Open Source Meeting/2008-07-10

From Second Life Wiki
Jump to: navigation, search
  • [14:02] Rob Linden: hi all
  • [14:02] Carjay McGinnis: it's appreciated
  • [14:02] Carjay McGinnis: greetings, Rob
  • [14:02] Squirrel Wood: old look? not for me :p
  • [14:02] Rob Linden: transcript will be posted on wiki.secondlife.com, so keep that in mind as you type
  • [14:02] Joeseph Albanese: lol
  • [14:02] Soft Linden: The blue was counterimmersive to me (to coin a word, I think!). The neutral gray is easier to focus past.
  • [14:03] Squirrel Wood: I prefer a minimalistic UI
  • [14:03] Aimee Trescothick: yeah, definitely
  • [14:03] Squirrel Wood: the less you see of it, the better
  • [14:03] Jacek Antonelli: waves to Poppy
  • [14:03] Jacek Antonelli: Don't shoot, I'm unarmed!
  • [14:04] Squirrel Wood: To not waste precious time, what's the topic for today?
  • [14:04] Poppy Linden: hello jacek
  • [14:04] Jacek Antonelli: I'm here to answer any questions about my menu patch :D
  • [14:05] Carjay McGinnis: ah, right, read on sldev about it
  • [14:05] Michelle2 Zenovka: http://wiki.secondlife.com/wiki/Open_Source_Meeting/Agenda
  • [14:06] Michelle2 Zenovka: looks like its my "VWR-5082[c
  • [14:06] Rob Linden: M2 wins :)
  • [14:06] Rob Linden: let's get started
  • [14:07] Michelle2 Zenovka: hehe
  • [14:07] Poppy Linden: is kinda busy, will be half-here
  • [14:07] Michelle2 Zenovka: SO yea, does what it says. massive time saver and builders friend
  • [14:07] Jacek Antonelli: Whoa. Does your patch work, M2?
  • [14:08] Michelle2 Zenovka: seems to
  • [14:08] Jacek Antonelli:  ! :-O *downloads as fast as she can*
  • [14:08] Tomiko Magic: Eh soooorrry to stand here like a mellon, I had to afk
  • [14:08] Tomiko Magic: I'll get out of your way! : D
  • [14:08] Rob Linden: yeah, Tofu, Qarl and I discussed this a little bit this morning (hence the update on "Last Triaged"). Looks good in principle.
  • [14:08] Michelle2 Zenovka: probably needs some polish
  • [14:08] Carjay McGinnis: nice one
  • [14:08] Rob Linden: we wanted to make sure Benjamin and the Rx team gets a chance to take a look
  • [14:08] Michelle2 Zenovka: cool
  • [14:09] Rob Linden: Michelle2, was there anything specific you wanted some feedback on here?
  • [14:09] Michelle2 Zenovka: no just to bring it to attention, unless any one had any issues with the implementation
  • [14:10] Poppy Linden: have a testplan?
  • [14:10] Poppy Linden: is reading diff now...
  • [14:12] Rob Linden: let's move along to the next item, and we can circle back if there's more to discss
  • [14:12] Michelle2 Zenovka: ack
  • [14:13] Rob Linden: next up: Jacek Antonelli's Advanced menu to XUI cleanup (VWR-2896[c]VWR-8056[c
  • [14:13] Michelle2 Zenovka: Jacek, can you give a quick overview
  • [14:13] Jacek Antonelli: 2896 is terribly out of date -- the stuff on the mailing list is the more recent work
  • [14:14] Jacek Antonelli: Basically.... I ported the Advanced menu to XUI
  • [14:14] Jacek Antonelli: Last time I did it (2896) it got run over by some internal menu changes and never got applied, and then got out of date -- so now I've updated it :D
  • [14:14] Rob Linden: (as an aside: can you update VWR-2896 if its not up to date?)
  • [14:14] Jacek Antonelli: Will do
  • [14:15] Carjay McGinnis: Maybe I overread it, what are the benefits of having it as XUI?
  • [14:15] Michelle2 Zenovka: Does this need to go to Benjimins or some other UI member for a peek
  • [14:15] Michelle2 Zenovka: consistancy
  • [14:15] Jacek Antonelli: From a user's perspective, it's almost exactly the same, behavior and look
  • [14:16] Michelle2 Zenovka: easier to move items around and even move out of advanced to other menus, just by XUI changes
  • [14:16] Jacek Antonelli: But I can poke Ben about it later today
  • [14:16] Jacek Antonelli: Yes, an example of what you could do with this, is to move the "Rebake Textures" tool into one of the main menus
  • [14:16] Jacek Antonelli: (Without having to recompile every time you change the menus)
  • [14:16] Michelle2 Zenovka: also one code path is got to be better than a seperate menu in c++ for one dropdown
  • [14:17] Carjay McGinnis: ok, but you still have to recompile if you want to add an entry
  • [14:18] Jacek Antonelli: correct, you'd have to program in the new functionality with C++
  • [14:18] Ann Otoole: how feasible would it be to make some hot bars with user configurable command via this XUI capability? (just a wish list item. 3D UIs tend to have user configurable hot bars these days)
  • [14:18] Jacek Antonelli: Really, the feedback I need on this one is some eyeballs to look at it -- it's a bit patch, because there's a lot of menu items that I had to convert. And I want to make sure the patch actually applies :)
  • [14:18] Jacek Antonelli: *a big patch
  • [14:18] Michelle2 Zenovka: it seems to apply for me ok
  • [14:18] Jacek Antonelli: Yay :D
  • [14:19] Carjay McGinnis: hm, guess it won't for me as I have added menus in there
  • [14:19] Carjay McGinnis: would need to find the patches and remove them to test it
  • [14:20] Jacek Antonelli: it shouldn't be too hard to just copy and paste most of it, it's pretty much all additions near the end
  • [14:21] Jacek Antonelli: I'll make a JIRA for this and link it to the old one, if the patch seems alright for other people
  • [14:21] Rob Linden: Jacek: it'll be easier if you update the attachment in the existing issue rather than create a new one
  • [14:21] Carjay McGinnis: the patch is in the sldev mail? I see 2 patches in there
  • [14:21] Rob Linden: (assuming the description of the fix is the same)
  • [14:22] Jacek Antonelli: sure
  • [14:22] Jacek Antonelli: Carjay: there's one patch for the C++ side, and another for the XML side (plus I included my whole XML file, for convenience)
  • [14:22] Rob Linden: it's ok to leave both patches there if you feel like that's appropriate (e.g. want to make sre there's something there for those wanting to patch earlier viewers). Just make sure that it's clear which patch applies to which version of the viewer.
  • [14:23] Jacek Antonelli: gotcha, Rob
  • [14:23] Rob Linden: yay! :)
  • [14:24] Rob Linden: are we ready to move onto the next item?
  • [14:24] Jacek Antonelli: I think so
  • [14:24] Rob Linden: Next item: Where are we with the 1.20.X branch, how close to release, whats held up in QA, any other issues? etc Michelle2 Zenovka 05:54, 10 July 2008 (PDT)
  • [14:24] Rob Linden: my understanding is that it's getting pretty close
  • [14:25] Michelle2 Zenovka: certainly .12 is the most stable for ages
  • [14:25] Ann Otoole: is everyone using the rc to evaluate it? I haven't had any issues with the last 2.
  • [14:25] Jacek Antonelli: 12 and 13 have been really good for me
  • [14:25] Q Linden: I'm using 13, and I'm very happy with stability
  • [14:25] Mm Alder: What about the SmartHeap crashes? Some people seem to still see them.
  • [14:26] Mm Alder: I think Torley had one while filming Mitch Kapor. :-)
  • [14:26] Rob Linden: Mm, I'm not sure. what JIRA issues are those....I can look p the internal chatter in real time on those
  • [14:26] Lars Donardson: waves..I was about to say that
  • [14:26] Carjay McGinnis: yes, stability is quite good for me, too
  • [14:26] Lars Donardson: smart heap still hits pretty frequently with me
  • [14:27] Aimee Trescothick: hmm, just trying to compile the latest svn ... where'd paths.xml go? :D
  • [14:27] Ann Otoole: i haven't seen a smartheap in ages. then again i fried up the old dual sli 6800 rig and have a new vista 64 box with 4gb ram
  • [14:27] Michelle2 Zenovka: chanced fate
  • [14:27] Lars Donardson: whispers "and Torley as well"
  • [14:27] Soft Linden: Aimee: Looking.
  • [14:27] Aimee Trescothick:  :D
  • [14:28] Mm Alder: VWR-3943
  • [14:28] Rob Linden: loks
  • [14:28] Rob Linden: looks even
  • [14:28] Tomiko Magic: lol
  • [14:29] Carjay McGinnis: Aimee: no paths.xml in there, is it Mac-specific? Because I compiled 2 hours ago and it worked ok
  • [14:29] Aimee Trescothick: hmm, could be just mac ...
  • [14:30] Soft Linden: newview/*.xml is exported, but not newview/skins/*.xml - that's a problem.
  • [14:30] Jacek Antonelli: ah
  • [14:30] Soft Linden: I'll nudge that in and reexport by the end of the meeting.
  • [14:30] Aimee Trescothick: ahh :)
  • [14:30] Carjay McGinnis: hm, right, found it in an older branch it just contains the xui paths
  • [14:31] Carjay McGinnis: but what is it for?
  • [14:32] Rob Linden: Re: VWR-3943, we seem to have a failure to communicate.
  • [14:32] Soft Linden: I believe that paths.xml tells the order in which to search xui files for elements. Basically just - search the current language for an element, then fall back to en-us if it didn't exist.
  • [14:32] Q Linden: paths.xml appears to define a set of search paths based on ... what he said
  • [14:33] Aimee Trescothick: lol
  • [14:33] Carjay McGinnis: ah, ok
  • [14:33] Rob Linden: I think we have a long term fix for that problem, but the internal discussion looks way too stale to be relevant to this conversatoin. Let's make sure this is on the agenda for next week
  • [14:34] Rob Linden: (where "that problem" is the Smartheap stuff)
  • [14:34] Carjay McGinnis: hm, I have heard from people about that "out of memory" problem
  • [14:34] Carjay McGinnis: was wondering what is causing it, those experiencing it simply click "ignore" until they crash, lol
  • [14:35] Rob Linden: I have to believe that this isn't anything new to 1.20, though, right? Not that that is great or anything, but I'm hoping this isn't an RC blocker in everyone's mind.
  • [14:35] Carjay McGinnis: which obviously is not the solution
  • [14:35] Rob Linden: (I say that because I'm guessing we're not going to fix the smartheap problems in the 1.20 cycle
  • [14:35] Carjay McGinnis: I only use Linux so cant say anything about it
  • [14:36] Mm Alder: It's a long standing problem, but it was marked as fixed in 1.20
  • [14:37] Rob Linden: well, it looks like it was marked as fixed by Gabriel Spiesel (who I don't know)
  • [14:38] Q Linden: I'm hearing that smartheap probably won't make 1.20
  • [14:38] Q Linden: but hopefully 1.21
  • [14:38] Rob Linden: thanks Q
  • [14:38] Rob Linden: we shoudl probably get that linked up wit the right internal issue so we know to fix it up....but that's just AR bookkeeping on my part
  • [14:39] Carjay McGinnis: AR?
  • [14:39] Rob Linden: anal retentive :)
  • [14:39] Carjay McGinnis: ah, ok, lol
  • [14:39] Rob Linden: anyhoo
  • [14:39] Carjay McGinnis: thought it was some specific internal Linden expression
  • [14:40] Q Linden: only when we talk about Rob
  • [14:40] Jacek Antonelli: lol
  • [14:40] Rob Linden: Q :-P
  • [14:40] Rob Linden: next up: Any web kit news? http://meta.nuanti.com/ which was linked from original news seems to not been updated. Any chance of getting some pre alpha code to play with?
  • [14:40] Rob Linden: sadly, nothing new to report. we're all on the edge of our seats
  • [14:41] Poppy Linden: i heard about the memory leaks, i *think* bao could repro it but it seems *very* config dependent (specific win, ati card, etc) - if any of you have insight please share
  • [14:41] Michelle2 Zenovka: supporting xulrunner is getting *hairy*
  • [14:42] Mm Alder: Poppy: I came to the same conclusion. I wa not able to reproduce it with my configureation. I was wondering if the crash logs could shed any light on it.
  • [14:42] Rob Linden: Michelle2: we feel your pain, really.
  • [14:43] Squirrel Wood: My RC13 behaves quite nicely so far
  • [14:43] Michelle2 Zenovka: at least you ship a well defined version ;-) i build on etch and support up to debian unstable
  • [14:44] Squirrel Wood: + Second Life | 14,06% | | 15:55 | 3:13 | 993 | 271.460 | 249.680 | 417.872 | 25 | Norm | 1.483 | 19 | 67 | 131 | Linden Lab | Not Signed | F:\Games\SecondLifeRC\SecondLifeReleaseCandidate.exe (270mb inmem, 249mb private, 410mb total)
  • [14:44] Rob Linden: since there is really not much to talk about on webkit, we should move on (unless we want to move back and talk about smartheap some more)
  • [14:45] Rob Linden: next item: OGP/PYOGP/Open Grid Beta? Not saying that I'm going to talk about all of these just thinking there should be a time to present what is going on regarding these projects to the rest of the open source folk. Saijanai 13:32, 10 July 2008 (PDT)
  • [14:45] Saijanai Kuhn: waves
  • [14:45] Saijanai Kuhn: /me
  • [14:45] Saijanai Kuhn: everyone's seen that blog announcement http://blog.secondlife.com/2008/07/08/ibm-linden-lab-interoperability-announcement/
  • [14:45] Saijanai Kuhn: so the big news is the webpage for the beta grid and WHump announced this morning a signup procedure for the beta
  • [14:46] Mm Alder: I liked the video. :-)
  • [14:46] Saijanai Kuhn: http://wiki.secondlife.com/wiki/Open_Grid_Public_Beta and
  • https://lists.secondlife.com/cgi-bin/mailman/listinfo/gridnauts
  • [14:46] Saijanai Kuhn: interesting embedded paragraph
  • [14:46] Rob Linden: brb...
  • [14:47] Saijanai Kuhn: as far as pyogp goes, Tao and Locklain are injoying designing the design at this point ;-)
  • [14:47] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Pyogp
  • [14:47] Saijanai Kuhn: Tao is talking about implementing a faux Agent Domain server in Plone
  • [14:48] Saijanai Kuhn: what is lacking is coordination between the beta grid folk and pyogp and ogp and the rest of the opensource community
  • [14:48] Q Linden: Saij : what kind of coordination are you thinking of?
  • [14:49] Saijanai Kuhn: well, one thing stands out. The pyogp is supposed to test the protocols. Apparently they're not far enough along to put into pyogp, which suggests that they're notfar enough along to use in the beta test either
  • [14:49] Saijanai Kuhn: the TP I mean
  • [14:49] Q Linden: TP protocols, you mean?
  • [14:49] Saijanai Kuhn: we should have some test of the protocols using the test harness before we tell everyone they're ready for general testing
  • [14:50] Q Linden: As I understand it, the point of the test harness is to test the protocols by referring only to the spec, so that we can prove the spec is good
  • [14:50] Saijanai Kuhn: can have 100 python bots or whatever bouncing back and forth easier and with less egg on the face than with all the opensim folk trynig to test something that is still totlly in PoC
  • [14:51] Saijanai Kuhn: thats part of its purpose. I believe its meant as a QA tool byitself as well
  • [14:51] Q Linden: Yes, agreed
  • [14:51] Q Linden: certainly we want the test harness running before it's deployed for real, but does it matter whether a c++ or python implementation comes first?
  • [14:51] Saijanai Kuhn: my main point is that if its not far enough along to add to a python script, its not far enough along to open to anyone for "grid" testing
  • [14:52] Baba Yamamoto: libsl can test anything
  • [14:52] Saijanai Kuhn: shouldn't, but would be nice for the OGP to show even a preliminary spec
  • [14:52] Baba Yamamoto: we'll crash it for you no problem
  • [14:52] Tomiko Magic: Isn't the 'test harness' Enus Linden's thing?
  • [14:52] Carjay McGinnis: hehe
  • [14:52] Q Linden: Enus and Locklainn
  • [14:52] Saijanai Kuhn: he's on vacation til tomorrow I think
  • [14:52] Tomiko Magic: Enus is back next week, from Vacation
  • [14:52] Rob Linden: is still a little confused about what Saijanai is asking for
  • [14:53] Tomiko Magic: (Ebus will be having Office hours weekly for that thingy) : D
  • [14:53] Saijanai Kuhn: eh just some way for the pyogp folk to start testing I guess
  • [14:53] Tomiko Magic: Enus*
  • [14:54] Baba Yamamoto: has there been any discussion of the new build system that went out in the source drop?
  • [14:54] Baba Yamamoto: with package installer
  • [14:54] Tomiko Magic: has to go, be well
  • [14:54] Q Linden: see ya
  • [14:54] Jacek Antonelli: take care
  • [14:56] Q Linden: baba, you mean cmake?
  • [14:56] Baba Yamamoto: it's really great for the most basic stuff but it seems to really be an impediment to anything that needs to change the build process even in a minor way
  • [14:56] Baba Yamamoto: no
  • [14:56] Baba Yamamoto: the new install.py
  • [14:56] Baba Yamamoto: cmake is great
  • [14:56] Baba Yamamoto: i figured that out last month
  • [14:57] Nyx Linden: the library change?
  • [14:57] Rob Linden: ah....ok. lets' wrap up on Saijanai's topic first.
  • [14:57] Baba Yamamoto: it's how to work around the linden proided libraries.. so say i am using boost 1.35 rather than boost 1.32 or whatever?
  • [14:57] Baba Yamamoto:  ;)
  • [14:57] Baba Yamamoto: sai talks about his problems all the time
  • [14:57] Baba Yamamoto: all the time ;P
  • [14:58] Rob Linden: well, I do have to admit this is the wrong crowd to resolve the pyogp issue now
  • [14:58] Poppy Linden: Baba, you mean the library downloader then? install.py?
  • [14:58] Saijanai Kuhn: sure its an internal thing. Just pointing out what I see as a possible issue
  • [14:58] Baba Yamamoto: yes
  • [14:58] Baba Yamamoto: and the changes to support it
  • [14:58] Rob Linden: so...ok, install.py. what in particular about it?
  • [14:58] Baba Yamamoto: how do i work around it so i can modify those provided libraries?
  • [14:58] Poppy Linden: sounds like it is blasting stuff
  • [14:59] Baba Yamamoto: it doesnt seem trivial
  • [14:59] Poppy Linden: i think you change hashcodes for it, then make sure you replace the temp file... i'm not super familiar with it
  • [14:59] Baba Yamamoto: what?
  • [14:59] Poppy Linden: jenn linden is the one to ask
  • [14:59] Baba Yamamoto: that sounds confusing
  • [14:59] Baba Yamamoto: jenn eh?
  • [14:59] Michelle2 Zenovka: looks like you need a hole where you can drop user provided libraries and these are taken in preference to anything else
  • [14:59] Rob Linden: Baba...can you send mail to sldev. I'll make sure that we ping Jenn and Phoenix about it
  • [15:00] Carjay McGinnis: when is cmake going to hit viewer-2?
  • [15:00] Rob Linden: I'd like to make sure it's easy to mod those libraries
  • [15:00] Baba Yamamoto: i just want to say flip a bit to tell the generator to not use install.py for this one thing but insead use xyzmylib
  • [15:00] Rob Linden: Baba: sounds reasonable to me, but we'd need to ask about it
  • [15:00] Michelle2 Zenovka: baba, just call cmake directly, don't use the .py scripts
  • [15:00] Baba Yamamoto: rob yeah.. as long as it's easy to change it's not an issue
  • [15:01] Jacek Antonelli: Gotta run. Take care, all
  • [15:01] Baba Yamamoto: michchelle, cmake calls install.py
  • [15:01] Rob Linden: I've also gotta run as well.
  • [15:01] Rob Linden: thanks all!