Open Source Meeting/2009-06-25

From Second Life Wiki
< Open Source Meeting
Revision as of 15:12, 25 June 2009 by Rob Linden (talk | contribs) (transcript)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Open source meeting - Thursday, 2pm PT.

Teleport to the Linden Open Source Project headquarters.

Agenda

  • I'd like to see discussed how to handle issues pertaining to both 1.23.4 and Snowglobe. For instance VWR-13868 exists in both viewers but there is no issue entered under the SNOW project, so would a fix implemented also be put into Snowglobe? Another example is SNOW-61 and VWR-13975 here we have the same issue under both projects, should SNOW-61 be resolved as a dupliciate of VWR-13975?. A third example is VWR-14357 a Snowglobe issue entered under VWR but resolved as a duplicate of VWR-13302 a 1.23.4 issue. And an example of a change that made it into 1.23.4 but not Snowglobe is "Groups" on the avatar pie menu. --Harleen Gretzky 16:56, 25 June 2009 (UTC)

Transcript

  • [13:58] Rob Linden: hi folks
  • [13:58] Harleen Gretzky: Hi Rob :)
  • [13:59] Aimee Trescothick: hey :)
  • [13:59] Rob Linden: thanks for the agenda item Harleen....very timely and appropriate
  • [13:59] Rob Linden: [1]
  • [13:59] Harleen Gretzky: yw :)
  • [13:59] Winne Woodget: hi bob, merov :)
  • [13:59] Merov Linden: hi
  • [14:00] Rob Linden: one easy thing while we're waiting on everyone: https://jira.secondlife.com/browse/SNOW-62
  • [14:00] Morgaine Dinova: Hi Rob, Merov :-)
  • [14:00] Merov Linden: Hi Morgaine
  • [14:00] Morgaine Dinova: Oh cool, we're tagged :P
  • [14:00] Rob Linden: probably shortly after this meeting I'll be tagging Snowglobe 1.0 in svn, creating a "trunk", and opening up the trunk for new edits
  • [14:01] Rob Linden: if anyone hates the locations, say so in JIRA
  • [14:02] Aimee Trescothick: sounds fine, at least till the rocket reaches Mercury
  • [14:02] Rob Linden:  :)
  • [14:02] Rob Linden: so....we have a quorum to really begin?
  • [14:02] Morgaine Dinova: We're in LEO currently :-)
  • [14:03] Rob Linden: Item #1 from the agenda: I'd like to see discussed how to handle issues pertaining to both 1.23.4 and Snowglobe. For instance VWR-13868[c
  • [14:03] Morgaine Dinova: This place feels so empty without the tiny protesters ;-)
  • [14:03] Harleen Gretzky: lol
  • [14:04] Aimee Trescothick: :D
  • [14:05] Rob Linden: so, the original thought behind having the SNOW project was to be able to track issues separately
  • [14:05] Rob Linden: however, we've been pretty hesitant to create duplicates
  • [14:05] Robin Cornelius: i think being to track seperatly is a good idea
  • [14:05] Rob Linden: ...and it hasn't been strictly necessary in the grand scheme of things
  • [14:05] Robin Cornelius: and i think duplicates are the only real way
  • [14:06] Robin Cornelius: i think if we track just via VWR its actually the identical situation in reverse
  • [14:06] Harleen Gretzky: So how do we insure changes made in one project make it to the other?
  • [14:06] Harleen Gretzky: *not changes *fixes
  • [14:06] Rob Linden: Harleen hit upon the essential question here
  • [14:07] Rob Linden: forking the issues does create a new level of burden
  • [14:07] Robin Cornelius: its more difficult as not all issues have patches comming from this side of the firewall
  • [14:07] Thickbrick Sleaford: marking as duplicate doesn't necessarily means resolving as duplicate
  • [14:07] Thickbrick Sleaford: (?)
  • [14:07] Morgaine Dinova: Given that main will cherry-pick from Snowglobe rather than simply follow, I think there's no option but to keep the issues separate. Something could get fixed in Snowglobe without getting fixed in main, simply because there are community commits in one and not in the other.
  • [14:08] Harleen Gretzky: But not all issues have a corresponding issue in the other project also to be marked duplicate, probably cause until now mostly techie ppl were using Snowglobe
  • [14:09] Harleen Gretzky: Like Robin's avatar is nothing but prims to me on Snowglobe, but I do not think there is a SNOW issue for that only a 1.23.4
  • [14:10] Rob Linden: so, I think those of us working on Snowglobe will inevitably have the burden of looking through the patches in VWR, and deciding what to pull into Snowglobe
  • [14:10] Rob Linden: however, we'll also need to leave those issues as unresolved when we fold them into Snowglobe, to avoid any confusion as to whether the patch has been folded into the mainline Second Life viewer
  • [14:11] Merov Linden: may be we should create the duplicate JIRA record *only* when cherry picking a patch?
  • [14:11] Rob Linden: +1 Merov
  • [14:11] Merov Linden: so we can mine the JIRA for history and keep svn messages useful
  • [14:11] Morgaine Dinova: One of the issues that has made Jira duplication a problem in the past is that dupes have been set to "Resolved" when actually all that has meant is that it was linked to another. This is problematic because clearly the issue is not "Resolved" from the PoV of the poster. An additional resolution state of "Linked" is needed I think.
  • [14:13] Rob Linden: Morgaine: that's an age-old problem that I don't think we need to try to solve now
  • [14:13] Merov Linden: The problem Morgaine is that, as a dev and committer, you want to sort out precisely what is outstanding, what has been done, etc...
  • [14:13] Rob Linden: let's take a look at a couple of Harleen's specific examples:
  • [14:14] Rob Linden: https://jira.secondlife.com/browse/VWR-13868
  • [14:14] Morgaine Dinova: Merov: sure, but if you had a resolution state of "Linked" then you could treat it as "done" quite trivially.
  • [14:15] Rob Linden: that's an example of one we can add "Affects Version" to "Snowglobe 1.0", and then look at that list when we're looking for stuff to address
  • [14:16] Rob Linden: Harleen, does that make sense for that one?
  • [14:16] Rob Linden: (so, in other words, no duplicate issue, just use "affects version")
  • [14:16] Rob Linden: btw, I just added "Snowglobe 1.0" to the list of options
  • [14:16] Harleen Gretzky: yes, and some are using Snowglobe test build in that way now, because they do not relize there is a SNOW project
  • [14:17] Rob Linden: Harleen: that's fine, we can make sure we look at those too
  • [14:17] Rob Linden: in fact, we'll often ask "is this also a problem in 1.23", because that has a big impact on what the Lindens involved in Snowglobe decde to do with it
  • [14:18] Harleen Gretzky: makes sense
  • [14:18] Rob Linden: cool...make sense to everyone else as well?
  • [14:18] Merov Linden: /nods
  • [14:18] Rob Linden: waits a second....
  • [14:18] Aimee Trescothick: think so
  • [14:19] Rob Linden: ok, how about https://jira.secondlife.com/browse/SNOW-61
  • [14:19] Harleen Gretzky: That is technically a dupe of https://jira.secondlife.com/browse/VWR-13975
  • [14:20] Rob Linden: so, since someone already went through the trouble of filing a separate issue in SNOW, there's no sense in resolving it as a duplicate
  • [14:20] Rob Linden: however....
  • [14:20] Rob Linden: we can set the priority lower
  • [14:21] Rob Linden: since it's not necessarily a high priority issue for the Snowglobe developers specifically to address
  • [14:21] Rob Linden: that seem about right?
  • [14:21] AOS-HUD 3.5.182: [WEAR ME!]: - The scanner on steroids - [2]
  • [14:22] Harleen Gretzky: yes
  • [14:22] Rob Linden: Next up: https://jira.secondlife.com/browse/VWR-14357
  • [14:23] Rob Linden: (btw...I'm not capturing any of these resolutions, other than the eventual transcript)
  • [14:23] Harleen Gretzky: Using SNOW-61 as an example, this should really be moved to SNOW, it is currently resolved as dupe of https://jira.secondlife.com/browse/VWR-13302
  • [14:24] Rob Linden: yup....that's an acceptable way of handling that
  • [14:24] Dirk Talamasca: Hiya Q
  • [14:24] Q Linden: yo
  • [14:24] Aimee Trescothick: lo
  • [14:25] Aimee Trescothick: tries to look behind the LInden's group tags, to see if it says "LL Back" on the other side
  • [14:25] Rob Linden: ok....so, we've been through the list of examples that Harleen provided
  • [14:25] Rob Linden: lol Aimee
  • [14:25] Darv Linden: We're just the front for LL, we're not actually them :P
  • [14:25] Aimee Trescothick: =)
  • [14:26] Rob Linden: hadn't considered how baffling that would be to Residents
  • [14:26] Harleen Gretzky: Well last thing I brought up didn't have a specific JIRA (too many) is that Groups was returned in 1.23.4 but not Snowglobe
  • [14:26] Rob Linden: yeah, that basically is our mark of being Howard Linden's minions ;-)
  • [14:26] Q Linden: prefers to think of himself as a maxion
  • [14:27] Aimee Trescothick: nah, just makes you sound like escapees from a monty python sketch
  • [14:27] Morgaine Dinova: lol
  • [14:27] Rob Linden: looks for his Ministry of Funny Walks animation override
  • [14:27] Morgaine Dinova: rofl
  • [14:28] Aimee Trescothick: LOL
  • [14:28] Merov Linden: thinks Rob is gonna open shop IW
  • [14:28] Rob Linden: I'd probably buy that if I could find it in world
  • [14:28] Rob Linden: admittedly hasn't looked
  • [14:28] Rob Linden: anyway...that's the only thing on our official agenda
  • [14:29] Rob Linden: one closely related topic I wanted to talk about:
  • [14:29] Rob Linden: looking through the VWR project for patches to fold into Snowglobe
  • [14:30] Rob Linden: here's a list of the patches that we look through on at least a weekly basis: https://jira.secondlife.com/secure/IssueNavigator.jspa?reset=true&&priority=1&priority=2&priority=3&priority=4&priority=5&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
  • [14:30] Rob Linden: (rather, we look at that query on a weekly basis, not every patch in it
  • [14:31] Morgaine Dinova: One thing that popped up in Tuesday's AW Groupies meeting was the question of patches for interop. Is there likely to be opposition to patches submitted for interop purposes, given that you may not find any great interest in cherry-picking such patches for the main viewer?
  • [14:31] Rob Linden: there's nothing stopping the people with commit access from committing the ones that we've already reviewed into Snowglobe
  • [14:32] Rob Linden: Morgaine: if someone steps up to actually maintain the code, and the changes are generally consistant with what we're doing on OGP, and they don't break the production grid experience, then no problem
  • [14:32] Morgaine Dinova: Sounds fair :-)
  • [14:34] Rob Linden: is there anyone who isn't a committer yet that's interested in shepherding patches into Snowglobe?
  • [14:35] Morgaine Dinova: No doubt more will appear, Rob :-)
  • [14:35] Rob Linden: k...well, anyway, let me ask a different easier question:
  • [14:35] Rob Linden: anything on that list stand out as somethign that really needs to get into Snowglobe sooner rather than later?
  • [14:37] Aimee Trescothick: VWR-11898 I got asked to "adopt" the other day by it's writer catherine pfeffer as she's too busy to follow it up
  • [14:37] Twisted Laws: has used the patch in [3] on Snowglobe with no adverse effects and seems to solve at least the one issue.
  • [14:37] Rob Linden: given that all of these are patches, if there exists examples in VWR that are candidates for Snowglobe, then a second issue in SNOW would be appropriate
  • [14:37] Aimee Trescothick: however it will need testing on a French OS X install, which mine isn't, looks pretty straight forward though
  • [14:37] Merov Linden: ducks
  • [14:38] Rob Linden: k....yup, those seem like two great examples
  • [14:38] Aimee Trescothick: heheh
  • [14:38] Aimee Trescothick: quack
  • [14:38] Merov Linden: well, I'm gonna have to do some localization merges and fixes in 1.1 I think
  • [14:39] Rob Linden: k....so, back on Harleen's original topic: should we create the duplicate issues now, or wait until there's a dev that's committed to do make the Snowglobe commit?
  • [14:39] Aimee Trescothick: the two line history bugs on there I posted this week I'll be looking to follow up shortly too
  • [14:40] Rob Linden: if the answer is "let's do it now", then anyone want to volunteer to create the issues?
  • [14:40] Aimee Trescothick: I'll go ahead and do those two now, as I'll be doing them anyway
  • [14:41] Aimee Trescothick: that's VWR-14267 and VWR-14278
  • [14:41] Rob Linden: cool
  • [14:41] Rob Linden: thanks Aimee
  • [14:41] Aimee Trescothick:  :)
  • [14:42] Rob Linden: for any others, let's consdier that the process for nomiating patches for Snowglobe
  • [14:42] Rob Linden: (VWR patches, that is)
  • [14:42] Aimee Trescothick: oh, and I'll pick up VWR-13227 too, as it was me that broke it *blushes*
  • [14:42] Rob Linden:  :)
  • [14:43] Harleen Gretzky: And what about VWR fixes that were not done by submitted patches, but by LL themselves
  • [14:43] Rob Linden: additionally, of course, before actually making a move to check something in, a quick note to sldev calling for review is the way to go
  • [14:43] Aimee Trescothick: sure
  • [14:44] Rob Linden: we'
  • [14:44] Rob Linden: we'd love to get more committers than we've got today, so ping me if you're interested
  • [14:44] Morgaine Dinova: Where do you typically store the patches prior to checkin, for SLdev access and review? Or just inline?
  • [14:45] Merov Linden: diff attached to JIRA
  • [14:45] Morgaine Dinova: kk
  • [14:45] Rob Linden: Morgaine: attached in JIRA is the formal process
  • [14:45] Morgaine Dinova: nods
  • [14:46] Rob Linden: we'll alternately look more at linking to Mercurial as another means, but we need to make sure it's unambiguous that the patch is actually being contributed
  • [14:47] Rob Linden: in the brave new future of Mercurial development, we'll need to work out norms
  • [14:47] Rob Linden: (different norms, that is)
  • [14:48] Rob Linden: anyway...we've talked that one out it seems. anything else on your minds?
  • [14:49] Soft Linden: Man, it's just low-key day.
  • [14:49] Rob Linden: I was just about to say
  • [14:49] Merov Linden: post partum depression...
  • [14:49] Thickbrick Sleaford: watches a tumbleweed go by
  • [14:49] Morgaine Dinova: <crickets>
  • [14:49] Aimee Trescothick: turns up the Ambient volume slider
  • [14:50] Rob Linden: I guess one thing I neglected to do was to say "w00t!" about Snowglobe 1.0
  • [14:50] Soft Linden: Speaking of which, if anyone wants to usher that wind noise optimization patch into Snowglobe... eh? eh?
  • [14:50] Thickbrick Sleaford: yes!
  • [14:50] Aimee Trescothick: :D
  • [14:50] Rob Linden: thanks everyone here for making that happen!
  • [14:50] Merov Linden: that's the baby I was thinking about...
  • [14:50] Soft Linden: Could still be worth something for the really low end systems.
  • [14:50] Aimee Trescothick: heh, was just looking at that
  • [14:51] Aimee Trescothick: it's gone a bit stale, needs a rewrite to work with the OpenAL stuff
  • [14:51] Merov Linden: Aimee's minimap featured in Torley's video :)
  • [14:51] Aimee Trescothick: ^^
  • [14:51] Robin Cornelius: though all the wind optimisation was done on openAL wind?
  • [14:52] Morgaine Dinova: Well if we're out of topic, perhaps another word or two on interop? A lot of people like to use LL viewers on Opensim, and Snowglobe is no exception. Would adding more grid support on the login screen of Snowglobe be a possible candidate?
  • [14:52] Rob Linden: Morgaine: possibly....need a dev volunteer to own that, though
  • [14:52] Morgaine Dinova: Understood
  • [14:52] Aimee Trescothick: no, the OpenAL wind is still generated the old way, but used a bit different, it may be that the optimizations don't provide much improvement now anyway
  • [14:53] Aimee Trescothick: it's been a while since I looked at it
  • [14:54] Robin Cornelius: ah ok, i know A whole chain of people worked on that and after i did wind in openAL gigs and others optimised it then backported that for fmod then tofu merged openAL and i forget...
  • [14:55] Aimee Trescothick: heh
  • [14:55] Rob Linden: okee doke. well, that's definitely a good candidate to get in at some point. I think there may still be a little bit of wondering aloud going on about whether it represents a significant cpu improvement, so some updated benchmarks would probably be part of the process
  • [14:56] Rob Linden: well, anyway, I'm eager to get SNOW-62 done, so if there's nothing pressing, should we wrap this up?
  • [14:57] Thickbrick Sleaford: one question: what's the policy on patches that are already in the linden pipeline (maint-viewer etc.) for inclusion in snowglobe?
  • [14:57] Rob Linden: (I'd like to declare the Snowglobe trunk open for commits)
  • [14:57] Morgaine Dinova: \o/
  • [14:57] Aimee Trescothick: that's a good question
  • [14:57] Rob Linden: Thickbrick: good question. I'd like to see them go in one way or another....
  • [14:57] Rob Linden: one rather cool way of dealing with it:
  • [14:58] Rob Linden: if someone wanted to volunteer for the harder task of merging in one of those branches that has the change it....
  • [14:58] Rob Linden: for example, we publish many of the maint-viewer branches
  • [14:59] Robin Cornelius: The whole process is made much harder by the fact we only see bulk pushes not individual commits and notes
  • [14:59] Rob Linden: yup, I know...
  • [14:59] Aimee Trescothick: nods in agreement with Robin
  • [14:59] Morgaine Dinova: That's interesting. My FPS rises when I mute the wind / Ambient. I've tested it 3 times.
  • [15:00] Robin Cornelius: yes, that itsself is an improvment Morgaine, it use to keep wasting CPU time even when muted
  • [15:00] Soft Linden: That would be really weird, because I think the current code still generates that sample when muted.
  • [15:00] Rob Linden: well, that's probably a good topic for sldev discussion and/or our meeting next week
  • [15:00] Robin Cornelius: not on openAL i don't think
  • [15:00] Soft Linden: That's in 1.23 ?
  • [15:01] Robin Cornelius: i believe the mute/no wind got in to the inital openAL merge
  • [15:01] Soft Linden: Cool cool
  • [15:01] Robin Cornelius: doe snot effect fmod though
  • [15:01] Aimee Trescothick: I think it gets turned off when using OpenAL yeah
  • [15:01] Morgaine Dinova: Robin: slight problem though --- I made a typo. My FPS *falls* when I mute Ambient. :-) That's why I said "interesting", as I'd expect the opposite
  • [15:01] Soft Linden: hrm
  • [15:01] Robin Cornelius: oh really? odd
  • [15:02] Morgaine Dinova: I don't believe it ... going to make another 10 tests when I'm in an empty zone
  • [15:02] Aimee Trescothick: o.O
  • [15:02] Thickbrick Sleaford: it looks like muting reduces the variablity in frame time for me
  • [15:03] Rob Linden: Morgaine: sounds like something to try to narrow down and maybe get a bug filed about...very odd
  • [15:03] Morgaine Dinova: LOL, it's highly repeatable here, 1.5-2.0 FPS drop on mute. Funny :-))
  • [15:04] Aimee Trescothick: that's really weird, hmm
  • [15:04] Morgaine Dinova: scratches head
  • [15:04] Rob Linden: thanks for coming everyone!
  • [15:04] Morgaine Dinova: Thanks Rob, and all :-)
  • [15:04] Thickbrick Sleaford: see you next week everybody
  • [15:04] Aimee Trescothick: maybe when you mute it, it's getting blocked and building up wind pressure
  • [15:04] Merov Linden: bye
  • [15:04] Morgaine Dinova: Haha
  • [15:04] Aimee Trescothick: I wouldn't leave it muted, the mixer might explode
  • [15:04] Harleen Gretzky: tc all
  • [15:05] Rob Linden: lol
  • [15:05] Aimee Trescothick: bye :)
  • [15:05] Rob Linden: bye all
  • [15:05] Morgaine Dinova: Tc all :-)