User:Andrew Linden/Office Hours/2009 10 20

Transcript

[10:56] Sahkolihaa Contepomi: Don't look what? oO
[10:56] Questar Utu: 'Nude'.
[10:56] Sebastean Steamweaver: He said nude, I think
[10:57] Moon Metty: hi Ardy
[10:57] xstorm Radek: when i first TP in humans are all nude
[10:57] Ardy Lay: Hi
[10:57] Sahkolihaa Contepomi: Heh...I wouldn't do that in real life, so why in second life? :P
[10:57] xstorm Radek: lol
[10:57] Sahkolihaa Contepomi: Haha.
[10:57] Questar Utu: He meant while the attachments load. :p
[10:57] Questar Utu: We have fur coats.
[10:57] Sahkolihaa Contepomi: Yeah, fur is much better. ^^
[10:57] Sebastean Steamweaver: I concur :>
[10:58] xstorm Radek: no i can see furys just fine
[10:58] xstorm Radek: just women and men are nude to me
[10:58] Sahkolihaa Contepomi: o_o
[10:58] xstorm Radek: and i must say some skins have odd tatoos
[10:58] Ardy Lay: Hehe
[10:58] xstorm Radek: in odd places
[10:58] Ardy Lay: "SAMPLE"
[10:59] Sebastean Steamweaver drags down his draw distance.
[10:59] xstorm Radek: im not taking screen shots ill get slap
[10:59] Moon Metty: hey rex, you're early!
[10:59] Rex Cronon: hi moon
[10:59] Moon Metty: :)
[11:00] Rex Cronon: i am living proof that miracles do happen:)
[11:00] Moon Metty: indeed
[11:00] xstorm Radek: lol
[11:00] Welcome to Linden office hours
[11:00] xstorm Radek: eeeeppp
[11:00] Moon Metty looks around for Lindens
[11:00] xstorm Radek: it talks
[11:00] Sahkolihaa Contepomi: Ah I keep forgetting how good Linux is with memory.
[11:01] Sebastean Steamweaver: I'm sure they'll appear shortly.
[11:01] Sahkolihaa Contepomi: 50MB swap usage. :)
[11:01] xstorm Radek: may i blow it up ?
[11:01] Rex Cronon: hello everybody
[11:01] xstorm Radek: there
[11:02] xstorm Radek: ooppsss i moved it
[11:02] Ardy Lay: Hi Simon
[11:02] Moon Metty: hi Simon :)
[11:02] Sahkolihaa Contepomi: Hey Simon.
[11:02] xstorm Radek: hi Simon
[11:02] Simon Linden: Hello
[11:02] lonetorus Habilis: had a freebsd heavy used freebsd server, with 2gb ram, and 1gb swap, and after a uptime just short of 1 year, it had not used any of its swap
[11:02] Rex Cronon: hi simon
[11:02] Sebastean Steamweaver: I've been geeking out about this lately: http://stevencreech.com/images/posters/Chaotic_Evil_Barney.jpg
[11:02] Sebastean Steamweaver: Ack, wrong link :P
[11:03] Simon Linden: waiting for a floor to rez...
[11:03] Sebastean Steamweaver: http://www.vimeo.com/3365942
[11:03] Changed Event tester: Hello, Avatar!
[11:03] Moon Metty: :)
[11:03] lonetorus Habilis: hey simon
[11:03] Rex Cronon: lol. evil barney
[11:03] Changed Event tester shouts: Denby
[11:03] Sebastean Steamweaver: Barney is indeed evil, however, that was the wrong link ;P
[11:03] Sebastean Steamweaver: The right one is the vimeo.com
[11:04] Rex Cronon: u have been geeking with him;)
[11:04] Arawn Spitteler: The changed event seems to be working
[11:04] Simon Linden: Still rezzing, but let's see if I can do annoucements...
[11:04] Simon Linden: First, Andrew is out today so I don't expect him to come to office hours
[11:04] Moon Metty: ok
[11:04] Sahkolihaa Contepomi: Aww.
[11:04] lonetorus Habilis: arawan, whats that plywood box?
[11:04] Arawn Spitteler: It's testing the CHANGED_REGION event
[11:05] Simon Linden: I don't think there's anything going on with server releases this week
[11:05] Moon Metty: hi Fisher :)
[11:05] Questar Utu: Vuh?
[11:05] Arawn Spitteler: I suppose I don't have to worry, about it's getting returned, for a few hours
[11:05] Questar Utu: Show transparencies doesn't show it.
[11:05] Questar Utu: Yeah.
[11:05] lonetorus Habilis: can we get more detail on what happend with 1.30.2?
[11:05] Fisher Linden: Hey Moon
[11:05] lonetorus Habilis: last meeting was mid roll, so no info was available
[11:05] Skills Hak: hi
[11:06] Simon Linden: We announced some changes in the viewer policies on a blog today : https://blogs.secondlife.com/community/community/blog/2009/10/20/third-party-viewer-policy
[11:06] lonetorus Habilis: hey skills
[11:06] Questar Utu: llSetPos only goes 10m in any direction.
[11:06] Questar Utu: llSetPrimitiveParams would be more likely. :p
[11:06] Arawn Spitteler: Something about a security matter, that caused a couple of screwy releases
[11:06] Simon Linden: I'm not familiar with the details, but it seems to be basically "we don't want other viewers to do bad things"
[11:06] Skills Hak: yea just saw it and was like :O
[11:07] Sahkolihaa Contepomi: Just saw that blog in the IRC channel.
[11:07] Questar Utu: llSetPrimitiveParams([PRIM_POS,position, with the PRIM_POS and position repeated as many times as there are 10m in distance.
[11:07] Arawn Spitteler: When editing a notecard, on the unupdated snowglobe, the notecard wouldn't show an update, until I reopened the edit window
[11:07] Questar Utu: >_> Isn't that against the ToS?
[11:07] Arawn Spitteler: What?
[11:07] Questar Utu: Hm.
[11:07] Sebastean Steamweaver: Simon, I actually had a question about attachment permissions.
[11:08] Simon Linden: OK, sure, let's open it up to questions ... what's up with attachments?
[11:08] Sahkolihaa Contepomi: You appear to be twitching, simon.
[11:09] Sebastean Steamweaver: Is there anything that can be done about the quirk which caused a restriction to be placed on putting non-transfer inventory into fullperm/transferable attachments?
[11:09] Sahkolihaa Contepomi: I know. :P
[11:09] Questar Utu: That, or it's seisure time.
[11:09] Simon Linden: need more coffee
[11:09] Moon Metty: hehe
[11:09] Sahkolihaa Contepomi: I'm multi-tasking between here and IRC right now.
[11:09] Sahkolihaa Contepomi: Haha.
[11:09] Questar Utu: Wouldn't that make it worse?
[11:09] Simon Linden: I can only try and see...
[11:09] Arawn Spitteler pours Coffee into Simon's client
[11:10] Simon Linden: Sebastean - what's your question?
[11:10] Sebastean Steamweaver: Is there anything that can be done about the quirk which caused a restriction to be placed on putting non-transfer inventory into fullperm/transferable attachments?
[11:10] Sebastean Steamweaver: At first I thought it was a bug, but the wording of the error involved indicated otherwise, and someone said it was a security issue.
[11:11] Sebastean Steamweaver: You can put non-trans objects in an attachment, if you drop it first, or rez it on the ground, but you can't do it while wearing it.
[11:11] Simon Linden: Hmm, I don't understand the issue ... you put a no-trans item into a full perm/trans attachment, and what happens?
[11:11] Sebastean Steamweaver: You can't
[11:11] Moon Metty: you can't
[11:11] Sebastean Steamweaver: It won't allow you to.
[11:11] Simon Linden: ah, I see
[11:12] Moon Metty: it's a protection, because the permissions aren't slammed in attachments
[11:12] Skills Hak: yes you have to take it off an rez it
[11:12] Arawn Spitteler: I wonder, if a new state of not receivable, but simply keep things from happening in unauthorised inventories, but would allow them to be contained in products. I think a lot of vendors wuld find that useful
[11:12] Simon Linden: Hmm, I don't know why that's blocked specifically ... normally putting into another object would just make that one no-trans
[11:12] Sebastean Steamweaver: Like Moon said, I heard it was a security issue. Something about being able to transfer no-trans things by putting them into full attachments and then detaching them.
[11:13] Simon Linden: Moon - right, that makes sense. If we're not slamming the permissions, it would be a way to copy itmes
[11:13] Sebastean Steamweaver: But, making it so that you can't put no-trans objects into full perm attachments seems like a "Oh, it's broken? Let's cut it off! :D" solution, instead of a fix.
[11:14] Simon Linden: Can you build what you want when the object isn't attached?
[11:14] Sebastean Steamweaver: Yes, but sometimes, especially in the cases of scripts, it's easier if you put things in while they're worn.
[11:15] Arawn Spitteler: As a UI issue, would it be possible, to clarify whether an object is no-trans by native perms, or by perms of contents?
[11:15] Sebastean Steamweaver: Or, to have the attachment 'refresh' it's permissions when detached?
[11:15] Simon Linden: yeah, but I think I know why it's done that way ... when you attach something, if it's copyable, we just make a new copy in the simulator. If it's no-copy, it gets removed from the asset system
[11:16] Simon Linden: When you take it off, the copyable item can just be deleted, while the no-copy item has to be put back into the DB
[11:16] Sebastean Steamweaver: Why is the bar on transfer objects then?
[11:17] Simon Linden: Now if you attached the copyable item, then put a more restrictive item inside it, we end up with the DB and in-world states being out-of-sync ... the in-world one has less permissions (and the new object inside it) than what was there before
[11:17] Simon Linden: So the updating is pretty ugly ... I'm _guessing_ that's why it's blocked
[11:17] Sebastean Steamweaver: In my tests, no-copy items didn't cause an issue. The only stipulation was that you couldn't put no-trans items into transferrable attachments.
[11:17] Moon Metty: SVC-3771 is another strange case
[11:17] Content of object lost while sending objects between avatars
[11:17] Content of object lost while sending objects between avatars
[11:17] Content of object lost while sending objects between avatars
[11:17] Content of object lost while sending objects between avatars
[11:18] Sahkolihaa Contepomi: lol, too many JIRA helpers.
[11:18] Arawn Spitteler: We have two?
[11:18] Moon Metty: someone has a jira helper attached :)
[11:18] Sahkolihaa Contepomi: Ah, Sebastean has.
[11:18] Sebastean Steamweaver: I need to script it so that it auto-detaches when I come here :P
[11:18] Simon Linden: Just reading that one is going to make me twitch again
[11:19] Sahkolihaa Contepomi: LOL
[11:19] Moon Metty: i have a new version for the triage, it doesn't react to objects
[11:19] Sebastean Steamweaver: If it were possible to have an attachment update its perms when being detached, wouldn't tha fix the transfer-security issue?
[11:19] Sahkolihaa Contepomi: Moon, could I get a copy of that script please?
[11:19] Sebastean Steamweaver: Thanks Moon :)
[11:19] Sahkolihaa Contepomi: Thank you!
[11:20] Skills Hak: oh me too please
[11:20] Simon Linden: It might ... I've avoided permissions issues for the last 6 months or so as my earlier attempts at fixes all had bad side-effects
[11:20] Moon Metty: yeah
[11:20] Simon Linden: It's probably best to save them for when Andrew gets back
[11:20] Sebastean Steamweaver: Will he be back on thursday?
[11:20] Sasy Scarborough: hi all
[11:20] Simon Linden: I think so
[11:20] Arawn Spitteler: Let him have some unpleasant side effects
[11:20] Sahkolihaa Contepomi: Oh dear.
[11:21] Rex Cronon: hi
[11:21] Sebastean Steamweaver: Well, we might as well do it Arawn.
[11:21] Sebastean Steamweaver: Would you like to go first?
[11:21] s opened http://jira.secondlife.com/browse/SVC-4930 in the past week: Does anyone know a reason, that detection volumes should be insensitive to the ground?
[11:21] Volume Detect Insensitive to land_collision_start
[11:22] Simon Linden: reading...
[11:22] Sahkolihaa Contepomi: Oh good I got the url script set correctly for snowglobe...
[11:22] Moon Metty: there was a bug that could make your avatar fall into the ground
[11:22] Sebastean Steamweaver: SVC-22 SVC-93
[11:22] ROTATION and llSetRot incorrectly implemented for child prims
[11:22] Vehicles crossing region borders aren't always treated as vehicles and can get incorrectly returned if the destination parcel is no-entry or parcel-full
[11:22] Moon Metty: i wouldn't be surprised if that's related
[11:22] Arawn Spitteler: Essentially, I was helping somebody's hive drip honey, and the collisions weren't happening, and the ground wasn't useful, so I had to just go phntom
[11:23] Simon Linden: So the land_collision_start events weren't happening?
[11:23] Sebastean Steamweaver: Yeah, it's been the case for a while that items with VolumeDetect enabled, drop through the ground.
[11:23] Arawn Spitteler: The collision with the honey pot wasn't happening, so I had to use the ground, but Detection Volumes aren't sensitive to the ground
[11:24] Moon Metty: it would make your avatar phantom as well, if you set volume detect true, while sitting on the object
[11:24] Arawn Spitteler: It's not that they drop, which could also be fixed, but that they don't notice, so we can't fix
[11:25] Sebastean Steamweaver: So, as long as the script "knew' it was dropping through the ground, it'd be all right/
[11:25] Arawn Spitteler: In this case, that would serve
[11:25] Rex Cronon: u coudl check if the prim is above or below:)
[11:25] Simon Linden: FWIW It would help if you could add a bit more detail there ... steps like "make a box, set physical, add this script..." and then "when I do this, I expect it to ..." to nail down what's wrong
[11:26] Arawn Spitteler: collision_end, checking height, didn't work, which left me to wonder about the server
[11:26] Simon Linden: I workd on some collision event problems before ... there were two types of toubles, usually
[11:26] Simon Linden: We use 'collision groups' to determine what collides with what ... sometimes they aren't set up properly
[11:27] Simon Linden: The other problem was the event dispatching ... we'd detect the collision, but the event queue got messed up
[11:27] Simon Linden: This sounds like the first issue if you never get the event
[11:28] Arawn Spitteler: Possibly, neither new servers, nor Burning Life, are trustworthy environments
[11:28] Simon Linden: Is this new in 1.30?
[11:29] Arawn Spitteler: I don't know; others haven't reproed the failure, and I was only working last weekend
[11:29] Twisted Laws: using volume detect in physical objects has long been a way to keep objects working when under the ground, changing this will break a bit of content, i'd think
[11:29] Arawn Spitteler: My boat worked under 1.28, but I only tested once, before the rollover happened
[11:30] Simon Linden: yeah, we don't want to change any existing behavior unless it's obviously a bug
[11:30] Arawn Spitteler: I'm not talking about the ability to sink into the ground, but just the detection of the ground
[11:30] Simon Linden: But when stuff breaks, that's a problem
[11:30] office hours is half over
[11:30] Simon Linden: Right, just triggering the event
[11:31] Moon Metty: hmmm, more event-related stuff
[11:31] Twisted Laws: but it ignores the ground now and changing that to make the event would affect that?
[11:32] Simon Linden: Are there 2 bugs here? The first part says it was 'ignoring collision_end'. Do you mean you weren't getting that event? That sounds like one problem
[11:32] Arawn Spitteler: What wuld it effect? Does someone use land_collision_start, to ascertain whether Volume Detect is active?
[11:33] Arawn Spitteler: It wasn't getting the collision end, under height, as a cue, but it was late, and I couldn't handle a lot of debug
[11:33] Sebastean Steamweaver: Well, there could be other situations, where a person is counting on it not triggering the land collision even. Say, the object wasn't -always- volume detect, and they wanted to detect "land bounces" and then turned on volume detect to make it drop through.
[11:34] Twisted Laws: well if it currently ignores the ground when physical and volumedetect and that is changed to not ignore the ground and now generate an event, it would seem to change an underlying concept of volumedetect physical prims
[11:34] Simon Linden: ok, well, it sounds like this needs some deeper investigation and thought, so I'll point it out to Andrew when he gets back
[11:34] Moon Metty: maybe we need llGroundDetect, or something
[11:35] Arawn Spitteler: Possible, but if the object detects the land collision event again, and sets volumedetect true, again, who's to notice?
[11:35] Twisted Laws: don't use volumedetect in a physical prim unless you test for the ground?
[11:35] Simon Linden: yeah, it sounds like there are some conflicts between the defined behavior and how VolumeDetect, phantom and the ground all interact
[11:35] Simon Linden: ... and phyiscal
[11:35] Arawn Spitteler: I've filed as New Feature, but I don't know that it ever wasn't a feature
[11:36] Twisted Laws: its been discussed in past office hours here as well... when 1.26 came out i think...
[11:36] Twisted Laws: maybe we can search the old notes and see that discussion
[11:37] Twisted Laws: except i never find anything i'm looking for when i search the wiki for office hour stuff :(
[11:38] Techwolf Lupindo: Every LL search feature is horriably weak or broken. I don't even bother with it, I try to use google site: feature instead.
[11:38] lonetorus Habilis: wikimedia built in search is notoriously bad
[11:38] Simon Linden: Google does tend to search well :)
[11:38] Moon Metty: i search my own chatlogs :)
[11:39] lonetorus Habilis: so yes, use google to do sl wiki specific searches
[11:39] Techwolf Lupindo: LL does contract out stuff, like the viewer 2.0 UI, why not contract out google to fix the inworld searches and other places searches?
[11:39] Twisted Laws: http://jira.secondlife.com/browse/SVC-727
[11:39] Objects with llVolumeDetect(TRUE) don't fall through ground
[11:39] Sebastean Steamweaver: If Google is willing :P
[11:39] lonetorus Habilis: i thought in world search is driven by a google appliance box?
[11:40] Twisted Laws: fixed ... by allowing them to drop thru again
[11:40] Simon Linden: Yes, this is wiki / web site searching, which is a different set of data
[11:40] Questar Utu: I've used plenty of volume detect things that fall through the ground recently, so yeah.
[11:41] Simon Linden: Things are still falling through OK, right? But you're not getting the expected event?
[11:41] Arawn Spitteler conceives of STATUS_FALL_THROUGH_GROUND
[11:42] Twisted Laws: there never was an event except maybe a while when it was broke early in havok4
[11:42] Questar Utu: The only thing that bothers me, is that land_collision events aren't triggered, unlike collision events.
[11:42] Twisted Laws: if there was a collision, wouldn't havok try to stop it from falling thru?
[11:42] Arawn Spitteler: height detection events would also be generally useful, but the lack of land_collision is strange
[11:43] Questar Utu: Potential collisions, not actual ones.
[11:43] Simon Linden: No, we get to specify how it handles the collision
[11:43] Simon Linden: The collision tells you that the two objects are touching or overlapping, but the code can specify if they need to be 'resolved' or not
[11:44] Simon Linden: If we want, they're set to 'resolve' and get bounced so they don't interpenetrate
[11:44] Questar Utu: One thing to note, is that collison events only trigger for moving objects. Not ones that are overlapping but standing still.
[11:44] Simon Linden: Otherwise we let them overlap, which is how phantom works
[11:44] Twisted Laws: different code for land collision tho? everything else that collides with ground bounces
[11:44] swedishfox Ghost: how nice, i teleport to a random location and i find this
[11:44] Questar Utu: Overlapping events would be useful.
[11:44] Simon Linden: So phantom objects only collide with the ground, so they don't drop off the world
[11:44] Arawn Spitteler: Halvea seat, Swedish
[11:45] Simon Linden: VolumeDetect works in a similar way
[11:45] Twisted Laws: phantom objects bounce off the ground
[11:45] Arawn Spitteler: Detective folumes fall off world
[11:45] Questar Utu: Phantom doesn't trigger collision events, but they hit the ground, volume detect does the opposite.
[11:45] swedishfox Ghost: so whats the topic(s)?
[11:45] Simon Linden: Right - the difference is we are telling one to resolve the ground collision, the other doesn't
[11:45] Arawn Spitteler: http://jira.secondlife.com/browse/SVC-4930
[11:45] Volume Detect Insensitive to land_collision_start
[11:46] Simon Linden: It's open for general questions, but we're talking about collisions now
[11:46] Questar Utu: .. Heh. :p
[11:46] Questar Utu: But yeah, it would be nice to have those triggered as well.
[11:46] Moon Metty: yep
[11:46] Questar Utu: That way, you could have an object that falls through the ground, and can tell when it does.
[11:47] Moon Metty: it would be crude, probably
[11:47] Moon Metty: ,
[11:47] Arawn Spitteler: http://wiki.secondlife.com/wiki/Office_Hours if Swedish hasn't seen this
[11:48] Simon Linden: It _might_ be possible to do that if we could get it set up with the collision groups properly - detect the ground collision but don't 'resolve' it and bounce off. Thus you'd get the event as well as the fall-through
[11:48] Twisted Laws: volumedetect physical items require their own scripted movement tho to be useful so that same script can check ground level now if needed, maybe
[11:48] Questar Utu: ^^
[11:48] Twisted Laws: if they are not scripted, they get returned to you from 0
[11:48] Simon Linden: Changing behavior scares me tho, because it seems whenever we do someone else's object suddenly starts doing new things
[11:49] Sahkolihaa Contepomi: Doesn't that usually show someone doing something they shouldn't be doing in the first place? :P
[11:49] Questar Utu: I have a board game of sorts that has 'units', and they're volume detect so that they don't run into things, but can generate events, and they move by their scripting.
[11:49] lonetorus Habilis: no guts no glory? ;)
[11:49] Twisted Laws: land bot things do sometimes... so they don't get caught in mountains
[11:49] Simon Linden: No, it's just they built something based on the current behavior, and then that behavior changes
[11:49] Sebastean Steamweaver: Speaking of new behavior...
[11:49] Simon Linden: ?
[11:49] Sebastean Steamweaver: Simon, before office hours is over, may I ask if there's been any progress on SVC-2885, SVC-2105 or MISC-3077 ?
[11:50] llSetObjectScale and llGetObjectScale
[11:50] for setting multiple, specific linked items
[11:50] List of Linden-Confirmed Easy Changes/Additions With Large Returns
[11:50] Questar Utu: If someone has something that's volume detect and scripted to do land collision stuff, then they should take better care in cleaning up their scripts, because it's not currently possible either way. :p
[11:50] Simon Linden: Not that I know off ... I'm still off on The-Project-That-Must-Not-Be-Named, and Andrew's been busy as well.
[11:51] swedishfox Ghost: seems like scripting, which not my issue, so i'll take my leave
[11:51] Simon Linden: We have been doing some recruiting and have new people coming into the server group
[11:51] Sahkolihaa Contepomi: :o
[11:51] Sebastean Steamweaver: New server devs? :o
[11:51] Sebastean Steamweaver: Take care Swedish :)
[11:51] Arawn Spitteler: We mostly discuss programming of the server, what're you loking for?
[11:51] Simon Linden: So the staffing is ramping up ... no short-term fixes, but that will help in the long run
[11:52] Twisted Laws: cool
[11:52] Simon Linden: I did get an email today from Babbage about the Mono script rezzing issue ... he's putting some time into that now
[11:52] Arawn Spitteler: You'll debug lots, when you show th4 noobs how easy it is
[11:52] Questar Utu would really like the use of the first two Jira's linked there, personally.
[11:52] lonetorus Habilis: i suppose it will take some time for them to get up to speed, or have some of these ppl been involved with sl code before?
[11:52] Sebastean Steamweaver: I've been pushing them for a long time Questar ;) I think at least one fo them has been around longer than I have.
[11:52] Questar Utu: Heh.
[11:53] Simon Linden: It will take some time to ramp up, but they typically can fix their first bug within a week or so. Start on an easy one and go from there...
[11:54] Questar Utu: The ability to change individual prims based on the primitive parameters thing would be the most useful. No more needing a script in each prim with link messages just to change a few things in different ways. Like the position arrangement of linked prims. :o
[11:54] lonetorus Habilis: thats really positive news imo
[11:54] Sebastean Steamweaver: Exactly
[11:54] Questar Utu: Although I have an idea there, I don't know if it would help any.
[11:54] lonetorus Habilis: though i have seen dev projects fail, because they ahd too many ppl
[11:54] lonetorus Habilis: was unmanagable
[11:55] Simon Linden: Yes, that would be a nice feature. We end up with way too many tiny scripts just to set one thing or another
[11:55] Arawn Spitteler: Communications have to be organized, and its an art
[11:55] Questar Utu: To impliment a piece in each parameter that specifies the prim, rather than a prim specification at the beginning, and then a list.
[11:55] Sebastean Steamweaver: And no more scritps in each prim for resizers.
[11:55] lonetorus Habilis puts on the rose tinted galsses for now ;)
[11:55] Sebastean Steamweaver: scripts*
[11:55] office hours is almost over
[11:55] Questar Utu: It would of course have to be implimented as an entirely new function as to not destroy other scripts.
[11:55] Sebastean Steamweaver: Simon, I think those things are "features" that are "needs" as you've heard me say before.
[11:56] Sebastean Steamweaver: If you want us to be efficient in our coding, to do our part in decreasing lag, we need ways to do so!
[11:56] Arawn Spitteler: Getting object scale is just llGetBoundingBox, but llGetLinkPrimitiveParams would be very useful. Also, llGetAgentTargte
[11:56] Simon Linden: Agreed, if we can add features that lower system load or building complexity, it's a very good thing
[11:56] Sebastean Steamweaver: I don't know about you Arawn, but I don't want to be doing extra vector math every time I want the scale of an object.
[11:56] Sebastean Steamweaver: Plus, the bounding box return isn't consistent.
[11:57] lonetorus Habilis: questar, i do that by offloading llsetlinkprimparm stuff to slave scripts, so it is still fast, but wont have a need for a script in each prim, and often i work from a pre-recorded dataset, so only need to getprimparams in the setup of the object
[11:57] Sebastean Steamweaver: It changes based on the rotations of its prims.
[11:57] Questar Utu: But, something like llSetMultiLinkPrimitiveParams([PRIM_COLOR,LINK_SET,ALL_SIDES,<1,0,0>,1.0,PRIM_SIZE,5,<3,1,1>]); where, in that example, the first set of stuff would be the color of all prims, and the second would be the size of link 5.
[11:57] Arawn Spitteler: We haven't got llGetLinkPrimitivePArams, is the probllem
[11:57] Sebastean Steamweaver: I don't think llGetLinkPP is a solution to all issues.
[11:58] Sebastean Steamweaver: You're still working with the delays, which cripple things, and people would use multiple slave scripts to get around that anyway.
[11:58] Sebastean Steamweaver: The idea is to reduce the number of scripts -and- the time needed to do these things.
[11:58] lonetorus Habilis: thats why i will hardcode a initial list of data, with default values, so i know the relative scale difference from the root prim for instance
[11:58] Questar Utu: Get params.. That would be a mess to read, haha.
[11:58] Sebastean Steamweaver: Not create more loops, which Simon has already confirmed creates more work for the server.
[11:59] Simon Linden: Reducing the number of scripts would be good ... I did a performance test with 3000 scripts, and found half our 'script time' was running through the list of scripts, not executing them
[11:59] lonetorus Habilis: but i agree its all currently quirky workarounds
[11:59] Sahkolihaa Contepomi: Ouch.
[11:59] Questar Utu: Here's an example I have, although it's a rather big item.
[11:59] Sebastean Steamweaver: llGetLinkPP would be a good thing.
[11:59] Moon Metty: SVC-4573
[11:59] Idle scripts still take server time
[11:59] lonetorus Habilis: seb, aya
[11:59] Sebastean Steamweaver: However, I think there needs to be a direct solution for things like setting a linkset scale, which Andrew has already said would be easy.
[11:59] Questar Utu: I think I rezzed a smaller version of this a few meetings back.
[12:00] Sebastean Steamweaver: And setting the linked list primitive params.
[12:00] Rex Cronon: crashed:(
[12:00] Thank you for coming to Linden office hours
[12:00] Questar Utu: The pods are what would benefit from an individual link param thing.
[12:00] Sebastean Steamweaver: That's a really cool looking contraption Questar.
[12:00] Sahkolihaa Contepomi sees a Fish.
[12:00] xstorm Radek: may i shoot that pod ship ?
[12:00] Simon Linden: It would be intersting to study all the little scripts that are in use on prims and figure out ways to eliminate them ... linked prim settings is an obvious one, but there are probably some others
[12:01] Questar Utu: Heh, thanks. :p
[12:01] Sebastean Steamweaver: There are Simon, and I have an example I can show you sometime if you're interested.
[12:01] Questar Utu: Also, 'Fish', hon? o.o
[12:01] Sahkolihaa Contepomi: Fisher Linden.
[12:01] Arawn Spitteler: svc-93 would be a good one to fix, but there's a lot to do. I don't think a lot of devs know about linear algebra
[12:01] ROTATION and llSetRot incorrectly implemented for child prims
[12:02] Questar Utu: Oh! Didn't see with all the name tags, haha.
[12:02] Simon Linden: I'm interested but as usual don't have time ... I hope to get back onto server issues like this in January; keeping my fingers crossed
[12:02] Moon Metty: :)
[12:02] Sebastean Steamweaver: Also, eliminating the script delays is another :P People do more work to get around the delays than they would if there were no delay to deal with as a general rule.
[12:02] Questar Utu: Ah, yeah, PRIM_ROTATIOn does it locally.
[12:02] Questar Utu: N*
[12:02] Questar Utu: That is a problem in some cases.
[12:02] Arawn Spitteler: Rmeber to sign the non disclosure agreement, before working on corporate transparency
[12:03] Sebastean Steamweaver: Hah, nice one
[12:03] Moon Metty: lol
[12:03] Arawn Spitteler: Prim Rot is borked]
[12:03] Arawn Spitteler: I think the root prim wouldnb't be rotated correctly, if you tuned the sim
[12:03] Moon Metty: i'm not allowed to say i signed the NDA
[12:04] Simon Linden: I have to run ...thanks everyone for coming today
[12:04] Sebastean Steamweaver: "The NDA does not exist. Yet you will still die if you transgress or mention it."
[12:04] Moon Metty: thank you simon
[12:04] Sebastean Steamweaver: Take care Simon
[12:04] Questar Utu: See ya, Simon. :o
[12:04] Sahkolihaa Contepomi: See you simon.
[12:04] Melchizedek Blauvelt: thanks Simon
[12:04] lonetorus Habilis: yes, thanks simon
[12:04] Sahkolihaa Contepomi: S*
[12:04] Simon Linden: See you next time, take care...
[12:04] Moon Metty: see you thursday
[12:04] Questar Utu: Glad I could actually make a bit of input this time. <_<;

Generated with SLog Wikifier

User:Andrew Linden/Office Hours/2009 10 20

Transcript

Navigation menu

Search