Chat log from the meeting on 2013-11-26

From OpenSimulator

(Difference between revisions)
Jump to: navigation, search
(Created page with "<pre style="white-space: pre-wrap; white-space: -moz-pre-wrap; white-space: -pre-wrap; white-space: -o-pre-wrap; word-wrap: break-word"> [11:02] Richardus Raymaker: maby after ...")
 
m
 
Line 1: Line 1:
<pre style="white-space: pre-wrap;
+
<pre>
white-space: -moz-pre-wrap;
+
white-space: -pre-wrap;
+
white-space: -o-pre-wrap;
+
word-wrap: break-word">
+
 
[11:02] Richardus Raymaker: maby after the meeting
 
[11:02] Richardus Raymaker: maby after the meeting
 
[11:02] Mircea Kitsune: ok
 
[11:02] Mircea Kitsune: ok
Line 350: Line 346:
 
[12:01] Andrew Hellershanks: Bye, justin
 
[12:01] Andrew Hellershanks: Bye, justin
 
</pre>
 
</pre>
 +
 +
 +
[[Category:Office Hour Logs]]

Latest revision as of 20:57, 24 October 2015

[11:02] Richardus Raymaker: maby after the meeting
[11:02] Mircea Kitsune: ok
[11:02] Andrew Hellershanks: Richardus, ah, ok
[11:02] Richardus Raymaker: unles sit gives problem s with opensim ?
[11:02] Richardus Raymaker: hi justin
[11:02] Tiffany Magic: Hi Justin
[11:02] Freaky.Tech @freakytech.spdns.de:8002: hi Justin
[11:02] Mircea Kitsune: Hi Justin
[11:02] Justin Clark-Casey: hi folks
[11:02] Andrew Hellershanks: Mircea, depends who is here. Sometimes a viewer dev is at these meetings
[11:02] Andrew Hellershanks: Hey, Justin
[11:03] Andrew Hellershanks: Justin, I found one more bug this past week.
[11:03] Justin Clark-Casey: oh goody
[11:03] Andrew Hellershanks: yea, I thought you would like to hear that. ;)
[11:03] Selea.Core @selea-core.com:8002: hi everyone :)
[11:03] Justin Clark-Casey: hi selea
[11:03] Andrew Hellershanks: Justin, I plan to file them first, then I can see if I can fix any of them.
[11:03] Tiffany Magic: Hi Selea
[11:03] Nebadon Izumi is online.
[11:04] Selea.Core @selea-core.com:8002: hi Justin and Tiffany
[11:04] Justin Clark-Casey: andrew: ok, would be great if you could note if you are looking at them
[11:04] Justin Clark-Casey: so there's no duplicate work
[11:04] Andrew Hellershanks: Justin, another day or two and I'll be finished the script I'm working on. After that I will take a break and start dealing with bug reports.
[11:04] Freaky.Tech @freakytech.spdns.de:8002: hi Neb
[11:04] Selea.Core @selea-core.com:8002: is this a dev's meeting?
[11:04] Nebadon Izumi: hello
[11:04] Andrew Hellershanks: Justin, sure.
[11:04] Justin Clark-Casey: selea: yes, though anyone can attend
[11:04] Justin Clark-Casey: topic is opensimulator issues though
[11:04] Selea.Core @selea-core.com:8002: okay then, thank you
[11:04] Richardus Raymaker: hi nebadon
[11:05] Justin Clark-Casey: hi neb
[11:05] Selea.Core @selea-core.com:8002: hi Nebadon
[11:05] Justin Clark-Casey: I am white on purpose btw - I had a little accident
[11:05] Freaky.Tech @freakytech.spdns.de:8002: you got wrapped in RL?
[11:05] Andrew Hellershanks: I have to find where I kept my notes on the bugs I've found. Too many pieces of paper on my desk
[11:05] Vivian Klees: milk does a body good
[11:05] Tiffany Magic: I thought you just went back to wearing long johns. *shrug*
[11:05] Justin Clark-Casey: lol
[11:05] Mircea Kitsune: Back. Odd crash of sorts
[11:06] Justin Clark-Casey: appearance got lost but I needed to see myself to look at some bugs, so I just started creating blank clothing
[11:06] Justin Clark-Casey: and body bits
[11:06] Mircea Kitsune: Ah, yes... there was something I was wanting to ask at the next office hour...
[11:07] Justin Clark-Casey: mircea: I still need to look at your patches btw, sorry, got caught up with other stuff this week
[11:07] Mircea Kitsune: Is the change hard to add here too? I'd imagine it's pretty easy
[11:07] Mircea Kitsune: I heard that the Main grid got an improvement, which properly sorts objects to be loaded based on distance from avatar. So instead of loading primitives on sims randomly, closest to avatar are loaded first. Opensim still seems to load them at random though
[11:08] Mircea Kitsune: Justin: That's ok. Hope you get to look at them soon, but no hurry if you can't
[11:09] Justin Clark-Casey: I've never looked at this area of code. There are a number of options for it in OpenSim.ini but I don't know how well they work
[11:09] Mircea Kitsune: ok. I think this should probably be done automatically if possible. An opensim.ini option could be good though, if current behavior would still have any use
[11:09] Andrew Hellershanks: Justin, has the patch for attachments to group notices been applied to git master yet?
[11:09] Justin Clark-Casey: possibly UpdatePrioritizationScehem in [InterestManagement]
[11:10] Justin Clark-Casey: don't know if that also governs the first appearance of objects
[11:10] Justin Clark-Casey: andrew: yes,. some time ago afair
[11:10] Richardus Raymaker: Mirea, you mean intrest list viewer
[11:10] Richardus Raymaker: and code
[11:10] Andrew Hellershanks: ok, I haven't been checking git master while working on my project.
[11:10] Justin Clark-Casey: Fri Aug 2
[11:10] Mircea Kitsune: http://www.youtube.com/watch?v=W7AC1RojXOk What I'm talking about BTW
[11:10] Simulator Version v0.5 shouts: OpenSim 0.8.0 Dev          f94d07f: 2013-11-16 03:02:57 +0000 (Unix/Mono)
[11:11] Andrew Hellershanks: I haven't rejoined mailing list since I was taken off it due to bounced mail.
[11:11] Richardus Raymaker: I know what you mean. but sofar i know opensim dont have it
[11:11] Mircea Kitsune nods
[11:13] Justin Clark-Casey: so over the last week, I've finally had a change to look at the camera problems with sitting on non-root prims
[11:14] Justin Clark-Casey: I think I have changes which now fix this (unless neb has found any new problems). However, a few other kinks it would be good to work out before the code goes in master
[11:14] Nebadon Izumi: no I have not found anymore issues
[11:14] Freaky.Tech @freakytech.spdns.de:8002: when does someone take a look at the HGFriends stuff?
[11:14] Mircea Kitsune: Glad to hear such a major problem was fixed
[11:14] Nebadon Izumi: just that one stand up issue
[11:14] Richardus Raymaker: nice
[11:14] Nebadon Izumi: I did not get a chance to test vehcile jitters yet
[11:14] Justin Clark-Casey: Freaky: Have to rely on Crista to do that, and she is very busy with classes I believe
[11:15] Nebadon Izumi: what is the problem with HG friends?
[11:15] Justin Clark-Casey: unless I get to it myself, but that's probably some time off
[11:15] Freaky.Tech @freakytech.spdns.de:8002: When you teleport to hyper grid at first It works
[11:15] Selea.Core @selea-core.com:8002: I'm baby sitting at the moment and out of 4 little ones, 2 of them require my attention ... auntie Selea has to go play cubes ... so thank you for having me for a wee bit ... take care everyone :)
[11:15] Freaky.Tech @freakytech.spdns.de:8002: but when you teleport to another region it does not
[11:15] Mircea Kitsune: Later
[11:15] Justin Clark-Casey: bye selea
[11:16] Tiffany Magic: Is anyone working on the vehicle sim crossings since it is apparent that that code will not be share?
[11:16] Freaky.Tech @freakytech.spdns.de:8002: and then you do not get that away friendship made
[11:16] Tiffany Magic: *will not be shared
[11:17] Richardus Raymaker: i have not got time to setup things. otyherwise am sure bumping in V2 problems.
[11:17] Richardus Raymaker: if there are
[11:17] Mircea Kitsune: Crossing sims with vehicles doesn't work yet?
[11:17] Richardus Raymaker: Tiffany, var regions can be the answhere
[11:17] Tiffany Magic: No
[11:17] Andrew Hellershanks: AFAIK, Melanie plans to provide code for veciclde crossings but hasn't yet.
[11:17] Mircea Kitsune: Ahh... I hope that will be fixed sometime
[11:17] Freaky.Tech @freakytech.spdns.de:8002: any object can cross sims as far as I have noticed
[11:17] Andrew Hellershanks: s/veciclde/vehicle/
[11:18] Tiffany Magic: Melanie has been saying that for a year and a half.
[11:18] Mircea Kitsune: Freaky: That's why I'm surprised
[11:18] Mircea Kitsune: IIRC vehicles are pretty much like any other physical object, only with some special physics and controlled by an avatar
[11:18] Andrew Hellershanks: yes. The initial delay was waiting for the code to stabilize and a release made.
[11:19] Tiffany Magic: Right now, you can't ride a vehicle across a simline.
[11:19] Richardus Raymaker: Well. readed some discussion days ago in chat. not sure whats going on. but not suprissed if var regions work sooner then verhicle crossing. and maby there's reason todo verhicle crossing after var to make it compatible
[11:19] Kayaker Magic: I have lots of non-physical vehicles I want to be able to sail across a SIM border, so not just physics
[11:19] Kayaker Magic: My vehicles will all be using llSetKeyframedMotion to move soon
[11:20] Richardus Raymaker: Kayaker. use mega region ?
[11:20] Tiffany Magic: Var's are only a solution if the same person owns the area. But, doesn't help for sailing into other water sims.
[11:20] Justin Clark-Casey: what happens when you do try to cross a vehicle over sims?
[11:20] Richardus Raymaker: Yup. but 2 different opensim owners with different places. not sure if you want to cross a simborder then
[11:20] Tiffany Magic: You crash and the boat/etc, poofs
[11:20] Freaky.Tech @freakytech.spdns.de:8002: I had a few runaway objects they mostly got lost
[11:21] Justin Clark-Casey: crash as in viewer terminates?
[11:21] Tiffany Magic: Yes
[11:21] Mircea Kitsune: Sounds bad. Do any of the sims involved crash too?
[11:21] Tiffany Magic: Or, other times, you just get thrown off and the boat poofs
[11:21] Kayaker Magic: All kinds of problems, lost prims, avatars stuck and can move.
[11:21] Kayaker Magic: *cannot
[11:22] Vivian Klees: eh that still happens in SL
[11:22] Freaky.Tech @freakytech.spdns.de:8002: or it puts you in a sim border and keeps you stuck there
[11:22] Kayaker Magic: But it happens EVERY time you cross on a vehicle here
[11:22] Justin Clark-Casey: vehicle crossing is going to be a very complex undertaking
[11:23] Freaky.Tech @freakytech.spdns.de:8002: For vehicle crossing it must be definitely smoother than current avatar sim crossing
[11:23] Richardus Raymaker: And lots of lag problems to.. voteing for var then
[11:23] Freaky.Tech @freakytech.spdns.de:8002: looks odd when a car or a boat jumps forth and back
[11:23] Kayaker Magic: Yes, even on the grids where it is possible, (Avaination, inWorldz, SL) there are frequent glitches on crossings
[11:24] Justin Clark-Casey: you're having to remove and insert objects into different scenes and physics simulations running on different threads, which I have to think is extraordinarily complex to do smoothly
[11:24] Kayaker Magic: Var will not fix existing grids with lots of 256 regions that want to connect together
[11:24] Justin Clark-Casey: ultimately, variable regions may be a better solutions but that has its own issues atm
[11:24] Freaky.Tech @freakytech.spdns.de:8002: I never assumed it to be easy
[11:24] Tiffany Magic: Again.... vars are only good for use on land with one owner.
[11:24] Mircea Kitsune: What are var regions exactly?
[11:25] Justin Clark-Casey: again?
[11:25] Vivian Klees: variable size regions
[11:25] Richardus Raymaker: Justin, var still somves lots of the need for large regions without border lagg.
[11:25] Mircea Kitsune: Oh. Like megaregions that can change scale anytime? Sounds like a good feature
[11:25] Justin Clark-Casey: they are like var regions but require explicit viewer support
[11:25] Justin Clark-Casey: sorry, like mega
[11:25] Richardus Raymaker: Tiffany. most own the var then.. there are exceptions.
[11:25] Justin Clark-Casey: this fixes some of the issues with mega but means you need to have a viewer which explicitly supports them
[11:26] Richardus Raymaker: But crossing between a sim in europe and sim in USA. that can be the worse simborder crossing i hav eseen :O
[11:26] Freaky.Tech @freakytech.spdns.de:8002: viewers have to understand that a region is not anymore a fixed size
[11:26] Mircea Kitsune: BTW. I was wondering if Opensim has or will have a complete megaregion system. As in, one that makes a single region larger, without having to fakely combine multiple regions in one. So still a single database, a single grid position specified for the sim, etc
[11:26] Justin Clark-Casey: it will be extraordinarily difficult to get a smooth crossing like taht when you have significant lag between regions
[11:26] Richardus Raymaker: well. useing singularity and firestorm most.. :O
[11:26] Dahlia Trimble is online.
[11:26] Justin Clark-Casey: significant ping time, I should say
[11:26] Richardus Raymaker: only want 64bit FS linux
[11:27] Justin Clark-Casey: mircea: that's what var regions are like
[11:27] Mircea Kitsune: Awesome :)
[11:27] Andrew Hellershanks: I think a hesitation at sim border with vehicles might be acceptable.
[11:27] Richardus Raymaker: Mircea, we have mega regions that do what you say. but the new upcomeing vars are much better
[11:27] Richardus Raymaker: also var give lower system load sofar i hav eseen
[11:28] Mircea Kitsune: A custom viewer could even let you see the correct position of objects. Where normal regions range between -256 and 256, one could go between -512 and 512, with all scripts and viewer parts understanding the new ranges
[11:28] Richardus Raymaker: compared to mega
[11:28] Mircea Kitsune: Nice
[11:28] Andrew Hellershanks: aren't vars a special class of mega (or is it the other way around)?
[11:28] Justin Clark-Casey: mircea: that's the kind of explicit viewer support needed I believe
[11:28] Dahlia Trimble: hi
[11:28] Freaky.Tech @freakytech.spdns.de:8002: hi
[11:28] Richardus Raymaker: hi dahlia
[11:28] Mircea Kitsune: hi
[11:28] Justin Clark-Casey: hello dahlia
[11:28] Nebadon Izumi: vars and megas are completely different
[11:28] Andrew Hellershanks: Why would one be lower system load than the other
[11:28] Nebadon Izumi: vars are 1 single region
[11:29] Nebadon Izumi: megas are many regions
[11:29] Nebadon Izumi: megas use more resources
[11:29] Andrew Hellershanks: Won't vars eventually replace megas?
[11:29] Justin Clark-Casey: well, at low scales it come be lower since you are only running a single scene loop rather than many
[11:29] Richardus Raymaker: Mega's give tou problem with all the child connections
[11:29] Justin Clark-Casey: but at the same time, I'm interested in whether putting multiple regions worth of content on a single scene loop will overload that loop
[11:29] Mircea Kitsune: Will normal viewers be able to support var's too? Without modding the simulator size range limits
[11:29] Freaky.Tech @freakytech.spdns.de:8002: I guess it is mostly because that larger region has a single communication point
[11:29] Dahlia Trimble: megas work with any viewer
[11:29] Freaky.Tech @freakytech.spdns.de:8002: megas seem to have multiple
[11:30] Justin Clark-Casey: mircea: no, that's essentially why we have megas in the first place
[11:30] Freaky.Tech @freakytech.spdns.de:8002: to be compatible with any viewer
[11:30] Mircea Kitsune: Ok, makes sense
[11:30] Andrew Hellershanks: There could be a number of scripts that might fail when coordinates can exceed 256m
[11:30] Justin Clark-Casey: freaky: yeah, that's an issue too - the single var region is served by a single inbound and outbound udp connection
[11:30] Justin Clark-Casey: sorry, I mean thread or port, really
[11:30] Freaky.Tech @freakytech.spdns.de:8002: that then may serve load issues
[11:30] Richardus Raymaker: Andrew. i need to make everything from scratch. only maby nmot start on var :(
[11:31] Mircea Kitsune imagines var regions might be the first thing to break OpenSim with the default SL viewer. Though after Linden cut loginuri support off, it's discussable how much that matters any more
[11:31] Richardus Raymaker: only saw a mega pushed the linux system load to werid levels. and the same size var keeps load nice at normal levels
[11:31] Justin Clark-Casey: cutting loginuri support made the tpvs a bit more prominent, I think
[11:31] Dahlia Trimble: does anyone use Linden default viewer?
[11:31] Justin Clark-Casey: I still use ll 3.3.4 in places
[11:31] Andrew Hellershanks: Dahlia, only for SL.
[11:31] Richardus Raymaker: i use it only in sl beta
[11:31] Mircea Kitsune: Yeah. Talking about Opensim viewers possibly having to become different from original SL viewer in the future to still work
[11:32] Dahlia Trimble: I used a hacked one to implement materials
[11:32] Freaky.Tech @freakytech.spdns.de:8002: mircea they already started to do so
[11:32] Richardus Raymaker: FS beta have materials
[11:32] Freaky.Tech @freakytech.spdns.de:8002: look at Firestorm
[11:32] Mircea Kitsune: yeah
[11:32] Richardus Raymaker: the only dojnt have a 64bit linux viewer
[11:32] Dahlia Trimble: Richardus, it didnt when I was working on it
[11:32] Freaky.Tech @freakytech.spdns.de:8002: but for licensing issues
[11:33] Mircea Kitsune runs my viewer from Lemmy's repository, who maintains it for openSUSE
[11:33] Richardus Raymaker: hmm, ok. i hav enot tested materials in OS
[11:33] Richardus Raymaker: think singularity have materials to
[11:33] Dahlia Trimble: please do
[11:33] Mircea Kitsune: I have. They look fine, but I don't know if they're presisted on the sim yet
[11:33] Mircea Kitsune: In the database storage
[11:34] Dahlia Trimble: they persist
[11:34] Andrew Hellershanks: Singularity alpha has a reference to materials on the texture tab
[11:34] Mircea Kitsune: Nice
[11:34] Dahlia Trimble: if you have it enabled
[11:34] Richardus Raymaker: Andrew, your running behind
[11:34] Andrew Hellershanks: I need to read up on that at some point.
[11:34] Richardus Raymaker: singularity 1.8.4 is out
[11:35] Andrew Hellershanks: Richardus, I've been running Alpha version for a while now instead of releases.
[11:35] Richardus Raymaker: still need to hard config viewers in xml files
[11:35] Andrew Hellershanks: Is anyone here good with quaternion math?
[11:36] Mircea Kitsune: I don't even understand what it is yet :p
[11:36] Freaky.Tech @freakytech.spdns.de:8002: they are a bit pesky when you try to understand them at first try
[11:36] Kayaker Magic: Anyone else? Besides whom?
[11:36] Freaky.Tech @freakytech.spdns.de:8002: they do not share the angle lock issue of Eulerian angles
[11:36] Andrew Hellershanks: I was working on a PHP script to rotate a region. I thought the math was correct but the results weren't. I'll have to run some more experiments.
[11:36] Dahlia Trimble: gimbal lock
[11:37] Freaky.Tech @freakytech.spdns.de:8002: the term you find within flight simulators ;)
[11:37] Andrew Hellershanks: Eulerian -> Euclidean
[11:37] Freaky.Tech @freakytech.spdns.de:8002: at least I made it understandable on the terms
[11:38] Richardus Raymaker: i not understand the whole rotation stuff compleet. euler ok. but thats it
[11:38] Freaky.Tech @freakytech.spdns.de:8002: Quaternions do resize and rotation
[11:38] Dahlia Trimble: just rotation
[11:38] Freaky.Tech @freakytech.spdns.de:8002: that is why you need unity quaternions for rotation only
[11:40] Justin Clark-Casey: andrew: did you ever look at the rexx tool to do region rotations?
[11:40] Andrew Hellershanks: From what I've read I need to multiply by <0, 0, .707, .707> to do 90 degree rotation about Y
[11:40] Freaky.Tech @freakytech.spdns.de:8002: and the whole thing has a length of 1 as a quaternion
[11:40] Freaky.Tech @freakytech.spdns.de:8002: otherwise resizing happens
[11:40] Andrew Hellershanks: justin, Yes. I have the 0.1 and 0.2 versions but neither runs with the version of rexx I now have.
[11:41] Andrew Hellershanks: Justin, I was looking at those routines to work out the changes needed re: moving objects before rotating them.
[11:41] Freaky.Tech @freakytech.spdns.de:8002: that is body-local coordinates vs. word coordinates
[11:42] Andrew Hellershanks: I'll test moving only then worry about rotation to see if I have a problem on the moving side or just on the rotation side.
[11:42] Justin Clark-Casey: I would have thought you want to translate the co-ordinate by the rotation, not move before rotate
[11:42] Andrew Hellershanks: The PHP script is pretty fast
[11:42] Andrew Hellershanks: There are two different parameters in the object XML files of an OAR.
[11:43] Andrew Hellershanks: Group position has the location of the root prim, then there is the parameter for object rotation.
[11:43] Dahlia Trimble: my guess is you would want to rotate the root prims, then you would want to rotate the position around some point like 128,128
[11:44] Justin Clark-Casey: ah yes, I guess it depends what you mean by rotate around 90 degrees
[11:44] Justin Clark-Casey: if your doing that to the whole region in place
[11:44] Dahlia Trimble: probably wouldnt need to change the child prims at all
[11:44] Andrew Hellershanks: Like the zoe programs, I'm only doing 90 degree rotations. Moving things is simple. Just minor changes to the group x/y coordinate.
[11:44] Justin Clark-Casey: yeah, child prims are specified as offsets to the root prim
[11:45] Andrew Hellershanks: No, child prims are supposed to be relative to the root but you do need to update the relative position of the child prims.
[11:45] Freaky.Tech @freakytech.spdns.de:8002: use vector math on x/y and make 0,0 the center of the region
[11:45] Justin Clark-Casey: yes, relative position would need updating
[11:45] Andrew Hellershanks: Multiplying two quaternions is using vector math.
[11:45] Freaky.Tech @freakytech.spdns.de:8002: I know
[11:45] Dahlia Trimble: subtract 128,128 from the position, rotate it, add 128,128 back
[11:45] Freaky.Tech @freakytech.spdns.de:8002: But you can rotate the vectors from region center much more easily
[11:46] Andrew Hellershanks: Dahlia, I was tempted to convert quaternion to x/y/z degrees, update degrees, and convert it back to quaternion. :)
[11:46] Freaky.Tech @freakytech.spdns.de:8002: and you can actually translate local positions in world positions and back and have probably the same good base for rotation
[11:46] Dahlia Trimble: Andrew, no need to
[11:46] Freaky.Tech @freakytech.spdns.de:8002: Just use the quaternion for all rotations
[11:47] Dahlia Trimble: just make a 90 degree quaternion and use it
[11:47] Freaky.Tech @freakytech.spdns.de:8002: quaternions can be rotated with quaternions
[11:47] Andrew Hellershanks: Dahlia, If the Q math I'm doing is correct (it is AFAICT), then the problem is with updating position of child prims.
[11:48] Dahlia Trimble: Andrew, you should not need to change child prims at all
[11:48] Freaky.Tech @freakytech.spdns.de:8002: I would try whether converting them to world position and back would help
[11:48] Andrew Hellershanks: Items on the face of a triangular object are away from the face. I haven't figured out why yet.]
[11:48] Freaky.Tech @freakytech.spdns.de:8002: and rotate the world position in between
[11:48] Dahlia Trimble: child prims inherit parent rotation
[11:48] Andrew Hellershanks: Dahlia, the X/Y relative offset changes when you rotate the prim 90 degrees.
[11:49] Dahlia Trimble: child prims inherit parent rotation
[11:49] Dahlia Trimble: if you rotate the root, the child prims will rotate with it'
[11:49] Freaky.Tech @freakytech.spdns.de:8002: but from what I saw a child prim actually gets a rotated local position as well when trying to get doors script right
[11:49] Andrew Hellershanks: An object with +X rotated clockwise 90 degrees needs -Y offset instead.
[11:49] Andrew Hellershanks: That is what was done in zoe
[11:50] Dahlia Trimble: well good luck :)
[11:50] Andrew Hellershanks: Dahlia, I'll run some tests where I don't touch the child prim data and see what happens.
[11:50] Dahlia Trimble: if you change the child prims you end up with a mess
[11:51] Andrew Hellershanks: Perhaps zoe was too old but I would not have expected any changes in how data is handled for child prims.
[11:51] Justin Clark-Casey: andrew: If you look at SOG.UpdateRootRotation() you will see what opensim does when the root is rotated
[11:51] Andrew Hellershanks: Justin, ok. Thanks. I'll take a look at that later.
[11:51] Justin Clark-Casey: that updates the child prim positions
[11:52] Andrew Hellershanks: ok, so if it does it, I don't need to.
[11:52] Justin Clark-Casey: in opensim code, you can multiple the original local position by the quaternion rotation to get a new position
[11:52] Andrew Hellershanks: Its possible I misread the rexx code in zoe but I didn't think I did.
[11:52] Justin Clark-Casey: no, you do need to since in your oar it can't do that
[11:53] Richardus Raymaker: andrew, I have a question later for you about something else
[11:53] Freaky.Tech @freakytech.spdns.de:8002: if oar needs rotation of child's position explicitly, then simply convert it to world position and rotate that
[11:53] Justin Clark-Casey: in libomv code, you can see how this is done in Vector3.operator *
[11:54] Andrew Hellershanks: Justin, If I update the rotation/position of root, wouldn't it still apply the new data to child prims when the object is rezzed which means I wouldn't need to update the child data?
[11:54] Freaky.Tech @freakytech.spdns.de:8002: and then back to local position
[11:54] Justin Clark-Casey: andrew: the child positions will have moved
[11:54] Andrew Hellershanks: Justin, ok. I did look at some math sites about q's and how the multiply math is done so it should be ok.
[11:54] Justin Clark-Casey: hmm
[11:54] Andrew Hellershanks: justin, but the child positions are still relative to the root.
[11:54] Dahlia Trimble: if I rotate a linkset, the child prim offsets and rotations dont change in SOP
[11:55] Justin Clark-Casey: ah of course, I'm being dumb
[11:55] Justin Clark-Casey: the sog code applies when you rotate a root prim separately from all othe rprims in world
[11:55] Dahlia Trimble: correct
[11:55] Justin Clark-Casey: when you do need to change all the child positions because you're changing the object itself
[11:55] Andrew Hellershanks: ok, sounds like my problem was trying to apply updates to the child prims. I'll run more tests later and see how it goes.
[11:55] Freaky.Tech @freakytech.spdns.de:8002: then there is no need to apply updates to child prims.
[11:55] Justin Clark-Casey: yeah, better to listen to dahila, she knows this stuff
[11:55] Justin Clark-Casey: I should shut up really :)
[11:56] Dahlia Trimble throws a trout at justin
[11:56] Richardus Raymaker: i lost dahlia 50 linux up :O someday i get that rotation stuff
[11:56] Richardus Raymaker: uhmm ^ lines
[11:56] Freaky.Tech @freakytech.spdns.de:8002: Rich start with matrix rotation first for understanding not with quaternions
[11:56] Mircea Kitsune: Before people leave, I wanted to ask something else please. It's about a viewer matter if that's ok.
[11:57] Justin Clark-Casey: sure
[11:57] Richardus Raymaker: Freaky. i use euler rotations
[11:57] Richardus Raymaker: thats at least logic between 0-360
[11:57] Freaky.Tech @freakytech.spdns.de:8002: you can actually extract Euler angles from matrices and quaternions
[11:57] Justin Clark-Casey: it's easier to vizualization. whereas quaternions don't appear to be easily vizualized at all
[11:57] Dahlia Trimble suspects RIchardus enjoys pain
[11:57] Freaky.Tech @freakytech.spdns.de:8002: and make ones out of it
[11:57] Justin Clark-Casey: er, easier to visualize
[11:58] Mircea Kitsune: If anyone here uses Linux and the free video drivers (especially Mesa + OSS Radeon driver), I'd like to talk sometime later. All viewers have big shader issues with the free drivers in Linux, and shaders even need modding to work. They're also glitchy
[11:58] Freaky.Tech @freakytech.spdns.de:8002: quaternions are four-dimensional that is what makes them hard
[11:58] Justin Clark-Casey: there's a bunhc of lsl code to convert between eulers and quaternions after all, which is all in opensim
[11:58] Mircea Kitsune: Not now, but maybe talk about it later. Would like to know more
[11:58] Andrew Hellershanks uses nVidia
[11:58] Richardus Raymaker: FRee video under linux ? no way the qualioty is terrible for virtual worlds
[11:58] Mircea Kitsune: Not planning on getting a video card change way too soon
[11:58] Richardus Raymaker: free video drivers. i dont like them
[11:59] Mircea Kitsune: The free driver works well enough really... just some shader compiler issues
[11:59] Freaky.Tech @freakytech.spdns.de:8002: nouveau-gallium is partly doing well on 3D
[11:59] Mircea Kitsune: Anyway, if anyone uses them, do let me know
[11:59] Andrew Hellershanks: I use the drivers provided by the manufacturer of the video card.
[11:59] Freaky.Tech @freakytech.spdns.de:8002: especially the texture amount is pretty low
[11:59] Dahlia Trimble: free as in nouveau?
[11:59] Richardus Raymaker: Mircea, the free drivers costs me plenty of framespersecond
[11:59] Freaky.Tech @freakytech.spdns.de:8002: it is the name of the open source variant for nvidia chips
[11:59] Mircea Kitsune finds performance nearly the same in free and proprietary drivers, with latest versions. Just some shader issues with Mesa's GLSL compiler
[12:00] Richardus Raymaker: nouveau is a bit on the i hate it list. gives me always problems
[12:00] Freaky.Tech @freakytech.spdns.de:8002: texture amount and shaders is a problem with nouveau still
[12:00] Mircea Kitsune: On Linux, the proprietary drivers gave me most problems :P
[12:00] Richardus Raymaker: Mircea, budget graphics card ?
[12:00] Mircea Kitsune: Nah, pretty modern. Radeon HD 6870
[12:00] Andrew Hellershanks: Richardus, I think I had issues with it which is why I don't generally use it. Haven't tried it again for a long time.
[12:01] Freaky.Tech @freakytech.spdns.de:8002: nvidia proprietary drivers give you a hard time on multi screen too
[12:01] Freaky.Tech @freakytech.spdns.de:8002: but nouveau got that at least right for now
[12:01] Andrew Hellershanks: Freaky, send me a second screen and I'll be happy to test it. :)
[12:01] Justin Clark-Casey: ok, I need to get going. Thanks for the conversation folks, see you around
[12:01] Richardus Raymaker: bye justin
[12:01] Nebadon Izumi: i have zero issues with multiple screens using nvidia proprietary drivers myself
[12:01] Freaky.Tech @freakytech.spdns.de:8002: maybe you got an HDMI connector ;)
[12:01] Andrew Hellershanks: Who works on the particle system?
[12:01] Freaky.Tech @freakytech.spdns.de:8002: bye Justin
[12:01] Mircea Kitsune assumes no one actively uses the free drivers with SL. Guess I'll be one to test those mostly then :)
[12:01] Nebadon Izumi: see you Justin
[12:01] Strachan Ofarrel: Bye
[12:01] Mircea Kitsune: Later, take care
[12:01] Dahlia Trimble: bye jcc
[12:01] Justin Clark-Casey waves
[12:01] Andrew Hellershanks: Bye, justin
Personal tools
General
About This Wiki