Chat log from the meeting on 2014-05-13

From OpenSimulator

Jump to: navigation, search
[11:06]  Andrew Hellershanks: The fix for event listings does involve both the ossearch module and the query.php file.
[11:06]  Nebadon Izumi: hey heres Justin
[11:06]  Nebadon Izumi: ok
[11:06]  Richardus Raymaker: hi justin
[11:07]  Andrew Hellershanks: I need to double check things again. This business of determining what is expected as a separate in the landing point coordinates is a pain.
[11:07]  Andrew Hellershanks is Online
[11:07]  dan banner: hey justin
[11:07]  Mata Hari: hi Justin
[11:08]  Andrew Hellershanks: I wanted to use / for consistency with other tables and I thought it was working on my machine but isn't working for showing events on the map.
[11:08]  dan banner: seem to be a few issues with timeouts
[11:08]  Andrew Hellershanks: Justin, I answered your question in the bug report about where to find an offline PHP file.
[11:08]  Justin Clark-Casey: thanks andrew
[11:09]  Andrew Hellershanks: smxy replied to my reply making it seem like he didn't realize I was answering a question so I amended my comment to clarify it was a reply.
[11:09]  Nebadon Izumi: ya looks like maybe some issues with on of your very recent patches Justin
[11:09]  Nebadon Izumi: I have not experienced it, but Dan and Mata have it seems
[11:10]  dan banner: ya i updated a plaza as soon as it was mentioned to me
[11:10]  Nebadon Izumi: http://opensimulator.org/viewgit/?a=commitdiff&p=opensim&h=bbc1dc6bceb18cc1970109c5069724fba84e9c83
[11:10]  dan banner: had same/similar issue
[11:10]  Mata Hari: Dan happened to come online right after I made my new build so he tried it on a plaza to see if he got the same issue
[11:10]  Mata Hari: (and did)
[11:10]  Nebadon Izumi: oops that was wrong link sort of
[11:10]  dan banner: http://opensimulator.org/mantis/view.php?id=7165
[11:10]  Nebadon Izumi: http://opensimulator.org/mantis/view.php?id=7165
[11:10]  Nebadon Izumi: ya thats what i meant to paste
[11:11]  Dahlia Trimble is Online
[11:12]  dan banner: its like instant errors on startup
[11:12]  Mata Hari: yep
[11:12]  Mata Hari: also on login
[11:12]  Vivian Klees: tp's
[11:13]  Justin Clark-Casey: ok, that's surprising. I will look into it
[11:16]  Simulator Version v0.5 shouts: OpenSim 0.8.0 Dev          73891c7: 2014-05-09 11:06:33 -0700 (Unix/Mono)
[11:17]  Justin Clark-Casey: hi robert
[11:17]  Richardus Raymaker: hi robert
[11:17]  Dahlia Trimble: hi
[11:17]  dan banner: hi
[11:17]  Mata Hari: hi Dahlia
[11:17]  Daisy Duke: hi everyone!
[11:18]  dan banner: hello
[11:18]  Robert Adams: hi all
[11:19]  Nebadon Izumi: so anyone else have anything they wanted to bring up this week?
[11:19]  Richardus Raymaker: hi dahlia
[11:19]  vegaslon plutonian: so I hear complaints about people unsitting and being halfway in the floor. Has this been behavior all along or is it recent?
[11:19]  Vivian Klees: script/error warning with no ext?
[11:19]  Vivian Klees: text
[11:19]  Nebadon Izumi: its probably been around a while vegaslon
[11:20]  Nebadon Izumi: thats why i made this stool at last weeks meeting
[11:20]  Nebadon Izumi: it gives upwards thrust when you stand
[11:20]  Nebs Metal Bar Stool v1.5 (w/sit & launch): Goodbye..
[11:20]  Nebs Metal Bar Stool v1.5 (w/sit & launch): Hello Nebadon Izumi, enjoy your sit..
[11:20]  Mata Hari: oh....ground-sit when prim is elevated.....you sit on the ground, not the prim
[11:20]  Nebs Metal Bar Stool v1.5 (w/sit & launch): Goodbye..
[11:20]  Nebs Metal Bar Stool v1.5 (w/sit & launch): Hello Nebadon Izumi, enjoy your sit..
[11:20]  Richardus Raymaker: hold on mata
[11:20]  vegaslon plutonian: is this a problem that came with bulletsim or has people seen it in ode?
[11:20]  Justin Clark-Casey: I did want to bring up Mantis 7164
[11:21]  Richardus Raymaker: this one mata ? http://opensimulator.org/mantis/view.php?id=6813
[11:21]  Mata Hari: yes
[11:21]  Andrew Hellershanks: I'm hearing that a 0.8 release candidate was issued. People have been asking me about 0.8
[11:21]  Justin Clark-Casey: aren't you following the mailing lists?
[11:21]  Mata Hari: happened to me the other day with FS
[11:21]  Richardus Raymaker: only know 0.8rc1
[11:21]  Andrew Hellershanks: I thought I was on the mailing list. I haven't seen anything in a while.
[11:22]  dan banner: ya smaller the step the harder it is to get up
[11:22]  Dev Random: I've had another occurrence of Collision events stopping firing too.
[11:22]  Justin Clark-Casey: robert: still mad at me about the terrain stuff? :)
[11:22]  Dev Random: Cant' reliably reproduce it, but I have a simple script where it occurs (yesterday it took 12 hours or so).
[11:22]  Robert Adams: another.... argh
[11:22]  Andrew Hellershanks wonders if his subscription got dropped when the mailing list host was changed.
[11:22]  Richardus Raymaker: i have seen the rc1 on mailinglist
[11:22]  Mata Hari: I've tested it with mesh stairs, sculpt stairs and prims stairs....same issue each time
[11:23]  Mata Hari: if I increase the step there's a window where it's fine, then when it gets too large again it runs into trouble
[11:23]  Vivian Klees: what's the width being used?
[11:24]  Mata Hari: 2m x 0.5m step size was the prim set
[11:24]  Richardus Raymaker: that never worked very good 0.5
[11:24]  Mata Hari: no, 0.5 deep, not high
[11:24]  Mata Hari: stair was 2m wide, 0.5m deep, and the 0.15m high
[11:25]  Richardus Raymaker: aah. that must work fine.
[11:25]  Mata Hari: had trouble with 0.15 and 0.2....when I increased to 0.25 it was okay
[11:25]  Mata Hari: and 0.3 is fine
[11:25]  Mata Hari: then again around 0.45-0.5 it has trouble (as you'd expect
[11:26]  Dahlia Trimble is Offline
[11:27]  Mata Hari: had no trouble at all walking up the stairs in this building which look like they may be more than 0.25 (though maybe they have an invisible ramp up them)
[11:29]  Justin Clark-Casey: one thing I was curious about is how many ways there are to hypergrid right now.
[11:29]  Justin Clark-Casey: is it just entering URLs directly into the main map search box?
[11:29]  Richardus Raymaker: not sure if it still depends on the direction
[11:29]  Mata Hari: landmarks will work too
[11:30]  Robert Adams: justincc... not mad about terrain... although I think there is a bug in the DB serializer.... it seems to cast to int16.... I haven't tested though
[11:30]  Mata Hari: I can't do it via a text link but that might be lingering FS bug
[11:30]  Justin Clark-Casey: robert: it still is int16 there. As far as I could see, changing that would mean a new db serialization
[11:31]  Justin Clark-Casey: and tbh, I wanted to do as little change as possible for now so there's an opportunity to sort it out properly after 0.8
[11:31]  Justin Clark-Casey: mata: yeah, I cuoldn't get text link to work either
[11:31]  Robert Adams: there is a bunch of stairs tuning in BulletSim
[11:32]  Dev Random: justincc: osTeleportAgent mentions HG TP too..
[11:32]  Justin Clark-Casey: dev: good point
[11:32]  Justin Clark-Casey: I really want to know if changing certain config parameters successfully blocks all the ways that hg teleport can happen, or successfully restricts them
[11:32]  vegaslon plutonian: just nothing to set stair minimum right now. that is hard coded in based on avatar size
[11:32]  Robert Adams: http://opensimulator.org/wiki/BulletSim#Avatar_Walking_Up_Stairs
[11:34]  Robert Adams: you can change the stairs minimum in BUlletSim... it looks for a colision close to the bottom of the avatar and then adds up position and up force.... both tunable
[11:34]  Mata Hari: I use the default values
[11:34]  Robert Adams: it has something for making sure one is walking straight into the stairs.... the reason why walking up stairs backward doesn't work in BulletSim
[11:35]  Nebadon Izumi: I have found that if you walk up stairs on an angle it works better
[11:35]  Mata Hari: right, but walking straight into a 0.15m step should be flawless(ish)
[11:35]  Nebadon Izumi: straight up seems to create a lot of friction
[11:35]  Mata Hari: and more importantly, if 0.25m step is flawless then doing the identical thing with a 0.15m step certainly should be
[11:35]  Mata Hari: and it isn't
[11:36]  Mata Hari: it seemed to coincide with the change to the avi bounding box/shape a month or so ago
[11:37]  Robert Adams: ok... that could again be wrong computation of the avatar height... since avatar position is kept as the avatar middle, the stairs has to recompute the bottom of the capsule
[11:37]  Mata Hari: prior to that it was working fine for a 0.15m step
[11:37]  Justin Clark-Casey: everything was better in the good old days ;)
[11:37]  Mata Hari: lol
[11:37]  vegaslon plutonian: ya alot of broken amusement park rides in opensim these days
[11:37]  Robert Adams: I'll take a look into that again.... i also need to re-verify the avatar height stuff so feet are at least on terrain and prims
[11:38]  Justin Clark-Casey: robert: Is it worth delaying the next release candidate until then?
[11:38]  Robert Adams: when are you releasing it?
[11:38]  Justin Clark-Casey: there are also a few other things I need to fix
[11:39]  Justin Clark-Casey: robert: I was thinking either friday or early next week
[11:39]  Robert Adams: I will do avatar height before Friday
[11:40]  Justin Clark-Casey: cool. I think a new rc early next week is most likely but might be Friday if things look good
[11:40]  Justin Clark-Casey: got to resist trying to fix too many bugs, at the risk of introducing problems (like the one I just did yesterday)
[11:41]  Justin Clark-Casey: any other opensim topics today?
[11:42]  Justin Clark-Casey: I guess not, party time!
[11:43]  Mata Hari: no, though when you tackle the login issue again I have some other info that might be relevant
[11:43]  Justin Clark-Casey: login issue?
[11:43]  Justin Clark-Casey: oh, the latest problem?
[11:43]  Mata Hari: the inventory fetch getting stuck
[11:43]  Justin Clark-Casey: yes, that's a complicated one
[11:43]  Justin Clark-Casey: I suspect that the simulator has to send back a different HTTP code if it's overloaded to throttle the incoming requests
[11:44]  Justin Clark-Casey: but that's not going to be a simple thing to implement if that's the case
[11:44]  Mata Hari: on a busy sim it also induced massive chat lag and a total failure to fetch or even send existing cache textures for anyone in the region
[11:44]  Justin Clark-Casey: it might also be that the osgrid inventory service is getting overloaded and that causes problems without a retry mechanism in place
[11:44]  Mata Hari: or rather, that issue seems to happen immediately after someone logs in directly to the region vs login plaza and then tp
[11:45]  Robert Adams is also looking at the tilted feet problem
[11:45]  Mata Hari: so I'm guessing it's related
[11:47]  Dev Random: Robert: is there any debugging info,etc that would help with the Collision problem (http://opensimulator.org/mantis/view.php?id=7132) ?
[11:47]  Andrew Hellershanks: Which mailing list announced 0.8rc1? I'm wondering why I didn't see the message.
[11:48]  Justin Clark-Casey: andrew: you're aware we switching mailing lists recently?
[11:48]  Justin Clark-Casey: but you should have been autosubscribed
[11:48]  Robert Adams: if you're willing to have large log files, the detailed physics logging records all the collisions and who they are being sent to
[11:48]  Andrew Hellershanks: yes. I was just checking the info I had and confirmed I was already subscribed to dev and core on the new server.
[11:48]  vegaslon plutonian: I had to pull it out of the junkmail folder
[11:49]  Justin Clark-Casey: I announced to both users and dev
[11:49]  Justin Clark-Casey: and people replied, so it was seen
[11:49]  Justin Clark-Casey: by at least some folks
[11:49]  Mata Hari: I saw it
[11:49]  Andrew Hellershanks: Other people were asking me about 0.8rc1 and all I could say is I didn't know anything about it.
[11:49]  Dev Random: Robert: I'll take a look... yesterday it happened after 3 teleports total.. all me.
[11:49]  Robert Adams: Look in OpenSimDefaults.ini .... there if you turn on that logger and can see it stop working, I will look at them for you
[11:49]  Dev Random: will do
[11:50]  Justin Clark-Casey: usually I keep rc1 fairly quiet so that there's an opportunity to winkle out major issues with people who are familiar with how this kind of thing works
[11:50]  Justin Clark-Casey: but tbh, it's hard enough to get people to test release candidates anyway so probably not a worthwhile tactic
[11:50]  Andrew Hellershanks nods
[11:50]  Justin Clark-Casey: if people are asking you about it might be a good sign
[11:50]  Andrew Hellershanks: yes, but I started to wonder why they had heard about it and I hadn't.
[11:51]  Andrew Hellershanks: Strange.
[11:51]  Robert Adams: Dev: if you put the log files in a different dir, you can zip them all together
[11:51]  Dev Random: I do. read-only bin dir.
[11:52]  Justin Clark-Casey: our 'put everything in the bin dir' habit can be a major pita
[11:53]  stiofain nbmcmedia: hi folks
[11:53]  Justin Clark-Casey: hi stiofain
[11:53]  Richardus Raymaker: hi
[11:56]  UUID Speaker: paulie Flomar: 86003a74-b4d4-45fc-bfa7-46aab9949e94
[11:56]  Justin Clark-Casey: ok, anything else quickly?
[11:57]  Dev Random: grepped 65536 on *.cs, and LLClientView.cs line 4882 looks iffy...
[11:57]  Dev Random: int simulatorCapacity = (int)(((float)landData.SimwideArea / 65536.0f) * (float)m_scene.RegionInfo.ObjectCapacity * (float)m_scene.RegionInfo.RegionSettings.ObjectBonus);
[11:57]  Dev Random: looks like it'll inflate prim limit for vars
[11:57]  Dev Random: anybody care?
[11:58]  paulie Flomar: Howdeh, all :)
[11:58]  Richardus Raymaker: hi pauli
[11:58]  Justin Clark-Casey: Would need to be fixed - I expect it should pull the dimensions from RegionInfo rather than hardcoding them
[11:59]  Justin Clark-Casey: and one could argue that logic should be in a module rather than LLClientView but that's another issue
[11:59]  Dev Random: no point before release though, right?
[11:59]  Justin Clark-Casey: it would be a pretty minor fix... but maybe not a crticial issue
[11:59]  Andrew Hellershanks: Not sure about that. I have people waiting for 0.8 as they are interested in varregions.
[12:00]  Andrew Hellershanks: Has anyone done terrain saves/loads for varregions?
[12:00]  Richardus Raymaker: it saves lots of noise if its fixt before release i think
[12:00]  Justin Clark-Casey: perhaps a bit more important is the issue with mantis 7039 where some chat distances don't work properly on var regions
[12:00]  Robert Adams: I wasn't sure about that computation.... there are a lot of odd limits for objects and the computation is not obvious
[12:00]  Justin Clark-Casey: I may try to take a look at that if i get time
[12:00]  Sarah Kline: i have
[12:00]  Justin Clark-Casey: yeah, always a risk of triggering something nasty in the viewer
[12:00]  Sarah Kline: works ok
[12:01]  Sarah Kline: heightmap
[12:01]  Justin Clark-Casey: though one would think dividing by a bigger rather than a smaller number should be okay
[12:01]  Robert Adams: yes justincc, there is a need to fix distance commputation of neighbors.... there are still some code places that have that problem
[12:01]  Andrew Hellershanks: Sarah, Does it save one file for each 256x256 section of the varregion or does it make a huge file?
[12:01]  Justin Clark-Casey: robert: not going to be a simple fix?
[12:01]  Robert Adams: related to nebadon's mulitple terrain rezzing also
[12:01]  Sarah Kline: oh i havent saved out
[12:02]  Dev Random: Andrew: I have saved and reloaded, but only really using Terragen format.
[12:02]  Andrew Hellershanks: Sarah, if you are loading, do you have to provide a much bigger file than normal to account for the difference in sizes?
[12:02]  Sarah Kline: oh my let me think
[12:02]  Dev Random: Andrew: It was one big file.
[12:02]  Andrew Hellershanks: Dev, ok. That would make sense. The standard raw format isn't set for bigger than 256x256
[12:03]  Sarah Kline: yes you have to lol
[12:03]  Robert Adams: 'save oar; saves the whole varregion
[12:03]  Andrew Hellershanks: I'm wondering how varregions will affect the ability to use a viewer to save/load terrain.
[12:03]  Robert Adams: the file input for terrain should accept larger images for the larger varregions (a 1kx1k image for a 1kx1k varregion)
[12:04]  Robert Adams: if the sizes don't match, odd defaulting will happen
[12:04]  Sarah Kline: yes
[12:05]  Robert Adams: I changed most of the image save/load code to save the larger images.... mistakes could have been made
[12:05]  Robert Adams: loading a legacy oar into a varregion with the --displace paramter will also displace the terrain
[12:05]  Andrew Hellershanks: ok. I'll have to try that myself at some point just to see it in action.
[12:06]  Andrew Hellershanks: ok.
[12:06]  Justin Clark-Casey: ok, I need to go get a bite to eat. Will look at the request bug I introduced after that
[12:07]  Richardus Raymaker: bye justin
[12:07]  Sarah Kline: byes
[12:07]  Justin Clark-Casey: bye folks
[12:07]  Mata Hari: bye Justin
[12:07]  Andrew Hellershanks: I'm going to look at the mailing list archives to look for messages I never received.
[12:07]  dan banner: later
[12:07]  Dev Random: I'm going to fill my drive with physics logs
Personal tools
General
About This Wiki