Content Creation/Scripting User Group/Transcripts/2011 09 19

From Second Life Wiki
Jump to navigation Jump to search

List of Speakers

Ape Piaggio
Cale Flanagan
flexi campfire
Leonel Iceghost
Liisa Runo
Sera Lok
Stickman Ingmann
Techwolf Lupindo


[08:56] Koli DeAllura Contepomi (sahkolihaa.contepomi): Morning, Kelly.

[08:56] tehKellz (kelly.linden): good morning.

[08:56] Ima Mechanique (nielarcher): hi Kelly

[08:56] Liisa Runo: afternoon

[08:56] Kaluura (kaluura.boa): Evening...

[08:56] Koli DeAllura Contepomi (sahkolihaa.contepomi): Or well afternoon, heh

[08:57] Liisa Runo: yea, actually it is evening here

[08:57] tehKellz (kelly.linden):

[08:58] Liisa Runo: is it that time of the year already?

[08:58] tehKellz (kelly.linden): Apparently so

[08:58] Koli DeAllura Contepomi (sahkolihaa.contepomi): Oh dear.

[09:00] tehKellz (kelly.linden): Small group today. Give it a few more minutes for people to wander in.

[09:04] tehKellz (kelly.linden): All right, lets get started then.

[09:04] tehKellz (kelly.linden): Hello.

[09:04] Sera Lok: hi^^ hi everyone

[09:04] Ape Piaggio: hello

[09:05] Cale Flanagan: hi

[09:05] Stickman Ingmann: 'Scuse me.

[09:05] Cale Flanagan: thanks we alreadyhave fire:)

[09:06] Techwolf Lupindo: I got the food on a stick here...:-)

[09:06] Kaluura (kaluura.boa): =^_^=

[09:06] tehKellz (kelly.linden): For some reason I just can't find the wiki page for this user group this morning

[09:07] tehKellz (kelly.linden): got it

[09:07] tehKellz (kelly.linden):

[09:07] tehKellz (kelly.linden): anyway! News first.

[09:08] tehKellz (kelly.linden): "DRTSIM-92" (server maintenance) seems to be doing okay in RC, but I don't think is getting promoted this week. I think the mesh maintenance project is.

[09:08] tehKellz (kelly.linden): 92 is on Le Tigre now and I think Blue Steel is what will be promoted.

[09:09] tehKellz (kelly.linden): Andrew's script performance work is in DRTSIM-96 which is on deck for an RC this week or next, I'm not sure.

[09:10] tehKellz (kelly.linden): While not technically part of this meeting's topic, there is a fix for 'timewarps' that we are testing this week on the beta grid I believe. A kernal upgrade. Hopefully that goes well.

[09:10] Koli DeAllura Contepomi (sahkolihaa.contepomi): :o

[09:11] Kaluura (kaluura.boa): I'm still wondering what those time warp are...

[09:11] Kaluura (kaluura.boa): +s

[09:11] tehKellz (kelly.linden): That is about all the release news I have. Aside from Andrew's changes nothing immediately pending for scripting.

[09:11] tehKellz (kelly.linden): The whole host seems to stall.

[09:11] tehKellz (kelly.linden): They aren't fun.

[09:12] Ima Mechanique (nielarcher): Kaluura, they take a jump to the right...

[09:12] Kaluura (kaluura.boa): Never seen...

[09:12] Techwolf Lupindo: I've been tp to regines that had one or two agents, good .99 TD and then suddelly start spiking in UDP traffice sent, buying stuff gets delayed and so on, but TD remains .99. Starnge.

[09:12] tehKellz (kelly.linden): It is kinda rare but particular to some hosts which may have them frequently.

[09:13] tehKellz (kelly.linden): On Friday I built another HTML Hud demo:

[09:13] tehKellz (kelly.linden): This one uses javascript, ajax/jquery and html5/canvas

[09:13] tehKellz (kelly.linden): It is more of a base for something useful than really useful by itself.

[09:14] Kaluura (kaluura.boa): Is it reasonable to think that one day we'll be able to use semi-transparent background for MoaP?

[09:14] tehKellz (kelly.linden): But pretty small. Updates in real time (~ 2hz)

[09:15] tehKellz (kelly.linden): Kaluura: I dunno. I don't think it is being actively worked on, and it would be a viewer feature.

[09:15] tehKellz (kelly.linden): would be nice though.

[09:15] Kaluura (kaluura.boa): Very useful for HUDs...

[09:15] tehKellz (kelly.linden): I agree.

[09:16] tehKellz (kelly.linden): I'm eager to find other html huds, and to encourage their creation if I can.

[09:16] Liisa Runo: got any comment on SCR-199 ?

[09:16] flexi campfire:

[#SCR-199] Reduction/removal of llCastRay limitations

[09:17] tehKellz (kelly.linden): Falcon took a look at that. "server-maintenance next" has adjustments that allow more cpu time for cast rays.

[09:17] Liisa Runo: good

[09:17] tehKellz (kelly.linden): That one isn't ready for internal QA yet though so it will be a bit before it sees a release

[09:17] Liisa Runo: frien also said he found a bug in ray, but i have not found time to look at it yet, ill JIRA it if it is real bug

[09:17] tehKellz (kelly.linden): k

[09:17] Kaluura (kaluura.boa): Apart from some combat sims, it's not as if a ton of ppl will use it all at the same time.

[09:18] Kaluura (kaluura.boa): Who came with this idea of limitation?

[09:19] Liisa Runo: when i use rays, i will need several rays on every cycle, for example 9 rays for a critter to wander around, or 4 rays for NPV car

[09:20] tehKellz (kelly.linden): Some limits are necessary else you could eat up all the script time with cast rays from very few scripts.

[09:21] tehKellz (kelly.linden): I think he erred on the side of restrictive so that we could adjust them later with less risk of breaking content.

[09:21] Techwolf Lupindo: And that new how?

[09:21] Techwolf Lupindo: Sercially, there are more CPU hogging commands then cast ray that have less restrictions on them.

[09:21] Kaluura (kaluura.boa): "he"? Names, we want names! :P

[09:22] Stickman Ingmann: Sounds like it was Falcon himself. Which seems odd.

[09:22] Stickman Ingmann: I'll get the lynching sticks. We can start before we have facts.

[09:22] tehKellz (kelly.linden): Actually there aren't more cpu hogging commands.

[09:22] tehKellz (kelly.linden): Depending on how you use llCastRay

[09:23] tehKellz (kelly.linden): It will be better. :) I just reviewed the change this morning.

[09:23] Liisa Runo: it will be fine

[09:23] Kaluura (kaluura.boa): I think that most of the time ppl will just want the first hit.

[09:23] tehKellz (kelly.linden): In that case you should use almost no time and you should be able to do lots of those.

[09:24] tehKellz (kelly.linden): The idea with this limit is to distinguish between the easy and hard calls.

[09:24] tehKellz (kelly.linden): Rather than some hard # of calls limit that would have to take the expensive ones into account

[09:24] tehKellz (kelly.linden): It is too restrictive now, but that will be fixed. If the next adjustment isn't enough we will adjust again.

[09:25] Kaluura (kaluura.boa): OK... We'll keep on complaining until the limits are lifted. :P

[09:25] tehKellz (kelly.linden): All right.

[09:25] tehKellz (kelly.linden): Any other super hot issues before triage?

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

[09:26] Kaluura (kaluura.boa): Let's triage!

[09:26] tehKellz (kelly.linden): Apparently thats the first thing on triage.

[09:26] Liisa Runo: ^^

[09:26] tehKellz (kelly.linden): I've bugged lots of people. Is it fixed on BlueSteel do you know?

[09:27] Liisa Runo: good to hear

[09:27] Liisa Runo: next

[09:27] tehKellz (kelly.linden): Hrm. No I don't see anything for it in bluesteel

[09:28] Object: Hello, Avatar!

[09:29] Object: Bienvenue à bord

[09:29] tehKellz (kelly.linden): Alright I acknowledged it, imported it and I will bug more people about it.

[09:30] tehKellz (kelly.linden): SCR-172

[09:30] flexi campfire:

[#SCR-172] need scripts to compile with accented character strings

[09:30] Kaluura (kaluura.boa): Can't reproduce. Must be OS specific...

[09:30] Object: Hello, Avatar!

[09:30] Kaluura (kaluura.boa): ...or viewer specific.

[09:31] Object: Hello, Avatar!

[09:31] Liisa Runo: i cant repro it eitehr

[09:31] Object: Bienvenue à bord

[09:32] Liisa Runo: reported 2007, maybe it was bugged back then

[09:32] tehKellz (kelly.linden): okay, next.

[09:33] tehKellz (kelly.linden): moon just moved it over so maybe they can still repro it?

[09:33] tehKellz (kelly.linden): SCR-133

[09:33] flexi campfire:

[#SCR-133] Script Info reporting on discontiguous parcels is not working as expected.

[09:33] tehKellz (kelly.linden): arg

[09:33] tehKellz (kelly.linden): that one is a viewer bug.

[09:33] tehKellz (kelly.linden): viewer but!

[09:35] tehKellz (kelly.linden): well I imported it. We do need to fix that.

[09:36] tehKellz (kelly.linden): SCR-115 is next.

[09:36] flexi campfire:

[#SCR-115] State changes considerably slower under Mono-LSL than LSO-LSL

[09:36] Object: 0.022222 seconds/state-change

[09:36] Object: 0.000326 seconds/state-change

[09:36] tehKellz (kelly.linden): Still true. Not sure that is going to get fixed though.

[09:36] Kaluura (kaluura.boa): I've never noticed any spectacular slow downs...

[09:37] tehKellz (kelly.linden): Under the hood state changes are much more involved in mono.

[09:38] Liisa Runo: i noticed with one toy that got around limits with state change spam, but nowdays we dont have those limits anymore, but still tthe state change is slower, and easy to confirm with the provided script

[09:38] tehKellz (kelly.linden): I like this one because it is easy to confirm and I like investigating easy to repro performance issues.

[09:39] tehKellz (kelly.linden): Though I'm skeptical about my ability to do much about it.

[09:39] Kaluura (kaluura.boa): The numbers are impressive...

[09:40] tehKellz (kelly.linden): switching states in LSL is just a couple pointer changes, mono isn't designed for states so it involves self-inspection, looking at all the defined functions and assigning them as event handlers as appropriate.

[09:41] tehKellz (kelly.linden): This is one case where C# (and writing code designed to be C#) would be faster.

[09:41] Cale Flanagan: ok, solution found, introduce c#

[09:42] Techwolf Lupindo: Just about anything would be better the the current LSL. LUA comes to mind.

[09:42] tehKellz (kelly.linden): LUA doesn't meet our concurrency requirements.

[09:43] tehKellz (kelly.linden): SCR-112

[09:43] flexi campfire:

[#SCR-112] Using PARCEL_MEDIA_COMMAND_AGENT with llParcelMediaCommandList() when parcel's media URL is set causes previous video replay

[09:44] tehKellz (kelly.linden): interesting, very specific

[09:45] Techwolf Lupindo: Qucik google shows a lot of reports on LUA concurrency.

[09:45] Techwolf Lupindo: I need to go. bye all.

[09:46] tehKellz (kelly.linden): cya techwolf

[09:47] tehKellz (kelly.linden): Okay I imported that one so normal server triage will see it .

[09:48] tehKellz (kelly.linden): Next is SCR-107.

[09:48] flexi campfire:

[#SCR-107] llTakeControls(0, TRUE, TRUE); triggers unexpected error msg

[09:49] tehKellz (kelly.linden): I'm not sure either. Are they trying to release controls?

[09:49] Liisa Runo: i guess it is just part of an experiment to do wild things and see what happens, sometimes you just have to divide by zero to see if anyting new happens

[09:50] Koli DeAllura Contepomi (sahkolihaa.contepomi): Heh...

[09:50] Liisa Runo: every now and then we find neat haxes this way

[09:51] Ima Mechanique (nielarcher): seems they use it to force the permission aquisition, and specify actual controls desired later

[09:51] Liisa Runo: like pushing self with zero force, makes no sense, but result is good and usefull

[09:52] Kaluura (kaluura.boa): I suppose they are trying to take no control in a HUD just to make the script work on no-script parcels.

[09:53] Kaluura (kaluura.boa): HUD or attachment...

[09:53] tehKellz (kelly.linden): Looks like they thought you had to do something when permissions were granted to hold them. But you don't.

[09:53] tehKellz (kelly.linden): llTakeControls(0,FALSE,TRUE) will work still though.

[09:54] Liisa Runo: or trying to take perms, and then take controls later when needed, without triggering new request to give perms

[09:54] tehKellz (kelly.linden): but you don't need to for that.

[09:54] tehKellz (kelly.linden): I mean you can request perms, then call llTakeControls whenever you want after you have them.

[09:55] Liisa Runo: yea

[09:55] Kaluura (kaluura.boa): NMI...

[09:55] tehKellz (kelly.linden): nmi?

[09:56] Liisa Runo: NeedMoreInfo

[09:56] Kaluura (kaluura.boa): Need More Info

[09:56] Liisa Runo: or just make it emit some silly error message and mark it as resolved

[09:56] Kaluura (kaluura.boa): hehehe

[09:57] Leonel Iceghost: O, 0, 1 should work the same as 0, 1, 1.. I wonder why it is different

[09:58] tehKellz (kelly.linden): Okay. For the last few minutes I'd like some feedback, if you have any, on these triages. Are the useful? Boring? Just biding your time until we get down to the feature request ones?

[09:59] Ima Mechanique (nielarcher): mixture of useful and biding time ;-)

[09:59] Kaluura (kaluura.boa): Useful, some times even fun...

[09:59] Leonel Iceghost: they are good, if it helps to figure out something.. otherwise they have their own discussion

[09:59] Charlie (charlar.linden): sorry, gotta run. Thanks Kelly for doing such a great job owning this group.

[09:59] Liisa Runo: i kinda like the triages, maybe not as intresting as some steamy debate over some specific issue, but gives me the feeling that every JIRA get atleast some attention.

[10:00] Ima Mechanique (nielarcher): although I'd say more biding time for pet entries more than features specifically

[10:00] Stickman Ingmann: I'd like it if they went a little faster. Cover more issues. But if we're going to cover things well, that may not be possible.

[10:00] tehKellz (kelly.linden): I'm gonna try and do my own quick triage before the next public one. I think it will be better if I clear out some of the obvious ones.

[10:01] tehKellz (kelly.linden): Once we get to the bugs with 0 votes we can decide if we want to skip to the feature requests.

[10:01] Liisa Runo: yay, some golden things in feature requests

[10:01] tehKellz (kelly.linden): Thanks everyone for coming.

[10:02] tehKellz (kelly.linden): See you next week.

[10:02] Liisa Runo: and we can actually get there some day, not that many bugs in LSL,

[10:02] Ima Mechanique (nielarcher): thanks kelly

[10:02] Koli DeAllura Contepomi (sahkolihaa.contepomi): See you, Kelly.

[10:02] Liisa Runo: thanks everybody

[10:02] Sera Lok: ty take care