User:Andrew Linden/Office Hours/2010 09 03

Transcript

[16:03]	Simon Linden:	Hello
[16:03]	Moundsa Mayo:	heee. Yeah, Sion too
[16:03]	Simon Kline:	Hi Moon hi Simon
[16:03]	Moundsa Mayo:	Simon
[16:03]	eighthdwarf Checchinato:	hi
[16:04]	reddot99 Republic:	theres andrew
[16:04]	Moundsa Mayo:	We busted!
[16:04]	Simon Kline:	oops i just realised reddot and Moundsa were talking to the other simon >.<
[16:04]	Simon Linden:	Ah, I was about to say I wasn't sur e if Andrew was coming or not
[16:04]	Moundsa Mayo:	I mean, er, Hi Andrew!
[16:04]	Andrew Linden:	Hello. Sorry I'm late.
[16:04]	reddot99 Republic:	its fine, lol
[16:04]	eighthdwarf Checchinato:	uh oh
[16:04]	Rex Cronon:	hi everybody
[16:04]	s greeter	says: Hello Pauline, Leesa, Kiptic, Moundsa, Techwolf, Moon, reddot99, Slee, eighthdwarf, Simon, Simon, Liisa, Andrew, Kaluura, Identity, and Ardy.
[16:04]	Andrew Linden:	Falcon and I were stepping through some disassembly and having too much fun.
[16:04]	Moon Metty:	hi hi
[16:04]	Moundsa Mayo:	Hey Rex
[16:05]	Leesa Donner:	hi guys
[16:05]	Rex Cronon:	hi moundsa
[16:05]	Rex Cronon:	hi
[16:05]	Moundsa Mayo:	Hiya, Leesa
[16:05]	Leesa Donner:	are you on voice
[16:06]	Pauline Darkfury:	Hi folks :)
[16:06]	Andrew Linden:	Lets see... I don't have much news to report today.
[16:06]	Rex Cronon:	hi
[16:06]	Moundsa Mayo:	Hi Pauline
[16:06]	Simon Linden:	No, we do these office hours on text, Leesa
[16:06]	Andrew Linden:	I'm still working on some optimizations for linkability rules.
[16:06]	eighthdwarf Checchinato:	I hope not, I have currently absolutely no sound onthis computer :(
[16:06]	Leesa Donner:	good I have not figured out voice yet on v2
[16:06]	eighthdwarf Checchinato:	ah thanks Andrew
[16:06]	Leesa Donner:	does not appear to be in preferences
[16:06]	Andrew Linden:	Adit has a "slightly" updated server that now has tcmalloc enabled.
[16:07]	Andrew Linden:	tcmalloc is a memory allocator that is part of the google perftools
[16:07]	Andrew Linden:	which allows easier debugging of memory problems
[16:07]	Kaluura Boa:	The avatar presence thing is much more interesting... for us...
[16:08]	reddot99 Republic:	dont they include it in the dev tools for several systems?
[16:08]	Simon Linden:	The lockout you sometimes get?
[16:08]	Kaluura Boa:	Yeah! Let's get rid of the lockout!
[16:08]	Kaluura Boa:	"You're too fast... 5 minutes of penalty"
[16:08]	Andrew Linden:	I haven't yet been able to test my latest maint-server work because it has a build error that Gulliver Linden is fixing.
[16:09]	reddot99 Republic:	build error?
[16:09]	Andrew Linden:	However, when I get that up on aditi I'll announce it here. It has SVC- 93 fixed therein.
[16:09]	ROTATION and	llSetRot incorrectly implemented for child prims
[16:09]	Simon Linden:	Ambroff Linden has been working on it ... some fixes are on the way
[16:09]	Andrew Linden:	Hey Simon, I think the jirabot is kinda busted with the new jira system
[16:09]	Andrew Linden:	SVC-93
[16:09]	Simon Linden:	(the lockout bugs)
[16:09]	omar Mosswood:	HI GUYS
[16:09]	reddot99 Republic:	[16:09] Meeter: http://jira.secondlife.com/browse/SVC-93 [#SVC-93] llSetPrimitiveParams PRIM_ROTATION and llSetRot incorrectly implemented for child prims
[16:09]	ROTATION and	llSetRot incorrectly implemented for child prims
[16:09]	Leesa Donner:	I had issues downloading the new version tonight
[16:09]	Moon Metty:	no no, Sue changed the jira title back
[16:09]	omar Mosswood:	I`M NEW
[16:09]	reddot99 Republic:	it already set it once
[16:09]	Rex Cronon:	hi
[16:09]	Leesa Donner:	almost could not use V2
[16:09]	Leesa Donner:	it hung
[16:09]	Andrew Linden:	oh, it looks fixed
[16:10]	Simon Linden:	Slow scripts ;0
[16:10]	Andrew Linden:	er... jira bot helper looks fixed
[16:10]	Moon Metty:	i will send out a new update for the jira helper after all changes to the jira are done
[16:10]	reddot99 Republic:	it did it twice
[16:10]	Andrew Linden:	ok thanks Moon. Is that your project? I wasn't sure who wrote the jira bot.
[16:11]	Moon Metty:	yeah, i originally made it for in-world triages
[16:11]	Liisa Runo:	Welcome to SL. This is very confusing first, but will be awesome when you find the stuff you like. Use search and explore. SL has everything. Dont stress about getting money yet. Ask for help when you need. Most people are happy to help.
[16:11]	Kaluura Boa:	=^_^=
[16:12]	Moon Metty:	i did a simple test with script-time yesterday
[16:12]	reddot99 Republic:	also, andrew, i heard somebody talking about an update that was rolled out unannounced, what was that about?
[16:12]	Moon Metty:	SVC-4573
[16:12]	Idle scripts	still take server time
[16:12]	Simon Linden:	a server update?
[16:13]	Andrew Linden:	reddot99, a server update?
[16:13]	reddot99 Republic:	supposedly, server release number had changed, at least the build number did
[16:13]	Moon Metty:	a script that's busy with sleeping uses far less time than an idle script :)
[16:14]	Moundsa Mayo:	ick
[16:14]	Moundsa Mayo:	Won't respond to any event until sleep over
[16:14]	reddot99 Republic:	that would be because a sleeping script doesnt have any events triggered, even if they are empty
[16:14]	Andrew Linden:	Hrm... well whatever update you're talking about... it didn't hit Denby here.
[16:14]	Moon Metty:	right
[16:15]	Andrew Linden:	Denby appears to have had a long uptime... since 08.20
[16:15]	Psi Merlin:	The 1.42.0 pilot was re-r olled (208156->208314) to fix an issue with twisted prims going phantom
[16:15]	Gooden Uggla:	that explains why we have to restart sims daily now
[16:15]	Andrew Linden:	reddot99, are you talking about the main grid (agni)? or aditi?
[16:15]	Simon Linden:	yes, this is on the 'pilot roll' list so it got 1.42.0 early on
[16:15]	Pauline Darkfury:	There's something going on with script times. Based on situations where I've seen 40ms of scripts reported, but the server still reporting 22-25ms frames, and lag not feeling too bad, I think it's at least partly a stats issue. SVC-6259 is probably related.
[16:15]	Simon Linden:	... and looks like it hasn't crashed
[16:15]	reddot99 Republic:	psi might have gotten it,
[16:15]	Total script	time shown in 'Top Scripts' is inconsistent with script time shown in the statistics bar
[16:16]	Moon Metty:	is that a new jira?
[16:16]	Pauline Darkfury:	it's quite new
[16:16]	Andrew Linden:	Sometimes we'll update targeted regions with a special build
[16:16]	reddot99 Republic:	pauline, that would be because the top scripts is a single snapshot
[16:16]	Simon Linden:	Script time measurement is a long conversation ... you can't compare the 'Top Scripts' time with the ctrl-shift-1 panel, for example
[16:16]	reddot99 Republic:	of a single frame
[16:17]	Andrew Linden:	such as regions that tend to get attacked by griefers. Sometimes we'll fix the crash bug and only update the regions that are being targeted.
[16:17]	Moon Metty:	i've always found the statistics window is the most accurate
[16:17]	Moon Metty:	all the numbers add up in my tests
[16:17]	Simon Linden:	Top Scripts is also an average of the last 5 times that script ran ... it might not run every frame if the region is busy
[16:17]	reddot99 Republic:	wait, why would a script run at all if no events were called?
[16:18]	Gooden Uggla:	thinig is, that top scripts number is inching up to the same number as the stat window lately (last roll)
[16:18]	Simon Linden:	Use Top Scripts to identify what's using the most, and consider the comparisons relative. The rankings can change a lot depending on what the script is doing
[16:18]	Andrew Linden:	a script with a timer or a sensor will run constantly
[16:18]	Rex Cronon:	to check for events:)
[16:18]	Andrew Linden:	or a listen
[16:18]	Pauline Darkfury:	well, the numbers on it seem to be a bit more meaningless on 1.42 than 1.40, frequently seeing llSitTarget and llParticleSystem scripts which normally sit at 0.002-ish getting reported as 0.100-1.000 (and very occasionally even higher)
[16:18]	reddot99 Republic:	what about a collision event?
[16:18]	Moon Metty:	really Pauline?
[16:19]	reddot99 Republic:	shouldnt those let the script run only when havok says it has a collision?
[16:19]	Pauline Darkfury:	yup
[16:19]	Moon Metty:	that sounds like the old SVC-3819
[16:19]	in estate	tools is inaccurate
[16:19]	Andrew Linden:	collision and touch events are optimized such that the script is not checked constantly
[16:19]	Gooden Uggla:	the mono script time detector (moon's) that i use to indicate when to restart sims is going apeshit lately, 3.2 ms across the board after a day with 2 events
[16:19]	Pauline Darkfury:	it happens when we get busy, and when there seems to be external lag (i.e. something external to the sim) at the same time (say 20+ AVs, and about 10-11k scripts)
[16:20]	Gooden Uggla:	looks to be a measurement problem, no change in scripts
[16:20]	claudinha Zarco:	oiiiiiii
[16:20]	reddot99 Republic:	but i can see every bullet in the top scripts window,
[16:20]	Moon Metty:	so maybe topscripts get totally confused by the load, you mean?
[16:20]	claudinha Zarco:	oiiiiii
[16:21]	Gooden Uggla:	the recent asset server problems propagating something?
[16:21]	Moon Metty:	i only have access to topscripts on aditi, so i rarely look at it
[16:21]	Moon Metty:	,
[16:21]	reddot99 Republic:	also, do colllsion filters get carried out at havok or does it call a collision event and ignore it if it matchs
[16:21]	Pauline Darkfury:	Yeah, when there's no grid or asset server issues, the script times are reasonable with 15-20 AVs, some heavily scripted, but occasionally spike to quite bizarre numbers. Does seem like Top Scripts becomes rather meaningless under heavy load
[16:22]	Andrew Linden:	yes, sounds like a problem with the measurement. I would not be surprised.
[16:22]	Pauline Darkfury:	possible hypothesis is that net time is getting added into per script time - high reported net time usually coincides with bizarre numbers
[16:23]	Gooden Uggla:	i was wondering, perhaps http textures is somehow adding to the load? like a double server call?
[16:23]	Andrew Linden:	reddot99, Havok filters the collision events that it passes -- basically it throttles the events that get to the scripts
[16:23]	Simon Linden:	I wonder if some download time is getting inadvertantly added into that script time
[16:23]	Moon Metty:	yes Simon
[16:23]	Gooden Uggla:	it's recent, seemed to be the last roll, anecdotally
[16:24]	Pauline Darkfury:	It does seem to me that scripts which I know have both HTTP requests and HTTP servers are more prone to getting listed abnormally high (potentially 10x their normal time) when things are misbehaving
[16:24]	Simon Linden:	We shouldn't be doing that, obviously, but maybe it's not stopping the clock until llRezObject() fully completes, or something like that
[16:24]	reddot99 Republic:	so not stoping the clock for sleep penalties would explain it
[16:24]	Pauline Darkfury:	don't think llRezObject relates to what I'm seeing, but could well be in the overall picture
[16:25]	Moon Metty:	the sleep times were fixed in svc-3819
[16:25]	Moon Metty:	that one is ok, i think
[16:25]	Simon Linden:	I just picked llRezObject since that's my main headache ... I can imagine the http functions having that problem
[16:25]	Gooden Uggla:	thing is... it's not just a measurement problem, performance drops too
[16:25]	Gooden Uggla:	everything slows down in the sim
[16:25]	reddot99 Republic:	wish top scripts actually showed what scripts in each object
[16:26]	Pauline Darkfury:	hippoVEND v2.7 panel vendors (typical 0.040ms approx) and Apez iVend v2.3 (typical 0.075ms approx) tend to show frequently very high if there is something else going on (such as possible network congestion or asset server slowness)
[16:26]	Gooden Uggla:	hippovends are notorious script hogs, the mono versions are tools of griefing
[16:26]	Pauline Darkfury:	Both of those vending systems are pretty stable on script times normally, and it's not related to them updating or people using them.
[16:27]	Pauline Darkfury:	Apez is twice as bad a hippoVEND, frankly, based on script times when the sim is quiet, and stable around 0.040ms or less.
[16:27]	Gooden Uggla:	yes pauline
[16:27]	Pauline Darkfury:	(hippo is stabel at about 0.040)
[16:27]	Gooden Uggla:	but there's far fewer apez
[16:27]	reddot99 Republic:	can we get a filter collision by type as well?
[16:28]	Gooden Uggla:	rather than the measurement itself, i'd like to know what slows down the sim whrn those measurements start to climb
[16:28]	Andrew Linden:	reddot99, enhancing the collision filters is probably not very hard
[16:28]	Gooden Uggla:	are sims self-regulating? is it swapping memory to dick when it doesn't have to?
[16:29]	Gooden Uggla:	oops disk*
[16:29]	Leesa Donner:	hehe
[16:29]	Pauline Darkfury:	I think the everything slowing down in sim is part of the symptoms of something external happening, along with the bizarre script times being reported, but it's not the actual scripts producing the slowdown.
[16:29]	Andrew Linden:	if you have some jiras go ahead and link them to MISC-3777 and I'll review them.
[16:29]
[16:29]	reddot99 Republic:	lol
[16:29]	reddot99 Republic:	jira has killed it
[16:29]	Gooden Uggla:	andrew? do the sims swap memory based on those measurements?
[16:29]	Moon Metty:	the jira helper is right lol
[16:30]	Gooden Uggla:	or on internal measurements that we can't access?
[16:30]	reddot99 Republic:	wrong misc though
[16:31]	reddot99 Republic:	http://jira.secondlife.com/browse/SVC-6181
[16:31]	List of	Linden-Confirmed Easy Changes/Additions With Large Returns
[16:31]	Andrew Linden:	Hrm... did I misremember the jira number?
[16:31]	Moon Metty:	i'll contact Sue about that one
[16:31]	reddot99 Republic:	yeah
[16:31]	Simon Linden:	Gooden - we don't control any memory swapping, the VM is all handled by the system
[16:32]	Gooden Uggla:	ok
[16:33]	reddot99 Republic:	also, whats this mean by aknowledged on SVC-2885
[16:33]	llSetObjectScale and	llGetObjectScale
[16:33]	Pauline Darkfury:	How difficult would it be to get some idea of actual virtual memory stats (real memory in use, virtual allocated, page ins/outs), somewhere that an estate manager could look at them?
[16:33]	Moon Metty:	when a jirqa issue is imported, it's marked as "acknowledged"
[16:34]	Moon Metty:	after the change to jira4, next week, all previously imported issues will follow
[16:34]	Psi Merlin:	I suggested something like taht last week Pauline - Falcon liked the idea.
[16:35]	Gooden Uggla:	i can't wait for the "it's the sims you're sharing the server with" excuse to make a comeback
[16:36]	Name Pinger:	Listening for "ardy" - Click to disable.
[16:36]	Moundsa Mayo:	Hi Falcon
[16:36]	Falcon Linden:	Hey folks
[16:36]	Moon Metty:	hey Falcon :)
[16:36]	Rex Cronon:	hi
[16:36]	reddot99 Republic:	hey falcon
[16:36]	reddot99 Republic:	any news on raycasting?
[16:36]	Andrew Linden:	I think Linden Homes are getting fewer CPU cycles. Those regions have smaller prim limits, right?
[16:37]	Liisa Runo:	smaller? or the other way round
[16:37]	Gooden Uggla:	johnathan when mainland regions aren't restarted unless they're rolled or they crash, the performnace drops
[16:37]	Andrew Linden:	lower prim limits
[16:37]	Simon Linden:	No, I'm pretty sure they are full regions
[16:37]	Gooden Uggla:	and what andrew said
[16:38]	Andrew Linden:	What's the name of one "Linden Home" region?
[16:38]	Simon Linden:	Some of the builds are slow to draw because the builders went a bit wild with sculpties
[16:38]	reddot99 Republic:	lol
[16:38]	Jonathan Yap:	Clarinet
[16:38]	Gooden Uggla:	if the lab wanted to improve mainland performance, they'd schedule a restart every day at a specified time
[16:38]	reddot99 Republic:	or once a week
[16:39]	Gooden Uggla:	weekly isn't enough for an active sim
[16:39]	Andrew Linden:	Clarinet is a full 15k-prim region and runs on its own CPU
[16:39]	reddot99 Republic:	for most sims it is enough though,
[16:39]	Moundsa Mayo:	Bloomindell is another
[16:39]	reddot99 Republic:	also, food, back in a minute
[16:39]	Jonathan Yap:	Andrew, for people who were not at Oskar's OH, do you want to repeat the new testing plan?
[16:40]	Gooden Uggla:	let's say your sim has a merchaant in the ongoing "twisted hunt"... you'd see hundreds of serialize/deserialize events ever day
[16:40]	Gooden Uggla:	you'd want daily
[16:40]	Andrew Linden:	Same story for Bloomindell
[16:40]	Andrew Linden:	Jonathan. you're asking about yesterday's memory profiling during Oskar's office hours?
[16:40]	Moundsa Mayo:	So, extrapolating from a small sample ... B^)
[16:41]	Jonathan Yap:	No, I was thinking about having 2 or 3 rollouts on 10% of the grid each and more rolling restarts
[16:41]	Pauline Darkfury:	Personally I'm seeing rather good performance out of my Linden Homes region (Tempo), but you do need to set draw distance short due to very high prim density within draw distance - I find 64m draw dist is fast there, where normally 128m is fast enough
[16:42]	Andrew Linden:	Oh, you're asking if I would summarize the new plan for how to deploy and test new server versions?
[16:42]	Jonathan Yap:	Yes, if that is a subjec for this meeting--for people who were absent
[16:42]	Andrew Linden:	I know there is a new plan, but I'm vaugue on the details actually.
[16:43]	Gooden Uggla:	beta grid meeting log not up ?
[16:43]	Andrew Linden:	I'll understand it better when I actually get some of my own work through it.
[16:43]	Andrew Linden:	Until then I'm not going to study it too much -- I've got too much other stuff on my mind.
[16:43]	Jonathan Yap:	What brought it to mind was mention of region restarts -- and in the meeting how they will be more like once a week vs once a month now
[16:43]	Moon Metty:	hehe
[16:44]	Simon Linden:	I can't describe it very well, but essentially we're trying to have more and faster updates, plus multiple channels of new versions on the main grid
[16:44]	Andrew Linden:	However, the plan is to do test deploys of new servers on the main grid before that work ever gets merged into the official version for the rest of the world.
[16:44]	Moundsa Mayo:	But rolling across regions, so most should not actually see restarts on a weekly basis.
[16:45]	Kiptic Horsley:	/me adds SVC-5419 to 6181
[16:45]	Add flag	parameter [vector slice] to llSetPrimitiveParams PRIM_TYPE flags PRIM_TYPE_BOX, PRIM_TYPE_CYLINDER, and PRIM_TYPE_PRISM
[16:45]	Andrew Linden:	It isn't clear to me how this is going to be faster -- I foresee a traffic jam at the experimental deploys, but I'll wait and see if it really happens.
[16:45]	Andrew Linden:	More frequent updates for those who want to live on the bleeding edge, yes.
[16:46]	Simon Linden:	I think the bigger change is putting the QA responsibilty closer to the developer groups ... hopefully it means we can iterate on finding and fixing problems faster
[16:46]	Moon Metty:	yes
[16:46]	Andrew Linden:	Simon and I were going over the list of "distinct projects" that our team has on the "to be tested" queue
[16:47]	Moundsa Mayo:	And deploy successful projects more rapidly
[16:47]	Andrew Linden:	and were wondering how best to coalesce them so that we have fewer test deploys
[16:47]	Pauline Darkfury:	as long as there's estate manager control to go to/from those experimental releases, it sounds good (estate owner is not enough - I rent from a major estate, and can't constantly bug the single owner for something like that)
[16:47]	Moundsa Mayo:	So smaller progress increments, but possibly better overall progress rate.
[16:48]	Moundsa Mayo:	And deploying individual projects MAY reduce some interaction defects.
[16:48]	Andrew Linden:	right, I think the system is only going to work if estate owners can opt into whatever versions they want (experimental or stable)
[16:48]	Jonathan Yap:	I don't remember if it was mentioned that you would have the option to be in a test deploy or not
[16:48]	Psi Merlin:	As a long term pilot region owere I'd certainly like to be able to join the testing for a channel of interest.
[16:48]	Simon Linden:	I'm pretty sure the estates were normally off the pilot releases ... it was going to be the mainland, Linden regions like this one, and regions that wanted to be on the program
[16:48]	Gooden Uggla:	i'd be inclined to test more thoroughly on agni if i could revert off the bleeding edge on estate controls
[16:48]	reddot99 Republic:	where'd falcon go?
[16:49]	Gooden Uggla:	that would allow testing with 30 avs present
[16:49]	Pauline Darkfury:	Yup, likewise, but only if I can rapidly opt back out with EM rights only, and not have to bug the owner)
[16:49]	Moon Metty:	right Gooden
[16:49]	Moundsa Mayo:	Falcon's better judgement prevailed B^)
[16:49]	Simon Linden:	It would be good to let owners have more control and choice on where their regions go
[16:50]	Kiptic Horsley:	/me links another JIRA issue *blush*
[16:50]	Gooden Uggla:	simon, if i could hold events that specifically announced testing, i could actually theme around that... however, i'd rather not do that every night, ya know?
[16:50]	reddot99 Republic:	which is more efficient, collision or collision start, out of curiousity?
[16:50]	Gooden Uggla:	sometimes people just want a wet t-shirt contest :)
[16:51]	Simon Linden:	I don't think we have features in the pipeline for that, Gooden :)
[16:51]	Moundsa Mayo:	Only use collision if you need rapid multiple events.
[16:51]	Gooden Uggla:	lol we already have that covered :)
[16:51]	Andrew Linden:	reddot99, I think collision_start() is more efficient overall -- you'll actually get fewer callbacks
[16:52]	reddot99 Republic:	it kills itself at the end of the event
[16:52]	Gooden Uggla:	but if i gave away prizes for finding bugs on the bleeding edge, or something like that, i could tyheme around the preview versions
[16:52]	Gooden Uggla:	theme*
[16:53]	Gooden Uggla:	i think a lot of people could find ways to encourage testing on agni IF they had the ability to revert immeditaely if the version was too unstable
[16:53]	Andrew Linden:	right Gooden, that is required, and I think it is in the plan
[16:54]	reddot99 Republic:	ah, falcon just crashed
[16:54]	Moundsa Mayo:	WB, FAlcon!
[16:54]	Moon Metty:	i think a public sandbox with an "experimental" server version would be popular
[16:54]	reddot99 Republic:	welcome back
[16:54]	Rex Cronon:	wb
[16:54]	Falcon Linden:	Thanks
[16:54]	Gooden Uggla:	i just hope the plan doesn't mean everyone on the 4th oldest server version because the newer ones are unusable
[16:54]	Liisa Runo:	++Moon
[16:54]	Jonathan Yap:	Moon, I think that was mentioned in the plan last night
[16:54]	Andrew Linden:	Moon, I think you're right. A public sandbox devoted to the bleeding edge would be popular.
[16:54]	Moon Metty:	yeah
[16:55]	Kiptic Horsley:	/me votes for that sandbox
[16:55]	Falcon Linden:	Andrew: And would keep all of us busy fighting the griefers
[16:55]	Pauline Darkfury:	sandboxes on main grid with test releases is an excellent idea :)
[16:55]	Simon Linden:	The sandboxes should end up in the pilot programs ... at least any that are on the mainland
[16:55]	Moon Metty:	hehehe
[16:55]	Pim Peccable:	/me Laughs
[16:55]	Moundsa Mayo:	I'd visit frequently just for the sheer ambience of it!
[16:55]	Moon Metty:	think of it as a griefer trap :)
[16:55]	Andrew Linden:	No, the griefers have other sandboxen to play with
[16:55]	reddot99 Republic:	lol
[16:55]	Kiptic Horsley:	honeypot for griefers :)
[16:55]	Falcon Linden:	But not with new crashes to find
[16:56]	Simon Linden:	or to test if your old griefer objects still work :)
[16:56]	Liisa Runo:	we have automated gadgets to fight griefers, they are just kids with 5 year old scripts. minor annoyance
[16:56]	Moundsa Mayo:	But if it's running debug, that might speed up defect detection.
[16:56]	Andrew Linden:	Yeah, we'll just spread the rumor that the Bleeding Sandbox is under tighter scrutiny.
[16:56]	Falcon Linden:	Andrew: Extra logging?
[16:56]	Falcon Linden:	:)
[16:56]	Simon Linden:	or announce a new server that "wont crash" :)
[16:56]	reddot99 Republic:	or give a few residents estate powers there
[16:56]	Moundsa Mayo:	HAR!
[16:56]	Moon Metty:	"hey! this server breaks my griefing tools!"
[16:56]	Gooden Uggla:	:)
[16:57]	Rex Cronon:	who r u going to sue:)
[16:57]	Pauline Darkfury:	or announce that a new test version lets an old attack they loved work again ;)
[16:57]	Moundsa Mayo:	'may' let and old ...
[16:57]	Moundsa Mayo:	[an]
[16:57]	Falcon Linden:	How about we just announce than anyone identified as griefing the sandbox will be banned?
[16:57]	reddot99 Republic:	lol, make two, one with a grey goo fence disabled,
[16:58]	Andrew Linden:	Ah. MISC-3777 has moved to SVC-6181. I'm getting emails about updates.
[16:58]	Pauline Darkfury:	how about not announcing it and just banning?
[16:58]	Rex Cronon:	can u defin griefing first?
[16:58]	Falcon Linden:	Rex: Nope!
[16:58]	Falcon Linden:	so be careful
[16:58]	Falcon Linden:	:)
[16:58]	reddot99 Republic:	and grey goo fence disabled one logs you off if you trigger it too many times or for too long
[16:58]	Simon Linden:	I was (as a joke) suggesting the simulator automatically add some attachments to your AV if it detects griefing activity
[16:58]	Moundsa Mayo:	Fair enough, Falcon. I think I'll just stay home B^)
[16:58]	Falcon Linden:	:)
[16:58]	Rex Cronon:	the problems is sometimes if u "experiment" thing can get out of control. unintentionally:)
[16:58]	reddot99 Republic:	that actually would be amusing simon, and useful
[16:58]	Kiptic Horsley:	dummy cone on top of head?
[16:58]	Moundsa Mayo:	Or on Aditi
[16:58]	Andrew Linden:	hehe, that would be funny Simon.
[16:58]	Simon Linden:	exactly!
[16:59]	Kiptic Horsley:	:D
[16:59]	Slee Mayo:	so create a group, only group members can enter SB
[16:59]	reddot99 Republic:	simon, make sure it cant be detached either
[16:59]	Liisa Runo:	experimental things should always have failsafe in them ;)
[16:59]	Andrew Linden:	but then the dunce cap would be a little "badge of honor"
[16:59]	Moon Metty:	lol
[16:59]	Moundsa Mayo:	yup!
[16:59]	Andrew Linden:	greifers would flock to get their dunce caps
[16:59]	Pauline Darkfury:	Yup, need to separate true griefers from "oops, that went into an exponential loop" bugs
[16:59]	Moon Metty:	the golden corn cob
[16:59]	Simon Linden:	right, maybe some new attachment points that are only controlled server-side :)
[17:00]	Moundsa Mayo:	ah, movie title - "True Griefers"
[17:00]	Ardy Lay:	Concrete boots
[17:00]	reddot99 Republic:	no, llsettext visible at infinite distance saying ban me
[17:00]	Gooden Uggla:	nah, just put the word "fail" on the cap, they'd never want to wear that
[17:00]	Simon Linden:	The Scarlet G
[17:00]	Moundsa Mayo:	Simon, sometimes it appears there is one deep inside teh pelvis.
[17:00]	reddot99 Republic:	in a way that couldnt be done with scripts
[17:00]	Pim Peccable:	They might wear "Epic Fail"
[17:01]	Jonathan Yap:	Not "fail" but "Demo"
[17:01]	Gooden Uggla:	LOL
[17:01]	Andrew Linden:	Demo? I don't get it.
[17:01]	eighthdwarf Checchinato:	/me giggles
[17:01]	Pim Peccable:	hehe Jonathan
[17:01]	Jonathan Yap:	Andrew, you see that for test hairs, etc
[17:01]	Andrew Linden:	Oh I see. Hehe.
[17:01]	Pim Peccable:	No one wants to admit to getting a Demo item - worse than a bade freebie
[17:01]	Simon Linden:	ah, right, sample objects and textures
[17:01]	Pim Peccable:	*bad
[17:01]	reddot99 Republic:	the no image texture should make a comeback
[17:02]	Liisa Runo:	/me will get "Demo" RL tattoo some day
[17:02]	reddot99 Republic:	lol
[17:02]	Moundsa Mayo:	Just turn them to grimy cloud.
[17:02]	Pauline Darkfury:	lol
[17:02]	Liisa Runo:	"Demo Skin"
[17:02]	Pim Peccable:	Demo Cloud
[17:02]	Moon Metty:	:)
[17:02]	Kiptic Horsley:	or just perma-ruth them :)
[17:02]	reddot99 Republic:	could actually do a lot of things to them, that would make them identifable
[17:02]	Jonathan Yap:	For Demo Skin do you go to see a Demotologist? ugh
[17:02]	Gooden Uggla:	liisa those are called "magic markers"...
[17:03]	Moundsa Mayo:	Naw, don't do that to Ruth. Turn them to snotty-nosed child avatar.
[17:03]	Kiptic Horsley:	:D
[17:03]	Moundsa Mayo:	With fixed display name Script Kiddie
[17:03]	Gooden Uggla:	BTW, people were seeing ruth again last weekend
[17:03]	Pim Peccable:	Ruth is too iconinc already and dhild avs don't need more bad press
[17:03]	Simon Linden:	ah, lots of fun ideas here :) I have to run today ... thanks everyone for coming
[17:03]	reddot99 Republic:	that and make sure it is persistent, and survives relogs,
[17:04]	Jonathan Yap:	Bye Simon
[17:04]	Rex Cronon:	tc simon
[17:04]	Moon Metty:	have a good weekend :)
[17:04]	Kiptic Horsley:	thanks Lindens
[17:04]	Psi Merlin:	Thanks Simon
[17:04]	Gooden Uggla:	later simon
[17:04]	Pim Peccable:	Phil's Old Av?
[17:04]	Andrew Linden:	Yes, I have to go too. Need to bicycle home before it gets dark.
[17:04]	Falcon Linden:	Ciao
[17:04]	Pauline Darkfury:	thanks, Simon :)
[17:04]	Simon Linden:	I'm on vacation next Tuesday, so see you in a week
[17:04]	Moundsa Mayo:	Thanks Simon
[17:04]	Pauline Darkfury:	and Andrew :)
[17:04]	Pim Peccable:	Be Careful Andrew
[17:04]	Rex Cronon:	tc andrew
[17:04]	Simon Linden:	bye
[17:04]	Andrew Linden:	Have a good weekend.
[17:04]	Moundsa Mayo:	And Andrew
[17:04]	Moon Metty:	ah, Simon on vacation, you don't hear that often
[17:04]	Psi Merlin:	Tc Andew and Falcon
[17:04]	Moon Metty:	have fun!
[17:04]	Moundsa Mayo:	AND FAlcon!
[17:04]	Rex Cronon:	enjoy your time off:)

Generated with SLog Wikifier