Open Source Meeting/2008-05-08

From Second Life Wiki
Jump to navigation Jump to search
  • [13:55] Elvis Marlowe: hola... is this the FLOSS triage?
  • [13:58] Michelle2 Zenovka: Hey everyone
  • [13:58] Opensource Obscure: hi michelle and all
  • [13:58] Elvis Marlowe: hola
  • [13:59] Carjay McGinnis: evening
  • [13:59] Elvis Marlowe: hola
  • [14:00] Opensource Obscure: anyone got a complete and up-to-date list of Debug Settings or know where I should look for it?
  • [14:01] Rob Linden: hi all
  • [14:01] Carjay McGinnis: Hello Rob
  • [14:01] Rob Linden: transcripts will be posted later (assuming I get to it; I'm behind)
  • [14:01] Rob Linden: would love to have a transcript volunteer
  • [14:02] Rob Linden: anyhoo...be aware that what you type here will show up on wiki.secondlife.com
  • [14:02] Michelle2 Zenovka: Hi ya Rob
  • [14:02] Rob Linden: while I'm trolling for volunteers; it'd be nice if someone could be agenda wrangler
  • [14:04] Rob Linden: alright....no one is immediately stepping forward, so I'll randomly pick someone
  • [14:04] Rob Linden: just kidding
  • [14:04] Magnet Homewood: Sorry, pick for what?
  • [14:04] Infinity Linden: Rob... for what are you seeking a volunteer?
  • [14:04] Magnet Homewood: If its money tree tester, I am in 100%
  • [14:05] Rob Linden: looking for two volunteers actually:
  • [14:05] Stephen Posaner: to do what
  • [14:05] Rob Linden: 1. Agenda wrangler for next week, so that we don't find ourselves agendaless
  • [14:05] Rob Linden: (or with a stale agenda, like this week)
  • [14:05] Rob Linden: 2. Transcript poster
  • [14:05] Michelle2 Zenovka: will wrangle an agenda for next week
  • [14:06] Infinity Linden: If agend-ing is just to capture agenda items, I'll volunteer
  • [14:06] Rob Linden: xlnt, thanks Michelle2
  • [14:06] Rob Linden: of course, everyone is welcome to chip in. it is a wiki and all
  • [14:06] Infinity Linden: beaten to the punch
  • [14:06] Stephen Posaner: ok
  • [14:06] Stephen Posaner: I take it transcripts are just the caht log
  • [14:07] Rob Linden: yup
  • [14:07] Rob Linden: there's a transcript reformatter tool that Tree Kyomoon wrote
  • [14:07] Rob Linden: link is up on the Open Source Meeting page
  • [14:07] Stephen Posaner: omg all i would have to do is hack out the invites i get from things anf I could help withthat
  • [14:07] Stephen Posaner: ok
  • [14:08] Q Linden: tool may do that for you?
  • [14:08] Rob Linden: tool actually does that
  • [14:08] Saijanai Kuhn: just type in what you want to filter out
  • [14:08] Rob Linden: Stephen, you got it?
  • [14:08] Rob Linden: if so, we can move on from administrivia
  • [14:08] Stephen Posaner: let me play with the tool a week then I will know for sure i can
  • [14:09] Stephen Posaner: but i will hand do it this week
  • [14:09] Saijanai Kuhn: taks 30 seconds once you get it
  • [14:09] Stephen Posaner: kool
  • [14:09] Rob Linden: k...that's cool. I'll make sure I save off my copy in case you don't get around to it
  • [14:09] Ellla McMahon: our apologies .. we were looking for Soft Linden re bug triage ... have a great meeting
  • [14:09] Rob Linden: alrighty: updates from Lindens
  • [14:09] Rob Linden: I'll start
  • [14:10] Saijanai Kuhn: soft is on #opensl irc .efnet.net
  • [14:10] Rob Linden: I spent some time on the discussion of our contribution agreement, both internal and public conversation
  • [14:10] Saijanai Kuhn: llsoft
  • [14:11] Rob Linden: if that's a burning issue for folks, let's tack that on as an agenda item today
  • [14:12] Rob Linden: we're more closely investigating webkit, and I've been doing some stuff there (intentionally vague about that)
  • [14:12] Squirrel Wood: /ao off
  • [14:13] Rob Linden: PWIKI/PJIRA should be in good shape now....let me know if you're still seeing microoutages
  • [14:13] Rob Linden: that's it for me
  • [14:13] Liana Linden: I've been lending a hand with the CMake project. Other than that, it's been a bunch of behind-the-scenes lifting and carrying -- nothing exciting to report today.
  • [14:14] Q Linden: Yeah, i'm still doing notifications work, great progress; I had hoped to post to SLDev this week but didn't get there yet.
  • [14:14] Saijanai Kuhn: Rob, it might be me or my ISP, but about 3 out of 10 times I get a error message when I try to access the wiki
  • [14:15] Rob Linden: what's the error?
  • [14:15] Saijanai Kuhn: server not found or unable to connect
  • [14:15] Saijanai Kuhn: I try again 2-3 times and it pops up the page.
  • [14:15] Rob Linden: how long has that been happening?
  • [14:15] Opensource Obscure: i had some 'cant reach the website' yesterday. it solves itself after a few minutes, for me.
  • [14:15] Saijanai Kuhn: last couple of weeks. I knew you were working on it, so I didn't say much, but it happened again thismorning
  • [14:16] Rob Linden: when was the last time it happened?
  • [14:16] Saijanai Kuhn: this morning about 2 hours ago
  • [14:17] Rob Linden: ok....hrm. I'll check in with our service provider.
  • [14:17] Saijanai Kuhn: it happens with an internal page but 100% fix if I first go to the main page
  • [14:17] Celierra Darling: I think I saw that just now clicking on the link you sent in #opensl, rob
  • [14:18] Rob Linden: okee dokee.....looks like there's more work to be done.
  • [14:18] Rob Linden: my understading from them is that there's some ARP storm problem they're dealing with that's been a game of whack-a-mole
  • [14:18] Saijanai Kuhn: it often happens when I'm saving content :-/
  • [14:19] Rob Linden: Saijanai: I'm kinda thinking the problems you're seeing may be a combination of problems on our end and on your end
  • [14:19] Saijanai Kuhn: wouldn't surprise me. My ISP has been busy messing up my service recently
  • [14:19] Michelle2 Zenovka: wiki server just reset connection on me
  • [14:20] Saijanai Kuhn: but it seems to be the wiki that shows the intermittent problem most often
  • [14:20] Kooky Jetaime: sai - same here..
  • [14:20] Rob Linden: hmmm...I have't seen the timeout since Monday or Tuesday
  • [14:21] Rob Linden: anyway more Linden updates? also, any non-Lindens working on patches or other contributions that they want to give an update on?
  • [14:21] Saijanai Kuhn: Enus and I have been working on test harness issues for the new login protocol
  • [14:22] Michelle2 Zenovka: indirectly: xmlrpc-epi is uploaded to debian, but i rememebr you said you may drop this anyway at some point
  • [14:23] Squirrel Wood: My terraformer is at a point where I can actually call it "Release Candidate" :) (It's not open sauce though :p)
  • [14:23] Rob Linden: Michelle2: that's cool nonetheless. I have no idea what our plans are on that component
  • [14:23] Rob Linden: (any other Lindens know?)
  • [14:23] Michelle2 Zenovka: So only mozlib in my way now of pushing the viewer debians way really.
  • [14:24] Rob Linden: Callum is at the Mozilla Embedders meetup today
  • [14:24] Rob Linden: if there's any chance of getting our issues addressed such that we don't need a fork, we'll likely know today
  • [14:25] Michelle2 Zenovka: good timing
  • [14:25] Rob Linden: we would love help from you all to bridge the gap there
  • [14:26] Rob Linden: we don't have an acromonious relationship with Mozilla, but this particular use of it hasn't been on their radar
  • [14:26] Kooky Jetaime: hehh
  • [14:26] Rob Linden: ....so they've not been motivated to either take our patches or suggest an alternate path that would be accepted upstream
  • [14:26] Rob Linden: hence the fork
  • [14:27] Michelle2 Zenovka: yea well please do report back the news on that front
  • [14:27] Kooky Jetaime: Love that forkin.
  • [14:27] Michelle2 Zenovka: i've got a fork of the fork ;-(
  • [14:28] Rob Linden: well, ultimately, the fork isn't a sustainable strategy for us, which is why webkit is on the table
  • [14:28] Opensource Obscure: *crash*
  • [14:29] Rob Linden: what are the other pain points on the debian front, M2?
  • [14:29] Michelle2 Zenovka: only final trademark issues now
  • [14:30] Liana Linden: Michelle2, you'll know more as soon as I do!
  • [14:30] Liana Linden:  :-)
  • [14:30] Michelle2 Zenovka: All other dependencies are met and artwork (apart from TMs) are acceptable now for main
  • [14:31] Michelle2 Zenovka: hehe Thanks Liana!
  • [14:32] Rob Linden: Saijanai and Enus, did you want to talk more about the test harness?
  • [14:32] Enus Linden: well.... sure.
  • [14:33] Enus Linden: AWG is starting to move forward with design of a test harness that will be focused on testing the open grid protocols as they come to fruition
  • [14:33] Enus Linden: starting to think about how it will work, what it will be written in, how to extend it etc etc etc
  • [14:33] Enus Linden: i posted a started to sldev, and to wiki
  • [14:34] Enus Linden: (will get wiki link in a sec)
  • [14:34] Enus Linden: looks like a python lib
  • [14:34] Enus Linden: much like libsl
  • [14:34] Enus Linden: is where we hope to head
  • [14:34] Enus Linden: need to work out repository/licensing details as well.
  • [14:35] Enus Linden: but, we're just gettings started.
  • [14:35] Saijanai Kuhn: Enus and I are looking at that sample login script I put up. I spent the weekend getting my mac to run more than the default 100 processes, and got 300 micro-POST-servers running as their own python scripts. The idea being you could control a bunch of scripts via POST to each script
  • [14:35] Enus Linden: anyone interested, please lemme know...
  • [14:35] Rob Linden: [1]
  • [14:35] Enus Linden: we start with login, which sai has been looking at (see above:)
  • [14:35] Saijanai Kuhn: so in theory you could run 300 or more login scripts at once to stress test the login protocol on the beta grid and send various control commands to each individual script
  • [14:36] Rob Linden: very cool
  • [14:36] Saijanai Kuhn: so as you guys extend the new protocols, we can add new functions to the login scripts and control them via POST
  • [14:36] Saijanai Kuhn: Which says the technique should work for any number of scripts distributed on a LAN
  • [14:37] Saijanai Kuhn: so we could spam any sized grid you wanted to test that way
  • [14:37] Enus Linden: libsl folks pointed out: why not use libsl?
  • [14:37] Enus Linden: i'll raise that at the next awg meeting as an option
  • [14:37] Saijanai Kuhn: overkill at this point, I think
  • [14:38] Saijanai Kuhn: easier to mod a python script and run it multiple times than to recompile libsl
  • [14:38] Q Linden: I'd rather not anoint libsl as the "right" API in any way. :-)
  • [14:38] Poppy Linden: it does compile really quickly
  • [14:38] Poppy Linden: in the order of seconds
  • [14:38] Enus Linden: why Q? not that i disagree, but reasons?
  • [14:38] Saijanai Kuhn: (not a C# guru either -> Mac user)
  • [14:38] Q Linden: it's accidental, not designed
  • [14:38] Enus Linden: right
  • [14:39] Enus Linden: that's my thought too.
  • [14:39] Enus Linden: we want code written to test the protocol as published
  • [14:39] Rob Linden: I suppose it would be good to have an implementation that didn't evolve from our existing protocol
  • [14:39] Q Linden: would rather build a toolkit like you're doing, thinking about it carefully
  • [14:40] Saijanai Kuhn: I used the python script to test my login protocol docs, so any mods would reflect the docs, rather than the other way around
  • [14:41] Saijanai Kuhn: Zha wanted the docs to be language neutral, so she tested teh docs by writing a java applet to do the same thing just by reading the docs
  • [14:42] Kooky Jetaime: Nice
  • [14:43] Saijanai Kuhn: we'd like to use that technique all the way through. That way the OGP doc that Zero is writing gets stressed tested in at least two langauges other than C++
  • [14:43] Rob Linden: any other contributions or Linden-updates to discuss?
  • [14:44] Twisted Laws: I have a question when/if its appropriate.
  • [14:44] Rob Linden: go for it
  • [14:44] Twisted Laws: I have been compiling and running an alternate version of SL on Windows Vista based on open source and changes of my own since 1.18.3. I'm currently using Visual C++ 2008 Express. All worked good until we got the the 1.20.x versions. Now I can compile but get an immediate infinite loop in a visual c++ location that is called from boost. Issue appears to be a head pointer that points to self, when Visual C++ library expects the head pointer to be null at head of list. Any pointers for working around this or who to talk to?
  • [14:44] Rob Linden: ooh....hmm. well, first thing, file something in JIRA if you haven't already
  • [14:45] Q Linden: sounds like mixed libraries
  • [14:45] Q Linden: rebuild all?
  • [14:45] Twisted Laws: yep... it probably is an issue with boost version and compiler version...
  • [14:45] Rob Linden: oh....yeah, listen to Q. he actually knows what he's talking about
  • [14:45] Q Linden: definitely make sure you rebuild boost
  • [14:46] Twisted Laws: ok, so the boost libraries that come with open source probably do need recompiling then?
  • [14:46] Michelle2 Zenovka: i would to make absolutley sure
  • [14:47] Michelle2 Zenovka: boost is very tetchy with compiler versions as its a c++ library
  • [14:47] Q Linden: right
  • [14:47] Rob Linden: I think it's one of those "in theory, no; in practice, yes"
  • [14:47] Twisted Laws: did LL gain an advantage by going with boost (only using a small part of it) ?
  • [14:49] Poppy Linden: as far as I've heard, using the whole thing is overkill for download size and memory size
  • [14:49] Poppy Linden: sits next to two people who work with it quite a bit, but doesn't much himself
  • [14:49] Rob Linden: hmmm...interesting
  • [14:50] Rob Linden: I didn't realize boost was monolithic like that
  • [14:50] Twisted Laws: yes, i spent some time looking at how to work around it (not use boost)... it appears that i can but want to keep with what LL is doing.
  • [14:50] Carjay McGinnis: it's a collection of libraries, guess nobody ever uses everything
  • [14:50] Poppy Linden: i mean, if we only use regex, why have the rest? :P
  • [14:50] Infinity Linden: Rob: another case of... in theory: no; in practice: yes.
  • [14:50] Rob Linden:  :)
  • [14:50] Michelle2 Zenovka: youve got signals and program options too
  • [14:50] Michelle2 Zenovka: as of 1.20
  • [14:51] Q Linden: we're using more boost over time
  • [14:51] Rob Linden: lord knows its better than doing our own homespun versions
  • [14:51] Michelle2 Zenovka: debian has about 18 boost sub packages
  • [14:51] Carjay McGinnis: the smart pointers are nice :)
  • [14:51] Q Linden: boost bind, smart ptrs, regex, signals -- all important stuff
  • [14:52] Carjay McGinnis: ah, right, bind is cool, too, never have to mess with nasty function pointer callbacks
  • [14:52] Rob Linden: anything else on this topic, or should we move on?
  • [14:52] Twisted Laws: ok, i will compile my boost for my compiler and try that approach, thanks. btw, wiki doesn't give help on compiling boost last i saw.
  • [14:53] Q Linden: see boost docs
  • [14:53] Poppy Linden: i think it's normal boost
  • [14:53] Poppy Linden: ^^^
  • [14:53] Michelle2 Zenovka: should be well documentated for windows
  • [14:53] Twisted Laws: i meant for what libraries are necessary, etc..
  • [14:53] Michelle2 Zenovka: normal boost works for us on linux ;-)
  • [14:53] Poppy Linden: Q is for Quick ;)
  • [14:53] Carjay McGinnis: the boost libs also have an installer where you can select the compiler you want
  • [14:54] Twisted Laws: yes
  • [14:54] Carjay McGinnis: but I'm not sure if they used SP1
  • [14:55] Rob Linden: Twisted: if you aren't on irc.efnet.org #opensl, you might want to be for help in getting up to speed faster
  • [14:55] Rob Linden: most of us also hang out there
  • [14:55] Twisted Laws: ok, will look at that... only on sldev digest atm
  • [14:56] Twisted Laws: thank you
  • [14:56] Rob Linden: no prob....good topic actually
  • [14:56] Rob Linden: we're just about to the end of the hour
  • [14:56] Rob Linden: any last minute stuff?
  • [14:56] Opensource Obscure: last Linux RC crashes for me more often than previous version. .. do you think these logs are helpful enough, should I submit them to PJIRA? I don't even know what to search to see if they are known issues .. [2] - [3]
  • [14:57] Rob Linden: looks
  • [14:58] Kooky Jetaime: Heading for Ben's.. later all :)
  • [14:58] Michelle2 Zenovka: got some more openjpeg beating with a stick to be done
  • [14:58] Liana Linden: thanks, all.
  • [14:58] Rob Linden: OO: I'm not sure. If you have repro cases, definitely post to JIRA
  • [14:59] Opensource Obscure: which title should i use for similar reports ?
  • [14:59] Rob Linden: thanks everyone
  • [14:59] Carjay McGinnis: hm, too bad the stack trace is useless
  • [14:59] Infinity Linden: thx Rob...
  • [14:59] Poppy Linden: i'm beatinng up on cmake mac a lot right now, changing the way libs are being done yet again, fyi mac cmakers
  • [14:59] Opensource Obscure: carjay: ah! :\
  • [14:59] Carjay McGinnis: that looks almost like the multithread string issue I found happening on my machine
  • [14:59] Stephen Posaner: I will get the transcript posted with in the hr
  • [15:00] Rob Linden: OO: well, the titel should relate to the repro case
  • [15:00] Carjay McGinnis: one thread deletes the string while another is trying to append to it
  • [15:00] Michelle2 Zenovka: Is that an unhandled SIGINFO?
  • [15:00] Opensource Obscure: ok. thanks, rob
  • [15:00] Rob Linden: thanks Stephen
  • [15:00] Rob Linden: bye everyone
  • [15:00] Twisted Laws: bye
  • [15:00] Q Linden: bye
  • [15:00] Stephen Posaner: bye
  • [15:00] Carjay McGinnis: bye and thanks
  • [15:01] Opensource Obscure: how may I provide a better output?
  • [15:01] Celierra Darling: waves
  • [15:01] Saijanai Kuhn: laters
  • [15:01] Carjay McGinnis: hm, I am not familiar with the way these stack traces are generated
  • [15:01] Michelle2 Zenovka: What would really help are the debug symbols for the offical clients
  • [15:01] Opensource Obscure: is something i can turn on on the fly, or should I recompile from sources, M2?
  • [15:01] Poppy Linden: hmm
  • [15:01] Carjay McGinnis: indeed
  • [15:02] Michelle2 Zenovka: You are building your self?
  • [15:02] Opensource Obscure: usually not
  • [15:02] Saijanai Kuhn: heading to benjamen's to see how that project is going later all
  • [15:02] Opensource Obscure: i mean: is there a way to get better output from the official builds?
  • [15:02] Carjay McGinnis: they are probably stripped
  • [15:03] Michelle2 Zenovka: afraid not, not sure if the lindes have a tool to reassemble these traces?
  • [15:03] Michelle2 Zenovka: *lindens
  • [15:03] Opensource Obscure: ok, thanks all. and ciao all
  • [15:03] Carjay McGinnis: well, if there was a map file one could at least guess the position
  • [15:04] Saijanai Kuhn: LOL google calander gives option to bring up map of USA to find "Linden Village"
  • [15:04] Poppy Linden: logs are the best right now afaik
  • [15:04] Opensource Obscure: ok
  • [15:04] Michelle2 Zenovka: Right i'm off to 1000m to debug textures
  • [15:05] Celierra Darling: poofs!
  • [15:05] Poppy Linden: ahh
  • [15:05] Celierra Darling: also melts from really bright sunlight >.>
  • [15:05] Carjay McGinnis: do you need an orbiter, Michelle? *g*
  • [15:05] Michelle2 Zenovka: hehe
  • [15:05] Michelle2 Zenovka: i'll be back on irc with any results
  • [15:05] Carjay McGinnis: i was attacked 2 days ago, always thought havok4 had stopped them but they seem still to work
  • [15:06] Michelle2 Zenovka: ohhh
  • [15:06] Michelle2 Zenovka: well sim crossings and even getting off veichels seems to sometimes orbit too *lol*
  • [15:07] Carjay McGinnis: bye everyone, am off, too
  • [15:09] Sara Gyoza: kay, buh bye everyone.
  • [15:09] Sara Gyoza: ttfn