov: cya selby.Evans
[12:16] Tess JL: That depends on how heavy the regions are. I've seen sme that take more than 10 minutes. Mine only take seconds
[12:16] Arielle Popstar: been a while since i tried kitely's solution but thought it was like a minute or less
[12:16] Ubit Umarov: don't think those systems do start nearby regions
[12:16] Ubit Umarov: but not sure
[12:16] Gavin.Hird Is it a full loading, or do they just restore a dormant virtual machine?
[12:16] Vincent.Sylvester Technically you don't need a maptile at all at registration you can generate that later. We have way to set static ones too, so if nothing much changed on the region you can just use a static one instead
[12:16] Ubit Umarov: on dream.. is a full load
[12:17] Gavin.Hird Kitely
[12:17] Ubit Umarov: no idea :)
[12:17] Andrew Hellershanks: A child avatar would exist in a neighbouring region so it would depend on how the auto start feature is written. Does it only start a region upon the presence of a root avatar or would it also start it for child avatars.
[12:17] Ubit Umarov: guess kitely is also abotu getting a cpu
[12:18] Vincent.Sylvester Easiest I think would be just to edit the call to put child agent into the region and have that tell some remote script to start the region first
[12:18] Ubit Umarov: kitely maybe using cloud with dinamic number of virtual machines
[12:18] Vincent.Sylvester You then only have to write something to wait and try again to put child agent there
[12:18] Gavin.Hird I am sure they do Ubit
[12:18] Ubit Umarov: only payed the ones in use
[12:18] Vincent.Sylvester fancy overpriced aws thing
[12:18] Gavin.Hird so I suspect they kick off a suspended vm which will be faster than loading the full simulator in most cases
[12:19] Andrew Hellershanks: That would make sense to do it that way.
[12:20] Tess JL: Would an alternative be to keep all regions online but drop the frame rate for ampty ones to a minimum?
[12:20] Andrew Hellershanks: Joe, you mentioned you were wroking on a viewer. Care to mention which one and where to get it (if it is available for others to use)?
[12:21] Joe Magarac: Oh, sure.
[12:21] Vincent.Sylvester Tess, it depends what's on them. If they contain only a few scripts and objects the most problematic aspect is the memory usage over time. Cpu is only used if it is processing script events or avatars, physics etc. Otherwise they sit pretty much at 0%
[12:22] Jamie.Jordan Have a great week yall
[12:22] Joe Magarac:
[12:22] Vincent.Sylvester Question is more how many people do you expect to roam around and what scripts and object load is going to be there
[12:22] Andrew Hellershanks: Bye, Jamie. See you later.
[12:22] Joe Magarac: A few people have the password to use it.
[12:22] Arielle Popstar: looks good
[12:22] Arielle Popstar: and thats on Rust?>
[12:23] Joe Magarac: Yes.
[12:23] Motoko.Karu no linux huh? :(
[12:23] Joe Magarac: If you want to download it, use "devs"/"thread"
[12:23] Joe Magarac: Yes, Linux and Windows.
[12:23] Motoko.Karu ah. yes.. sorry.. under fixes... didn't see
[12:24] Joe Magarac: Single region, no varregions, avatars are blocks, few menus.
[12:24] Joe Magarac: Most objects work. Email problems to "".
[12:24] Gavin.Hird blocks?
[12:24] Gavin.Hird like lego blocks?
[12:24] Vincent.Sylvester lol
[12:25] Joe Magarac: Avatars are just rectangular solids. Not implemented yet.
[12:25] Misterblue Waves: very good Joe! The avatars in Basil are still ducks :))
[12:25] Andrew Hellershanks: :)
[12:25] Arielle Popstar: will an Opensim version need to be a fork?
[12:26] Joe Magarac: No. Same code.
[12:26] Joe Magarac: It does Opensim now, but varregions are only half implemented.
[12:26] Arielle Popstar: oh better then Lumiya  which doesnt  :)
[12:27] Ubit Umarov: Oh hi joe
[12:27] Joe Magarac: You can reach OSGrid home and Kitely home, which gets me started on debugging OS grid side.
[12:27] Joe Magarac: Hi Ubit.
[12:27] Ubit Umarov: no var regions right?
[12:28] Arielle Popstar: can varregions be made more compatible with standard s/l viewers?
[12:28] Andrew Hellershanks: Joe, Good luck with the work on the viewer. You are taking on a big project.
[12:29] Joe Magarac: It's been two years of work so far. I was really bored during lockdown.
[12:29] Misterblue Waves: the problem with varregions and the SL viewer is that the viewer code has constants for region size everywhere and an implicit assumption that regions are square
[12:30] Ubit Umarov: wel by spec we also only have square regions
[12:30] Joe Magarac: Right. And I have neither assumption. I just don't have the parsers for the modified UDP messages yet.
[12:30] Ubit Umarov: even fear some of our old code also assumes that
[12:30] Andrew Hellershanks: Misterblue, that should be one of the first things to fix when working on a viweer.
[12:30] Ubit Umarov: so let them be square :)
[12:30] Joe Magarac: I
[12:31] Ubit Umarov: well lludp is a growing problem
[12:31] Joe Magarac: It's not squareness that's the  problem. It's what to do when a region has two neighbors along one edge.
[12:31] Andrew Hellershanks: It would be interesting to see non-square regions.
[12:31] Ubit Umarov: ll is removing lludp, some we will not remove..
[12:31] Joe Magarac: The neighbor-finding system assume one neighbor at each edge.
[12:31] Vincent.Sylvester Joe, that's easy, it just crashes :)
[12:31] Ubit Umarov: and a old problem is server side bakes
[12:32] Misterblue Waves: adjacent, different sized regions is hard
[12:32] Gavin.Hird non-square .... was not that mega regions we abandoned?
[12:32] Joe Magarac: lludp is not that bad. I have a compiler that turns message-template.msg into Rust code.
[12:33] Ubit Umarov: in same cases it is better than the ll replacements.. but well...
[12:33] Joe Magarac: I just wish that the open simulator UDP messages that are different had different message numbers.
[12:33] Vincent.Sylvester non square vars were a thing briefly, but yeah not good all round
[12:33] Vincent.Sylvester Mixing regions of same size is best
[12:33] Andrew Hellershanks: Misterblue, it would be worse still if you have a 3x3 with two adjacent 1x1s located at the NW and SW corner.
[12:34] Vincent.Sylvester Least amount of explosions
[12:34] Ubit Umarov: what messages are diferent??
[12:34] Ubit Umarov: only remember one case..
[12:34] Tess JL: I have to go now. Have a great whatever-time-of-day everybody!
[12:34] Ubit Umarov: the minimap dots on large regions
[12:34] Andrew Hellershanks: ok, Tess. Good to see you. Drop by any Tuesday.
[12:35] Ubit Umarov: and the message is the same..  just a tricky rescaling
[12:36] Joe Magarac: Ah, the ones that changed are blacklisted now.
[12:36] Joe Magarac: TeleportFinish and CrossedRegion
[12:36] Ubit Umarov: well ours where as sl
[12:37] Ubit Umarov: but some.. as sl we did more to event
[12:37] Ubit Umarov: eventqueue
[12:37] Ubit Umarov: i meant moved..
[12:37] Joe Magarac: It looks like that's legacy code from some earlier large region system. Aurora?
[12:38] Joe Magarac: I see code for those in Firestorm.
[12:38] Ubit Umarov: seems we do send crossedregion as lludp
[12:39] Ubit Umarov: nahh aurora is related to var regions only
[12:39] Joe Magarac: CrossedRegion with
[12:39] Joe Magarac: {   RegionSizeX     U32     }   // Open Simulator only
        {   RegionSizeY     U32     }   // Open Simulator only
[12:39] Ubit Umarov: thatr is var region support yes
[12:40] Joe Magarac: Right, and that is a different lludp message format used in Open Simulator.
[12:40] Ubit Umarov: but those are not on the packet let me check
[12:41] Joe Magarac: Firestorm can parse them but I don't know if they are still used.
[12:41] Ubit Umarov: public sealed class RegionDataBlock : PacketBlock
            public uint SimIP;
            public ushort SimPort;
            public ulong RegionHandle;
            public byte[] SeedCapability;
[12:43] Joe Magarac: Firestorm expects two more fields if in OpenSimulator mode.
[12:44] Gavin.Hird CrossedRegion is not really used for anything. Even my old version 2.x code does nothing with it
[12:44] Ubit Umarov: yeah guess we do not use it either..
[12:44] Gavin.Hird there is no functioning code for it at all in the viewer, only definitions
[12:45] Ubit Umarov: oops we do
[12:45] Ubit Umarov: lol
[12:45] Ubit Umarov: if (m_eqModule != null)
                    neighbourRegion.RegionHandle, pos, vel2 /* agent.Velocity */,
                    endpoint, capsPath, agentUUID, agent.ControllingClient.SessionId,
                    neighbourRegion.RegionSizeX, neighbourRegion.RegionSizeY);
                m_log.ErrorFormat("{0} Using old CrossRegion packet. Varregion will not work!!", LogHeader);
                agent.ControllingClient.CrossRegion(neighbourRegion.RegionHandle, pos, agent.Velocity,
[12:45] Ubit Umarov: as i said.. long moved to http eventcap
[12:46] Orbert.Tatham cd ..
[12:46] Orbert.Tatham oops
[12:46] Ubit Umarov: lludp variant never got varregions
[12:46] Andrew Hellershanks: Orbert, access denied
[12:46] Motoko.Karu RL calls... have a good week everyone
[12:46] Lyr Lobo: bye Motoko!
[12:46] Arielle Popstar: Tc Moto
[12:46] Andrew Hellershanks: ok, Motoko. See you again next week.
[12:47] Joe Magarac: OK, so you don't send those. Firestorm still has code for them.
[12:47] Joe Magarac:
[12:47] Ubit Umarov: we we do..
[12:47] Ubit Umarov: but via http
[12:47] Gavin.Hird FS has a lot of old code that could be purged
[12:47] Joe Magarac: Search for "region_size_x".
[12:48] Joe Magarac: OK. So that's dead code and I don't need to implement it. Good.
[12:48] Andrew Hellershanks: Joes, Not useful unless FS does something with that information.
[12:48] Joe Magarac: OK. That's why I come to these meetings. I can read the code but do not know all the lore.
[12:49] Ubit Umarov: teleportfinish
[12:49] Ubit Umarov: we also send via cap now
[12:49] Vincent.Sylvester You can poke on IRC or the mailing list too if you have questions
[12:49] Gavin.Hird there is code for teleportfinish, yes
[12:50] Arielle Popstar: there is an opensim-dev irc  though dont know if it is very active anymore
[12:50] Ubit Umarov:  OK, send TPFinish to the client, so that it starts the process of contacting the destination region
            if (m_eqModule != null)
                m_eqModule.TeleportFinishEvent(destinationHandle, 13, endPoint, 0, teleportFlags, capsPath, spUUID,
                            finalDestination.RegionSizeX, finalDestination.RegionSizeY);
                sp.ControllingClient.SendRegionTeleport(destinationHandle, 13, endPoint, 4,
                                                            teleportFlags, capsPath);
[12:51] Joe Magarac: OK. so that's all over HTTP capabilities. Thanks.
[12:51] Joe Magarac: Bye for now. Thanks all.
[12:51] Gavin.Hird I have to go too
[12:51] Andrew Hellershanks: ok, Joe. Good to see you. Come back any time.
[12:51] Gavin.Hird cheers
[12:52] Lyr Lobo: take care
[12:52] Lyr Lobo: heading out *smiles*  have a great week
[12:52] Andrew Hellershanks: I should also get going. Almost two hour long meeting today. I'll close it out as I usually do.
[12:52] Jagga Meredith: great meeting - love when it gets into the weeds
[12:52] Andrew Hellershanks: That will do it for another week. Thank you all for coming. See you again next week.
[12:52] Cuga.Rajal TC Lyr
[12:52] Cuga.Rajal Take care all, I should get going as well
Personal tools
About This Wiki