#Ubuntu-Meeting

Ubuntu-Meeting Log 2004-11-13

Topic: Something which will pass for a tech board meeting

04:53   mdz     Keybuk: I received an SMS from a number that I don't recognize saying that they won't be able to make the tech board meeting due to being on a plane
04:55   Keybuk  what's the country code?
04:59   mdz     Keybuk: so shall we do this, or punt to the next meeting?
05:00   mdz     this isn't necessarily tech board stuff, I suppose, but we do need to talk about it
05:00   Keybuk  we can have a chat about it, sure
05:00   Keybuk  only the one thing on the agenda?
05:00   mdz     yep
05:01   Keybuk  suppose we better grab a couple of people
05:02   Keybuk  I can sense this is going to be of record length
05:03   daniels Keybuk: er, no, hoary kickoff meeting
05:03   mdz     elmo: the only thing we need to discuss is how to manage the ongoing merges of new packages from Debian
05:03   Keybuk  daniels: what was last week, dude
05:03   thom    daniels: the other direction
05:04   elmo    mdz: ?
05:04   mdz     elmo: Debian, big place with lots of packages
05:04   elmo    mdz: ?
05:05   mdz     elmo: ...
05:05   thom    *giggle*
05:05   mdz     Keybuk: so anyway, what kind of shape are the tools in for use on an ongoing basis?
05:05   daniels elmo:
05:05   Keybuk  ok
05:05   Keybuk  well, as I understand things
05:05   Keybuk  packages we haven't modified, elmo's magic will pull from Debian
05:05   mdz     Keybuk: is it feasible to automate at least the generation of the best-effort merges, and file bugs for their review?
05:06   Keybuk  yeah, I think so
05:06   Keybuk  if he can flag that a package we've modified has a new Debian version, my stuff can pull previous and current Debian, try to apply the diff to warty, and do it's dropped stuff
05:06   Keybuk  automating a bug filing ... I have no idea about ... I imagine it's doable though
05:07   mdz     Keybuk: there's a module in debzilla for filing bugs in bugzilla; it's dead easy
05:07   Keybuk  should we flag them all, or just the ones which the debian patch doesn't apply to?
05:08   mdz     based on the last round, i think at least an eyeball review is called for
05:08   mdz     and a test build
05:08   Keybuk  yeah, I tend to agree
05:08   Keybuk  there were a few where patch did odd things
05:08   mdz     though, these are much smaller
05:08   elmo    Keybuk: how do you want me to flag that to you?
05:08   mdz     the ones where it did truly odd things seemed to cause build failures anyway, so they'd be caught
05:09   mdz     having elmo's stuff send notifications sounds hairy; maybe Keybuk's stuff should just pull a Sources file from Debian on its own
05:09   Keybuk  the main issue is that I'm running on rookery
05:09   elmo    mdz: I can ultra-trivially dump the lorraine output to a text file which keybuk can wget?
05:09   elmo    Keybuk: why's that an issue?
05:10   mdz     elmo: works for me
05:10   Keybuk  just a text file of source package names elmo thinks Debian have newer than warty would be ideal
05:11   Kamion  sorry I'm late
05:11   mdz     Keybuk: as part of your output, can you include a debdiff from Hoary->the merged version?
05:11   Keybuk  yeah, if elmo installs that on rookery <g>
05:11   mdz     good point
05:12   mdz     I've been nagging him for a week to put it on jackass :-P
05:12   elmo    err, you so haven't
05:12   elmo    anyway, it's on rookery
05:12   mdz     Oct 29 12:41:07 <mdz>   elmo: debdiff on jackass, please?
05:13   Keybuk  would you prefer debdiff from base -> {old hoary, new hoary} as well?  instead of just ordinary patch?
05:13   elmo    mdz: ... one IRC message [that I missed]  is nagging for a week?
05:13   mdz     elmo: I could dig up more :-)
05:13   mdz     Oct 26 15:25:44 <mdz>   could I get debdiff+interdiff on jackass?
05:14   Keybuk  mdz: so, this debzilla thing, what's that?
05:14   mdz     Keybuk: that's the thing which creates bugs in bugzilla based on debbugs bugs
05:14   elmo    do you want just main, or main +universe?
05:15   mdz     Keybuk: matt.zimmerman@canonical.com--2004/debzilla--mainline--0
05:16   mdz     elmo: ideally both, but filterable
05:16   mdz     e.g., mark them as main or not in the text file
05:16   mdz     so we could generate the output for universe, but not file bugs
05:17   elmo    ok
05:17   mdz     Keybuk: work for you?
05:18   Keybuk  yup
05:19   mdz     Keybuk: yeah, base->new hoary is good too
05:19   mdz     I'm hoping those will be small and easily reviewable
05:19   Kamion  has the problem that we often don't have the right base been fixed?
05:19   Keybuk  yeah, hopefully
05:19   Kamion  that bit me with a lot of the last round of merges
05:19   Keybuk  Kamion: theoretically it's seen the right base now :p
05:20   Keybuk  but yeah, I'll write some code to string ubuntu* and look on snapshot for that
05:20   Kamion  Keybuk: certainly won't've done for some of the ones I'm doing
05:20   Kamion  ok
05:20   Keybuk  string? strip!
05:20   Keybuk  that's actually easy ... the only reason I didn't do it before is because I'm fallible and didn't think of that <g>
05:20   elmo    keybuk: jackass/lorraine/needs-merged.txt
05:20   elmo    missing universe atm. but is that okay?
05:21   Keybuk  cool
05:21   Keybuk  yup
05:21   Keybuk  I assume it'll be s/main/universe/ :p
05:21   elmo    for universe ones, yes
05:22   elmo    it's "%s %s" % (pkg, component)
05:22   mdz     so I guess we'll find out in the course of doing this, how much manual work it will truly be
05:23   mdz     I think it will be fairly reasonable
05:23   mdz     but hopefully just enough to provide the necessary incentives to push our changes upstream :-)
05:23   mdz     speaking of which, herbert's kernel-package stem patch needs to go upstream
05:23   Keybuk  what are we doing wrt to the current patches, btw?
05:24   Keybuk  have they all been pushed, or do we need to push some, etc.
05:25   mdz     we don't have those figures
05:25   Kamion  I've been going over d-i stuff as I upload the merges, pushing upstream as appropriate
05:26   mdz     of course we do have a fair volume of stuff that upstream isn't going to want
05:26   mdz     maybe after hoary when we have better tools, we can do a comprehensive review and make sure that everything has been submitted upstream
05:27   Kamion  indeed, by far the greatest volume of the d-i patches is branding
05:27   mdz     maybe we'll even have someplace to store that metadata :-P
05:27   mdz     and mark patches as should-go-upstream or not
05:27   mdz     anything else to discuss as part of the meeting?
05:28   Kamion  seeds?
05:28   mdz     good call
05:28   Keybuk  I'm going to finish brutalising hct's cli today ... will work on the continual merge thing tomorrow
05:28   Kamion  (viz., having some)
05:28   mdz     should we temporarily use some other repository for the master seed lists until the wiki is fixed?
05:28   Kamion  I'm happy with a world-writable text file on rookery, personally :-)
05:28   mdz     they could be part of germinate's source tree
05:28   Keybuk  we could always use the old wiki?
05:28   Kamion  Keybuk: that might be a plan
05:29   mdz     Keybuk: only if we can enable changes on a per-page basis
05:29   mdz     what about a group-writable arch archive?
05:30   Kamion  as long as we can get at it from {my dev boxes, little, jackass} that'd work
05:30   Kamion  I don't think it should be part of germinate's source tree
05:31   mdz     is the wiki really the best place for it long-term?
05:31   mdz     maybe when we were editing it every day, it was
05:31   mdz     we can keep proposals in the wiki, and move them into the official seeds by other means
05:32   mdz     needs a change history -> arch
05:32   elmo    the nice thing about the wiki is it was a http get for germinate
05:33   elmo    or anything else that wanted to fuck around with seeds
05:33   Kamion  elmo: we can do that with arch too
05:33   mdz     a tla get isn't much different, no?
05:33   Kamion  elmo: dump a checkout on rookery
05:33   mdz     can't tla dump a file for you in one step?
05:34   elmo    Kamion: yeah, true
05:34   mdz     even if not, a checkout isn't a big deal, I suppose
05:34   Kamion  fewer dependencies would be good, but I don't see a problem with having a cronned tla update on people/~cjwatson/seeds/ or whatever
05:34   mdz     I don't see a problem with germinate doing a tla get
05:34   Kamion  it's slow enough already :P
05:35   mdz     yeah, exactly :-)
05:35   mdz     what should we call the archive?
05:35   Kamion  and where should it go? do we want non-canonical staff to be able to write to it?
05:36   Kamion  (I assume the answer is "eventually, yes")
05:36   mdz     right
05:36   mdz     but it isn't a big deal if they can't quite yet
05:36   mdz     it's not something which changes often, and we can propagate things easily enough when there's a consensus
05:36   Kamion  so just chinstrap as before then, I guess
05:36   mdz     yeah
05:37   mdz     more a question of a personal archive vs. something else
05:37   mdz     Kamion: fine with me if you want to keep them in your archive next to germinate
05:37   Kamion  ubuntu-devel@lists.ubuntu.com/seeds--warty--0 etc.?
05:37   Kamion  mdz: that's a bit technically awkward for group-writability, because my archive on chinstrap is only a mirror
05:38   Kamion  and y'all don't have logins on my home server
05:38   mdz     ah, separate then, I suppose
05:38   mdz     does anyone know if we have an official naming scheme for public archives from the arch guys?
05:38   Keybuk  we do for imports, yes
05:38   mdz     anything which we could apply here as well, for consistency?
05:39   Keybuk  <product>@arch.ubuntu.com[--<source>-<year>] /<source>--<branch>--<version>
05:39   Keybuk  Kamion's idea seemed sound to me
05:39   mdz     the ubuntu-devel one?
05:39   Keybuk  yeah
05:40   mdz     works for me
05:40   Kamion  if there are no objections, I can create that after the meeting, import warty's seeds, and tag them onto hoary
05:40   mdz     we can always change it if the arch secret police object
05:40   mdz     Kamion: sounds good
05:40   mdz     we have a bunch of seed changes to review and effect
05:41   mdz     I have a list from the kickoff meeting
05:41   mdz     and I believe there is stuff in the wiki for hoary
05:42   mdz     arything else to discuss?
05:42   Keybuk  not from me
05:43   mdz     ok, adjourned
05:44   mdz     thanks, guys