Chat log from the meeting on 2011-08-16

From OpenSimulator

Revision as of 12:14, 16 August 2011 by Justincc (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
[10:04 AM]  John Flower is Online
[10:04 AM]  Aaron Duffy has entered chat range (6.3m)
[10:04 AM]  Richardus Raymaker has entered chat range (7.9m)
[10:04 AM]  Nebadon Izumi has entered chat range (6.2m)
[10:04 AM]  John Flower has entered chat range (8.8m)
[10:04 AM]  Richardus Raymaker: hi justin
[10:04 AM]  Nebadon Izumi: hello Justin
[10:04 AM]  Justin Clark-Casey: hi richardus, nebadon, folks
[10:05 AM]  John Flower: Hello justin
[10:05 AM]  Master Dubrovna: Hi Justin
[10:05 AM]  Richardus Raymaker: Justin, have you done something with the wrong rendered sculpts. (square)
[10:05 AM]  Justin Clark-Casey: wrong rendered sculpts?
[10:06 AM]  Nebadon Izumi: physics right Richardus?
[10:06 AM]  Nebadon Izumi: they are colliding like a cube
[10:06 AM]  Richardus Raymaker: well, i have lots of sculpts that are square to walk
[10:06 AM]  Richardus Raymaker: yes nebadon
[10:07 AM]  Justin Clark-Casey: uh, again?
[10:07 AM]  Richardus Raymaker: sometimes the sculpts are fine, but most after restart the are wrong.
[10:07 AM]  Richardus Raymaker: if you set them phantom and back noram the work fine to until resart. sometimes a restart works fine.
[10:07 AM]  Justin Clark-Casey: are these new sculpts or old or both?
[10:08 AM]  Nebadon Izumi: i have not seen it happen myself
[10:08 AM]  Richardus Raymaker: and i see it only on my windows test pc with latest osgrid 13-8
[10:08 AM]  Richardus Raymaker: very old ones justin.
[10:08 AM]  Justin Clark-Casey: very latest code?
[10:08 AM]  Justin Clark-Casey: I think it's possible that old sculpts actually contained sculpt data with the region data itself
[10:08 AM]  Richardus Raymaker: i use on windows this OSgrid OpenSimulator - [zip] [25.8mb] 08-13-2011
[10:08 AM]  Justin Clark-Casey: instead of storing them as an asset, which might possibly lead to race condition behaviour
[10:08 AM]  Justin Clark-Casey: though I don't immediatley know how that could occur
[10:09 AM]  Richardus Raymaker: but if you set it phantom and normal it works fine
[10:09 AM]  Justin Clark-Casey: yes, that would be the remshing that this triggers
[10:09 AM]  Richardus Raymaker: i rezzed a new one and that seems after a restart ok to. need to try that again
[10:09 AM]  Justin Clark-Casey: when the physics mesh happens, if the sculpt data can't be found for some reason then you end up with a square physics proxy
[10:09 AM]  Richardus Raymaker: so, can that trigger not be done at startup ?
[10:10 AM]  Justin Clark-Casey: well, as you know, sometimes it is
[10:10 AM]  Richardus Raymaker: yes, sometimes.... :)
[10:10 AM]  Justin Clark-Casey: and sometimes it isn't possibly. This would be a bug
[10:10 AM]  Justin Clark-Casey: If you have a particular sculpt on which this is occuring I woulod like to see it as an iar
[10:11 AM]  Richardus Raymaker: can decodesculptmappath setting help ?
[10:11 AM]  Justin Clark-Casey: no
[10:11 AM]  Richardus Raymaker: ok. then i dont need to change that
[10:12 AM]  Richardus Raymaker: i can give you the sculpt private justin. copy from my sim
[10:12 AM]  Justin Clark-Casey: ok. I probably won't get time to look before end of this week, though
[10:12 AM]  Richardus Raymaker: must eb carefull with the rocks
[10:12 AM]  Richardus Raymaker: hold on.
[10:13 AM]  Nebadon Izumi: I did some testing with NPC's Justin
[10:13 AM]  Nebadon Izumi: did you see the video?
[10:13 AM]  Justin Clark-Casey: not yet
[10:13 AM]  Nebadon Izumi: let me play it
[10:13 AM]  Void Pipe: that was great
[10:13 AM]  Nebadon Izumi: something seems very wrong though
[10:13 AM]  Nebadon Izumi: when the NPC's have attachments they use an ungodly amount of memory
[10:13 AM]  Andrew Hellershanks: That voice sounds familiar
[10:14 AM]  Void Pipe: i love the dance scene
[10:14 AM]  Nebadon Izumi: it took over 4gb to do this
[10:14 AM]  Void Pipe: lol
[10:14 AM]  Justin Clark-Casey: there's at least one bug I need to fix with npc attachments
[10:14 AM]  Nebadon Izumi: this same test with avatars with no attachments
[10:14 AM]  Nebadon Izumi: took 40mb ram
[10:14 AM]  Justin Clark-Casey: hmmmm, that is odd
[10:14 AM]  Nebadon Izumi: with these avatars it took over 4gb
[10:14 AM]  Justin Clark-Casey: interesting
[10:14 AM]  Justin Clark-Casey: we must be holding on to textures or something somehow
[10:14 AM]  Nebadon Izumi: video playing for you/
[10:14 AM]  Nebadon Izumi: ?
[10:15 AM]  Justin Clark-Casey: no
[10:15 AM]  Nebadon Izumi: these avatars only use one texture
[10:15 AM]  Nebadon Izumi: ah
[10:15 AM]  Nebadon Izumi: bummer
[10:15 AM]  Justin Clark-Casey: my imprudence might not be set right
[10:15 AM]  Justin Clark-Casey: one moment
[10:15 AM]  Nebadon Izumi: k
[10:15 AM]  Sarah Kline: ok for me
[10:15 AM]  Void Pipe: for me too
[10:15 AM]  Nebadon Izumi: its really choppy for me
[10:15 AM]  Justin Clark-Casey: where is it meant to be playing - main screen?
[10:15 AM]  Nebadon Izumi: not sure why
[10:15 AM]  Nebadon Izumi: its playing like 1 fram every 5 seconds
[10:15 AM]  Void Pipe: its fine
[10:16 AM]  Nebadon Izumi: yes main screen Justin
[10:16 AM]  Nebadon Izumi: you change your settings?
[10:16 AM]  UUID Speaker: little bunnie: c891873a-5190-40b0-8424-bd8931d779eb
[10:16 AM]  Justin Clark-Casey: yeah, not working for me for some reason
[10:16 AM]  Nebadon Izumi: let me restart it
[10:16 AM]  Nebadon Izumi: ok restarted
[10:16 AM]  Nebadon Izumi: anything?
[10:16 AM]  Justin Clark-Casey: no, maybe this isn't working well with linux and this viewer
[10:17 AM]  Nebadon Izumi: k
[10:17 AM]  John Flower: i see the vid ok in here
[10:17 AM]  Nebadon Izumi: well here is link to version on web
[10:17 AM]  Nebadon Izumi:
[10:17 AM]  Nebadon Izumi: it seems unlikely to me though the memory issue is realted to textures
[10:17 AM]  Richardus Raymaker gave you rock.
[10:17 AM]  Nebadon Izumi: i dunno could be
[10:17 AM]  Justin Clark-Casey: richardus: you have to e-mail me an iar
[10:17 AM]  Nebadon Izumi: thing is though these avatars only use 1 texture
[10:18 AM]  Nebadon Izumi: they are 1000's of prims each, but its only 1 texture
[10:18 AM]  Sarah Kline: just a few prims thats all
[10:18 AM]  Richardus Raymaker: ohh. uhmm. that need a few more hands. ill ask you when im done
[10:19 AM]  Justin Clark-Casey: nebadON: I presume you created the video after the memory fix on sta?
[10:19 AM]  Justin Clark-Casey: sat?
[10:19 AM]  Nebadon Izumi: yes
[10:19 AM]  Nebadon Izumi: i made that on sunday
[10:19 AM]  Nebadon Izumi: on master git at the time
[10:19 AM]  Justin Clark-Casey: lol :)
[10:20 AM]  Nebadon Izumi: its 140,000 prims
[10:20 AM]  Nebadon Izumi: heh
[10:20 AM]  Andrew Hellershanks: nebadon, you found the memory leak?
[10:20 AM]  Nebadon Izumi: its like 3500 prims per avatar
[10:20 AM]  Nebadon Izumi: ya did a git bisect and located the break
[10:20 AM]  Justin Clark-Casey: god, 140,000 prims
[10:20 AM]  Nebadon Izumi: Justin fixed it up
[10:20 AM]  Sarah Kline: dont prims take up memory a bit?
[10:20 AM]  Nebadon Izumi: oh and also
[10:21 AM]  Nebadon Izumi: i tried doing this same test with avatar wearing a Rigged mesh avatar
[10:21 AM]  Justin Clark-Casey: love the dancing aliens though :)
[10:21 AM]  Nebadon Izumi: and like 50% of the time the were a gas cloud
[10:21 AM]  Andrew Hellershanks: If nebadon is going to do something he usually goes big (prim count wise). :-)
[10:21 AM]  Void Pipe: lol me too
[10:21 AM]  Nebadon Izumi: hehe
[10:21 AM]  Justin Clark-Casey: yeah, don't know why there would be a mesh avatar issue off the top
[10:21 AM]  Richardus Raymaker: hmm, iar slow for 1 rock. lol
[10:22 AM]  Nebadon Izumi: ya me either Justin
[10:22 AM]  Nebadon Izumi: it could just be issues with locking of some kind
[10:22 AM]  Nebadon Izumi: one thing i noticed when doing these tests
[10:22 AM]  Nebadon Izumi: is the scripts fail a lot
[10:22 AM]  Nebadon Izumi: it seemed like doing inventory functions or a datastore
[10:22 AM]  Nebadon Izumi: could completelt stop the NPC rezzer script
[10:23 AM]  Justin Clark-Casey: that's pretty odd
[10:23 AM]  Justin Clark-Casey: was this under conditions of using a lot of memory?
[10:23 AM]  Void Pipe: is it asyncronous connections?
[10:23 AM]  Nebadon Izumi: ya if you even have 1 attachment
[10:23 AM]  Nebadon Izumi: it uses ungodly ammounts of memory and resources
[10:23 AM]  Justin Clark-Casey: ok, that needs fixing
[10:23 AM]  Justin Clark-Casey: that shouldn't be happening
[10:23 AM]  Nebadon Izumi: i had to hack it into 64 bit mode in windows
[10:23 AM]  Nebadon Izumi: just so i could do that test
[10:23 AM]  Nebadon Izumi: i couldnt even do it at all in 32 bit mode
[10:24 AM]  Nebadon Izumi: but like i said same test with no attachments
[10:24 AM]  Nebadon Izumi: the sim only uses like 39mb ram
[10:24 AM]  Nebadon Izumi: hehe
[10:24 AM]  Justin Clark-Casey: yeah, attachments must be holding on to massive chunks of asset data somehow
[10:24 AM]  Justin Clark-Casey: or it's some NPC specific issue when using them, though can't think what off the top of my head
[10:25 AM]  Amanda rose is Online
[10:25 AM]  Nebadon Izumi: ya
[10:25 AM]  Nebadon Izumi: good start though
[10:25 AM]  Nebadon Izumi: i was suprised i could even do that test honestly
[10:25 AM]  Nebadon Izumi: considering the viewer and simulator were running on the same machine
[10:25 AM]  Nebadon Izumi: and so was fraps
[10:25 AM]  Justin Clark-Casey nods
[10:25 AM]  Justin Clark-Casey: 140,000 prims, that's ungodly :)
[10:25 AM]  Nebadon Izumi: hehe ya
[10:26 AM]  Andrew Hellershanks: I don't think that was a record for Nebadon
[10:26 AM]  Nebadon Izumi: its the equivalent of 9 Second Life sims dancing around
[10:26 AM]  Nebadon Izumi: lol
[10:26 AM]  Sarah Kline: no wonder a bit choppy
[10:26 AM]  Nebadon Izumi: actually i think that is my record
[10:26 AM]  Sarah Kline: )
[10:26 AM]  Nebadon Izumi: for prims rezzed in world it was like 122000 prims
[10:26 AM]  Richardus Raymaker is Offline
[10:26 AM]  Richardus Raymaker is Online
[10:26 AM]  Andrew Hellershanks: how many prims were in the sierpinski triangle?
[10:26 AM]  Nebadon Izumi: 122k
[10:27 AM]  Richardus Raymaker: last TP made me invisible. loked like i where on 2 sims at the same time. because iar i could start.
[10:27 AM]  Richardus Raymaker: justin can you IM the mail adress ?
[10:27 AM]  Ruben Haan: hi
[10:28 AM]  Justin Clark-Casey: richardus:
[10:28 AM]  Andrew Hellershanks is fighting with the cat for control of the chair he is sitting on
[10:28 AM]  Justin Clark-Casey: you and your cat :)
[10:29 AM]  Nebadon Izumi: memory is looking a lot better here
[10:29 AM]  Andrew Hellershanks sighs
[10:29 AM]  Nebadon Izumi: i still think its leaking a bit though
[10:29 AM]  Andrew Hellershanks: the cat won
[10:29 AM]  Nebadon Izumi: lol
[10:30 AM]  Richardus Raymaker: ok. its rolling to you now justin
[10:30 AM]  Justin Clark-Casey: ok
[10:30 AM]  John Flower: Hi Amanda
[10:30 AM]  Nebadon Izumi: but i think the leaks we are seeing now are minimal, and were there for very long time
[10:30 AM]  Amanda rose: Hi John
[10:30 AM]  Justin Clark-Casey: yeah, probably the previous ones
[10:30 AM]  Nebadon Izumi: ya
[10:30 AM]  Justin Clark-Casey: I made a whoopsie a month back, so it was my fault :)
[10:30 AM]  Nebadon Izumi: right now its showing 999mb on show stats
[10:30 AM]  Nebadon Izumi: thats actually lower than i would expect
[10:30 AM]  Richardus Raymaker: one of my sims is up to 800MB to
[10:30 AM]  Andrew Hellershanks: nebadon, you are getting good at using biesct
[10:31 AM]  Nebadon Izumi: heh ya bisecting actually isnt very hard
[10:31 AM]  Nebadon Izumi: its just time consuming and dull
[10:31 AM]  Andrew Hellershanks nods
[10:31 AM]  Nebadon Izumi: it took 9 steps to find the break
[10:31 AM]  Nebadon Izumi: basically had to recompile and log in to this sim 9 times to find the break
[10:31 AM]  Nebadon Izumi: took about 3 hours
[10:31 AM]  Richardus Raymaker: under windows i see the memory usage jump up a little bit when you TP to the sim. i think simcrossing does the same.
[10:32 AM]  Andrew Hellershanks: 9 steps was pretty good
[10:32 AM]  Richardus Raymaker: but it drops back to
[10:32 AM]  Justin Clark-Casey: nebadon: thanks Neb. I know it's a pain but it's actually the quickest way
[10:32 AM]  Nebadon Izumi: ya
[10:32 AM]  Justin Clark-Casey: the profiling just doesn't give enough informatio nright now
[10:32 AM]  Richardus Raymaker: hmm i think. anyway opensim consume more and more in tim. also on window but i use it to short for that
[10:32 AM]  Nebadon Izumi: the hardest part was actually finding a good version
[10:32 AM]  Justin Clark-Casey: I knew it was a byte[] holding the data but i couldn't tell which object was actually holding it
[10:33 AM]  Richardus Raymaker: so the memory leak is fixt. need to upgrade that 800<MB sim.
[10:33 AM]  Nebadon Izumi: i was suprised that it started on July 4 though
[10:33 AM]  Nebadon Izumi: i swore it was going on longer than that
[10:33 AM]  Richardus Raymaker: hi odd
[10:33 AM]  Nebadon Izumi: felt longer anyway
[10:33 AM]  Nebadon Izumi: hehe
[10:34 AM]  Andrew Hellershanks: what caused the leak?
[10:34 AM]  Justin Clark-Casey: yeah, it can be tricky to rely on perceptions with this kind of stuff
[10:34 AM]  Nebadon Izumi: it was Flotsam cache
[10:34 AM]  Nebadon Izumi: it was loading Memory cache even when it was disabled
[10:34 AM]  Andrew Hellershanks: Oh. So something completely different from what was first suspected
[10:34 AM]  Justin Clark-Casey: andrew: When I was changing code, I made a mistake where if a file was retrieved from the file cache, it was then dumped in the memory cache even if the memory cache was disabled
[10:34 AM]  Richardus Raymaker: i found a other small problem. opensim dont use its own directory path when it loads ini files.
[10:34 AM]  Nebadon Izumi: ya
[10:34 AM]  Nebadon Izumi: none of us guessed right
[10:34 AM]  Nebadon Izumi: hehe
[10:34 AM]  Andrew Hellershanks: Justin, ah, ok.
[10:34 AM]  Richardus Raymaker: so if you start from D: drive opensim on the C: drive it fails
[10:35 AM]  Nebadon Izumi: what is really odd though
[10:35 AM]  Justin Clark-Casey: richardus: I'm quite surprised to hear that
[10:35 AM]  Nebadon Izumi: is that the problem was way worse when there was neighbor sims
[10:35 AM]  Nebadon Izumi: does the sim your in cache neighbors content ?
[10:35 AM]  Justin Clark-Casey: There's no C: drive wired in anywhere in opensim, otherwiseit wouldn'#t work on other systems
[10:35 AM]  Justin Clark-Casey: nebadon: no, not at all
[10:35 AM]  Richardus Raymaker: i hope the mantis explains it clear
[10:35 AM]  Nebadon Izumi: ya thats really odd
[10:35 AM]  Nebadon Izumi: cause when i moved Wright Plaza away
[10:35 AM]  Nebadon Izumi: memory leak was gone
[10:35 AM]  Nebadon Izumi: bring it back it was horrible again
[10:35 AM]  Nebadon Izumi: makes no sense
[10:35 AM]  Andrew Hellershanks: nebadon, yeah, that's why we were guessing the leak was in a completely different part of the code.
[10:35 AM]  Richardus Raymaker: uhmm, justin . i mean opensim dont collect its location.
[10:36 AM]  Richardus Raymaker: so it assume the ini's are on the place you start opensim
[10:36 AM]  Justin Clark-Casey: nebadon: yeah, I'#m not sure what the connection would be
[10:36 AM]  Richardus Raymaker: maby this helps
[10:36 AM]  Justin Clark-Casey: no texture information is routed through from the neighbours
[10:36 AM]  Nebadon Izumi: only thing i could think is the viewer is also asking the sim your in for the same content?
[10:37 AM]  Justin Clark-Casey: it all goes directfrom the child aganet to the viewer. So only the neighbouring regions would suffer the memory penalty, not the root region
[10:37 AM]  Nebadon Izumi: its asking the neighbor sim and the sim your in?
[10:37 AM]  Nebadon Izumi: for the neighbor sims content?
[10:37 AM]  Justin Clark-Casey: no, the sim doesn't do that at all
[10:37 AM]  Justin Clark-Casey: there's no mechanism for it
[10:37 AM]  Nebadon Izumi: no i meant he viewer is doing it
[10:37 AM]  Justin Clark-Casey: however, if there were surrounding regions then child agents from neighbours would be logged in
[10:37 AM]  Justin Clark-Casey: and they would trigger the problem when they requested assets
[10:37 AM]  Justin Clark-Casey: so yes, neighbouring regions would contributed
[10:37 AM]  Justin Clark-Casey: contrbute
[10:38 AM]  Nebadon Izumi: hmm ya i cant say i recall if there were any child agents at the time
[10:38 AM]  Richardus Raymaker: lag ?
[10:38 AM]  Nebadon Izumi: anyway, glad its behind us
[10:38 AM]  Nebadon Izumi: that one sucked
[10:38 AM]  Nebadon Izumi: hehe
[10:39 AM]  Justin Clark-Casey: yeah, I was so relieved to see you'd pinnred it down
[10:39 AM]  Justin Clark-Casey: after wasting too much time on friday, but at least I learnt more about the current state of mono profilingg
[10:39 AM]  Nebadon Izumi: ya
[10:39 AM]  Nebadon Izumi: we should use that more
[10:39 AM]  Nebadon Izumi: do it again see if we can figure out some of these other leak sources
[10:39 AM]  Justin Clark-Casey: it doesn't really return neough info though
[10:40 AM]  Justin Clark-Casey: and it still has an impact on thre sim even with some of the most resource consuming profiling turned off
[10:40 AM]  Justin Clark-Casey: hopefully that will improve in the future
[10:40 AM]  Nebadon Izumi: ya i doubt it
[10:40 AM]  Nebadon Izumi: its not much better doing it under .net either
[10:40 AM]  Justin Clark-Casey: profiling changed a lot between 2.10 and 2.6
[10:40 AM]  Justin Clark-Casey: profiling on .net?
[10:40 AM]  Nebadon Izumi: the sim is pretty much unuseable everytime we did it before
[10:40 AM]  Nebadon Izumi: ya
[10:41 AM]  Justin Clark-Casey: you could try turning a lot of the process call overhead off
[10:41 AM]  Nebadon Izumi: we even rented a 16 core amazone node with 15gb ram
[10:41 AM]  Nebadon Izumi: and it wasnt enough
[10:41 AM]  Justin Clark-Casey: the only thing we're really interested in is a dump of heap memory usage
[10:41 AM]  Justin Clark-Casey: wow, when was this?
[10:41 AM]  Nebadon Izumi: while back
[10:41 AM]  Nebadon Izumi: i forget the app we used now
[10:41 AM]  Nebadon Izumi: adam recommended it
[10:41 AM]  Nebadon Izumi:
[10:41 AM]  Nebadon Izumi: this
[10:41 AM]  Nebadon Izumi: dot trace
[10:42 AM]  Justin Clark-Casey: yeah, I've heard mention of that
[10:42 AM]  Richardus Raymaker: maby stupid question: can you add the ini paths hard in the opensim.ini files ? offcorse thats a terrible way
[10:42 AM]  Justin Clark-Casey: tbh, probably the more important thing to address slowly is that http server
[10:42 AM]  Nebadon Izumi: the reason you didnt hear much about our tests is because they didnt really produce any useable data either
[10:42 AM]  Justin Clark-Casey: really beginning to bug me now
[10:42 AM]  Justin Clark-Casey: nebadon: I see
[10:42 AM]  Richardus Raymaker: or can opensim have option to add the base directory in opensim.ini ?
[10:42 AM]  Nebadon Izumi: 1 avatar would log in under profile
[10:42 AM]  Nebadon Izumi: and the sim would freeze
[10:42 AM]  Nebadon Izumi: it would take an hour for the sim to fully rez
[10:42 AM]  Justin Clark-Casey: richardus: it might work
[10:42 AM]  Andrew Hellershanks: I wonder if valgrind under Linux would help at all with memory profiling
[10:42 AM]  Nebadon Izumi: for 1 user
[10:42 AM]  Justin Clark-Casey: Andrew: no, that would work at too low a level
[10:43 AM]  Justin Clark-Casey: we need a dump of C# objects
[10:43 AM]  Justin Clark-Casey: nebadon: tbh, I think the real way would be going through with some understanding of the code and making sure that resources are being properly released
[10:43 AM]  Richardus Raymaker: whats mono profiling for use if you need bigblue ? :)
[10:43 AM]  Justin Clark-Casey: but that's time consuming work
[10:43 AM]  Nebadon Izumi: ya
[10:44 AM]  Nebadon Izumi: no doubt alot of the problem is some of the code in OpenSim
[10:44 AM]  Andrew Hellershanks: understanding the code? That's a tall order ;-)
[10:44 AM]  Justin Clark-Casey: heh
[10:44 AM]  Justin Clark-Casey: nebadon: most probably, though possible some of the 3rd party libs or even mono itself could be to blame
[10:44 AM]  Justin Clark-Casey: which is why it's sometimes best to just wait....
[10:44 AM]  Nebadon Izumi: ya, more likely all of the above
[10:44 AM]  Nebadon Izumi: hehe
[10:44 AM]  Justin Clark-Casey: and, er update the 3rd party libs
[10:44 AM]  Justin Clark-Casey: that http server specially
[10:44 AM]  Nebadon Izumi: http server probably is a big part of the problem
[10:44 AM]  Nebadon Izumi: heh ya
[10:45 AM]  Nebadon Izumi: hopefully diva has some time soon
[10:45 AM]  Nebadon Izumi: we can discuss that more with her
[10:45 AM]  Justin Clark-Casey: I have a feeling she will be busy with her forthcoming conference for quite some time
[10:46 AM]  Void Pipe: what is the problem with the http server?
[10:46 AM]  Justin Clark-Casey: void: occasionally it just blows up
[10:46 AM]  Justin Clark-Casey: with little useful error information.
[10:46 AM]  Justin Clark-Casey: nebadon: any idea what happened to the patches that diva slapped on it?
[10:47 AM]  Justin Clark-Casey: any idea if those made it into opensim-libs or even upstream?
[10:47 AM]  Andrew Hellershanks: Any way of hooking in a different http server to do some testing?
[10:47 AM]  Justin Clark-Casey: andrew: I'm sure there is, if one changes the code
[10:47 AM]  Void Pipe: exactly my thought
[10:47 AM]  Nebadon Izumi: hmm, good question, i am not sure Justin
[10:47 AM]  Nebadon Izumi: i suspect it did make it into opensim-libs, but i am not 100% sure
[10:48 AM]  Justin Clark-Casey: it's hoving in to view on my radar... slowly :)
[10:48 AM]  Justin Clark-Casey: but at some point we need to release 0.7.2
[10:48 AM]  Nebadon Izumi: one of the problems with the http server we are using is i beleive the main project for it is sort of defunct
[10:48 AM]  Nebadon Izumi: and no longer maintained
[10:49 AM]  Justin Clark-Casey: I took a brief look recently and I thought that I saw some activity
[10:49 AM]  Nebadon Izumi: thats good, i think for a long time it sat pretty dormant
[10:51 AM]  Nebadon Izumi: trying to think if there was anyting else I wanted to mention, cant really think of anything
[10:51 AM]  Nebadon Izumi: sorta feel like im forgetting something though
[10:51 AM]  Andrew Hellershanks: which http server project is being used with OS?
[10:52 AM]  Justin Clark-Casey:
[10:52 AM]  Justin Clark-Casey: alright, so not a huge amount of activity, juts a series of bug fixes on June 30th :)
[10:52 AM]  Justin Clark-Casey: but they would be good to try out
[10:53 AM]  Richardus Raymaker: yes
[10:53 AM]  Nebadon Izumi: ya bug fixes are good
[10:53 AM]  Nebadon Izumi: heh
[10:53 AM]  Nebadon Izumi: i have a feeling that wont be very easy though
[10:53 AM]  Nebadon Izumi: i think there are a lot of customizations in our version
[10:53 AM]  Void Pipe: there seems to be some exceptions fault in there
[10:53 AM]  Justin Clark-Casey: yeah, this is the problem
[10:53 AM]  Justin Clark-Casey: if those customizations didn't go upstream that makes it a pita
[10:54 AM]  Nebadon Izumi: ya, something i guess we'll need to discuss
[10:54 AM]  Nebadon Izumi: Melanie and Diva might have a better idea about that
[10:55 AM]  Justin Clark-Casey: anyway, briefly, any other opensim topics today?
[10:56 AM]  Richardus Raymaker: i told what i now.. uhmm.
[10:56 AM]  Nebadon Izumi: I cant think of anything else at the moment
[10:56 AM]  Void Pipe: how about My Notes... in profiles
[10:56 AM]  Void Pipe: ;o)
[10:57 AM]  Andrew Hellershanks: Notes should work
[10:57 AM]  Void Pipe: nope
[10:57 AM]  Andrew Hellershanks: Void, are you talking about not working here or in your own standalone/grid?
[10:57 AM]  Richardus Raymaker: put something in mine. so need to check after TP
[10:57 AM]  Nebadon Izumi: seems i can make notes in my own profile
[10:57 AM]  Void Pipe: not anywhere
[10:58 AM]  Nebadon Izumi: but no one elses
[10:58 AM]  Nebadon Izumi: everyone else here just says Loading..
[10:58 AM]  Justin Clark-Casey: It sounds like profiles is enormously hard to set up
[10:58 AM]  Justin Clark-Casey: yeah, that's true for me too
[10:58 AM]  Andrew Hellershanks: um... yeah, that does appear to be a problem, Void. I'll look in to it.
[10:58 AM]  Void Pipe: minor but still anoying
[10:58 AM]  Void Pipe: new commers asks about it
[10:58 AM]  Richardus Raymaker: same for your profile neb
[10:59 AM]  Void Pipe: its in the Mantis
[10:59 AM]  Andrew Hellershanks: It was working at one point. At least now its been long enough that I can ignore support for 0.6.9 in osprofile
[10:59 AM]  Justin Clark-Casey: yeah
[10:59 AM]  Void Pipe: also big problems with inconsistant attachment state
[11:00 AM]  Andrew Hellershanks: I really need to get back to that soon. I need to determine which of several version is the most recent code.
[11:00 AM]  Void Pipe: with v2 code
[11:00 AM]  Justin Clark-Casey: I suspect ias (inconsistent attachment state) is due to packet drop
[11:00 AM]  Richardus Raymaker: oh. thats why i almost neve rsee it.
[11:00 AM]  Sarah Kline: I have kept a diary and I had one case of it for whole week
[11:00 AM]  Void Pipe: just v2 viewres + singularity
[11:00 AM]  Andrew Hellershanks: Void, what bug #?
[11:00 AM]  Void Pipe: one sec
[11:00 AM]  Sarah Kline: I relogged on the sim and then was ok
[11:00 AM]  Void Pipe: 0005644
[11:01 AM]  Void Pipe: i posted today
[11:01 AM]  Sarah Kline: so probably right justin
[11:01 AM]  Sarah Kline: i carry on with my diary
[11:01 AM]  Justin Clark-Casey: sarah: thanks, that's good data
[11:01 AM]  Richardus Raymaker: i have sometimes after a failed teleport that all sims around me are gone. then you need to relog or your in jail
[11:01 AM]  Justin Clark-Casey: I suspect we need to be less agressive on packet sending on initial login, then ramp up after that
[11:01 AM]  Justin Clark-Casey: it's the same problem that causes terrain holes on first login
[11:01 AM]  Void Pipe: well it is much more frequent than that actually
[11:02 AM]  Void Pipe: with kirsten or foirestorm or v2 v3
[11:02 AM]  Justin Clark-Casey: frequence would depend on connection quality and possible changes in the viewer code
Personal tools
About This Wiki