Chat log from the meeting on 2018-01-30

From OpenSimulator

Jump to: navigation, search

[10:46] Ubit Umarov: im breaking xinventory etc
[10:47] Bill.Blight wonders how much more broken inventory could be, but I'm sure Ubit will find a way
[10:48] Ubit Umarov: but htink version of folders is broken
[10:48] Ubit Umarov: just not sure if any uses it really
[10:49] Ubit Umarov: fact is that i do see inv fetches that do not happen at sl
[10:49] Ubit Umarov: just after log with clean cache, you can check inv items and there is not fetch..
[10:50] Ubit Umarov: you relog again and there are fetches
[10:50] Ubit Umarov: at sl no...
[10:50] Ubit Umarov: thats version control
[10:50] Ubit Umarov: we where not even sending version on all cases
[10:50] Ubit Umarov: and we have it as ushort and it is a int
[10:50] Ubit Umarov: bahhh
[11:03] Ubit Umarov: btw kay appdomaind do make a dif on mono 5, unlike what you said
[11:04] Kayaker Magic: Is search down? I can't get it to find someone I want to invite, or to find anyone for that matter.
[11:04] George Equus: Problem with Blamgates solved for me, added a hg. in front of destination. some change in new server...
[11:09] Andrew.Hellershanks Bill is a shadow of his former self because of the complexity of his avatar.
[11:12] Andrew.Hellershanks This past week hasn't been too product on most fronts. I did finish most of the wiring I needed to ressurect an old electronics project.
[11:13] James.atLLOUD electronics? how 1980's lol
[11:13] Andrew.Hellershanks James, :) If it wasn't for electronics we wouldn't have computers and wouldn't have this grid. :)
[11:14] James.atLLOUD (my soldering iron has gotten quite dusty)
[11:14] Kayaker Magic: I have a question about scripts on crossings: When I ride a vehicle across a border, the script that moved the vehicle usually survives now, but HUDs and other scripts often have problems on crossings. Any hope for improvement on this?
[11:14] Andrew.Hellershanks My project is from 1980. :)
[11:14] Andrew.Hellershanks @hg.osgrid.orgAndrew.Hellershanks looks at Ubit for a response to Kayakers question
[11:15] Andrew.Hellershanks @hg.osgrid.orgAndrew.Hellershanks wonders if Ubit is awake
[11:16] Kayaker Magic: He said something earlier about appdomain. I thought that mostly had an effect on Win .NET systems, but not on Linux.
[11:16] Andrew.Hellershanks It has an effect on Linux too.
[11:17] Bill.Blight yes it does
[11:17] Sheera Khan: it has an effect on Linux boxes ... quite a bit of an effect^^
[11:17] Kayaker Magic: What is the recommended value for appdomain now? FALSE?
[11:18] Andrew.Hellershanks Yes. The script runtime is affected greatly by the setting of appdomain
[11:19] Andrew.Hellershanks Kayaker, IIRC, setting it to true is better for memory management but you take a hit on script performance.
[11:20] Ubit Umarov: huds do have issues.. because of taken controls
[11:21] Ubit Umarov: the order the take is reapplied may break them
[11:21] Andrew.Hellershanks Ubit, what about HUDs that don't take any controls
[11:21] Ubit Umarov: i did tried to improve that in releation to a mantis...
[11:22] Ubit Umarov: well then i need a repo case to check :)
[11:22] Kayaker Magic: I have a HUD that does not take controls, but it does have llListens, I'm considering re-listening on every crossing.
[11:23] Kayaker Magic: And I have one HUD that just stops running on some crossings. It cannot be reset. The only way to get it working again is to re-compile it.
[11:24] Ubit Umarov: if you can try make a simple case that repos it , to easy testing
[11:24] Ubit Umarov: and mantis it :)
[11:25] Kayaker Magic: LOL, yes that is the challenge, to find a way to repeat the failure every time.
[11:26] Ubit Umarov: better you than me :p
[11:26] Ubit Umarov: hihiih
[11:26] Andrew.Hellershanks @hg.osgrid.orgAndrew.Hellershanks nods
[11:27] Andrew.Hellershanks Trying to repro a problem where the script needs resetting on a crossing would be "interesting".
[11:27] Ubit Umarov: appdomains is simple.. DO NOT use
[11:28] Kayaker Magic: I checked on my INI file, it is false.
[11:28] Ubit Umarov: if memory goes up bc of dead scripts, just restart the instance
[11:28] Ubit Umarov: whats the state now
[11:28] Andrew.Hellershanks So take the hit on memory
[11:29] Ubit Umarov: appdomains jsut make scripts 2 damm slow ( and do use more memory )
[11:29] Ubit Umarov: thats a open opensim issue..
[11:30] Andrew Hellershanks: I wish I knew more about XEngine. It doesn't make sense to me that you can't have a way to keep scripts separate while not affecting performance.
[11:31] Kayaker Magic: How about mantis 8273 CHANGED_TELEPORT is triggered on some border crossings? That one looks straightforward, any progress?
[11:33] Andrew Hellershanks: My computer performance sucks when I have a viewer running. I can't check the mantis right now.
[11:33] Ubit Umarov: missed that mantis.. well i don't know either :(
[11:33] Ubit Umarov: so still open :)
[11:34] Ubit Umarov: guess changed_region should trigger but not TP one on crossings
[11:35] Kayaker Magic: It pays to come here and make noise about mantuses!
[11:35] Andrew Hellershanks: Kayaker, having the CHANGED message trigger on a border crossing would help you with your script problem on a region crossing.
[11:36] Ubit Umarov: hmm and there is code for that really
[11:36] Ubit Umarov: odd
[11:36] Kayaker Magic: The HUD problem is that when the script stops running, it can no longer respond to changed messages either. So putting in a changed event to catch the crossing doesn't help.
[11:37] Andrew Hellershanks: Kayaker, it does.
[11:37] Kayaker Magic: You don't understand: The script stops getting changed events, and touch events, or timer events. It is NOT RUNNING ANY MORE>
[11:38] Andrew Hellershanks: Kayaker, it won't help if the script stops running but it would help in the other cases.
[11:38] Bill.Blight That is odd, I have not seen that behavior , I have some huds that will reset on crossings but have not seen one stop all together in a loooooonnnnggg time
[11:38] Andrew Hellershanks: Kayaker, are you crossing between regions running different versions of OS when the scripts stop?
[11:39] Bill.Blight Or instances with different appdomain settings?
[11:39] Kayaker Magic: All running 0.9.1 of some recent vintage.
[11:40] Kayaker Magic: I'll check that appdomain setting.
[11:40] Ubit Umarov: recent vintage ??
[11:40] George Equus: :)
[11:40] Andrew Hellershanks: I'm also going to assume that you aren't crossing in to a region with scripting disabled.
[11:40] Kayaker Magic: Recent vintage meaning built from the git hub since the release
[11:41] Ubit Umarov: oops vintage ok
[11:41] Kayaker Magic: No, the vehicle script keeps working, only the HUD stops.
[11:42] Ubit Umarov: err i did mantis a issue on change if i remember
[11:42] Ubit Umarov: we may need a & on same cases
[11:42] Kayaker Magic: Doesn't always happen, doesn't happen everywhere. Difficult to diagnose....
[11:42] Bill.Blight Huds being attachments, sounds like could be something in the attachment code not restarting/resetting attachments ..
[11:43] Bill.Blight maybe not directly related to the script engine
[11:43] Kayaker Magic: Hmm, what other scripts participate in the crossing? Scripts in child prims that are not moving the vehicle could be tested.
[11:43] Bill.Blight a way to test could be attach the hud to the vehicle, and see if it fails when not attached to the avatar
[11:43] Kayaker Magic: exactly
[11:45] Ubit Umarov: [11:44] Object: 256
[11:45] Ubit Umarov: ...
[11:45] Kayaker Magic: What happens different when a script is reset and when a scripted object is attached? I state_entry called in both cases?
[11:45] Ubit Umarov: [11:45] Object: 32768
[11:45] Object: 256
[11:45] Object: 16384
[11:46] Ubit Umarov: well the hud survives crossings
[11:46] Ubit Umarov: hmm funny windows repos one changed, linux repos 2
[11:47] Andrew Hellershanks: Does XEngine treat scripts in a HUD differently from other scripts? I wouldn't think it should need to other than a script for a HUD can get a couple of extra signals.
[11:47] Kayaker Magic: It gets attached() events.
[11:47] George Equus: Kayaker, sounds like an old car...
[11:48] Ubit Umarov: err one is 0x8000
[11:48] Kayaker Magic: I'll put that in my signature file: "OpenSim is like an old car"
[11:48] Andrew Hellershanks: Kayaker, right. Couple of signals that other scripts don't receive but I would think there shoudn't be any other difference.
[11:48] Ubit Umarov: other is 0x100
[11:48] Ubit Umarov: hmm what are those hmmm hmm
[11:50] Kayaker Magic: Yeah, there are a couple of poorly documented bits that are sent to changed()
[11:50] Kayaker Magic: I'm starting to see a lot of chat lag here.
[11:50] Ubit Umarov: ok 0x100 is changed region
[11:51] Ubit Umarov: 0x8000 is the non standard nonsense changed_position
[11:51] Kayaker Magic: right.
[11:51] Ubit Umarov: kay my basic hud seems working fine
[11:52] Ubit Umarov: [11:47] [Grid: The region you have entered is running a different simulator version.
Current simulator: OpenSim Snail Dev 4381f16 (Unix/Mono)
Previous simulator: OpenSim Snail Dev (Win/.NET)]
[11:47] Object: 32768
[11:47] Object: 256
[11:47] Object: 16384
[11:52] Ubit Umarov: [11:52] [Grid: The region you have entered is running a different simulator version.
Current simulator: OpenSim Snail Dev (Win/.NET)
Previous simulator: OpenSim Snail Dev 4381f16 (Unix/Mono)]
[11:52] Object: 256
[11:52] Object: 16384
[11:55] Ubit Umarov: that looks ok
[11:56] Kayaker Magic: If I recall, xengine translates LSL into C# and runs that, is there a way to compile the C# into byte codes and run that?
[11:56] Ubit Umarov: hmm just me or this region has lag ?
[11:56] Ubit Umarov: lag..
[11:57] Ubit Umarov: yes.. lsl -> c# -> normal .net assembly
[11:59] Ubit Umarov: yes.. lsl -> c# -> normal .net dll assembly
[11:59] Andrew Hellershanks: Ubit, the lag monster keeps following you around from region to region.
[11:59] Andrew Hellershanks: There is some chat lag here. :P
[11:59] James atLLOUD: The lag seems internet wide in my region.
[12:00] George EquusGeorge Equus checking Kayakers profile  :)

Personal tools
About This Wiki