Chat log from the meeting on 2012-03-13

From OpenSimulator

Jump to: navigation, search
[11:04] Andrew HellershanksAndrew Hellershanksoh, well.
[11:04] Nebadon Izumi: heh
[11:04] paulie Flomar: :)
[11:04] Nebadon Izumi: well maybe he can help you get that straight Andrew
[11:04] dan banner: hey justin
[11:04] Nebadon Izumi: did you erase your original branch or something?
[11:05] Justin Clark-Casey: hello dan, folks
[11:05] Nebadon Izumi: and then couldnt get the patches to reapply?
[11:05] Andrew Hellershanks: its going to have to wait for a few weeks or months now. I've got other things I need to work on.
[11:05] BlueWall Slade: Hi Justin
[11:05] Nebadon Izumi: i am sure i missed out on some of what actually happened
[11:05] Nebadon Izumi: hello Justin
[11:05] sim core: :-J Greets, Justin
[11:06] paulie Flomar: Howdeh, JCC. :)
[11:06] hippie balbozar: hi justin
[11:06] Richardus Raymaker: hello jcc
[11:06] Andrew Hellershanks: I got rid of the 0.7.2PF branch a while back. I don't think I've been developing on it for a while now.
[11:06] Nebadon Izumi: ya i do remember you saying that
[11:06] Nebadon Izumi: few weeks back
[11:07] Andrew Hellershanks: I dn't know how to make git tell me the branch a commit is in based on hash code unless you have the relevant branch checked out.
[11:07] Nebadon Izumi: hrmm Justin do you know how to do that
[11:07] Nebadon Izumi: must be a way
[11:08] Justin Clark-Casey: not off the top of my head.
[11:08] Nebadon Izumi: there also has to be a way to get it to merge into master gracefully
[11:08] Nebadon Izumi: git is great until somethign wierd like this happens
[11:08] Nebadon Izumi: hehe
[11:08] BlueWall Slade: I remember talking in IRC and I seem to remember that you saved the patches to files, then reset everything.
[11:09] BlueWall Slade: if that is the case, then you only have those files
[11:09] Andrew Hellershanks: yeah, I no longer hate git but I still don't like it.
[11:09] Nebadon Izumi: hehe
[11:09] Andrew Hellershanks: I can't tell from the patch file which version of git I patched.
[11:09] Nebadon Izumi: are the patches on mantis?
[11:09] Andrew Hellershanks: yes
[11:09] Andrew Hellershanks: Bug 5885
[11:09] Nebadon Izumi: k what #?
[11:09] Nebadon Izumi: thanks
[11:10] Andrew Hellershanks: They won't apply to master due to a conflict. I can't find a version they will apply to in the master branch. I tried rolling back a ways.
[11:10] BlueWall Slade: they should leave the patched files in the tree
[11:10] BlueWall Slade: and you shoudl be able to see those with git status
[11:11] BlueWall Slade: then edit the files and look for <<<<<<<
[11:11] Nebadon Izumi: ya that hash at the top must have been from your local branch
[11:11] BlueWall Slade: then resolve those and git add
[11:11] Nebadon Izumi: that is my guess
[11:11] Andrew Hellershanks: I had the two patches committed. I did a pull --rebase and got a conflict. After that, things went down hill.
[11:12] BlueWall Slade: do you still have the patches in a branch?
[11:12] Nebadon Izumi: what the heck
[11:12] Nebadon Izumi: the patch is dated Sept 17 2001
[11:12] Andrew Hellershanks: Blue, not any more. Lost them from my source tree while fixing up my repo.
[11:12] BlueWall Slade: wow
[11:12] Nebadon Izumi: whats up with that
[11:12] BlueWall Slade: andrew - do you own a time machine?
[11:12] Nebadon Izumi: in the patch header
[11:12] Nebadon Izumi: both of them
[11:13] Nebadon Izumi: From 86873054b78f2203792bc62c3e4187460809a091 Mon Sep 17 00:00:00 2001
[11:13] Andrew Hellershanks: I would just blame it on git.
[11:13] Nebadon Izumi: heh
[11:13] Andrew Hellershanks: :)
[11:13] Nebadon Izumi: though 2 lines down it says
[11:14] BlueWall Slade: um, nope
[11:14] Nebadon Izumi: Date: Fri, 17 Feb 2012 22:00:07 -0500
[11:14] Richardus Raymaker: but , i guess you have the patch still stored local to...
[11:14] Nebadon Izumi: so have you tried applying to a pull from that day?
[11:14] Nebadon Izumi: or maybe day before
[11:14] Andrew Hellershanks: yes, I have three am formatted patches saved, and two standard patches of work in progress.
[11:15] Andrew Hellershanks: Nebadon, yup. Git can't apply the patches
[11:15] Nebadon Izumi: hrmm
[11:15] Nebadon Izumi: have you tried like patch -p1 <
[11:15] Nebadon Izumi: instead of using git?
[11:15] Andrew Hellershanks: I have my MIT course work to do so I don't have as much free time any more and was going to just leave it for a while.
[11:15] Nebadon Izumi: ok
[11:15] Andrew Hellershanks: yea, patch -p1 generated some rejects
[11:16] Nebadon Izumi: ya what a pain, ok well hopefully it can be worked out
[11:16] Nebadon Izumi: sorry about the pain
[11:16] Nebadon Izumi: never fun
[11:17] sim core: :-J
[11:17] Andrew Hellershanks: No. I keep thinking it wouldn't have been as much of a headache if it was SVN.
[11:17] Nebadon Izumi: wish i was more of a git wizard
[11:17] Andrew Hellershanks: Lost a few hours to this already.
[11:18] Andrew Hellershanks: At least I saved my changes and work in progress. It will take a few hours to get back on track to get patches back in place, then figure out where I left off.
[11:18] Richardus Raymaker: nebadon is already a wizard with so many things you do
[11:18] BlueWall Slade: which patch is it - or both?
[11:18] Nebadon Izumi: its both
[11:19] Nebadon Izumi: heh ya unfortunately richardus i know enough about git to be real dangerous
[11:19] Nebadon Izumi: hehe
[11:19] Andrew Hellershanks: git is dangerous on its own. I've actually lost some work as a result of it.
[11:19] Andrew Hellershanks: Never had that happen with any other VCS I've used
[11:20] Justin Clark-Casey: I have never found git to be a problem myself
[11:20] paulie Flomar: Using the VCS is more comlex than writing the code? dang
[11:20] Andrew Hellershanks: When I first started using git I would often get in to a state where it said I'm not on any branch.
[11:20] Andrew Hellershanks: paulie, it almost feels like that
[11:21] Andrew Hellershanks: I would usually find the fastest way to fix the "not on a branch" problem was wipe out the repo and do a fresh clone.
[11:21] paulie Flomar: It scares the git out of ya, eh.
[11:21] Andrew Hellershanks: hehe
[11:21] BlueWall Slade: git rocks
[11:21] Richardus Raymaker: i can do 3 git things now with reading.. others i need to learn but not needed for now
[11:21] Andrew Hellershanks: I probably did that once or twice when I had commits in there that I hadn't saved seaprately.
[11:22] Andrew Hellershanks: git is full of rocks. :-)
[11:22] BlueWall Slade: lol
[11:22] Justin Clark-Casey: anyway, just to let you know, dan, I have not implemented the AppDomain = hybrid/default switch
[11:22] BlueWall Slade: well, opinion
[11:22] Richardus Raymaker: ANdrew, boulderdash rocks :)
[11:22] dan banner: ya i notice justin
[11:22] Justin Clark-Casey: because even running it on false still brings up an issue of scripts never being unloaded, which we can't just make the default option yet
[11:23] Nebadon Izumi: hey
[11:23] Nebadon Izumi: i got patch 001 to apply cleanly
[11:23] Justin Clark-Casey: will have to look at having such a switch also load attachment scripts into new appdomains and region scripts into the same appdomain
[11:23] dan banner: well you could leave true the default
[11:23] Nebadon Izumi: let me try patch 002
[11:23] dan banner: would be nice to set the osgrid builds to auto tho
[11:23] Justin Clark-Casey: true is the default right now
[11:23] Richardus Raymaker: that sounds intressting justin
[11:23] Nebadon Izumi: got patch 2 to apply cleanly Andrew
[11:23] Nebadon Izumi: try this git hash
[11:23] dan banner: since osgrid builds are set to false
[11:24] Nebadon Izumi: 6b867773a83e90407203d0f0f6f7fa36121265d9
[11:24] Andrew Hellershanks: Nebadon: what branch are you on?
[11:24] Nebadon Izumi: master
[11:24] Nebadon Izumi: and using that hash
[11:24] Nebadon Izumi: they applied cleanly
[11:24] Justin Clark-Casey: oh, I meant true is the default in opensim itself
[11:24] Nebadon Izumi: i did a fresh clone
[11:24] Nebadon Izumi: git reset --hard 6b867773a83e90407203d0f0f6f7fa36121265d9
[11:24] Nebadon Izumi: then i did
[11:24] Nebadon Izumi: git am --signoff 0001-Added-ability-to-exclude-inventory-items-or-folders-.patch
[11:24] dan banner: ya if it was true on osgrid builds it might not be an issue as much
[11:24] Nebadon Izumi: git am --signoff 0001-Added-ability-to-exclude-inventory-items-or-folders-.patch
[11:24] Nebadon Izumi: and got no errors
[11:25] Nebadon Izumi: i just took the last git hash from Feb 16
[11:25] Andrew Hellershanks: nebadon, patch does not apply
[11:25] Nebadon Izumi: are you getting an error?
[11:25] Nebadon Izumi: i didnt get any error at all
[11:25] Richardus Raymaker: andrew, are there some permissions your missimh as user to do some things maby ?
[11:25] Andrew HellershanksAndrew Hellershanks points to what he just said
[11:26] Nebadon Izumi: huh
[11:26] Nebadon Izumi: i wonder why i get no error
[11:26] Andrew Hellershanks: This is my user account and I'm the only one on this machine
[11:26] BlueWall Slade: workede here
[11:26] Nebadon Izumi: ok well thats semi-good news
[11:27] Andrew Hellershanks: Probably git has screwed up my repo. I'll do a fresh clone and start from there. That's what I usually have to do when I get in to problems with it.
[11:27] Nebadon Izumi: bizzaro they don't apply for Andrew though
[11:27] Nebadon Izumi: ok cool Andrew ya best to start fresh
[11:27] Nebadon Izumi: probbably need to apply them before creating a branch
[11:27] Nebadon Izumi: maybe?
[11:27] Nebadon Izumi: i dunno
[11:27] Nebadon Izumi: i didnt create a branch
[11:28] BlueWall Slade: both worked
[11:28] Andrew Hellershanks: No. Safer to start over.
[11:28] BlueWall Slade: at this point they are in
[11:28] Nebadon Izumi: ok thanks for testing too BlueWall
[11:28] BlueWall Slade: now, what you do is create a branch
[11:28] BlueWall Slade: git branch iar
[11:29] BlueWall Slade: then do git pull
[11:29] Andrew Hellershanks: I don't trust git. I'd rather start over so I know I have a clean tree.
[11:29] BlueWall Slade: that puts the master back in sync
[11:29] Andrew Hellershanks: Besides, I've alread rm'ed it.
[11:29] BlueWall Slade: then cherry-pick the 2 commits on the iar branch
[11:29] BlueWall Slade: resolve the conflicts
[11:29] BlueWall Slade: add / commit / done
[11:30] Andrew Hellershanks: Easier said than done. The conflicts are what caused the headaches in the first place.
[11:30] Nebadon Izumi: ya
[11:30] Nebadon Izumi: that can certainly send you into a tail spin
[11:30] BlueWall Slade: that is from the timespan
[11:30] sim core: :-? I wonder if it's normal, I saw objects with more than 300 linked prims a few times now on laggy situations
[11:30] Andrew Hellershanks: It does with git. Easier to handle in SVN.
[11:31] BlueWall Slade: that is an opinion
[11:31] Nebadon Izumi: sim core, generally those only lag the viewer of the person who selects it
[11:31] Nebadon Izumi: it doesnt lag the sim
[11:31] sim core: K, not even with physics for the avatars bounding box then
[11:31] Nebadon Izumi: nope
[11:32] Nebadon Izumi: it has to do wiht the highlighting of the prims
[11:32] Nebadon Izumi: when you select it
[11:32] BlueWall Slade: ok, I just pulled and it told of the conflicts on the master branch
[11:32] BlueWall Slade: they are listed
[11:32] Nebadon Izumi: OpenGL barfs
[11:32] Nebadon Izumi: cause of all those highlighted lines
[11:32] Richardus Raymaker: CUDA rokcs :)
[11:32] sim core: Right, I just had a lot of login lag for everyone and when I unlinked the 300+ prim object it seemed to go better for everyone
[11:33] Nebadon Izumi: oh unlinking and linking
[11:33] Nebadon Izumi: yes
[11:33] Nebadon Izumi: that can actually cause lag for a moment
[11:33] Richardus Raymaker: the object where not phyical ?
[11:33] Nebadon Izumi: but just selecting it no
[11:33] Andrew Hellershanks: linking/unlinking has some inefficient code.
[11:33] sim core: No, just not phantom
[11:33] Nebadon Izumi: its gotten better
[11:33] sim core: :-) Ok, thank-you nebadon
[11:33] Nebadon Izumi: it used to be a lot worse few months ago
[11:33] Nebadon Izumi: melanie improved it significantly
[11:33] Nebadon Izumi: still needs some love
[11:34] Andrew Hellershanks: I haven't had time to work on the linking issue. THe problem is the use of MapAndArray()
[11:35] Andrew Hellershanks: It has to rebuild the array everytime it (un?)linked a prim.
[11:35] Justin Clark-Casey: I'm thinking of maybe starting to send appearance updates periodicially
[11:35] BlueWall Slade: ok, they're done
[11:35] BlueWall Slade: about 4 small conflicts to resolve
[11:35] Richardus Raymaker: that would possible fix the cloud problem justin
[11:35] Justin Clark-Casey: despite all my previous efforts this hasn't got rid of grey avatars - I almost wonder if this isn't simply a viewer bug which the lindens overcome just by resending the data
[11:35] BlueWall Slade: and those are about the same as working in svn
[11:35] Justin Clark-Casey: richardus: it would not - that's a more fundamental rebaking issue on the client side
[11:35] Nebadon Izumi: Justin if LL does change to server side avatar baking
[11:36] Nebadon Izumi: how hard do you think it will be for us to implement that?
[11:36] Justin Clark-Casey: hard
[11:36] Nebadon Izumi: I just wonder if they will remove it from the viewer
[11:36] Andrew Hellershanks: won't that just add more load to the server?
[11:36] Nebadon Izumi: you would think
[11:36] sim core: Especially if just a ctrl-alt r will fix most problems
[11:36] Nebadon Izumi: but LL is supposed to be changing to server side
[11:36] Nebadon Izumi: ti create a more consistent environment for baking
[11:37] Nebadon Izumi: i guess because of all the TPVs and various hardware, some using OpenJPG some use KDU
[11:37] Nebadon Izumi: things are not quite as consistent as they would like
[11:37] Justin Clark-Casey: yes
[11:37] paulie Flomar: client baking has always been broken, hasnt it?
[11:37] Nebadon Izumi: somewhat
[11:38] Nebadon Izumi: there have always been issues with it yes
[11:38] BlueWall Slade: I see a lot fewer of them that I used to
[11:38] Nebadon Izumi: 2nd monitor all black avatar textures issues
[11:38] BlueWall Slade: greys
[11:38] Richardus Raymaker: it does weird things apulie thats for sure
[11:38] Nebadon Izumi: and many other driver related issues
[11:38] Richardus Raymaker: i think avatars look 95% of the time good here
[11:38] paulie Flomar: less dependence on flaky grafix drivers sound like a positive
[11:38] Nebadon Izumi: and the fact that TPVs are not allowed to iclude KDU or even allow it to be dropped in anymore
[11:39] Nebadon Izumi: forcing them to use OpenJpeg basically
[11:39] Justin Clark-Casey: howver, I have no idea how you would even start to do baking on the server side
[11:39] Nebadon Izumi: ya
[11:39] Dahlia Trimble: viewer uses opengl to bake which is kinda flakey
[11:39] paulie Flomar: rule 2k disallows KDU? :(
[11:39] Nebadon Izumi: i just hope they don't jump into it too soon
[11:39] Nebadon Izumi: it will be interesting if they completely remove it from the viewer at some point
[11:39] Dahlia Trimble: there's baking code in libomv
[11:39] Nebadon Izumi: thats good to know
[11:39] BlueWall Slade: easybake
[11:40] Nebadon Izumi: hehe
[11:40] paulie Flomar: :)
[11:40] Dahlia TrimbleDahlia Trimble wrote the original libomv baker
[11:40] BlueWall Slade: ok, I have that iar code updated
[11:40] paulie Flomar: it bakes likes a 60 watt bulb, baby
[11:40] Richardus Raymaker: maby intressting page, it also explains (with free code sofar i see) how to paint on mesh. etc.
[11:40] BlueWall Slade: Andrew - can you test it if I push
[11:40] BlueWall Slade: ?
[11:41] Andrew Hellershanks: Justin wanted me to do one other change before that code was to be pushed.
[11:41] Richardus Raymaker: and sofar i understand avatar is mesh... i guess linden labs is doing some texture paint before the send it
[11:41] BlueWall Slade: ok
[11:41] Nebadon Izumi: into a branch?
[11:41] BlueWall Slade: I will do that
[11:41] Andrew Hellershanks: Perhaps git screwed up my am patches when I saved them. I'll try the bug report attachments.
[11:41] BlueWall Slade: lol
[11:41] BlueWall Slade: they applied
[11:42] BlueWall Slade: there were only about 4 small conflicts
[11:42] Andrew Hellershanks: One patch I have (not in report) is just for some typos I found.
[11:42] BlueWall Slade: I'll push into a branch
[11:42] Andrew Hellershanks: ok, so they didn't apply out of the box.
[11:42] Nebadon Izumi: cool, thanks BlueWall
[11:42] Nebadon Izumi: crisis averted!
[11:42] Nebadon Izumi: :D
[11:42] BlueWall Slade: well, if the code moves out from under you - you'll need to resolve the conflicts
[11:42] BlueWall Slade: that's just life
[11:44] Andrew Hellershanks: sure but git makes life so much more difficult.
[11:44] Justin Clark-Casey: any other opensim topics today?
[11:44] Nebadon Izumi: heh
[11:44] Richardus Raymaker: hi ruben
[11:45] Ruben Haan: hi
[11:45] Nebadon Izumi: I dont really have anything more myself
[11:45] Richardus Raymaker: it seems dead neighborns can stillc reate lag.. it happend a while ago again to somebody..
[11:46] BlueWall Slade: it's in a branch "iar_mods"
[11:47] Andrew Hellershanks: k, thanks BlueWall. I'll look up my notes on how to checkout a branch.
[11:48] Simulator Version v0.5 shouts: OpenSim 0.7.4 Dev          0007711: 2012-03-01 03:23:10 +0000 (Unix/Mono)
[11:48] Dahlia Trimble: my viewer clock says 11:48 PST
[11:48] BlueWall Slade: ++
[11:48] sim core: :-J They might bot be 'Dead' just behind a firewall
[11:48] sim core: *Not
[11:49] Richardus Raymaker: well it create still problems on neighborn sims
[11:49] Nebadon Izumi: my guess is the sim maybe in a bad state
[11:49] Nebadon Izumi: and not full crashed
[11:49] Nebadon Izumi: i could see that being a potential source for lag
[11:50] Nebadon Izumi: especially if its stuck pounding the neighbor for child agent data
[11:50] VivK Lowlag: got any logs?
[11:50] sim core: :-J Would be good, to be able to dissable the neighbors for that
[11:50] Richardus Raymaker: well something like that seems to happen yes
[11:50] sim core: Can't receive their child agent requestsà
[11:50] Nebadon Izumi: well the only thing i can suggest is temporarily iptable ban them
[11:51] Nebadon Izumi: not a perfect solution, but it would stop the pounding
[11:51] Nebadon Izumi: if your on linux, you can look at iftop or iptraf
[11:51] Nebadon Izumi: to see if your getting excessive requests from the nieghbor
[11:52] Nebadon Izumi: there are probably some good tools for windows too
[11:52] Nebadon Izumi: im just not as up to date with windows monitoring tools
[11:55] Justin Clark-Casey: ok, so as some people probably know, daylight times are shifting again
[11:56] Justin Clark-Casey: so things will be out of sync for a few weeks. I think we'll formally shift this meeting when european daylight kicks in in a few weeks time, when everything will go back to 'normal'
[11:56] Nebadon Izumi: ya there is 1 more week of this nonsense?
[11:56] Nebadon Izumi: when do you change?
[11:56] Nebadon Izumi: 27th or 29th?
[11:56] Nebadon Izumi: something around there
[11:56] sim core: :-J
[11:56] Nebadon Izumi: maybe 2 more weeks
[11:56] Justin Clark-Casey: tyeah, one more week
[11:56] Nebadon Izumi: ok
[11:56] Justin Clark-Casey: clocks change on 25th march
[11:56] Nebadon Izumi: ok cool
[11:57] Nebadon Izumi: i'll still get here early next week
[11:57] Richardus Raymaker: so, next week same time 20:00 ?
[11:57] Nebadon Izumi: just in case
[11:57] Andrew Hellershanks: I think I have my PDA set to remind me based on UTC time.
[11:57] Nebadon Izumi: it will be 19:00 UTC
[11:57] Richardus Raymaker: what where it this week then ? i think 19:00 to
[11:58] Nebadon Izumi: right it will be the same next week
[11:58] Andrew Hellershanks: Ah, it was set based on PST
[11:58] Nebadon Izumi: then after that its to 18:00 UTC?
[11:58] Justin Clark-Casey: yes
[11:58] Nebadon Izumi: grr, freakin time makes my head want to explode
[11:58] paulie Flomar: heh
[11:59] Dahlia Trimble: gotta run, bye all :)
[11:59] Richardus Raymaker: ok. now im complete lost with time.
[11:59] Nebadon Izumi: kk see ya Dahlia
[11:59] Justin Clark-Casey: all you need to know is that if your clock has changed, then it's an hour earlier next week like tihs week
[11:59] Andrew Hellershanks: nebadon, the worst part is if an event is happening in the wee hours of the morning the night the clocks changes.
[11:59] Nebadon Izumi: hehe
[11:59] BlueWall.Slade By eDahlia
[11:59] Justin Clark-Casey: then after that it goes back to normal time
[11:59] Richardus Raymaker: bye dahlia
[11:59] Justin Clark-Casey: bye dahlia
[11:59] Nebadon Izumi: next week 19:00 UTC
[11:59] sim core: =) See-you, everyone and thank-you the devs for the meeting
[11:59] Nebadon Izumi: 2 weeks from now 18:00UTC
[11:59] Justin Clark-Casey: bye sim
[11:59] Richardus Raymaker: yes and the week after that its normal ?
[11:59] Nebadon Izumi: see ya Sim
[11:59] dan banner: bye everyone
[11:59] Ni Lemon: bye sim
[11:59] BlueWall.Slade bye sim
[11:59] Andrew Hellershanks: justin: this was an hour later for me.
[11:59] Ni Lemon: bye dan
[11:59] Justin Clark-Casey: hour plus, minus, whatever
[11:59] Andrew Hellershanks: hehe
[11:59] Nebadon Izumi: lol
[11:59] Andrew Hellershanks: one of the above. :-)
[11:59] BlueWall.Slade g'bye dan
[12:00] Richardus Raymaker: time zones i hate them. sure i get screwed one of this weeks
[12:00] Justin Clark-Casey: ok, I need to pop too
[12:00] Justin Clark-Casey: see you guys around
[12:00] Andrew Hellershanks: I'm also going to head out.
[12:00] Richardus Raymaker: bye justin
[12:00] Richardus Raymaker: bye andrew
[12:00] paulie Flomar: I'm gonna pop over to LBSA. Catch yall. :)
[12:00] Justin Clark-CaseyJustin Clark-Casey waves
[12:00] Ni Lemon: see you all and thanks
Personal tools
About This Wiki