ImpDev Meetups/2010-11-15

From Kokua Wiki
Jump to: navigation, search


ImpDev Meetup for November 15, 2010.

  • We discussed the Viewer 2 menu layout, and what Kokua's menu layout should be like.
    • Some parts of the V2 layout are good. Boroondas likes that there is no "File" catch-all menu (Jacek agrees).
    • But, the "Me" is also not that great. Jacek suggests a "Kokua" menu, like how apps on Mac do it.
    • Jacek says to experiment, try different things, see how they feel.
    • Regardless, we will probably have an option for a 1.x-style menu layout, to make it more comfortable for current users.
  • We discussed pie menus in Kokua.
    • Jacek says we will probably also (eventually) have an option for a 1.x-style pie menu, for the same reason.
    • But, licensing is an issue. There is no pie menu code in the LGPL source, so we would have to reinvent it.
    • Kakurady mentioned the mockups for a new type of pie menu he made long ago.
      • Jacek isn't opposed to trying new options, but reiterated that the immediate purpose for recreating the pie menus is to make the viewer more comfortable to current users.
  • We discussed branching / merging strategy. This was a long and dense discussion, so read the transcript (12:33 to 01:08). It did not reach a final resolution, so it will need to be discussed again later.
  • We discussed the status of the Windows and Mac Experimentals.
    • They now compile and run, but audio streams do not work on either, no doubt due to some issue with the Quicktime plugin.
    • We either need to fix/extend the Quicktime plugin, or replace/augment it with another solution, like the GStreamer plugin or a new VLC plugin.
  • We discussed relicensing of existing (GPL) patches to LGPL.
    • Jacek's opinion: If we have the patch author's permission, in most cases there is no problem. Code original written by the author can be relicensed by them, even if it used to be surrounded by some GPL code.
    • For changes to previously existing GPL code (rewritten lines), if an equivalent change can be made to the LGPL source, it's not an issue either.
    • If the function/class that the patch applied to has been removed or substantially rewritten, then we will probably have to rewrite the patch from scratch. But, the patch wouldn't have worked anyway, so no loss.
    • Other special cases will need to be resolved individually.

Agenda Items

  • Structure for the Kokua main menu (Boroondas)
    • Armin already changed it away from V2's structure
    • I think some aspects of V2 menu structure are good and worth keeping
  • Branching / Merging strategy for Kokua repos (Boroondas)


[12:14 PM] Boroondas Gupte: I think so, 2.4 beta should be cut soon
[12:14 PM] Jacek Antonelli: Hiya, sorry I'm late!
[12:14 PM] Jacek Antonelli: I got caught up in Blender, making a 3D version of our logo
[12:14 PM] Thoria Millgrove: will it blend?
[12:14 PM] Jacek Antonelli: It will!
[12:14 PM] Ron Overdrive: heh
[12:14 PM] Thickbrick Sleaford: hi jacek
[12:15 PM] Jacek Antonelli: Ok, let's get started
[12:16 PM] Jacek Antonelli: Item #1: Structure for the Kokua main menu (Boroondas)
* Armin already changed it away from V2's structure
* I think some aspects of V2 menu structure are good and worth keeping
[12:17 PM] Jacek Antonelli: Yes, I agree some parts of it are worth keeping.
[12:17 PM] Thickbrick Sleaford: I think most of it is pretty good.
[12:17 PM] Thickbrick Sleaford: at least stuff is not as random as 1.23 and there's no huge long menus.
[12:17 PM] Nicky Perian: i dont like the lst still menus
[12:18 PM] Jacek Antonelli logs in with V2 on the side to look at the menus
[12:18 PM] Boroondas Gupte: For example, I like the absense of a "File" menu (I know others disagree on that). A virtual world isn't a word processor. Even in browsers, "File" has become "stuff we don't know where else to put, but too important to hide in "tools" or "extras".
[12:18 PM] Thickbrick Sleaford: I guess File means "local computer-related stuff."
[12:18 PM] Boroondas Gupte: Though ... not everything currently in "Me" actually belongs there
[12:18 PM] Ron Overdrive: the File menu was something very traditional though
[12:19 PM] Ron Overdrive: yeah I'm not a fan of the Me menu
[12:19 PM] Jacek Antonelli: I'm not a fan of either File or Me
[12:19 PM] Thickbrick Sleaford: Kokua will have more file related stuff than the LL viewer (import-export, etc.) so it needs a sensible place.
[12:20 PM] Nicky Perian: i think upload belongs in file
[12:20 PM] Nicky Perian: not inventory
[12:20 PM] Kakurady Drakenar: I'm pretty sure someone has mentioned consolidating the menu bar, the sidebar and the toolbar into one by this point... has anyone?
[12:20 PM] Jacek Antonelli: I'd suggest "Kokua" for the first menu, and put things related to the program itself there. Preferences, Quit, other things. Pretty much like Mac does it.
[12:21 PM] Boroondas Gupte: It's now duplicated in "Build" (in viewer-dev ... probabla not yet in betas or releases)
[12:21 PM] Boroondas Gupte: upload, I mean
[12:21 PM] Jacek Antonelli: Hi Aleric
[12:21 PM] Thickbrick Sleaford throws an alarm clock at Aleric
[12:21 PM] Aleric Inglewood: hi
[12:21 PM] Boroondas Gupte: "Kokua" menu sounds good
[12:22 PM] Jacek Antonelli: Anyway, I think this is an area where we can just play around with the menus, try different things, see what works
[12:22 PM] Fix Cache Reminder Tree reminds you to fix the cache
[12:22 PM] Jacek Antonelli: We might end up creating a "legacy menu" layout that people can choose, too.
[12:23 PM] Nicky Perian: great
[12:23 PM] Jacek Antonelli: That will resemble the V1 menu layout, for people most comfortable with that
[12:23 PM] Jacek Antonelli: So... play around
[12:23 PM] Jacek Antonelli: Anything else about the menus we should talk about now?
[12:24 PM] Thickbrick Sleaford: Pie!
[12:24 PM] Jacek Antonelli: Mmmm pie
[12:24 PM] Jacek Antonelli: Good point
[12:24 PM] Boroondas Gupte: pie menus?
[12:25 PM] Jacek Antonelli: Eventually, we will probably want to offer pie menus as an option. For the same reason -- make it comfortable for current users
[12:25 PM] Jacek Antonelli: I think we might have to reinvent the pie menus though, unless for some reason LL left the code in the LGPL source.
[12:26 PM] Thickbrick Sleaford: eeep, that's a problem (and no, thy didn't)
[12:26 PM] Jacek Antonelli: Which part is the problem?
[12:26 PM] Thickbrick Sleaford: GPL vs LGPL
[12:27 PM] Aleric Inglewood: Hi Robin
[12:27 PM] Robin Cornelius: Hi ya
[12:27 PM] Robin Cornelius: sorry cassie took ages to go to sleep
[12:27 PM] Ron Overdrive: yeah if they left pie menus in the V2 codebase it would be easy
[12:27 PM] Jacek Antonelli: Yeah. I think there are going to be quite a few things we end up having to reinvent
[12:28 PM] Kakurady Drakenar: Well, finally there's a chance to put my pie mockups into code.
[12:28 PM] Nicky Perian: move away form xml to code?
[12:29 PM] Kakurady Drakenar: Umm, of course not.
[12:29 PM] Jacek Antonelli: I'm afraid not so much, Kakurady. The immediate purpose for recreating the pie menus is to make something quite similar to the V1 pie menus.
[12:29 PM] Boroondas Gupte: There are (or were) leftovers from pie menus (like the the context menu related code still has names with "pie" in them), but I don't think a functional implementation was still there.
[12:29 PM] Jacek Antonelli: But I certainly wouldn't object to people experimenting with new alternative menus
[12:30 PM] Aleric Inglewood: It's not on the agenda - but I'd like to get an update of the windows and Mac status of the current weekly, too.
[12:30 PM] Jacek Antonelli: Ok Aleric, we'll be sure to cover that
[12:30 PM] Jacek Antonelli: Any more menu stuff, or shall we move on?
[12:31 PM] Boroondas Gupte: I guess move on for now
[12:31 PM] Kakurady Drakenar: ...
[12:31 PM] Boroondas Gupte: we can pick up the topic again on future meetings on on IRC
[12:31 PM] Jacek Antonelli: What's up, Kakurady?
[12:31 PM] Kakurady Drakenar: Translations.
[12:32 PM] Jacek Antonelli: Is this a new topic, or related to menus?
[12:32 PM] Kakurady Drakenar: Yeah, it's a new topic. Let it pass.
[12:32 PM] Jacek Antonelli: Ok, we can discuss it later if there's time
[12:33 PM] Jacek Antonelli: Ok, next agenda item: Branching / Merging strategy for Kokua repos (Boroondas)
[12:33 PM] Jacek Antonelli: I'm not sure if this is exactly what you meant Boroondas, but I'd like to briefly cover the branch layout I'm suggesting for Kokua
[12:34 PM] Boroondas Gupte: sure
[12:34 PM] Jacek Antonelli: At the moment there are four permanent branches: dev, exp, release, and stable
[12:34 PM] Jacek Antonelli: dev is where the activity for the next release will be. It's equivalent to our current "next" branches.
[12:35 PM] Jacek Antonelli: exp is where work on the Experimentals will go. It's equivalent to "weekly".
[12:35 PM] Jacek Antonelli: release will be a reference to the latest non-Experimental release, including betas and RCs
[12:35 PM] Jacek Antonelli: stable will be a reference to the latest final release, not betas or RCs.
[12:36 PM] Jacek Antonelli: So, basically, each branch has varying levels of churn
[12:36 PM] Jacek Antonelli: Does that make sense?
[12:37 PM] Jacek Antonelli: dev flows into release (when we release), and release flows into stable (when we release a final).
[12:38 PM] Nicky Perian: exp floew to?
[12:38 PM] Jacek Antonelli: exp is... different. That's mostly McCabe's domain. Probably there will be some dev flowing into exp. And some exp flowing into dev.
[12:38 PM] Aleric Inglewood: Actually, I have a related remark... although not everyone is here, but ok.
[12:39 PM] Jacek Antonelli: One of the things that has bothered me about the current weekly branches, is that it tends to be pretty chaotic and messy. I'm hoping we can keep dev and exp more tidy.
[12:40 PM] Aleric Inglewood: I think it would be a very good idea if EVERY commit comment was prefixed with IMP-xyz... so that's it's clear which group of commits belong together and have an immediate reference to a center repository (redmine) that has all the ins and outs of that little "project".
[12:41 PM] Aleric Inglewood: Currently we have loads and loads of commit comments without any external reference :( .. that makes it very hard to port-- like we have to do now. And I'm convinced that this isn't going to be the last time we have to port the imprudence patches to a new source tree :p
[12:42 PM] Aleric Inglewood: So, what do you say? Can be make it a rule that every commit has reference it's related redmine issue?
[12:42 PM] Aleric Inglewood: Can we*
[12:42 PM] Jacek Antonelli: Ok. Not every commit will have an associated issue, but when they do, yes, please mention the issue number in the commit message.
[12:43 PM] Aleric Inglewood: Maybe an issue should be created, if it doesn't exist :p
[12:43 PM] Jacek Antonelli: For big things, sure. But I don't want to create some awkward process where you have to file an issue just to fix a typo
[12:44 PM] Boroondas Gupte: For stuff that already has an issue on LL's jira, should we duplicate the issue on Imp RedMine?
[12:45 PM] Aleric Inglewood: Yes, at least that's what I have been doing. Then I prefixed it with RED-xyz: SNOW-abc: blah blah
[12:46 PM] Jacek Antonelli: Boroondas: If it's significant, and it needs review or merging, then yes it would be a good idea to create a new issue to make sure it doesn't fall through the cracks.
[12:46 PM] Thickbrick Sleaford: I think trying to keep to this rule as much as makes sense also encourages people to not push branches that have ton of small WIP commits, but instead compress that once it's done into a single commit with a redmine number attached. (BTW, how do you do that?...)
[12:46 PM] Robin Cornelius: git squash, but it destroys history for anyone tracking that branch
[12:46 PM] Jacek Antonelli: But if the change/issue is pretty much identical to the one on LL's JIRA, I'd say use that issue number instead of the new one
[12:47 PM] Aleric Inglewood: I don't think it has any merit to push 'WIP' commits: only finished commits should be pushed to non-local, and they *should* be squashed into one then.
[12:48 PM] Jacek Antonelli: If there are WIP commits that need to be pushed, for testing or whatever, I'd suggest making a new branch with "wip" in the name somewhere
[12:49 PM] Robin Cornelius: Aleric i do have a problem with keeping wip local, as i work from multiple locations, but i'm happy to append -wip or something to my private branches
[12:49 PM] Jacek Antonelli: Then when it's tested, squash them all, push a finished branch, and push-delete the wip branch ('git push :mybranch-wip" <-- note the colon)
[12:50 PM] Aleric Inglewood: The "problem" is that you can't rewrite exported/pushed commits, and therefore you cannot first commit something without a IMP issue# and later decide you want to make a remark about the work for future reference. So, if people push a commit without IMP-... in the commit, without redmine issue, they are likely to not want to add a redmine issue later on for that work, even if it is needed. While if you just create a nearly empty redmine issue anyway (ie, a proxy for a jira) then you CAN add something there later, if need be.
[12:50 PM] Jacek Antonelli: Before we get too off track -- Boroondas, was there anything you wanted to ask, since this is your agenda item?
[12:51 PM] Jacek Antonelli: If you push a *-wip branch, everyone should assume that it should not be relied upon for a stable history
[12:54 PM] Boroondas Gupte: well, I currently have my standalone_ouf-of-source branch where I got all the changes I need to build Kokua on 64bit
[12:54 PM] Jacek Antonelli: While we're talking branching and merging... I'd really like to encourage the use of "topic branches". These should be branches based on the latest dev branch from kokua/kokua, with each topic branch related to one change/series of related changes.
[12:54 PM] Boroondas Gupte: so I wondered if I should now merge that into my dev brach or what
[12:54 PM] Jacek Antonelli thinks.
[12:55 PM] Boroondas Gupte: standalone_ouf-of-source is a topic branch ... kinda
[12:56 PM] Jacek Antonelli: I'm trying to decide whether it would be better to have each person's individual dev branch flow into the main dev branch, or if only topic branches should flow into main dev, and the indiividual dev branches are each person's playground to do with as they please
[12:57 PM] Thickbrick Sleaford: an expcted lifecycle of a commit is dev -> ext -> release -> stable ?
[12:57 PM] Thickbrick Sleaford: s/ext/exp/
[12:57 PM] Robin Cornelius: topic-> individual dev branch to ensure it merges/builds correctly before pulling to main dev
[12:57 PM] Jacek Antonelli: The former is easier/lazier, while the later is cleaner in the long run.
[12:57 PM] Aleric Inglewood: Jacek: I don't understand your remark in the light of git. I would for SVN.
[12:59 PM] Jacek Antonelli: Well, we have the central repository, kokua/kokua. I want to keep this clean and organized, so that it's easier to follow the history.
[12:59 PM] Kakurady Drakenar: That sounds like SVN.
[12:59 PM] Aleric Inglewood: Everyone works locally ... they might have some branch that they share with someone else, but that is not imprudence business. Once they are happy - they request a merge - either their "topic branch" or they merge it themselves into 'experimental' locally - and then that is merged into imprudence/experimental
[01:00 PM] Boroondas Gupte: i.e., we don't need "dev"
[01:00 PM] Boroondas Gupte: ?
[01:00 PM] Aleric Inglewood: well, whatever weekly is :p
[01:00 PM] Boroondas Gupte: weekly is exp IIRC
[01:00 PM] Robin Cornelius: hmm yea it makes no difference, either bring your topic branch upto date with latest then request a pull, or pull your topic branch into your dev then request a pull, you get the same on the master repository
[01:00 PM] Jacek Antonelli: Yes, it's a somewhat centralized system. I think it needs to be. Some people like to advocate total decentralization, but that sounds like utter chaos to me.
[01:02 PM] Jacek Antonelli: My idea for the flow is, start with a topic branch based on the central dev branch. That flows into both dev and exp. exp gets mish-mashed chaotically, while dev stays more sane. dev eventually flows into release, and then into stable.
[01:02 PM] Boroondas Gupte: how will dev stay more sane when the same branches are merged in?
[01:02 PM] Aleric Inglewood: What I think we should NOT do - is what mccabe did for a whole (he seems to have stopped doing that now): cherry pick everything, completely destroying the history... I think we should use git the way it was intended and just merge branches directly without wanting a "clean" history: that just isn't possible with so many branches and people working in parallel.
[01:03 PM] Aleric Inglewood: for a while*
[01:03 PM] Jacek Antonelli: Well, the idea behind the weekly branch, was that it could have anything in it, even things that would not necessarily make it into the final release.
[01:04 PM] Boroondas Gupte: cherry picking can almost completely be avoided if we do "daggy fixes" (base changes on the oldest commit that still makes sense, then merge them forward)
[01:04 PM] Jacek Antonelli: That's what I would like, Boroondas. Although McCabe is in charge of weekly/exp, and he tends to play fast and loose.
[01:05 PM] Aleric Inglewood: Boroondas: But base changes are only possible if you didn't publish your branch yet (or at least, when nobody is using it to continue development from)
[01:05 PM] Jacek Antonelli: Maybe we can pressure him to be more organized, now that it doesn't take him 2 hours to recompile every time he switches branches :P
[01:05 PM] Boroondas Gupte: daggy fixes tend to make the history *look* very chaotic and insane, but should actually make picking isolated changes easier (because they aren't based on unrelated stuff)
[01:05 PM] Boroondas Gupte: base changes?
[01:05 PM] Boroondas Gupte: you mean re-base?
[01:06 PM] Aleric Inglewood: I thought that was what you meant
[01:06 PM] Boroondas Gupte: you shouldn't re-base to make a daggy fix (or if, then only locally)
[01:06 PM] Boroondas Gupte: you try to start on the old revision right away
[01:06 PM] Aleric Inglewood: TWO ROBIN's!
[01:06 PM] Boroondas Gupte: (it's easier for bugs than for new features, as you can just start on the revision that introduced the bug)
[01:06 PM] Jacek Antonelli: Another possibility is to just agree that the exp history is subject to change, and can be rebased or cherry picked at will. Then changes in exp could be cherry picked into dev, whereupon they become non-rebaseable.
[01:06 PM] Thoria Millgrove: that's an interesting choice of name, Brokofy
[01:07 PM] Brokofy Neva: ty
[01:07 PM] Jacek Antonelli: But, I think we need to move on. I don't want these meetings to get super long.
[01:08 PM] Jacek Antonelli: We can continue this in IRC, mailing list, and/or next week. Ok?
[01:08 PM] Boroondas Gupte: alright
[01:08 PM] Aleric Inglewood: sure.. I'm only here for the next topic anyway :p
[01:08 PM] Jacek Antonelli: Alright, then let's get to it. Next topic: Update on Mac/Win weekly status.
[01:09 PM] Jacek Antonelli: Basically, audio streams don't work for either one. No doubt Quicktime (or the Quicktime plugin, or the way it's used) is to blame.
[01:09 PM] Jacek Antonelli: Audio streams seem to work well for people on Linux, using Gstreamer.
[01:10 PM] Thickbrick Sleaford: Quicktime has no audio problems when playing movies, right?
[01:10 PM] Jacek Antonelli: Right
[01:10 PM] Jacek Antonelli: It just doesn't seem to like the audio streams. I think someone even said that the Quicktime player wouldn't play MP3 streams either. :\
[01:10 PM] Aleric Inglewood: My understanding was that Robin got things to work on Mac and windows.
[01:11 PM] Aleric Inglewood: And I saw her say on IRC that she knew exactly what was wrong and how to fix it.
[01:11 PM] Thickbrick Sleaford: I tried playing some shoutcast streams in the standalone quicktime player on winXP, and that didn't seem to work
[01:11 PM] Jacek Antonelli: Robin, any info on that front?
[01:11 PM] Aleric Inglewood: And I tried hard to get here here this meeting and now she fell asleep :(
[01:11 PM] Nicky Perian: It doesnt on mine either
[01:12 PM] Thickbrick Sleaford: some streams:
[01:13 PM] Robin Cornelius: I've not got anything going on mac yet
[01:13 PM] Robin Cornelius: my latest build results are
[01:13 PM] Jacek Antonelli: If Robin hasn't got it to work with Quicktime.. or maybe even if she has... it would probably be beneficial to create a VLC (or whatever) plugin to replace Quicktime.
[01:14 PM] Jacek Antonelli: I'm not familiar with either the VLC API or creating plugins, though. x_x
[01:14 PM] Thickbrick Sleaford: maybe quicktime just doesn't like the way it's initialized for audio/* (i.e. 1x1 pixels shm, etc.)
[01:14 PM] Nicky Perian: Is the Codes directory needed?
[01:15 PM] Nicky Perian: *Codecs
[01:16 PM] Jacek Antonelli: I'm not sure what that is, Nicky. Where is the codecs directory?
[01:17 PM] Nicky Perian: usually in llplugin
[01:17 PM] Jacek Antonelli: Robin: have you got audio streaming working on Windows with Quicktime?
[01:18 PM] Robin Cornelius: no
[01:18 PM] Jacek Antonelli: Ok
[01:18 PM] Nicky Perian: C:\Users\Bill\weekly-wib\linden\libraries\i686-win32\lib\release\codecs
[01:18 PM] Nicky Perian: then staged froward
[01:19 PM] Aleric Inglewood: :/
[01:19 PM] Aleric Inglewood: Jacek: Did you get it to compile on Mac yet?
[01:19 PM] Jacek Antonelli: Yes. It compiles and runs fine now, aside from audio streaming silently not working.
[01:20 PM] Jacek Antonelli: (silent as in no error messages or indication of problems)
[01:20 PM] Aleric Inglewood: Also using quicktime right? Might be the same thing.
[01:20 PM] Jacek Antonelli: Yeah, I think so too
[01:20 PM] Robin Cornelius: can i just point out to people, that i have a 15 month old daughter, which is way way more important than being here, i did not just fall asleep
[01:21 PM] Nicky Perian: Robin ++
[01:21 PM] Jacek Antonelli: We appreciate you coming, Robin, and thanks for your insight
[01:21 PM] Thoria Millgrove: awwww ㋡
[01:22 PM] Robin Cornelius: Jacek, mac is building for you now?
[01:22 PM] Jacek Antonelli: Yes
[01:22 PM] Aleric Inglewood: Did you fall back to an older version of llqtwebkit, or did you compile it yourself?
[01:23 PM] Robin Cornelius: from last weekly its still failing for me
[01:23 PM] Jacek Antonelli: An older llqtwebkit, for now at least. 20100617
[01:23 PM] Jacek Antonelli: Check my weekly branch, Robin
[01:23 PM] Jacek Antonelli: It has a bunch of commits from after the "official" weekly
[01:23 PM] Aleric Inglewood: Well, it's the same as Armin wanted to use for linux :/ ... I guess MoaP is completely off the agenda for 1.4 then.
[01:23 PM] Nicky Perian: Need to go later folks
[01:24 PM] Jacek Antonelli: Take care Nicky
[01:24 PM] Thoria Millgrove: tc Nicky
[01:24 PM] Jacek Antonelli: Why do you say it's off the agenda for 1.4, Aleric?
[01:24 PM] Aleric Inglewood: Because everyone seems to be against using the latest llqtwebkit.
[01:24 PM] Thickbrick Sleaford: Aleric, first get the M working, then oaP
[01:25 PM] Thickbrick Sleaford: once we have a working baseline, we can start changing thing and immedialty know what broke
[01:25 PM] Jacek Antonelli: I'm not _against_ it. It just didn't work, so I fell back so it would compile and I could release something for Mac.
[01:25 PM] Aleric Inglewood: If you got to source code and compiled it yourself, it would have worked too :)
[01:25 PM] Jacek Antonelli: Yeah, in a week
[01:26 PM] Jacek Antonelli: Anyway.
[01:26 PM] Jacek Antonelli: Getting back on the subject at hand: Mac and Windows don't have streaming audio, but otherwise work.
[01:27 PM] Jacek Antonelli: We'll need to either fix the Quicktime plugin to make it work, or use Gstreamer or make a new plugin with VLC or something.
[01:27 PM] Jacek Antonelli: Gstreamer might be the best short-term solution.
[01:27 PM] Jacek Antonelli: (Sad to say)
[01:27 PM] Aleric Inglewood: Will experimental be released without it? Do we the experience to get it to work at all (apparently not?). If not, will you back out of plugins again throwing away everything done so far? And, what shall I do with all the work that I have ready for merging? I'm holding back until weekly is stable again.. but it takes long :/
[01:28 PM] Aleric Inglewood: Do we have*
[01:28 PM] Jacek Antonelli: The experimental was already released two days ago, on all platforms
[01:29 PM] Aleric Inglewood: I wouldn't want to use a viewer without music... not for a day :/
[01:29 PM] Jacek Antonelli: And, it's "stable" aside from the audio stream issue.
[01:29 PM] Ron Overdrive: my vote is either fix QuickTime or consider VLC if possible
[01:29 PM] Aleric Inglewood: I'm sure it's a simple issue
[01:29 PM] Jacek Antonelli: Sounds like you're volunteering to fix it then, Aleric?
[01:30 PM] Thickbrick Sleaford: heh
[01:30 PM] Jacek Antonelli: McCabe has been poking at it for weeks with no luck, but since it's a simple issue it will be no trouble for you
[01:30 PM] Thickbrick Sleaford: it is *possible* that it's not something that's easily solved
[01:31 PM] Thickbrick Sleaford: since noone used the quicktime plugin to play choutcast streams until now. LL viewer have FMod for that.
[01:31 PM] Aleric Inglewood: No, where did I say that? You know I don't have a windows machine... but if you insist, perhaps I should try... I did install windows on a dual boot last time out of frustrating when on snowglobe it was stalling long on windows too (the SNOW-713 thing)... Someone needs to hand-hold me though with how to set upa compile environment on windows (XP)
[01:32 PM] Jacek Antonelli: I guess I misunderstood. Since you were complaining about it and saying how easy it should be to fix, that sure sounded like volunteering to me.
[01:32 PM] Aleric Inglewood: A complete misunderstanding. I volunteer only for linux issues, and then I just fix them(tm).
[01:32 PM] Aleric Inglewood: Max time 1 week.
[01:33 PM] Jacek Antonelli: Since you don't use Windows anyway, you're not affected by the problem, so there is nothing holding you back.
[01:33 PM] Robin Cornelius: i suspec t the quicktime plugin as it stands is no use for audio only streams, it would need massive refactoring to call the correct quicktime API to play a audio only stream
[01:33 PM] Jacek Antonelli: Yes, that's my suspicion as well, Robin
[01:33 PM] Aleric Inglewood: There is: if you back out of plugins then I will have to revert everything else I do from now on too. So, I hesitate putting time into more work, unless I feel secure that that won't happen.
[01:34 PM] Jacek Antonelli: We're not backing out of plugins.
[01:34 PM] Robin Cornelius: worst case is we write a new plugin for VLC or god forbid, fmod
[01:34 PM] Aleric Inglewood: ok.. then I'll proceed as if windows works :/ .. or at least, as if what is in weekly now is there to stay.
[01:35 PM] Robin Cornelius: i guess you could even use windows media API and write a plugin to call that
[01:35 PM] Thickbrick Sleaford: <cough>phonon</cough>
[01:35 PM] Boroondas Gupte: on windows?
[01:35 PM] Thickbrick Sleaford: yes
[01:35 PM] Kakurady Drakenar: Yeah, except it doesn't seem to be fully working...
[01:35 PM] Robin Cornelius: i forgot about that
[01:35 PM] Thickbrick Sleaford: (in theory)
[01:35 PM] Robin Cornelius: i started work on that ages ago, did not look hard
[01:35 PM] Robin Cornelius: but dropped it
[01:36 PM] Jacek Antonelli: Whatever works.
[01:37 PM] Boroondas Gupte: Are there any toolkits left that the viewer doesn't depend on (yet)?
[01:37 PM] Boroondas Gupte: ;-)
[01:37 PM] Robin Cornelius: phonon is qt
[01:37 PM] Aleric Inglewood: (PS my conclusion that it had to be something simple was based on my idea that this worked in LL's viewer. But if, as thickbrick says, quicktime has never been used by any viewer - then I take back that conclusion)
[01:37 PM] Jacek Antonelli: I don't think we're using fltk yet
[01:37 PM] Kakurady Drakenar tries testing speakers on KDE4's settings panel on Windows. Sound plays and then the settings panel crashes.
[01:37 PM] Thickbrick Sleaford: Aleric, used for movies, not shoutcast streams.
[01:37 PM] Jacek Antonelli: Anyway, let's adjourn, this meeting is way overtime.
[01:38 PM] Jacek Antonelli: Kakurady, we can discuss translations next week, or you could send something to the list if you have a question/comment
[01:39 PM] Kakurady Drakenar: It's not really about translations as much as about them being one of the things that may need to be rewritten from scratch,
[01:39 PM] Kakurady Drakenar: due to licensing.
[01:40 PM] Jacek Antonelli: Ah. We can probably get permission from the authors, in cases where the translations aren't already in the LGPL source
[01:40 PM] Kakurady Drakenar: There are six translations that aren't.
[01:40 PM] Jacek Antonelli: Ok. Thanks for coming, all. Take care!
[01:41 PM] Thoria Millgrove: bye, tc
[01:41 PM] Boroondas Gupte: Btw., are daggy fixes an issue due to the license change?
[01:42 PM] Jacek Antonelli: No, the license change is just an issue for code that we didn't write ourselves, really.
[01:42 PM] Jacek Antonelli: In most cases the author's permission will be sufficient. In rare cases we'll need to rewrite from scratch.
[01:43 PM] Aleric Inglewood: Grrmbl, I need to figure out how to create a new popup with an OK and Cancel button :/
[01:43 PM] Jacek Antonelli: But, no GPL patches or commits from other people should be applied without the author's permission to relicense to LGPL.
[01:44 PM] Boroondas Gupte: So basing new changes on GPL versions of existing code and merging forward doesn't make the new code GPL?
[01:45 PM] Thickbrick Sleaford: Boroondas, read this:
[01:45 PM] Boroondas Gupte: thx
[01:45 PM] Boroondas Gupte: um, whut?
[01:46 PM] Thickbrick Sleaford: the term "kitbag question"
[01:46 PM] Thickbrick Sleaford: nevermind. I preffer to pretend licensing doesn't exist for as long as I can.
[01:47 PM] Boroondas Gupte: Though ... how long *can* we :-P
[01:47 PM] Thickbrick Sleaford: considering what Phoenix is doind, and getting away with, I think we're safe.
[01:47 PM] Boroondas Gupte: true
[01:48 PM] Jacek Antonelli: In my opinion (based on info from people who are familiar with the precedent here), code newly written by the author can be relicensed by them, even if it used to be surrounded by some GPL code. For any changes to existing GPL code (rewritten lines), if an equivalent change can be made to the LGPL source, it's not an issue either.
[01:48 PM] Kakurady Drakenar: Emphasis on "newly written", I suppose...
[01:48 PM] Thickbrick Sleaford: (i.e. linking to closed source, mixing GPL+GLOSS with LGPL, etc.)
[01:48 PM] Boroondas Gupte: Ah, that makes sense Jacek.
[01:49 PM] Jacek Antonelli: But if they changed some linse in function X, but function X was removed or rewritten in the new codebase, then the patch can't be used. Both legally and practically.
[01:50 PM] Jacek Antonelli: (Obviously)
[01:50 PM] Boroondas Gupte: So the only case when you may not merge your own change forward is when you cannot, anyway.
[01:50 PM] Boroondas Gupte: :-D
[01:50 PM] Jacek Antonelli: Hehe, pretty much
[01:51 PM] Jacek Antonelli: There might be some other cases, like copying and pasting Linden code then tweaking it. We'd have to evaluate those individually.
[01:51 PM] Kakurady Drakenar: That would be a problem with the Chinese translation, which is based on GPL sources and doesn't appear in the new codebase.
[01:51 PM] Kakurady Drakenar: Some of it, anyway.
[01:51 PM] Aleric Inglewood: I'm with thickbrick: lets just make things work, port it over. Let LL worry about license issues.
[01:51 PM] Kakurady Drakenar: s/and/that/
[01:52 PM] Jacek Antonelli: Kakurady, these are the translations you did? Did you write the new phrasings yourself, independent of the fact that they are embedded in GPLed XML?
[01:53 PM] Aleric Inglewood: Am I the only one who is being flooded with WARNING: LLError::NoClassInfo::decodeImpl: not enough data for requested discard level, setting mDecoding to FALSE, difference: 5 ?
[01:53 PM] Aleric Inglewood: I get that 10 times per second.
[01:53 PM] Jacek Antonelli: Not just you Aleric
[01:53 PM] Thickbrick Sleaford: I don't get it here
[01:53 PM] Boroondas Gupte: nope, I just see display_stats every 10 sec, ATM
[01:54 PM] Kakurady Drakenar: Jacek: Parts were adapted from the earlier translation. I guess those parts will have to be rewritten.
[01:54 PM] Thickbrick Sleaford: we might need to revert
[01:54 PM] Jacek Antonelli: Yeah, I think so Kakurady. Unless the earlier translation also exists in the LGPL source.
[01:55 PM] Kakurady Drakenar: I remember that warning too.
[01:55 PM] Aleric Inglewood: If that depends on a openjpeg patch that we don't have... then it doesn't make sense to have it.. but what does it attemp to fix, thickbrick?
[01:55 PM] Jacek Antonelli: Ok, now I really have to go, so let's adjourn. Carry on in IRC if you like, I'll be around-ish.
[01:55 PM] Thoria Millgrove: take care Jacek
[01:55 PM] Jacek Antonelli: Take care all
[01:55 PM] Aleric Inglewood: cya