Open Source Meeting/2008-11-13

From Second Life Wiki
Jump to: navigation, search
  • [13:59] Robin Cornelius: Hey everyone
  • [13:59] Connecting to: in-world Voice Chat...
  • [13:59] Morgaine Dinova: Hiya Robin
  • [13:59] Connected undefined:
  • [13:59] Carjay McGinnis: greetings Soft
  • [13:59] Morgaine Dinova: Hiya Rob
  • [13:59] Carjay McGinnis: Hi Robin
  • [14:00] Rob Linden: howdy
  • [14:00] Carjay McGinnis: greetings Rob
  • [14:00] Ellla McMahon: hello everyone :)
  • [14:01] Robin Cornelius: Sorry navigational issues in my text client
  • [14:01] Squirrel Wood: Tacos!
  • [14:01] Carjay McGinnis: Hi Ella
  • [14:01] Morgaine Dinova: No tacos here, but I have apple pie :-)
  • [14:02] Squirrel Wood: short agenda today
  • [14:02] Rob Linden: there doesn't appear to be anything new on the agenda this week
  • [14:02] Morgaine Dinova: Hiya Sheet
  • [14:02] Sheet Spotter: Allo, allo!
  • [14:02] Rob Linden: any unresolved issues from the sldev@ mailing list?
  • [14:02] Carjay McGinnis: Hello Sheet
  • [14:03] Rob Linden: has anyone here given the QtWebkit stuff a whirl?
  • [14:03] Robin Cornelius: i've checked it all out
  • [14:03] Robin Cornelius: the git:// you gave was actualy the wrong repto
  • [14:03] Rob Linden: really?
  • [14:03] Rob Linden: did you find the right one then?
  • [14:04] Robin Cornelius: its on the same server, but that one was just a git version of the LLMozlib stuff you have in SVN
  • [14:04] Robin Cornelius: yea its using git browse so it was trival to find the other repos on the server
  • [14:04] Carjay McGinnis: anything interesting?
  • [14:05] Robin Cornelius: but i'm not 100% sure what it is, it appears to be the qt webkit port? or is there more in there too. Its auto synced to the main webkit git
  • [14:05] Robin Cornelius: i've not tried building yet
  • [14:05] Rob Linden: yeah, so what it is is QtWebkit integrated into the existing LLMozLib API
  • [14:06] Rob Linden: of course, the API naming should probably be revisited if we move forward with this
  • [14:06] Soft Linden: Do we have some simple su-and-say instructions on the wiki yet? It's early, but it would be nice to have more people looking at this early on.
  • [14:06] Robin Cornelius: the main git tree is about 700MB
  • [14:06] Carjay McGinnis: hm, I missed the git reference it seems
  • [14:06] Carjay McGinnis: oh
  • [14:07] Q Linden: Hey, all
  • [14:07] Carjay McGinnis: 700?
  • [14:07] Carjay McGinnis: why is it so big?
  • [14:07] Carjay McGinnis: Hey Q
  • [14:07] Robin Cornelius: its all of webkit in there too
  • [14:07] Robin Cornelius: but thats from a git clone
  • [14:07] Robin Cornelius: all ref heads
  • [14:07] Q Linden: / I've brought a classroom of visitors today (piped through the camera of Void Svoboda here)
  • [14:07] Robin Cornelius: waves
  • [14:08] Q Linden: We've been talking about open source at LL
  • [14:08] Carjay McGinnis: think I need to refresh my git knowledge
  • [14:08] Aimee Trescothick: mexican waves
  • [14:08] Carjay McGinnis: *sine waves*
  • [14:08] Rob Linden: waves too!
  • [14:08] Khyota Wulluf: Soft : that issue i mentioned on monday was my fault in not updating the artwork
  • [14:08] Morgaine Dinova: /waves in a random direction
  • [14:08] Khyota Wulluf: hi hi!
  • [14:08] Soft Linden: Which - the issue of not packaging, or another?
  • [14:09] Q Linden: no agenda today?
  • [14:09] Khyota Wulluf: blah, somethin with thes cmake flags, dont even remeber
  • [14:09] Soft Linden: Gotcha
  • [14:10] Soft Linden: Not today, Q
  • [14:10] Soft Linden: We're just fishing for any unresolved mailing list issues right now
  • [14:10] Q Linden: So maybe if the class has any questions of OS contribs, they shoudl feel free to ask them
  • [14:10] Soft Linden: Ha, go for it :)
  • [14:10] Khyota Wulluf: hey whats the new opensource-drops branch about?
  • [14:11] Aimee Trescothick: oh yup, was about to ask that too
  • [14:11] Soft Linden: Ah, that's where Cg is doing experiments with an overhaul of the open source drops
  • [14:11] Rob Linden: yeah, go for it Q and class
  • [14:11] Soft Linden: We're still running with my stack of scripts, which I implemented about a year ago as a stop gap to get source published more frequently... which were built on top of Rob's stop gap scripts...
  • [14:11] Q Linden: overhaul of the open source drops == automation of publishing
  • [14:12] Q Linden: (I just talked about this in the class)
  • [14:12] Soft Linden: We're going to have something that's completely automated, tied to our internal build validation now. So all branches will be guaranteed buildable on all three platforms before export, and internal devs will know immediately when they check in something that will break the public build, not just the internal build.
  • [14:12] Soft Linden: Anyway, that directory is where test drops are happening right now.
  • [14:13] Soft Linden: afaik, that won't be a long-lived branch, but rather all the currently published branches will be replaced when Cg thinks his scripts are bulletproof.
  • [14:13] Khyota Wulluf: awsome!
  • [14:14] Squirrel Wood: nono. OSSM ^^
  • [14:14] Soft Linden: Something else that will probably come with that are build binaries as well. So people who want to track branches can snatch those, even if they're not comfortable building their own viewers. (Or even if they have painfully slow machines and don't want to build daily)
  • [14:15] Carjay McGinnis: sounds great
  • [14:15] Soft Linden: I've got something I could offer up as an agenda item -
  • [14:15] Rob Linden: go for it
  • [14:15] Soft Linden: When triaging issues, we're finding a few that have sat idle just because they could use a bit more vetting. We've been making that note on the issues themselves... is that the best move, or should those be going to sldev? Or should we be doing something else to highlight those issues in JIRA?
  • [14:16] Soft Linden: I'll pull up one example, just a sec...
  • [14:16] Soft Linden: http://jira.secondlife.com/browse/VWR-3878
  • [14:16] Robin Cornelius: a ping to sldev would probably do some good, i often catch up to issues because someone else is discussing or asking about
  • [14:16] Carjay McGinnis: yes, not much happening on the list usually
  • [14:17] Soft Linden: This issue is in a number of public viewers.
  • [14:17] Soft Linden: Rather, this patch is in a number.
  • [14:17] Robin Cornelius: ah
  • [14:17] Soft Linden: But when we looked at it ourselves internally, we didn't see it actually making a difference against the named problem
  • [14:17] Robin Cornelius: that patch is very interesting
  • [14:17] Khyota Wulluf: oh i saw this one today
  • [14:17] Robin Cornelius: it was not quite doing what was expected
  • [14:17] Soft Linden: What would help here is having someone who's experiencing the problem drop some meaningful metrics into the viewer to show that pauses are actually happening right in the named spot.
  • [14:18] Carjay McGinnis: oh, I think I saw that a while ago
  • [14:18] Robin Cornelius: if i recall the actual call to the purge_time_spliced() was missing
  • [14:18] Carjay McGinnis: for old viewers though
  • [14:18] Robin Cornelius: so it did no purging at all
  • [14:19] Soft Linden: Anyway - we don't need to necessarily resolve the issue here, but would you like to see these bounced to sldev? Or should we have some component or meta issue? What's most useful?
  • [14:19] Robin Cornelius: sldev ping
  • [14:20] Carjay McGinnis: yeah, sldev
  • [14:20] Soft Linden: So maybe a note on the JIRA saying the conversation's gone to sldev, just to be sure the original contributor sees this as well
  • [14:20] Soft Linden: Many contributors aren't on the mailing list
  • [14:20] Sheet Spotter: As an aside...Soft, do you want me to try testing your question on VWR-3878 tonight?
  • [14:20] Robin Cornelius: yes, i tend to watch all jira's i've commented on or contributed too as well but there are many that would be of interest and an additional ping would help
  • [14:20] Soft Linden: Sure, it would be great if you could show it's valid and still applicable to 1.21 or trunk
  • [14:21] Rob Linden: by the way, here's the list that we make sure we check every week: http://jira.secondlife.com/secure/IssueNavigator.jspa?reset=true&&status=1&status=3&status=4&customfield_10002=Patch+attached&sorter/field=updated&sorter/order=DESC&sorter/field=customfield_10000&sorter/order=ASC&sorter/field=customfield_10020&sorter/order=ASC
  • [14:21] Sheet Spotter: I'll take a run at it tonight.
  • [14:21] Rob Linden: that's sorted by "last triaged date", so that we make sure we cycle through the list once every few weeks
  • [14:22] Soft Linden: With the link Rob lists, we run down those in order every week, as many as we can do with a small group in an hour
  • [14:22] Soft Linden: Generally we're looking for anything where someone has asked a question or asked for a change that's made a previously unimportable patch importable
  • [14:22] Carjay McGinnis: ah, ok
  • [14:23] Soft Linden: Or we're finding things where a patch has been neglected and the internal owner could use a nudge, or where the issue should be taken back so someone with more availability can handle it
  • [14:24] Soft Linden: Anyway, that was it for my question. Ping on sldev, and a note to that effect on the issue.
  • [14:24] Robin Cornelius: yes i have talked about a few of these recently
  • [14:24] Robin Cornelius: people msg me on IRC with questions about old patchs not applying
  • [14:24] Khyota Wulluf: hehe
  • [14:24] Robin Cornelius: which are good patches and have been neglected
  • [14:24] Soft Linden: Yeah. We're narrowing the number of those, or at least giving feedback on ones which we think we can't use. It's a game of catchup still
  • [14:25] Soft Linden: If you see any really egregious ones which you think should go in asap, sldev is a good place to highlight those
  • [14:25] Robin Cornelius: yea thats what happedn this week with the save scripts/text editor to file patch
  • [14:27] Rob Linden: we also want to be pinged on sldev about issues that you may be thinking of updating, but you want to find out if that would help it get merged in
  • [14:28] Rob Linden: anyone here got stuff in the works that you're planning to submit?
  • [14:28] Aimee Trescothick: yup
  • [14:28] Aimee Trescothick: I have a math parser for the build tools window that just needs a bit of clean up first
  • [14:29] Soft Linden: Ha - so you'd enter like 1/3 instead of 0.33333333 for a scale?
  • [14:29] Aimee Trescothick: yes
  • [14:29] Robin Cornelius: cool
  • [14:29] Aimee Trescothick: and can do things like X+3
  • [14:29] Soft Linden: Innnnnteresting
  • [14:29] Aimee Trescothick: or make Yscale XS*2
  • [14:29] Morgaine Dinova: And 5 / 0 ? ;-)
  • [14:29] Aimee Trescothick: lol
  • [14:29] Morgaine Dinova: Megaprims!
  • [14:29] Soft Linden: 5/0 = return of megaprim
  • [14:29] Soft Linden: Yus :>
  • [14:29] Aimee Trescothick: haha
  • [14:29] Carjay McGinnis: hehe
  • [14:30] Khyota Wulluf: laughs
  • [14:30] Aimee Trescothick: has trig functions too, handy for rotations
  • [14:30] Robin Cornelius: what would also be cool is basic snap tools. like center of, end of etc like a cad app
  • [14:30] Morgaine Dinova: NaN prims would be fun too
  • [14:30] Rob Linden: that's very cool
  • [14:30] Soft Linden: bool LLPrim::isfinite() -- don't make me!
  • [14:31] Aimee Trescothick:  :D
  • [14:31] Morgaine Dinova: bool LLPrim::isImaginary() might help with prim torturing ;-)
  • [14:32] Carjay McGinnis: and add the orthographic view while you're at it
  • [14:32] Khyota Wulluf: wow
  • [14:33] Soft Linden: Q's class ought to have come on a week where we had something more contentious going on. I think this has turned more social.
  • [14:33] Aimee Trescothick: yes, we're not usually very sociable at all
  • [14:33] Robin Cornelius: Aimee what are you using for the maths parser?
  • [14:33] Aimee Trescothick: it's written using Boost Spirit
  • [14:33] Q Linden: hah
  • [14:33] Q Linden: loves boost spirit
  • [14:34] Carjay McGinnis: cool, I almost used it once but didn't get too look too close at it
  • [14:34] Aimee Trescothick: ooh, at least one person will understand my code then :D
  • [14:34] Q Linden: It's like the maximum possible template magic
  • [14:34] Q Linden: I'll definitely look that patch over when you're ready
  • [14:34] Q Linden: (the class left, by the way)
  • [14:35] Rob Linden: awww
  • [14:35] Aimee Trescothick: EBNF massaged into C++ lol
  • [14:35] Carjay McGinnis: ah
  • [14:35] Soft Linden: I don't understand where "maximum possible template magic" and "like" can belong to the same concept :(
  • [14:35] Aimee Trescothick: it's the first time I've done anything serious with it, so might need a little tidying up
  • [14:36] Soft Linden: I get four or more open angle brackets in a row and I want to start swatting 'em like angry bees.
  • [14:36] Aimee Trescothick: seems pretty solid though, even picks up errors and drops your cursor in the right place in the expression :D
  • [14:36] Carjay McGinnis: you could typedef em away ^^
  • [14:36] Soft Linden: Ha, now that's just showing off, Aimee.
  • [14:36] Aimee Trescothick:  :)
  • [14:36] Q Linden: Aimee -- is it integrated into the text component?
  • [14:37] Q Linden: Or is it part of just the build dialog?
  • [14:37] Morgaine Dinova: Soft: get ready with a code LLCode::deObfuscate()
  • [14:37] Aimee Trescothick: it's in the line editor
  • [14:37] Soft Linden: that would be nice to have for things like parcel pricing too. meters*5-1 or such
  • [14:37] Aimee Trescothick: well, that's where it's hooked into
  • [14:38] Soft Linden: Does it carry something like a flag in the xui that indicates whether it should be active?
  • [14:38] Aimee Trescothick: the build window is used to drag the variables out of
  • [14:38] Aimee Trescothick: erm, not right now it doesn't, it's just active in any spincontrol
  • [14:39] Soft Linden: To be really racecar awesome, it would be nice to have some way to define callbacks for variables that can be referenced in a given floater, and an attribute that says that math should be active in a given control
  • [14:40] Aimee Trescothick:  :D now that would be nice
  • [14:40] Rob Linden: now Soft is just pushing it :)
  • [14:41] Soft Linden: <pony />
  • [14:41] Aimee Trescothick: pony^2
  • [14:41] Morgaine Dinova: Soft: callback for the result too, and then lock them all together with a constraint, rerunning any time that one parameter changes :-)
  • [14:42] Soft Linden: Do constraints already exist in the xui for spinners? I hope those aren't hard-coded in dialogs...
  • [14:42] Soft Linden: s/dialogs/floaters/
  • [14:42] Aimee Trescothick: pretty sure they're in the xui
  • [14:42] Q Linden: not so sure that's a pony
  • [14:42] Q Linden: I think it might be just Good Design
  • [14:42] Aimee Trescothick: lol
  • [14:43] Soft Linden: It would sure make it easier to get the patch in. There's enough UI work going on that anything keeping specific implementation details out of a panel is a Good Thing.
  • [14:43] Rob Linden: ah, righto
  • [14:44] Rob Linden: anyone else cooking up anything else?
  • [14:44] Robin Cornelius: i've been having trouble with my clutter webkit
  • [14:45] Soft Linden: Trouble?
  • [14:45] Robin Cornelius: yea i seem to not be able to feed the textures back to the viewer, i guess its an elemental mistake but the llimagegl throws up as subimage is not in image
  • [14:45] Robin Cornelius: but it apears llimagegl thinks my image is 0x0
  • [14:46] Robin Cornelius: not had chance to hunt it down yet
  • [14:46] Soft Linden: good fun
  • [14:47] Robin Cornelius: kind of got distracted with my text client still ;)
  • [14:48] Rob Linden: your text client is libomv based, right?
  • [14:48] Robin Cornelius: yes
  • [14:50] Rob Linden: well, we seem to be running out of steam here
  • [14:50] Rob Linden: anything got anything else that we should cover here today?
  • [14:51] Robin Cornelius: one quick thing
  • [14:51] Rob Linden: go for it
  • [14:51] Robin Cornelius: should the server prevent the creation of inventory folders with a zero parentID eg the same level as my inventory and library
  • [14:51] Soft Linden: I'd think so, yeah.
  • [14:51] Soft Linden: You're able to?
  • [14:51] Robin Cornelius: yes
  • [14:52] Robin Cornelius: if you buy the contents of an object and pass UUID.Zero as the target folder you get a working top level folder
  • [14:52] Q Linden: hmmm....we have internal code to fix such things and reparent them.
  • [14:52] Soft Linden: yus. That should go in JIRA. Especially since there's talk of splitting Library and Inventory into two different windows. It's likely those top level folders would just vanish with a change like that.
  • [14:52] Q Linden: we consider it a bug if we have that
  • [14:53] Robin Cornelius: my top level folder is very much still in existance
  • [14:53] Q Linden: Well, if someone ran the script against your acccount they'd get moved.
  • [14:53] Morgaine Dinova: Could more than two top-level entries have a use?
  • [14:53] Q Linden: But we don't do it unless someone complains
  • [14:54] Robin Cornelius: ah
  • [14:54] Carjay McGinnis: guess things like these could make the whole SL universe implode?
  • [14:54] Robin Cornelius: any how i think i can move it with libomv where it belongs
  • [14:54] Q Linden: / so yeah, please submit JIRA on it
  • [14:54] Robin Cornelius: i was purposly trying to create one to simulate what someone was reporting
  • [14:55] Morgaine Dinova: What happens when interop starts? Perhaps there should be more top level entries, since Library actually means SL:Library, whereas Inventory might be mixed
  • [14:56] Sheet Spotter: I had a silly question...How do I submit objects that repro JIRA issues?
  • [14:56] Soft Linden: Sheet - if you give it a unique name and note that in the JIRA, we can rez it out of your inventory.
  • [14:57] Sheet Spotter: Oooo...sneaky. LOL
  • [14:57] Morgaine Dinova: And You might want X::Library, where X refers to each grid you might have objects from.
  • [14:57] Soft Linden: (Implicit in that is that you're granting us permission to look at it - we don't go spelunking through inventory as a matter of routine, of course!)
  • [14:57] Sheet Spotter: Understood.
  • [14:58] Rob Linden: Morgaine: not sure...that seems more of an AWG question
  • [14:59] Rob Linden: anyway, we're just about out of time here
  • [14:59] Morgaine Dinova: The standard answer in AWG is "Client? Not interested" ;-)
  • [14:59] Morgaine Dinova: Which is fair enough. But someone should be interested.
  • [15:00] Rob Linden: hmm...well, something for some other time
  • [15:00] Rob Linden: thanks everyone for coming
  • [15:00] Sheet Spotter: Thank you for this hour; always interesting and informative.
  • [15:00] Robin Cornelius: thanks Rob
  • [15:00] Morgaine Dinova: Cheers all.
  • [15:00] Soft Linden: thumbsup - "Thanks for coming!"
  • [15:00] Aimee Trescothick: thanks :)
  • [15:00] Q Linden: thanks
  • [15:00] Carjay McGinnis: thanks Q, Rob and Soft
  • [15:00] Rob Linden: bye everyone!
  • [15:00] Sheet Spotter: Have a great one!
  • [15:00] Home: Going: to home
  • [15:01] Aimee Trescothick: sin(thanks)*sqrt(2)