Open Source Meeting/2009-04-30

• [14:25] Q Linden: Simultaneously
• [14:25] Morgaine Dinova: Is there a wiki page for Philip's viewer, something with links off to the relevant data?
• [14:25] Soft Linden: yus :>
• [14:25] Melinda Latynina: this is all still very much a work-in-progress, no? I.E. can't have permanant rules & regulations yet, but the good news is that we all get to help drive the direction.
• [14:25] Morgaine Dinova: checks
• [14:26] Q Linden: Seriously, though, I think that we're moving toward "post a patch, post on SLDev for feedback, react to feedback, eventually check in patch."
• [14:26] Mm Alder: But none of the policy seems to be written down yet.
• [14:26] Merov Linden: when Philip was there remember , he didn't want to start the project with a long list of "processes" and "regulation"
• [14:27] Mm Alder: I'm trying to get in literally a one bit change and I'm finding it very difficult.
• [14:27] Merov Linden: so his idea was more: proposes something on list and we get from there
• [14:27] Q Linden: mm, you mean lip flap?
• [14:27] Merov Linden: which one Alder?
• [14:27] Soft Linden: I wouldn't anticipate that subsequent contributions have the overhead of the first dozen.
• [14:27] Mm Alder: yes
• [14:28] Morgaine Dinova: Well in theory Philip's branch was intended to make it *easier* to commit changes. If it's still not easy enough, maybe a third LL VWR project is needed :P
• [14:28] Morgaine Dinova: laughs
• [14:28] Soft Linden:  :)
• [14:28] Q Linden: facepalms
• [14:28] Soft Linden: Hold off on assuming that you know what the final process looks like until a few patches have gone through.
• [14:29] Merov Linden: Your contribution has been voted "in" Alder
• [14:29] Mm Alder: Does it have to be this hard?
• [14:29] Q Linden: we sure hope not
• [14:30] Q Linden: but we're still working it out...and everyone is still tentative
• [14:30] Robin Cornelius: i was very suprised by the reaction to such a simple change
• [14:30] Soft Linden: Anything touching on voice invites an awful lot of friction. I don't fully understand it myself.
• [14:30] Squirrel Wood: At least we know that one contribution method still involves bottles of scotch brought directly to the San Francisco, Brighton, Boston and Chicago offices. ^^
• [14:30] Q Linden: maybe we're just used to the surprise
• [14:30] Melinda Latynina: is not tentative :-)
• [14:31] Robin Cornelius: i would have 2nd my self but i've never evern seen teh code running as i almost never use voice, so i don't feel i can approve any code i can't test my self
• [14:31] Squirrel Wood: Or so I've heard.
• [14:31] Mm Alder: So when I want to contribute my next patch which has popup menus and writes files, what then?
• [14:31] Merov Linden: Alder: not everybody has commit privileges so it's assigned to someone who has
• [14:31] Morgaine Dinova: Might have been a good time to move to Hg actually.
• [14:31] Robin Cornelius: MM if its an area i know, then i would be happy to look
• [14:31] Merov Linden: this someone is me now
• [14:31] Aimee Trescothick: well Hg is there now at least :)
• [14:31] Soft Linden: Hg is coming. And we did talk about using it for this, but didn't want yet another thing that had to be working before launching.
• [14:32] Merov Linden: so you can kick my butt now if that doesn't move forward
• [14:32] Q Linden: morgaine, we're almost all the way to hg
• [14:32] Morgaine Dinova: Woohoo!
• [14:32] Q Linden: can't happen fast enough for me. :)
• [14:32] Mm Alder: Hg?
• [14:32] Q Linden: Mercurial source control
• [14:32] Aimee Trescothick: [1]
• [14:33] Q Linden: OK? "I wonder if someone could give a brief update on the status of plugin explorations." -- JB Kraft
• [14:33] JB Kraft: i found some discussion on the dev list from a couple of years ago, soft did some work on this i belive. just wondered if it stalled and why and what spark might ignite it again, if any?
• [14:33] Robin Cornelius: people have been contacting me recently about this as well
• [14:33] Q Linden: JB, it's smouldering as we speak
• [14:34] Morgaine Dinova: Mm: [2] , [3]
• [14:34] Soft Linden: I sooooo want plugins. :<
• [14:34] Merov Linden: me too
• [14:34] Soft Linden: oops gesture
• [14:34] JB Kraft: yeah, it would be nice :) is there anything form outside that might push it along?
• [14:34] Q Linden: the stall has been that this is one case where the hack is the true enemy of a good design, because we'll be stuck with it forever (*cough*skin layers*cough*)
• [14:35] Merov Linden: it's not easy to get it right from the host standpoint
• [14:35] Morgaine Dinova: Plugins in what? In viewer? If anyone's working on that, I'd love to talk to them, since that's my goal for Imprudence.
• [14:35] Q Linden: so we've been working on an event system that would act as the transport of information between plugins
• [14:35] JB Kraft: ah ok, that is good news
• [14:35] Q Linden: You can see a preview of some of these ideas in the Notifications event API, but that's not complete yet
• [14:36] Merov Linden: Q: using IPC?
• [14:36] Melinda Latynina: JB: weren't you looking into a plug-in archetecture for text editors?
• [14:36] Q Linden: boost signals
• [14:36] JB Kraft: yes, the editor is one of my pet projects and that was a part of it. thats why i wondered :)
• [14:37] Q Linden: and...not coincidentally, we actually wanted to talk about script editor improvements today
• [14:37] Q Linden: we thought it would be a good time to start the design discussion on sldev/here
• [14:38] Melinda Latynina: i believe we can have plug-ins in multiple places and deal with the merits and weaknesses of each rather than try to find a one-size-fits-all
• [14:38] Q Linden: projects like [4] (scratch for second life) need a plugin editor
• [14:38] JB Kraft: excellent! i have a bunch of patches for the editor but kind of got discouraged with process tbh. it would be great to hack out a path for it :)
• [14:39] Morgaine Dinova: [5] --- that gives a general view of our plugin architecture for Imprudence.
• [14:40] Squirrel Wood: improvements... one simple improvements would be to allow bulk uploads of .txt and .lsl files ?
• [14:40] Q Linden: yes, morgaine...but some of the details of how the sockets communicate and what the control model is are key to getting this right
• [14:40] Q Linden: so I like that chart, but I want more. :)
• [14:40] Morgaine Dinova: This is the general forum, everything marked as [Plugins
• [14:40] Morgaine Dinova: [6]
• [14:41] JB Kraft: squirrel, i have a patch on the jira that load and saves from the editor to disc, kudgly at the mo but it works for me
• [14:41] Morgaine Dinova: Q: yep, the internal viewer API is central, and clearly needs to be event-based.
• [14:41] Robin Cornelius: JB i think squirrel whats the File->Uload->Bulk to support text for notecards and scripts (with approprate expension)
• [14:42] JB Kraft: ah, right
• [14:42] Robin Cornelius: pretty simple to do actually
• [14:42] Squirrel Wood: there is a jira somewhere for it even ^^
• [14:42] Robin Cornelius: but loading and saving from the text editor is a godsend
• [14:43] Morgaine Dinova: Q: are you thinking as we are then, that decoupling plugins from the viewer monolith is the key to sanity?
• [14:43] Q Linden: yes
• [14:43] Q Linden: *I* am
• [14:43] Morgaine Dinova: Great to hear! :-)
• [14:43] JB Kraft: https://jira.secondlife.com/browse/VWR-5206
• [14:43] Q Linden: And we're moving in that direction...but honestly, it's a multi-year effort
• [14:43] Morgaine Dinova: Yep. not easy
• [14:43] Q Linden: A big chunk of our work this year is going to enabling that decoupling
• [14:43] JB Kraft: thats awesome Q!
• [14:44] Mm Alder: Please don't forget, "Is there anyone other than me who would be interested in open code reviews for this project? Mm Alder 15:57, 30 April 2009 (UTC) "
• [14:44] Morgaine Dinova: Just binding in a language would be pretty trivial, bit the wrong thing to do, if there is to be any hope of stability and portability.
• [14:44] Q Linden: right, morgaine; whatever we do, we do forever
• [14:45] Q Linden: so we really want to make sure it has a future
• [14:45] Q Linden: (while not living in analysis paralysis forever either!)
• [14:45] Morgaine Dinova: Yeah
• [14:45] Robin Cornelius: mm, i'm fully up for open code reviews
• [14:45] Aimee Trescothick: yes
• [14:46] Squirrel Wood: JB: https://jira.secondlife.com/browse/VWR-7959
• [14:46] Q Linden: mm, that's what I thought the idea was behind posting the patch on jira before applying it
• [14:46] Q Linden: is that not the same thing you're saying?
• [14:46] Morgaine Dinova: Q: if anything happens on Imprudence plugins (and that's not certain), I'd appreciate your feedback.
• [14:46] Mm Alder: No, I mean a code presentation in-world where people can ask questions and learn a bit about what the code does.
• [14:47] Robin Cornelius: sounds like what Phillip was taking about
• [14:47] Merov Linden: Alder: that's more of a "design review" then (as in "code design")
• [14:47] Mm Alder: Philip was talking about it, but then ....
• [14:47] Q Linden: The more of that we see, the better....but I'm not sure it needs to be a required part of every submission.
• [14:48] Q Linden: depends a lot on the scope of what's happening, I guess
• [14:48] Q Linden: I think a valid response to "hey, I've just attached a patch" would be "I'd like to have a conversation about this inworld."
• [14:48] Mm Alder: If I were to ask someone to review my patch for editing morphs, I would be surprised if anyone would understand it without explanation.
• [14:48] Merov Linden: Alder: are you talking about general part of the viewer architecture or specific submission?
• [14:49] Mm Alder: Specific submission. But you'd also learn about the viewer in the process just by listening.
• [14:49] Merov Linden: understood
• [14:49] Morgaine Dinova: Instead of conversations in-world, how about insisting on Doxygen-style documentation in patches instead? That can be anything, including design documentation.
• [14:49] Mm Alder: I can ask Doxygen a question. :-)
• [14:50] JB Kraft: +1 morgaine
• [14:51] Merov Linden: well, on one instance, (VWR-12748 submitted by Robin) I reviewed and proposed a modified patch, she modified, etc...
• [14:51] Mm Alder: I love Doxygen, but it can't replace a human who understands the code.
• [14:51] Merov Linden: so the code review was done in the JIRA through a back and forth of commented patches
• [14:51] Morgaine Dinova: Mm: hehe, well Doxygen only reflects what devs put it ... but if there's a long chuck of code without any human-added Doxygen headers, then I think it's reasonable to ask "Hey, how about some explanatory text for this?" :-)
• [14:52] Q Linden: me too
• [14:52] Merov Linden: I guess IW presentation would be good for general part of the architecture or really hairy patches
• [14:52] Q Linden: that's a part of a good code review -- it shouldn't be just "does it work" but "is it maintainable"
• [14:53] Melinda Latynina: likes the idea of IW code/design reviews
• [14:54] Q Linden: So Aimee asks "Procedure for getting supporting artwork in before committing a patch to http-texture? i.e. VWR-12748[c
• [14:54] Merov Linden: realized he mixed up Robin and Aimee
• [14:54] Soft Linden: Ah, because the artwork has to go into the tarball, gotcha
• [14:54] Aimee Trescothick: that's OK, I get myself mixed up all the time
• [14:54] Melinda Latynina: artwork is far more easily changed/fixed later and therefore lower risk
• [14:55] Aimee Trescothick: yep, if I commit the patch before it's present, things will break/the universe will implode etc.
• [14:55] Soft Linden: I think with a special case like that, you can just ask a Linden to commit the art for you when you'd be ready to commit code. Then go ahead and commit code when you see the export complete with the art in it.
• [14:55] Soft Linden: Just get to the point where the group's signed off on the code, and you can kick off that request for help from a Linden.
• [14:56] Aimee Trescothick: ok
• [14:56] Squirrel Wood: tosses code completion and auto indentation into the script editor improvement battle pit. Place your bet while you still can :p
• [14:56] Melinda Latynina: Don't know if it's been mentioned but Rob created this plug-in task: https://jira.secondlife.com/browse/VWR-5560
• [14:57] Q Linden: Heh, that's an external version of the one I wrote ages ago
• [14:58] JB Kraft: i vote elisp integration to the editor and then we are done! *ducks*
• [14:58] Q Linden: nudges conversation firmly bakc to plugins
• [14:58] Q Linden: I'll kick off the discussion on sldev with a reference to this JIRA. Thanks, Melinda!
• [14:58] Q Linden: any last comments? I've got a hard stop in 2 min
• [14:59] Melinda Latynina: JB: IIRC Richard proposed an OS editor that itself had a plug-in archetecture. do u remember what that was?
• [14:59] JB Kraft: i dont recall at the moment
• [14:59] Mm Alder: I'd still like more documentation of the open source viewer process on the wiki
• [15:00] Soft Linden: Mm - sometimes the best way to kick off documentation is to write the skeleton. People are often more motivated to fill in incomplete or wrong information than they are to start a fresh doc.
• [15:00] Morgaine Dinova: Great to see that there's a lot of internal interest in this general direction of plugins/external facilities . Our multicores are getting bored. :-)
• [15:00] Merov Linden: Alder: I think Rob fully intends to fluff up what he wrote so far
• [15:00] Soft Linden: That sums up the entirety of the wikipedia article lifecycle right there :)
• [15:01] Mm Alder: I could make something up :-)
• [15:01] Q Linden: And with that, I've gotta run. Thanks all, I'll post to wiki.
• [15:01] Morgaine Dinova: Thanks Q
• [15:01] Aimee Trescothick: Thank Q :)
• [15:01] Mm Alder: Thanks all
• [15:01] Soft Linden: Posting what you think is right, then dropping a link on sldev is likely to kick off the right discussions & motivate contributions.