User:Andrew Linden/Office Hours/2009 02 10

Transcript of Andrew Linden's office hours:

[11:01]	Arawn Spitteler:	I recall there's also a problem, with using redundant functions in vehicles.
[11:01]	xstorm Radek:	every thing started to lift off the ground
[11:01]	xstorm Radek:	the sim had to be restarted
[11:02]	Kitto Flora:	If what I describe dont fit - then ask Adrew L about it
[11:02]	Kitto Flora:	Here he is!
[11:02]	Rex Cronon:	hellow everybody
[11:02]	Kitto Flora:	Hi Andrew
[11:02]	Dakotta Karu:	hello rex
[11:02]	Andrew Linden:	Hello
[11:02]	xstorm Radek:	hi Andrew
[11:02]	Dakotta Karu:	hello andrew
[11:02]	Arawn Spitteler:	Hi, Rex, Andrew, and whoever I don't see yet.
[11:02]	Rex Cronon:	hiii
[11:02]	Andrew Linden:	Someone has some questions about vehicles?
[11:03]	Rex Cronon:	lol. popay the sailor man:)
[11:03]	Arawn Spitteler:	xstoirm had an adventure of phgysical objects flying away.
[11:03]	xstorm Radek:	ty :-)(
[11:03]	Arawn Spitteler:	I don't suppose it's possible, to set up diagnostics for unpredictable failures.
[11:03]	xstorm Radek:	i have been having problems with prims drifting off world
[11:04]	Andrew Linden:	xstorm, what sort of prims? scripted vehicles?
[11:04]	Rex Cronon:	nice. imagine Popeye and iron man sitting at same table:)
[11:04]	xstorm Radek:	in fact every thing in the sim is drifting up and out
[11:04]	Andrew Linden:	Ironman with a wizards hat.
[11:04]	Arawn Spitteler:	Spinache comes in Steel Cans
[11:04]	Rex Cronon:	lol
[11:05]	xstorm Radek:	no script just prims set to phys ..
[11:05]	Andrew Linden:	Looks like Rex has settled on an incongruous avatar.
[11:05]	Andrew Linden:	xstorm, perhaps some people are kicking your prims about?
[11:05]	Rex Cronon:	i think i might add thor's hammer, captain america's shield
[11:06]	xstorm Radek:	no they all shot out and up off the sim grid and they keep going
[11:06]	Rex Cronon:	they r sliding?
[11:06]	Andrew Linden:	hrm...
[11:06]	xstorm Radek:	its starting to happen to non phys prims too
[11:06]	Andrew Linden:	perhaps something is rezzing into penetration with the objects?
[11:06]	Rex Cronon:	oh. i think i know
[11:06]	Andrew Linden:	it might be that the "penetration resolver" is trying to separate your prims from some others
[11:07]	Rex Cronon:	r avatars also sliding?
[11:07]	xstorm Radek:	this is sim wide
[11:07]	Arawn Spitteler:	Static Prims wouldn't seperate so, would they?
[11:07]	Arawn Spitteler:	How reproducible?
[11:07]	xstorm Radek:	i have been see in a patter to it
[11:07]	Rex Cronon:	did somebody rez a megaprim on that sim?
[11:08]	xstorm Radek:	for over a week now
[11:08]	Andrew Linden:	hrm... megarim might do it
[11:08]	xstorm Radek:	even after sim restart]
[11:08]	Andrew Linden:	what region?
[11:08]	Rex Cronon:	if is in the next sim?
[11:09]	xstorm Radek:	the new London welcome place and ABC sandbox has the same error
[11:09]	Andrew Linden:	no, an object in a neighbor region wouldn't have any physics effect
[11:09]	xstorm Radek:	they are islands and i no longer see it on mainland
[11:10]	Arawn Spitteler:	Is New London welcome in Australia, like ABC?
[11:10]	Andrew Linden:	I don't see "new London" nor a region named "London"
[11:10]	Andrew Linden:	but there is a London2
[11:10]	xstorm Radek:	last time i seen this happen was back in 2003 when people drifted up and off world
[11:10]	xstorm Radek:	and a restart fix it
[11:10]	xstorm Radek:	hold on i will get the LM
[11:11]	Andrew Linden:	2003 was a long time ago, and a lot of bugs have been fixed and introduced since then
[11:11]	xstorm Radek:	thats the place
[11:11]	xstorm Radek gave you Hyde Park London England , London England UK (112, 88, 21).
[11:12]	xstorm Radek:	it starts to happen when there are over 60 people in the sim
[11:12]	Arawn Spitteler:	What's the limit on HIP?
[11:12]	Rex Cronon:	that is rubber-banding, and is associated with lag:(
[11:13]	xstorm Radek:	with ABC sandbox it starts with just 10 people in the sim so im at a lost why it happens
[11:13]	xstorm Radek:	it may be but why do they drift out from the center of the sim ?
[11:13]	Arawn Spitteler:	ABC might be a connection lag, if you aren't in Australia, wouldn't it?
[11:13]	Andrew Linden:	Ok, well I'll download a copy of the region stat for "London England UK" and see if I can scan for megaprims
[11:14]	Andrew Linden:	I wonder if I can remember the grep + awk magic for that.
[11:14]	xstorm Radek:	i know there are some in there but not that big
[11:15]	Andrew Linden:	Ok, well announcements... I wonder if I have any
[11:15]	xstorm Radek:	i try to keep megas small
[11:15]	Andrew Linden:	I don't have much news that I can think of.
[11:16]	Andrew Linden:	I'm sitting on a bunch of bug fixes in the two code branches I maintain: maint-server-6 and -7
[11:16]	Andrew Linden:	it is unclear whether any of those fixes will be able to get into the pending server-1.26 branch
[11:16]	Andrew Linden:	we're having a series of meetings about it
[11:16]	Andrew Linden:	and the decision process may be slow
[11:17]	Andrew Linden:	it certainly will be if the question about maint-server-* fixes needs to wait for the resolution of all the other stuff that people are trying to work out
[11:17]	Andrew Linden:	regarding best strategies for branching, testing, and deploying, and what changes need to be made
[11:18]	Andrew Linden:	meanwhile, if I were to fix any bug today then it may be months to get out the door
[11:19]	Andrew Linden:	so I've been trying to do other things that I've had to do for a while, until we know how fast the bugs fixes will be able to get into release
[11:19]	Arawn Spitteler has this question from NCI Scripters: [11:18] LindaB Helendale: hi all.. when i try to move a child prim for a small distance with llSetLinkPrimitiveParams, it ends up 1-3 mm off , and then i r-click the prim goes where it should have gone.. and it's systematic, but random.. does anyone know is there some fucntion to force the client to seriously draw the prim correctly?
[11:19]	Andrew Linden:	other things such as writing some scripts that will count the outstanding bugs in various categories and types
[11:20]	Andrew Linden:	Arawn, that souds like a bug. Anyone know it it has been filed and what the number is?
[11:21]	Arawn Spitteler:	Sounds to me like Viewer Side, similar to the flyaway text that was just fixed.
[11:21]	Andrew Linden:	As a workaround... if they can make a call that triggers a "full update" on the object then that will force the true position to get resent to the client
[11:21]	Andrew Linden:	it used to be that a color change would force a full update, but I'm not sure that is true anymore. I forget
[11:23]	Andrew Linden:	When you have the debug option on... AdvancedMenu --> ShowUpdatesOption it will cause these little red, blue, and green particle-like effects to appear whenever an update comes in
[11:23]	Andrew Linden:	the red ones are "full updates" the blue are "partial updates" -- just position and a few other bits
[11:23]	Andrew Linden:	green effects are for objects that get deleted
[11:24]	Andrew Linden:	Arawn, it could be viewer-side. It is hard to tell from what little info we have.
[11:25]	Arawn Spitteler:	Appearently, the error happens on large moves as well.
[11:25]	Andrew Linden:	well hopefully someone will file a jira bug about it.
[11:25]	Arawn Spitteler:	[11:22] Janet Rossini: lol, i don't know. did you try the big move trick?
[11:22]	LindaB Helendale:	yes
[11:23]	Janet Rossini:	and?
[11:23]	LindaB Helendale:	the big move puts it 1-3 mm off.. and the big-small move has the same tolarance
[11:23]	LindaB Helendale:	tried diffent calls and in same setlinkprimparams call
[11:23]	Janet Rossini:	that surprises me. i've always thought big ones went right
[11:23]	LindaB Helendale:	sounds like the movementn has some 3 mm tolerance.. at_target() kind of
[11:25]	LindaB Helendale:	and i need 2-3 mm movement.. arrgghhh
[11:25]	Andrew Linden:	Oh yeah, I remember something I was going to talk bout.
[11:26]	Andrew Linden:	I had mentioned in the past that when it comes to adding new LSL functions I was waiting for the MONO team to start adding their own
[11:27]	Andrew Linden:	I asked Babbage yesterday whether new functions could be added these days
[11:27]	Andrew Linden:	He said "Yes" and that new methods are currently in the pipe for some HTTP-in functionality or something.
[11:28]	Andrew Linden:	So, I was hoping you all could remember and list for me your favorite new function calls you'd like to see added. Bonus points for easy projects -- I'm not going to take on any big project that takes weeks of development and testing.
[11:28]	Andrew Linden:	But quick and easy features I'll consider.
[11:29]	Andrew Linden:	I'm sure I could search the public jira for LSL call suggestions.
[11:29]	Andrew Linden:	If anyone has a filter that catches many of them let me know.
[11:29]	Arawn Spitteler:	llDetectedTarget?
[11:29]	Andrew Linden:	What would llDetectedTarget() do?
[11:29]	xstorm Radek:	? any to do with built in cad tools and converters ?
[11:30]	Andrew Linden:	No xstorm, such features are outside the scope of what I could take on in my copious spare time.
[11:30]	Arawn Spitteler:	What llDetectedPos did, under H1, I think. It would allow sit targets to be set by detecting avatars.
[11:30]	Rex Cronon:	llgethealt(uuid avatarKey) return int
[11:30]	Rex Cronon:	llgethealth*
[11:31]	Andrew Linden:	llGetHealth() would be easy.
[11:31]	Arawn Spitteler:	llGetTheAlt(UUID) would be God-Mode Only
[11:31]	xstorm Radek:	i hope that never is given out
[11:31]	Andrew Linden:	I looked for llDetectedTarget on the public jira and came up empty
[11:31]	Arawn Spitteler:	I'll have to add it, then.
[11:31]	xstorm Radek:	i feel that will be a spy tool and griefer tool
[11:32]	Andrew Linden:	Hrm... we were just talking about some methods in a previous office hour just a few back.
[11:33]	xstorm Radek:	i have seen a lot of peoples UUID get taken with out the person doing it asking first
[11:33]	Andrew Linden:	I'll have to scan the transcripts in the archives
[11:33]	Arawn Spitteler:	UUID? That's just llSensor.
[11:33]	Arawn Spitteler:	Many legitimate uses.
[11:33]	xstorm Radek:	i see that as TOS rules getting bent
[11:35]	Rex Cronon:	uuis is the name that is unique for each avatar. why would u want that hidden?
[11:35]	Arawn Spitteler wonders if http://jira.secondlife.com/browse/SVC-3816 was off topic at CGHour
[11:35]	Rex Cronon:	uuid*
[11:36]	xstorm Radek:	there are police groups starting to grief people again and they are using that info in other ways
[11:36]	Andrew Linden:	An avatar's UUID is sorta like a computer-friendly unique "name" for them. We can't block access to that.
[11:36]	xstorm Radek:	the same as many other griefers that can track you from sim to sim
[11:37]	Andrew Linden:	SVC-3816's title is: SchitzoPermia: Perms different between editing and properties
[11:37]	xstorm Radek:	but it closer to them getting a way to spy on IM's and other things they can do in time
[11:38]	Opensource Obscure:	i thought you couldn't spy on IMs. can you?
[11:38]	Arawn Spitteler:	http://jira.secondlife.com/browse/SVC-3817
[11:38]	Rex Cronon:	i don't think u can be tracked more than one sim, if u go to next onethey loose u
[11:38]	xstorm Radek:	what if they use it to track AR's a person sends in ?
[11:38]	Arawn Spitteler can't track his own ARs
[11:38]	Opensource Obscure:	... and i thought you couldn't spy on ARs.
[11:39]	Andrew Linden:	I don't know how one resident could spy on IM's to another. If anyone knows I'd like to hear about it.
[11:39]	Rex Cronon:	oh yes. how can somebody track others when they fill an AR?
[11:39]	Andrew Linden:	AR's should be unspyable. I can't imagine how someone would track that.
[11:39]	Arawn Spitteler doubts that G-Team could track an AR
[11:39]	xstorm Radek:	one way was they use a tool they give to some one its like the old slave collars they has in SL
[11:40]	xstorm Radek:	they now do it with scripted items
[11:40]	xstorm Radek:	im trying to find some one to make a spy stopper
[11:40]	Arawn Spitteler:	That's part of a game, and some games require that. Keeping a game secret is another game.
[11:41]	Avatar Ghost:	You can't ban/prevent the use of UUIDs, llListen, llSensor, scripts... how much would that break? All of SL
[11:41]	Rex Cronon:	what is your spy stopper supposed to do?
[11:41]	xstorm Radek:	but secondlife is not a game
[11:41]	xstorm Radek:	i like to see private info stay that way
[11:41]	Arawn Spitteler:	Second Life is an environment, where games are allowed, within reason.
[11:41]	Arawn Spitteler:	If you think you're being tracked, type "WORN" in your inventory window.
[11:42]	Rex Cronon:	u can't really stop a spy scripted object, at most u can detect it
[11:42]	xstorm Radek:	its as bad as bloodlines going in to a non RP sim and bite people with out asking
[11:42]	Andrew Linden:	btw, not giant megaprims in "London England UK" region
[11:42]	Andrew Linden:	max size is 64 meters
[11:42]	xstorm Radek:	it ignores the TOS
[11:43]	xstorm Radek:	ty Andrew :-) i did not think it was
[11:44]	Arawn Spitteler:	Max-Size might be a nice function for sim owners.
[11:44]	Andrew Linden:	anyone have any new questions or issues to bring to the table?
[11:45]	Opensource Obscure:	will we ever see writeable notecards?
[11:45]	Opensource Obscure:	dorry if this offtopic / old question.
[11:45]	Andrew Linden:	Avatar Ghost, are you afk?
[11:45]	Opensource Obscure:	*sorry
[11:45]	Avatar Ghost:	I'm not afk
[11:45]	Rex Cronon:	how hard it will be to also have llsetheath(uuid userKey, int value)?
[11:45]	Opensource Obscure:	i mean a way to write notecards via lsl
[11:45]	LindaB Helendale:	hello
[11:45]	Rex Cronon:	hi
[11:46]	Dakotta Karu:	hello linda
[11:46]	LindaB Helendale:	hello Arawn :)
[11:46]	Arawn Spitteler recalls Notecards are unwritable, to prevent the exhaustion of UUIDs, but there are better wayls of doing that.
[11:46]	xstorm Radek:	hi LindaB
[11:46]	Andrew Linden:	ok, just curious. Looks like you've got a hacked client that does not send avatar info.
[11:46]	Rex Cronon:	i would also like what opensource wants:)
[11:46]	Arawn Spitteler:	Linda's the oe with Linked Prims htat don't update.
[11:46]	Avatar Ghost:	it's an animation (many metres underground)
[11:46]	Avatar Ghost:	and then an invisiprim, in case someone cams down there
[11:46]	LindaB Helendale:	yes, after child prim movement it ends up 1- 3 mm off
[11:47]	Avatar Ghost:	and a particle effect
[11:47]	LindaB Helendale:	and when i r-click it it pops where it belongs
[11:47]	LindaB Helendale:	i can show it
[11:47]	Andrew Linden:	Hrm... writable notecards.
[11:47]	LindaB Helendale:	may i ask.. .where are we?
[11:47]	Andrew Linden:	Yeah, we probably didn't implement that for worry about runaway asset creation.
[11:48]	Arawn Spitteler:	We're in Simon's House, in Denby
[11:48]	xstorm Radek gave you Snapshot : Linden Village - Simon Linden, Denby (212, 43, 33).
[11:48]	Rex Cronon:	i have wanted writable notecards for soooooooo long:(
[11:48]	Arawn Spitteler:	If you just spin aournd a pole, at the corner of a sim, every attachment you wear will create new assets.
[11:48]	Andrew Linden:	I'd have to ask around LL to see what people remember from the last time we considered writable notecards.
[11:49]	Andrew Linden:	sounds like a hard problem for me to solve... bigger than I could fit into my spare time.
[11:49]	Simon Linden:	Arawn - right, when you cross a sim boundary they are deleted from the old sim and re-created on the new one
[11:49]	Arawn Spitteler:	If you have an off grid server, you can write to those cards.
[11:49]	Kitto Flora:	The lack of write/read notecards has probably resulted in even more asset size to work around the problem.
[11:50]	Simon Linden:	scripts actually do the same thing - when you edit and update it's effectively a new script
[11:50]	Rex Cronon:	u can even put a limt like no more than 1 new notecard created by a script every minute
[11:50]	Arawn Spitteler:	What strikes me odd, about Linda's problem, is that you still get a displacement, if you move large distances, but it's a displacement less than half a centimeter.
[11:51]	Rex Cronon:	like there is a limit on how many emails an object can send
[11:51]	Simon Linden:	that might be under a small threshold of movement
[11:51]	Andrew Linden:	Rex, yes there are limits on outgoing emails
[11:51]	LindaB Helendale:	heres example.. it should move exactly aligned to the tube..and exactly it's own length onn side when you click
[11:51]	Rex Cronon:	i know, i was giving an example
[11:51]	Simon Linden:	Is that SVC-1945?
[11:52]	Andrew Linden:	oh ok, sorry I read it as a question
[11:52]	Rex Cronon:	np
[11:52]	Simon Linden:	That is aligned for me now
[11:52]	Rex Cronon:	to me that pencil works
[11:53]	LindaB Helendale:	i r-clciicked it
[11:53]	Arawn Spitteler:	Just dysaligned
[11:53]	LindaB Helendale:	now it should be it's own wifth off.. edges aligned
[11:53]	LindaB Helendale:	now it popped right
[11:53]	Rex Cronon:	have u tried using verbal/typed commands, instead of clicking?
[11:54]	Arawn Spitteler sees it still misaligned
[11:54]	LindaB Helendale:	i had the problem in command based
[11:54]	Rex Cronon:	now is back
[11:54]	LindaB Helendale:	now it should be aligned.. and it's like 1.5 mm off
[11:54]	Arawn Spitteler:	It might have aligned, when I ended Selection
[11:54]	LindaB Helendale:	and i had color change there , in case it works.. nope
[11:55]	Simon Linden:	Is there a jira for this?
[11:55]	LindaB Helendale:	now i r-clicked and it's ok
[11:55]	Arawn Spitteler:	Close look shows slight misalignment.
[11:55]	Rex Cronon:	the new position values might have been truncated
[11:55]	xstorm Radek:	her problems is like my problem in many ways its drift or looks like it
[11:55]	xstorm Radek:	its math is it not ?
[11:56]	LindaB Helendale:	it's not such math any truncation error accumulate
[11:56]	Simon Linden:	There are two things I can think of here : there's some tolerance limit in our sending updates, so we don't send a zillion updates for tiny shifts in position. This might be hitting that
[11:56]	LindaB Helendale:	Simon, i thought that, and tried to make big movement and then back big-offset
[11:56]	Simon Linden:	Also, the values that get sent over the wire for updates are sometimes stuffed into smaller data types (fewer bits) to make the messages smaller. That can indroduce errors
[11:57]	LindaB Helendale:	but it seems every movement ends up in 1-2 mm tolerance and then after a while goes right
[11:57]	Andrew Linden:	well, it deserves examination I think.
[11:57]	Simon Linden:	Have you made the entire object set larger? Say a 5m pencil?
[11:57]	Rex Cronon:	is still offset for me
[11:57]	xstorm Radek:	but if you add that up over time is a bigger error
[11:57]	Andrew Linden:	it has been a while since I've looked at the position data in updates for linked objects
[11:58]	Simon Linden:	Right -- my ideas weren't saying 'we shouldn't look at this' but 'here's what might affect it'
[11:58]	Andrew Linden:	I'd expect the prims to line up if they are really alligned
[11:58]	Andrew Linden:	I see a slight offset for the pencil eraser
[11:58]	LindaB Helendale:	there's like 2-10 seconds delay until it gets aligned
[11:58]	xstorm Radek:	like i did say it reminds me of a verry old problem
[11:58]	Rex Cronon:	is not aligned
[11:58]	Simon Linden:	Can I get a copy of it?
[11:58]	LindaB Helendale:	sure
[11:58]	Arawn Spitteler:	Misalignment is changing.
[11:59]	Simon Linden:	me too - it got worse, then better, but still not correct
[11:59]	xstorm Radek:	even if it starts out as a 0.0002 off set it adds up in the end
[11:59]	Rex Cronon:	this might be a problem for rezzers
[12:00]	Arawn Spitteler:	Clickable Pens would have a problem
[12:00]	Andrew Linden:	LindaB, if you've got a simple script that demonstrates the problem, one that you're willing to share with the world, file a jira bug about it and add that script to the comments
[12:00]	Andrew Linden:	there may be a jira bug already entered
[12:00]	Simon Linden:	I have to run ... sorry to be late and leave quickly, but thanks everyone for coming. See you next time...
[12:00]	Opensource Obscure:	about the writeable notecard thing - I would appreciate if you could give us an update at http://jira.secondlife.com/browse/SVC-402 or http://jira.secondlife.com/browse/SVC-3515 or http://jira.secondlife.com/browse/SVC-1406 (not strictly about notecards, 80 votes)
[12:01]	Andrew Linden:	but a simple script for reproducing the problem woudl be nice
[12:01]	Rex Cronon:	bye simon
[12:01]	Simon Linden:	SVC-1945 may be related
[12:01]	Opensource Obscure:	bye simon
[12:01]	Arawn Spitteler:	It takes an expert, to search the Jira so, so the rest of us just file, and let them link
[12:01]	xstorm Radek:	its as bad as them 0 prim rezzers that tells the server 200 prims is only 1 prim it makes errors all over a sim after time
[12:01]	LindaB Helendale:	ummm.. the object i need it is kind of .. non-public *blushes*
[12:01]	Andrew Linden:	I have another meeting that is starting now, so I have to run unfortunately.
[12:01]	Opensource Obscure:	thanks andrew - ciao
[12:01]	LindaB Helendale:	but that pencil shows it, when the width it the same as the offset
[12:01]	xstorm Radek:	thank you Andrew
[12:01]	Avatar Ghost:	ty andrew
[12:01]	Opensource Obscure:	ciao everybody
[12:01]	xstorm Radek:	bye
[12:01]	Rex Cronon:	bye andrew
[12:01]	Andrew Linden:	Obscure, perhaps I'll find the time to go back to the transcript and examine the bugs you mentioned.
[12:02]	Rex Cronon:	bye epensource
[12:02]	Opensource Obscure:	nice.
[12:02]	Kitto Flora:	Byebye all
[12:02]	Rex Cronon:	bye beverybody
[12:02]	xstorm Radek:	thank you LindaB
[12:02]	LindaB Helendale:	bye all, thank you for attending my problem :)
[12:02]	Rex Cronon:	have fun:)