Content Creation/Scripting User Group/Transcripts/2011 03 07

From Second Life Wiki
Jump to: navigation, search

List of Speakers

flexi campfire

Transcript

09:03] tehKellz (kelly.linden): okay. I think we are on triage today

[09:03] Kaluura (kaluura.boa): Calendar is broken because it sorts days alphabetically

[09:03] tehKellz (kelly.linden): https://wiki.secondlife.com/wiki/User:Kelly_Linden/script_jira_triage

[09:03] tehKellz (kelly.linden): Anyone in touch with haravikk?

[09:05] tehKellz (kelly.linden): Maybe if I start messing with his jiras he will magically show up.

[09:06] draconis.neurocam (draconis.neurocam): GodLoginAndTeleport heh

[09:06] tehKellz (kelly.linden): SVC-6432 seems like a no brainer. Only trick is that sometimes the displayname won't be available

[09:06] flexi campfire: https://jira.secondlife.com/browse/svc-6432

[09:06] MarkByron (markbyron.falta): not tiny friendly

[09:07] Kaluura (kaluura.boa): (Use your AO)

[09:07] Liisa Runo (liisa.runo): the sofa can be tiny friendly, just select from the menu to not animate you, will remember your setting so you dont need to do it every week

[09:07] Jonathan (jonathan.yap): Kelly, will you fix your fireplace jira-helper with the one I am about to give you -- the new one knows how to display the Summary correctly

[09:08] tehKellz (kelly.linden): SVC-1432

[09:08] flexi campfire: https://jira.secondlife.com/browse/svc-1432

[09:08] JIRA-helper: http://jira.secondlife.com/browse/SVC-1432

[#SVC-1432] llDetectedNormal() - provide normal vectors (or rotations) for collision events

[09:09] Dantia: ((dantia Gothly has entered the area.))

[09:09] Kaluura (kaluura.boa): Might eventually be useful... tho I don't for what right away...

[09:09] tehKellz (kelly.linden): I'd have to defer to Falcon for this one.

[09:10] draconis.neurocam (draconis.neurocam): would be good for anything really, probably mostly mesh

[09:10] NeoBokrug Elytis (neobokrug.elytis): https://jira.secondlife.com/browse/SVC-4632 ?

[09:10] flexi campfire: [#SVC-4632] People getting past Estate and Land bans

[09:10] tehKellz (kelly.linden): On the LSL side - extending the data returned in llDetected* functions is kind of tricky, unfortunately. It isn't a very extensible system.

[09:11] tehKellz (kelly.linden): Neo: I keep a triage wiki page here: https://wiki.secondlife.com/wiki/User:Kelly_Linden/script_jira_triage feel free to add yours to the end of the new features section.

[09:11] NeoBokrug Elytis (neobokrug.elytis): Thanks

[09:12] tehKellz (kelly.linden): For everyone else - I am working through that page now. :) moving on to the third.

[09:12] tehKellz (kelly.linden): SVC-377

[09:12] flexi campfire: http://jira.secondlife.com/browse/SVC-377

[#SVC-377] LSL functions to access the description of objects in an object's inventory

[09:12] tehKellz (kelly.linden): I wish that was easy.

[09:12] draconis.neurocam (draconis.neurocam): Is that even possible?

[09:12] Kaluura (kaluura.boa): Please, pretty please...

[09:12] tehKellz (kelly.linden): for some definition of possible, yes.

[09:13] tehKellz (kelly.linden): However the description is part of the asset, so we would have to fetch the inventory asset (akin to rezing it) to get that information, I think.

[09:13] draconis.neurocam (draconis.neurocam): dataserver call?

[09:14] tehKellz (kelly.linden): Actually I may be wrong.

[09:15] tehKellz (kelly.linden): I think we could get the description.

[09:15] tehKellz (kelly.linden): We couldn't set it, but we could get it. Is that really useful?

[09:15] NeoBokrug Elytis (neobokrug.elytis): Done and edited.

[09:15] Kaluura (kaluura.boa): That would be a start...

[09:16] draconis.neurocam (draconis.neurocam): isnt that how llRequestInventoryData for landmarks works?

[09:16] tehKellz (kelly.linden): well, it would be an 'end' too probably.

[09:17] Qie (qie.niangao): is the Set problem just as bad for the script's own description? (https://jira.secondlife.com/browse/SVC-2382)

[09:17] flexi campfire: [#SVC-2382] lsl functions llGetScriptDesc() and possibly llSetScriptDesc()

[09:17] tehKellz (kelly.linden): so you wouldn't be able to use it as a datastore. Yes Qie

[09:17] Qie (qie.niangao): okay, never mind. Thanks.

[09:18] Ashur Constantine (ashur.constantine): It'd be useful even without a setter function.

[09:18] tehKellz (kelly.linden): Ashur: can you give me some examples, all the ones in SVC-377 only make sense if you can set it as well I think.

[09:20] tehKellz (kelly.linden): Well, Haravikk makes a good case for storing parameter data for animations in their description.

[09:20] Ashur Constantine (ashur.constantine): I haven't read that JIRA, but I make my own vendor board when I need to sell something for example. A lot of times I'll butcher the objectname with a CSV to say...ItemName,VersionNumber

[09:21] Ashur Constantine (ashur.constantine): and if I could put some of the ugliness in description that'd be one case it'd make sense; where I had to manually set the description once anyway and not have to set it in any automated way.

[09:21] tehKellz (kelly.linden): There are two subtasks .... I'll comment on those specifically

[09:21] Ashur Constantine (ashur.constantine): I could have 5 objects named the same thing, with different metadata like version number.

[09:21] Ashur Constantine (ashur.constantine): alright

[09:21] tehKellz (kelly.linden): SVC-5892 and SVC-5893

[09:21] flexi campfire: http://jira.secondlife.com/browse/SVC-5892

[#SVC-5892] string llGetInventoryDesc(string name)

[09:21] flexi campfire: http://jira.secondlife.com/browse/SVC-5893

[#SVC-5893] llSetInventoryDesc(string name, string description)

[09:22] draconis.neurocam (draconis.neurocam): especially since you cant edit the name of an attachment while its on you and have it saved, so people would have to live with the butchered object name

[09:22] tehKellz (kelly.linden): You can't have identically named items in object inventory I think.

[09:22] Ashur Constantine (ashur.constantine): Right I forgot about that; so that wouldn't change that aspect.

[09:22] draconis.neurocam (draconis.neurocam): it adds numbers to the end i thought

[09:22] tehKellz (kelly.linden): but I like the idea of storing meta data or other properties in the description.

[09:22] Ashur Constantine (ashur.constantine): Yeah it does. My bad.

[09:26] Kaluura (kaluura.boa): (ping)

[09:26] tehKellz (kelly.linden): PONG

[09:26] Jonno Stromfield (jonno.stromfield): Would it be possible to have a new differently handled property that would be settable in an inventory, or would they all suffer from the same problem?

[09:26] tehKellz (kelly.linden): sorry, updating jiras as we go.

[09:26] Kaluura (kaluura.boa): Nope... Not disconnected..

[09:26] tehKellz (kelly.linden): They'd all suffer the same problem.

[09:26] Jonno Stromfield (jonno.stromfield): :( Thought you'd say that.

[09:27] draconis.neurocam (draconis.neurocam): im still curious how it knows the position from landmarks then

[09:27] tehKellz (kelly.linden): draconis: it does an inventory fetch with the dataserver event

[09:27] tehKellz (kelly.linden): that doesn't mean it is a great idea.

[09:27] draconis.neurocam (draconis.neurocam): ah

[09:27] Cummere Mayo (cummere.mayo): sorry for getting here late... is there an agenda?

[09:27] tehKellz (kelly.linden): I think I should hold off on SVC-6300 for Haravikk

[09:27] flexi campfire: http://jira.secondlife.com/browse/SVC-6300

[#SVC-6300] llStartAnimationSynced() - start an animation synchronised with a given timestamp

[09:28] tehKellz (kelly.linden): Working through triage today cummere https://wiki.secondlife.com/wiki/User:Kelly_Linden/script_jira_triage

[09:28] Cummere Mayo (cummere.mayo): ty

[09:28] tehKellz (kelly.linden): So, moving on to SVC-6441

[09:28] flexi campfire: http://jira.secondlife.com/browse/SVC-6441

[#SVC-6441] Allow us to decide which llDetected*() functions are populated to save on memory/processing

[09:29] Jonno Stromfield (jonno.stromfield): er, just a sec

[09:29] Jonno Stromfield (jonno.stromfield): what happened to start animation synched?

[09:29] Jonno Stromfield (jonno.stromfield): of sorry, you said

[09:29] tehKellz (kelly.linden): That is haravikk's baby. He is extremely invested in it - it would be unfair to discuss without him.

[09:29] Jonno Stromfield (jonno.stromfield): missed the comment

[09:30] Jonno Stromfield (jonno.stromfield): just want to say I'm really interested too

[09:31] tehKellz (kelly.linden): Feel free to comment/watch the issue. We'll get back to it some time when he is here.

[09:31] Jonno Stromfield (jonno.stromfield): kk

[09:32] tehKellz (kelly.linden): I think this jira over estimates the memory and processing involved in the llDetected* data

[09:32] Ashur Constantine (ashur.constantine): Would that actually result in a lot of savings? i.e. is llDetectedPos always populated even if I'm only ever getting llDetectedName?

[09:32] Ashur Constantine (ashur.constantine): probably.

[09:33] tehKellz (kelly.linden): That is correct - all the llDetected data is populated for each 'hit'

[09:33] Cummere Mayo (cummere.mayo): dumb question but ... why?

[09:33] tehKellz (kelly.linden): The biggest savings is probably the name - it is the only string value.

[09:33] Kaluura (kaluura.boa): Maybe that would be useful not to populate everything.. from the sim point of view...

[09:34] Qie (qie.niangao): I'd think there'd be some overhead in having variable format event records.

[09:34] NeoBokrug Elytis (neobokrug.elytis): You'd probably need a filter like llCastRay has.

[09:34] Jonathan (jonathan.yap): I would think the number of touch events/second is not going to be very large, or am I missing something?

[09:34] Ashur Constantine (ashur.constantine): stuff like collision() happens a lot, and sensors.

[09:35] draconis.neurocam (draconis.neurocam): even so changing any of how it works could possibly break existing content, so unless it just knew what was needed, it would just be a new set of functions anyway

[09:35] Ashur Constantine (ashur.constantine): but its a question of whether sim changing savings are actually there, doesn't sound like it.

[09:35] tehKellz (kelly.linden): Honestly it is about 100 bytes plust the name.

[09:36] Liisa Runo (liisa.runo): i guess the proper way to tune it is to make the compiler see what llDetected* functions are used when it comples, and the tunes go to the server. So that would not need any changes to the actual LSL code

[09:37] tehKellz (kelly.linden): that would be even trickier.

[09:37] draconis.neurocam (draconis.neurocam): you would have to unroll everything

[09:37] draconis.neurocam (draconis.neurocam): that sounds horribly unfun

[09:37] Jonathan (jonathan.yap): The server would have to keep a lookup list for every script and check it for every event

[09:38] draconis.neurocam (draconis.neurocam): i feel like this is just not worth it.

[09:38] Ashur Constantine (ashur.constantine): doesn't sound like it.

[09:38] Cummere Mayo (cummere.mayo): thats kinda my impression to

[09:38] Cummere Mayo (cummere.mayo): *too

[09:38] LaRPS Meter v0.25: ((dantia Gothly's meter has been detached.))

[09:39] tehKellz (kelly.linden): Ok.

[09:39] Liisa Runo (liisa.runo): all im saying, why should we decide what llDetected* is populatesd when the compiler already see that just fine, it can be automatic, and that way it would also help with old scripts

[09:39] tehKellz (kelly.linden): Because it is more difficult to do that. :)

[09:39] draconis.neurocam (draconis.neurocam): severly so

[09:39] Kaluura (kaluura.boa): Changing LSL btecode is a no-no...

[09:40] Kaluura (kaluura.boa): bytecote*

[09:40] Qie (qie.niangao): maybe the name could be retained internally as a handle, instead of a string, and only dereferenced when llDetectedName is called... if that would save enough mem to be worth the effort.

[09:40] Liisa Runo (liisa.runo): difficult for 3 hours for a linedn vs difficult for the future scripters to the end of the time

[09:40] tehKellz (kelly.linden): 3 hours?

[09:41] draconis.neurocam (draconis.neurocam): no its not the time it would take Liisa its the extra strain on everything it would place.

[09:42] tehKellz (kelly.linden): Ok, next

[09:42] tehKellz (kelly.linden): SVC-6532

[09:42] flexi campfire: http://jira.secondlife.com/browse/SVC-6532

[#SVC-6532] agent() event for detecting changes in avatar state, ideal for animation overriders

[09:42] Liisa Runo (liisa.runo): i dont see the difference in automated thing vs user settable thing, both create the same strain when being run on server, only difference is when the script get compiled

[09:43] draconis.neurocam (draconis.neurocam): agent() would be great

[09:43] Liisa Runo (liisa.runo): yes

[09:43] Kaluura (kaluura.boa): Yep!

[09:43] Qie (qie.niangao): yeah, this *would* be great

[09:43] Dantia (dantia.gothly): yeesss

[09:44] Kaluura (kaluura.boa): That or just a function to replace the default anims...

[09:44] Liisa Runo (liisa.runo): that

[09:44] draconis.neurocam (draconis.neurocam): its useful for more than animations though

[09:44] Liisa Runo (liisa.runo): lot more

[09:44] tehKellz (kelly.linden): Kaluura - yeah I'd love to see animation overrides more fully supported

[09:45] Cummere Mayo (cummere.mayo): okay so just to make crystal clear the only thing these could be used for is to better time and clean up animations?

[09:46] draconis.neurocam (draconis.neurocam): no, it could be used for objects changing anything about them if an agent was doing x

[09:46] draconis.neurocam (draconis.neurocam): theres more than just animations that this is useful for

[09:46] tehKellz (kelly.linden): The permissions issue isn't touched on here - who should a script be able to observe?

[09:46] Ashur Constantine (ashur.constantine): I think it'd be particularly useful if agent() had change flags for things we poll for in llGetAgentInfo

[09:46] draconis.neurocam (draconis.neurocam): probably who is worn or sat upon kelly

[09:47] Qie (qie.niangao): hmmm. perhaps a new permission, auto-granted for attachments and sat-upon things.

[09:47] NeoBokrug Elytis (neobokrug.elytis): Owners Only.

[09:47] Ashur Constantine (ashur.constantine): Could it bet setup to work like sensor? llSetAgentEvent(key agent)?

[09:47] Liisa Runo (liisa.runo): just make it use the same perms system everything else already have, autogrant when attached or sat and dialog grant otherwise

[09:47] Cummere Mayo (cummere.mayo): yeah... i -could- see where like huggers/sex huds could use it too.... but in that case it would need to trigger a dialogue

[09:49] Qie (qie.niangao): (just in passing: those skirt-sitter scripts, whatever they're called, also sample agent state frequently)

[09:49] Ashur Constantine (ashur.constantine): The use case I'm thinking of is say I want to detect if someone is flying, who's just in my sim. If I could attach an agent event to them and get a change flag of AGENT_FLYING, I could do something.

[09:49] tehKellz (kelly.linden): Well, is the pemissions needed? I don't think llGetAgentInfo and llGetAnimation require any permission to check

[09:49] Ashur Constantine (ashur.constantine): Or an AWAY, etc.

[09:49] NeoBokrug Elytis (neobokrug.elytis): THat's a good point, since it doesn't act upon an avatar.

[09:49] draconis.neurocam (draconis.neurocam): yeah

[09:50] Kaluura (kaluura.boa): Yes... No perms... That's even better

[09:50] Qie (qie.niangao): I think that's correct. no need for perms.

[09:50] Liisa Runo (liisa.runo): someone will cry about privacy, but yea i agree, no need to keep those secret. no need for permission check

[09:50] tehKellz (kelly.linden): This is a fairly large undertaking

[09:50] Cummere Mayo (cummere.mayo): llgetagentinfo is sometimes abused though isnt it?

[09:50] draconis.neurocam (draconis.neurocam): this is just a replacement for calling those anyway with some other form of call

[09:51] Kaluura (kaluura.boa): Ppl are already cryiing about privacy because we have radars who can say where they are...

[09:51] Qie (qie.niangao): don't see how, Cummere. although llRequestAgentInfo() is controversial because it gets online status remotely.

[09:51] Ashur Constantine (ashur.constantine): Well its just change events.

[09:51] Cummere Mayo (cummere.mayo): because of svc-4823 and others...

[09:51] flexi campfire: http://jira.secondlife.com/browse/SVC-4823

[#SVC-4823] block online status at server level

[09:51] NeoBokrug Elytis (neobokrug.elytis): Well, since Web Profiles don't show if people are on or offline...

[09:52] draconis.neurocam (draconis.neurocam): or since you can entirely stop your web profile from being seen

[09:52] Cummere Mayo (cummere.mayo): sorry but agent info and anything that replaces it should trigger a dialogue imo

[09:52] NeoBokrug Elytis (neobokrug.elytis): As an estate owner, it's hard to try to contact people if they all appear offline.

[09:52] draconis.neurocam (draconis.neurocam): indeed

[09:53] Cummere Mayo (cummere.mayo): you drop them an offline

[09:53] Cummere Mayo (cummere.mayo): problem solved

[09:53] Rhiannon (sista.mixemup): agrees with Cummere

[09:53] NeoBokrug Elytis (neobokrug.elytis): Not everyone forwards IMs.

[09:53] Cummere Mayo (cummere.mayo): so?

[09:53] NeoBokrug Elytis (neobokrug.elytis): ANd some things are time sensitive.

[09:53] draconis.neurocam (draconis.neurocam): like whether or not a sim is kept or sold

[09:53] Cummere Mayo (cummere.mayo): so drop them a notecard

[09:53] Cummere Mayo (cummere.mayo): even capped ims that is always forwareded

[09:53] Qie (qie.niangao): but two very different things, GetAgentInfo and RequestAgentData... unless we're thinking of agent() events seeing online state transitons, too.

[09:53] F L I P (flip.idlemind): If you're gonna block online status at the server level you'll have to block sensors as well. If a sensor sees you, you're online.

[09:53] Cummere Mayo (cummere.mayo): even if they dont forward ims

[09:53] Jonno Stromfield (jonno.stromfield): Would what' s being proposed make this worse?

[09:54] NeoBokrug Elytis (neobokrug.elytis): Or parcel, etc.

[09:54] draconis.neurocam (draconis.neurocam): no this evolved into something else, this would not change anything

[09:54] Kaluura (kaluura.boa): It would change anything

[09:54] draconis.neurocam (draconis.neurocam): its gotten off topic actually

[09:54] Jonno Stromfield (jonno.stromfield): That's what I thought, Drac

[09:54] Kaluura (kaluura.boa): wouldn't*

[09:54] draconis.neurocam (draconis.neurocam): it would be an improvement

[09:54] NeoBokrug Elytis (neobokrug.elytis): I guess the problem is I don't understand why it's a bit deal to hide if you're online or not.

[09:54] draconis.neurocam (draconis.neurocam): to have agent()

[09:55] draconis.neurocam (draconis.neurocam): so a timer could be free from polling other functions

[09:55] draconis.neurocam (draconis.neurocam): or anything else for that matter

[09:55] Cummere Mayo (cummere.mayo): yeah... but it needs to be restricted only to the owner or trigger a dialogue when its not (except when sat upon)

[09:55] Qie (qie.niangao): unless agent() events are raised for online status changes, I don't see the point... and that's not realistic, if I understand the presence system at all.

[09:55] Ashur Constantine (ashur.constantine): My point was that we already have llGetAgentInfo(), but to get the key we usually get it through other events like collision, touch, sensors etc.

[09:56] Ashur Constantine (ashur.constantine): So it could make some sense to set an agent key with 1 function and have that agent() event work thusly.

[09:56] draconis.neurocam (draconis.neurocam): it would only be per agent anyway, so it would be tied to a key

[09:56] Jonno Stromfield (jonno.stromfield): only on the current sim (and possibly the immediate surrounds)?

[09:57] Cummere Mayo (cummere.mayo): definately jonno

[09:57] Ashur Constantine (ashur.constantine): Same constraints as llGetAgentInfo; we can already get that stuff.

[09:57] Qie (qie.niangao): nice demo, flip. A plywood box is worth a thousand words. :D

[09:57] Ashur Constantine (ashur.constantine): Its just how we get it.

[09:57] Cummere Mayo (cummere.mayo): ild argue MORE constraints ashur

[09:57] tehKellz (kelly.linden): Ok, I think that is all we have time for today.

[09:57] tehKellz (kelly.linden): Any last minute topics?

[09:57] Cummere Mayo (cummere.mayo): and that box is a great example of why I would argue for more constraints

[09:58] Kaluura (kaluura.boa): llIntegerToHex()... Any chance?

[09:58] Jonathan (jonathan.yap): The proposal is only to do something more efficiently

[09:58] tehKellz (kelly.linden): Kaluura no not really

[09:58] Ashur Constantine (ashur.constantine): Not sure what you mean but llGetAgentInfo only shows us what we can see plain as day with our eyes; such as if someone is AFK, its not a privacy breech.

[09:58] Cummere Mayo (cummere.mayo): ild still like to see some-thing done with svc-4823 kelly

[09:58] Kaluura (kaluura.boa): Bleh...

[09:59] tehKellz (kelly.linden): Cummere, I know. I definitely don't have time to get into that today

[10:00] Cummere Mayo (cummere.mayo): next time?

[10:00] tehKellz (kelly.linden): It isn't on my triage list, but I'm also not sure how applicable it is to my triages

[10:00] Jonathan (jonathan.yap): Wil the remaining items carry over to next week

[10:00] Liisa Runo (liisa.runo): I know, when people disable their online visibility, automatically TP them to cornfield (or similar) and block all commun iocations and everything, so they can be very alone and in privacy

[10:00] draconis.neurocam (draconis.neurocam): heh

[10:00] tehKellz (kelly.linden): jonathan yup that is how it works. Well, in two weeks. Next week is freeform discussion

[10:00] Jonathan (jonathan.yap): Liisa, look at my profile picks

[10:00] MarkByron (markbyron.falta): there's plenty of valid uses for online status such as customer service in stores and so forth

[10:01] Liisa Runo (liisa.runo): (look at my picks)

[10:01] Cummere Mayo (cummere.mayo): mark i run a store, i dont need to know when a person is online... dropping them a notecard works well

[10:01] tehKellz (kelly.linden): Ok, thank you everyone for coming. See you all next week. :)

[10:01] draconis.neurocam (draconis.neurocam): Have a good day kelly.

[10:01] Ashur Constantine (ashur.constantine): thanks kelly

[10:01] Ashur Constantine (ashur.constantine): see you next week.

[10:01] Rhiannon (sista.mixemup): thank you Kelly

[10:01] Jonno Stromfield (jonno.stromfield): ty

[10:01] Qie (qie.niangao): Thanks Kelly!

[10:01] Cummere Mayo (cummere.mayo): ty kelly