[02:21] <jamesh> spiv: I'm pushing a fix for bug 4852
[02:21] <Ubugtu> Malone bug #4852: Cannot add a comment to a bug using only the keyboard (with epiphany) In: launchpad (upstream), Severity: Normal, Assigned to: James Henstridge, Status: Accepted https://launchpad.net/bugs/4852
[02:23] <spiv> jamesh: Thanks!
[03:29] <stub> lifeless: I'm creating the next production config. Is there an consensus yet on the 'double slash after the hostname means absolute path' issue?
[03:29] <lifeless> stub: yes, that is bogus. its %2f that makes it absolute
[03:29] <lifeless>  // was never legit
[03:30] <lifeless> right
[03:30] <lifeless> this has swayed my case hugely :)
[03:34] <stub> Is there a bzr command that does 'cat .bzr/parent' ?
[03:41] <lifeless> nope
[03:41] <lifeless> api branch.get_parent()
[03:42] <stub> Bug 5775
[03:42] <Ubugtu> Malone bug #5775: viewing magic file only way to determine default pull location In: bzr (upstream), Severity: Normal, Assigned to: Nobody, Status: New https://launchpad.net/bugs/5775
[03:43] <lifeless> stub: not sure thats a bug
[03:43] <lifeless> stub: whats the use case
[03:43] <stub> Me wanting to know what 'bzr pull' is going to do
[03:45] <lifeless> then I suggest the bug report would be better with that context :)
[03:45] <lifeless> i.e. 'I cannot tell what bzr pull is going to do'
[03:50] <jamesh> lifeless: I suppose the whole "//" vs. "/%2F" thing could be avoided for the Launchpad supermirror by making all users' home directories the root directory
[03:50] <jamesh> so sftp://host/foo is the same as sftp://host/%2Ffoo
[03:51] <lifeless> jamesh: yes. we could do that
[04:11] <lifeless> pqm is going down after the current merge
[04:18] <dilys> Merge to devel/launchpad: [r=BjornT]  add missing breadcrumbs for Souyz pages (r2914: Dafydd Harries)
[04:26] <stub> launchpad will be going down in 10 mins, which will also put the wiki's into read only mode. Estimated down time is 45 mins.
[04:38] <spiv> jamesh: That's precisely what the launchpad supermirror does.
[04:39] <jamesh> spiv: cool.
[04:42] <spiv> jamesh: Or rather, the current working dir upon login is the root of the virtual filesystem.
[04:42] <jamesh> spiv: So both interpretations of "sftp://host/foo" are the same, which is good
[04:42] <spiv> "home directory" is a possibly confusing term for the supermirror sftp server :)
[04:43] <spiv> It doesn't do anything clever with ~, so e.g. "cd ~" will fail.
[04:43] <jamesh> well, sftp doesn't have a "cd" command, right?
[04:44] <spiv> The openssh client does, not sure about the protocol level.
[04:44] <jamesh> what the client does is different to what happens over the wire
[04:44] <spiv> Indeed.
[04:45] <spiv> I'm not as familiar with what happens over the wire, though :)
[04:45] <spiv> But you appear to be right, at glance.
[04:45] <jamesh> I didn't see anything about changing directories in the spec :)
[04:46] <spiv> So, pretend I said "stat ~" rather than "cd ~" ;)
[04:47] <spiv> Anyway, "foo" and "/foo" are the same as far as the supermirror sftp is concerned :)
[04:50] <jamesh> I think you refer to the starting directory as "", actually
[04:50] <lifeless> IIRC wire level considers ~ to be abomination
[04:50] <lifeless> but foo is  ~/foo
[04:50] <lifeless> and .. is ~/foo/..
[04:51] <lifeless> etc
[04:56] <jamesh> lifeless: is the pqm upgrade for moving it to another box?
[04:56] <lifeless> yes
[05:04] <stub> lifeless: Would rewriting sftp://chinstrap.ubuntu.com/foo to sftp://stub@chinstrap.ubuntu.com/foo be the responsibility of config manager, or an external tool that produces a munged config that is then passed to config manager?
[05:05] <lifeless> I think that that might be a useful feature for config manager
[05:07] <spiv> stub: Or echo -e "Host chinstrap.ubuntu.com\n User stub" > .ssh/config  ;)
[05:07] <stub> spiv: Doesn't work for bzr yet, because paramiko doesn't support it
[05:08] <lifeless> stub: paramiko can use openssh
[05:08] <lifeless> but there is a general issue
[05:08] <stub> it can? .ssh/config meets my current use cases then if you are prioritizing features
[05:11] <lifeless> configs need to be changed, to be able to list multiple locastions that *might* be aliased, along with some verification means
[05:12] <lifeless> then we can list the local address in  pqm as an alias to the official location, and the revid (either the tip we should get, or a prefix we once saw)
[05:12] <lifeless> maybe overxmas
[05:13] <lifeless> jamesh: do you know the revid that your patch was merged in in?
[05:18] <stub> Production is live, with new layout
[05:29] <lifeless> wow thats, different
[05:29] <jamesh> lifeless: you mean the redirection() one?
[05:30] <lifeless> jamesh: the openssh using patch
[05:30] <jamesh> oh
[05:30] <lifeless> id, not revno ;)
[05:31] <jamesh> mbp@sourcefrog.net-20051130064100-50469cf3ee5f42b2
[05:32] <jamesh> lifeless: you probably want a newer version though.  there was an error in fixing a merge conflict when mbp merged my branch
[05:33] <jamesh> perhaps john@arbash-meinel.com-20051130145503-95fed0de893c86a1 (the revision after, which includes John's fixes)
[05:35] <lifeless> stub: ^^^ to get bzr using sshd
[05:35] <lifeless> erm ssh
[05:36] <stub> This is for production rollouts - I'll wait until the patch filters through.
[05:38] <lifeless> so I've been thinking
[05:38] <lifeless> why not build the tree on chinstrap and rsync it outwards
[05:39] <lifeless> better for security anyway
[05:55] <stub> lifeless: I thought pulling signed revisions straight from the RCS would be better for security. Building the tree on chinstrap and rsyncing worked well though when I did it that way last week. This week cm.py update did the trick happily.
[05:56] <lifeless> stub: well if we push from chinstrap we can disable ssh *from* galapagos to chinstrap
[05:56] <lifeless> so the vector of compromise launchpad, then attack inwards is reduced
[06:00] <stub> I see
[06:01] <jamesh> the new layout looks really weird until you reload the CSS
[07:08] <lifeless> stub: ping
[07:08] <stub> lifeless: pong
[07:09] <lifeless> should we run the lp tests as amd 64 or i386 ?
[07:09] <lifeless> we have chroots for both
[07:10] <stub> I think amd64 would be good - production systems tend to be moving in that direction and we are more likely to get glitches on that architecture we need to be aware of than i386
[07:11] <lifeless> ok. When I get pqm grokking dchroot commands, I shall make that happen :)
[08:02] <bob2> bah stub
[08:02] <bob2> I'm still getting LP junk
[08:04] <lifeless> what on?
[08:09] <bob2> epiphany-browser at least
[08:09] <bob2> evolution
[08:09] <bob2> pppoeconf
[08:13] <lifeless> yay nessus
[08:13] <lifeless> GPL->proprietary
[08:13] <lifeless> fuckers
[08:20] <matid> Are the changes or launchpad that I've noticed today a bug or is it just a little redesign?
[08:21] <lifeless> I think you may need to reload the css
[08:21] <lifeless> it is meant to look different, but not weird
[08:21] <matid> Maybe, 'cause my IS a bit weird...
[08:24] <jamesh> matid: if it appears that everything is squeezed into a thin column on the left, do a shift+reload
[08:25] <matid> No, that's not the problem
[08:25] <matid> http://developer.slashgeek.net/files/launchpad.png
[08:25] <matid> That's how it looks like
[08:25] <jamesh> matid: that's the new design
[08:25] <matid> Is the column on the right supposed to be highlighted?
[08:26] <jamesh> matid: there are still some changes being made (e.g. making the first box in the left column a bit more dynamic)
[08:27] <matid> jamesh: I'd say that the left column is ok, but the white highlight on the right just doesn't look good
[08:27] <jamesh> matid: the exact styling will probably change a bit
[08:27] <jamesh> matid: the idea is that the tabs along the top of pages have been moved to the menu on the right
[08:28] <jamesh> and the stuff that was in the menu on the right is now a "submenu" of the selected tab
[08:28] <jamesh> (although it isn't a tab anymore ..)
[08:28] <matid> jamesh: was the previous design considered less usable?
[08:28] <matid> jamesh: 'cuase I find this one a bit hard to navigate
[08:29] <matid> jamesh: Or maybe I'm just used to previous one
[08:29] <jamesh> matid: the details of the changes, including some of the rationale, can be found here: https://wiki.launchpad.canonical.com/NewPageLayout
[08:31] <matid> jamesh: BTW - is this change releated to upgrading Launchpad or is it just design change?
[08:33] <jamesh> matid: it is related to upgrading Launchpad, in the sense that this is the first production rollout since the redesign was merged.
[08:33] <matid> jamesh: Oh, ok
[08:33] <lifeless> matid: it weirds me out
[08:34] <lifeless> matid: but I figure we give it 3-4 weeks
[08:34] <matid> jamesh: Just one more question, 'cuase you seem to know quite a lot about launchpad ;)
[08:34] <matid> jamesh: When is the karma updated?
[08:34] <jamesh> matid: once a day, but I'm not sure what time
[08:34] <lifeless> and we should get keybuk and the other architects of the new design to give feedback/etc
[08:34] <matid> jamesh: I already know that, but anyway, thx for help
[08:35] <SteveA> lifeless: yes, but, when this design has completely landed
[08:35] <SteveA> there's still more to do
[08:35] <SteveA> jamesh has some work pending, and so does mpt
[08:36] <lifeless> SteveA: of course. thats why I said 3-4 weeks :)
[08:37] <jamesh> SteveA: do you have some time to talk about the site map stuff a little?
[08:50] <jamesh> SteveA: one of the issues I noticed was that some of the interfaces I wanted to check for in traversed_objects were not getting added.  This was because only types with registered Navigation classes ran the code path that appended things to traversed_objects
[08:50] <jamesh> I'm not sure if that means the hook needs to be placed elsewhere, or to register an empty navigation class for the interface
[08:56] <lifeless> SteveA: anyone getting pqm errors
[08:56] <lifeless> bah
[08:56] <lifeless> *all*: anyone getting pqm errors please forward them to me
[08:57] <SteveA> jamesh: the hook actually needs to go elsewhere, but i'm surprised it doesn't work in the navigation class
[08:57] <SteveA> which empty Navigation class would be needed?
[08:57] <jamesh> SteveA: it does work in the navigation class
[08:57] <jamesh> SteveA: one of the subpillars mpt listed for products was "milestones"
[08:58] <jamesh> and IMilestone has no Navigation class, so I couldn't check to see if a milestone had been traversed
[08:58] <SteveA> i see
[08:58] <SteveA> can you add a Navigation for IMilestone then, but mark it XXX for me to remove when i've done the other refactoring
[08:58] <jamesh> for the cases where things work, they work quite well
[08:58] <SteveA> that's good
[08:58] <jamesh> okay
[09:01] <jamesh> SteveA: I also ran into one subpillar whose only page is a view too: the "branches" subpillar for products
[09:02] <jamesh> there are content objects for branches, but they sit under the owner person in the URL space
[09:03] <jamesh> It is possible that the subpillar should be removed though, since it duplicates a facet link
[09:04] <SteveA> so, this is a subpillar for a collection of branches
[09:04] <SteveA> that is represented by a page that displays that collection
[09:05] <SteveA> can we get the sitemap landed without that "branches" map item, and add it later if we still need it?
[09:07] <jamesh> sure.  I've got a few minor changes to make, and I'll stick it up on the review page
[09:17] <SteveA> i guess you could make it work for a page, if that turned out to be important
[09:31] <sivang> Morning, my favorite channel :)
[10:01] <salgado> don't we have a meeting now?
[10:01] <lifeless> jamesh: you are about to get a test failure
[10:01] <lifeless> salgado: yes
[10:01] <lifeless> lets DoIt
[10:01] <lifeless> reviews meeting
[10:01] <lifeless> agenda:
[10:01] <lifeless>     *
[10:01] <lifeless>       Roll call
[10:01] <lifeless>     *
[10:01] <lifeless>       Agenda
[10:01] <lifeless>     *
[10:01] <lifeless>       Next meeting
[10:01] <jamesh> lifeless: do you want a copy of the email?
[10:01] <lifeless> wow that pastes bad. I shall fix for the next meeting
[10:01] <lifeless> jamesh: please
[10:02] <lifeless> So, whos a reviewer, or interested, and here ?
[10:02] <jamesh> me
[10:02] <BjornT> i'm here
[10:02] <salgado> me
[10:03] <lifeless> I appear to be
[10:03] <BjornT> lifeless: did you see that i added a proposed item on ReviewerMeetingAgenda?
[10:03] <lifeless> ahha, I needed a reload.
[10:03] <lifeless> yes.
[10:04] <lifeless> one second
[10:05] <lifeless> ok
[10:05] <lifeless> SteveA: are you 'here' ?
[10:05] <lifeless> BjornT: thats on the agenda for today.
[10:06] <BjornT> cool
[10:06] <lifeless> right. so we have jamesh, BjornT, salgado, me.
[10:06] <lifeless> next meeting - same time, one week ?
[10:06] <salgado> yes
[10:06] <BjornT> i won't be here next wednesday, i'll be flying to sweden
[10:06] <jamesh> okay
[10:07] <lifeless> BjornT: thats fine
[10:07] <lifeless> these are best effort anyway
[10:07] <lifeless> ok. BjornT your topoic
[10:07] <BjornT> ok
[10:08] <BjornT> so, the general queue still (since the last meeting) seems to be a place where branches get dumped until some of the reviewers is nagged to review it.
[10:08] <BjornT> that defeats the purpose of the general queue, since you're way better off putting your branch in the reviewer's queue directly, no need to let it rot in the general queue first.
[10:08] <BjornT> i think we either need to get rid of the general queue, or find a way of getting it processed quicker. i would prefer the latter.
[10:08] <BjornT> for example, lifeless, as head of the review team, could daily move branches from the general queue to a suitable reviewer. this would make it more obvious for the reviewer that he has some branches to review.
[10:08] <BjornT> thoughts?
[10:08] <lifeless> I committed to doing what I had percieved kiko as doing, which was directly nagging, but I've not been around *at the right time* as people to be effective at that
[10:09] <lifeless> so your suggestion works for me, if you guys will 'own' your own queues, I'm happy to schedule stuff from general to you.
[10:09] <lifeless> jamesh, salgado - opinions ?
[10:10] <BjornT> i think actually moving the branches to queues will work better than just nagging.
[10:10] <jamesh> lifeless: okay.
[10:10] <salgado> that sounds reasonable, but I can see some cases where we may need to reject (not review) some of these assigned branches
[10:11] <salgado> for instance, if you're on a tight deadline, like I was some time ago with shipit
[10:11] <lifeless> salgado: that will make it explicit which is better than the catch all
[10:11] <lifeless> so lets do two things.
[10:11] <salgado> I don't expect that to happen often, but wanted to raise the point
[10:12] <lifeless> if you are busy, put a /!\ I am busy, queue closed except for prior arrangement
[10:12] <lifeless> in your queue
[10:12] <lifeless> that will tell me, and anyone else, that you are busy.
[10:12] <lifeless> and if everyone is like that, its an obvious problem :)
[10:13] <lifeless> secondly, if you have to reject something, put it in a new queue I'll create called 'needs reassignment' with an explanantion
[10:13] <lifeless> I can imagine several: too busy, feel uncomfortable reviewing the specific code touched etc
[10:14] <lifeless> salgado: does that fit ?
[10:14] <salgado> yes, I think so
[10:14] <lifeless> jamesh, BjornT - objections ? thoughts ?
[10:15] <jamesh> lifeless: sounds okay.  What about the case of branches that a particular reviewer is best qualified to review, but are assigned to someone else?
[10:15] <jamesh> just coordinate with the other reviewer to take over?
[10:16] <lifeless> jamesh: yeah, common sense applies for 'normal' issues.
[10:16] <BjornT> lifeless: sounds good to me
[10:16] <lifeless> please tell me about specialities like that.
[10:16] <lifeless> *both when you reassign*
[10:16] <lifeless> and when you realise they exist.
[10:17] <lifeless> I ask this because one of the things we need as a team is a certain degree of cross training, so that we dont have areas of code suddenly unreviewable when one person is unwell
[10:17] <lifeless> or busy
[10:17] <lifeless> so I may assign outside of the specialities deliberately, for small reviews to built knowledge.
[10:17] <jamesh> sounds sensible
[10:17] <SteveA> lifeless: hello
[10:17] <lifeless> SteveA: hi.
[10:17] <SteveA> what's up?
[10:18] <SteveA> i'll be working on malone stuff with brad
[10:18] <lifeless> SteveA: reviewers meeting, was wondering if you where here for it ;)
[10:18] <SteveA> not really
[10:18] <lifeless> ok, read the minutes then :)
[10:18] <SteveA> thanks
[10:18] <lifeless> ok
[10:19] <lifeless> so thats the last agenda item
[10:19] <lifeless> we have time if theres new business, otherwise doom-it-twill-be
[10:19] <lifeless> 5
[10:19] <lifeless> oh, I'll be allocating items daily
[10:19] <lifeless> as a 'todo item' in my task list
[10:20] <lifeless> 4
[10:21] <lifeless> 3
[10:21] <lifeless> 2
[10:21] <lifeless> 1
[10:21] <lifeless> meeting over, thanks for playing
[10:26] <carlos> morning
[10:30] <daf> morning
[10:32] <jordi> hola
[10:33] <jordi> carlos: so translation uploads is merged, will it show up in the next updatE?
[10:33] <carlos> jordi, is it?
[10:33] <carlos> jordi, I'm fighting with PQM 
[10:34] <bradb> stub: around?
[10:34] <carlos> because it's rejecting the merge
[10:34] <jordi> i dunno
[10:34] <carlos> jordi, If I'm able to get it merged before tomorrow... it should land next week
[11:11] <lifeless> carlos: see topic
[11:13] <carlos> lifeless, well, the rejects come from more than one week ago
[11:13] <carlos> lifeless, most of them conflicts with the other merges
[11:13] <carlos> lifeless, and now, because problems with pqm that I suppose you are fixing now
[11:19] <lifeless> jamesh: ping
[11:23] <SteveA> daf: https://launchpad.net/projects/launchpad, go to the "specifications" menu, and get an oops page!
[11:23] <SteveA> daf: can you add that to your list to look into?
[11:25] <daf> done
[11:25] <SteveA> thanks
[11:25] <SteveA> also
[11:26] <SteveA> on https://launchpad.net/projects/launchpad/+specs, i'm logged in
[11:26] <SteveA> but i see no traceback
[11:26] <daf> noted
[11:33] <jblack> lifeless: ping
[11:33] <lifeless> pong
[11:33] <jblack> > Right. They *should not use rsync to push*. But the place they are
[11:33] <jblack> > specifying is done via *rsync syntax*.
[11:33] <jblack> I'm very confused now. 
[11:33] <lifeless> ok
[11:33] <lifeless> tell me where you are at
[11:33] <jblack> Rsync is dangerous. Its safe. No, its actually dangerous. But its not rsync at all, just rsync syntax.
[11:34] <lifeless> ok
[11:34] <lifeless> two cases
[11:34] <lifeless> push
[11:34] <lifeless> pull
[11:34] <lifeless> many use cases
[11:34] <lifeless> 1 use case is 'making a mirror of something someone else works on'
[11:34] <jblack> safe.
[11:34] <lifeless> rsync is safe for this, but it will TRASH local data. by design.
[11:35] <jblack> Ok. I think I get you now.
[11:35] <jblack> specify that we're using rsync. Here's the places you need to be careful? 
[11:35] <lifeless> another use case 'copy a tree from A to B'. rsync will preserve trees - it wont corrupt or anything. But it will overwrite data without warning.
[11:36] <lifeless> one specific case is 'publishing a branch'. Rsync is particularly bad for this, because its easy to overwrite a remote branch without warning using rsync directly.
[11:36] <lifeless> bzrtools has a rsync-under-the-hood 'push' plugin
[11:36] <jblack> yup. I use it
[11:36] <lifeless> that kicks in when you specify a branch using rsync syntax rather than url syntex.
[11:37] <lifeless> so we want people to use that rsync-using-bzrtools-push plugin
[11:37] <lifeless> rsync syntax; not using rsync directly.
[11:37] <lifeless> the bzrtools plugin is safe because it checks the revision-history first.
[11:38] <lifeless> unconfused now ?
[11:38] <jblack> Yup. clear as a bell.
[11:38] <lifeless> cool
[11:38] <jblack> Thanks
[11:39] <lifeless> jamesh: ping
[11:41] <SteveA> hi stub 
[11:41] <SteveA> i sent you email
[11:42] <SteveA> also, brad and i want to work on the whiteboard stuff  in 4-5 hours
[11:42] <stub> SteveA: The lpnotification branch is still in your review queue
[11:42] <SteveA> oh, poo
[11:42] <SteveA> can you put it into someone else's ?
[11:42] <SteveA> i'm working with brad this week
[11:43] <SteveA> maybe bjorn or jamesh can look at it?
[11:43] <stub> Sure. General or are you designating a victim?
[11:43] <lifeless> stub: does it need someone specific ?
[11:43] <stub> lifeless: Nope.
[11:44] <lifeless> stub: please put it in any empty queue, or in the shortest non empty queue
[11:45] <stub> lifeless: punted to your queue
[11:45] <lifeless> heh. b'stard
[11:49] <ddaa> hey lifeless
[11:49] <ddaa> could you remind me when I can find the bzr branch to make launchpad to like
[11:49] <ddaa> hey stub
[11:50] <stub> Hi
[11:50] <ddaa> did sending you the email to ask for the db access for bzrsyncd was the right way to get it?
[11:51] <stub> Sure. I just was incompetent and forgot.
[11:52] <stub> Bah... can't find the email now :-(
[11:52] <ddaa> Want me to forward it again?
[11:53] <lifeless> stub: can you forward me the pqm failure you had
[11:53] <lifeless> ddaa: /home/warthogs/archives/robertc/bzr/importd
[11:54] <stub> lifeless: I don't recall a PQM failure recently... you thinking of someone else?
[11:54] <lifeless> nope
[11:54] <lifeless> you
[11:54] <jamesh> lifeless: pong
[11:55] <lifeless> jamesh: got that failure yet ?
[11:55] <jamesh> lifeless: no
[11:55] <lifeless> hmm
[11:56] <stub> Can't find bzrsyncd in my mail archive. Please resend.
[11:56] <stub> ddaa: ^^^
[11:56] <stub> I've got roberts emails
[11:57] <ddaa> sent
[11:58] <jamesh> lifeless: the last pqm message I've received was from yesterday
[12:00] <lifeless> stub: you should have had mail from balleny
[12:00] <lifeless> elmo: ping - help needed 
[12:00] <stub> lifeless: Nope. No mail.
[12:01] <lifeless> stub: garh
[12:01] <lifeless> 5/7 pqmjobs tested on balleny have failed
[12:01] <lifeless> and neither you nor jamesh have recieved email
[12:03] <Znarl> lifeless : Can I help you?
[12:03] <lifeless> Znarl: please
[12:03] <lifeless> 5 emails should have been sent from pqm on ballent
[12:03] <lifeless> *balleny*
[12:03] <lifeless> in the last 10 hours
[12:03] <lifeless> but noone has recieved any AFAICT
[12:03] <lifeless> I need to track them down for starters.
[12:04] <stub> ddaa: Haven't seen that email yet. I'm receiving messages from other sources though.
[12:04] <ddaa> ha ... ha...
[12:05] <Znarl> Dec 14 09:55:36 balleny postfix/smtp[30173] : 48088201D1B: to=<stuart@stuartbishop.net>, relay=none, delay=3198, status=deferred (connect to mail4.bur.st[218.214.5.115] : Co)
[12:05] <lifeless> Znarl: thats a good start! thanks
[12:05] <Znarl> Looks like our fault, give me a minute to fix this.
[12:09] <cprov> morning all
[12:09] <lifeless> morning
[12:11] <ddaa> stub: _really_ sent
[12:12] <Znarl> Fixed.
[12:14] <cprov> lifeless: hi, is there anything wrong with PQM ? is is processing a job of mine from yesterday 12:20. Is it blocked ?
[12:16] <matsubara> good morning!
[12:17] <jamesh> cprov: he's been moving it to its own machine
[12:20] <cprov> jamesh: uhm, ok then, PQM has a better home now ...
[12:27] <lifeless> cprov: /topic
[12:27] <lifeless> stub: please forward me that email :)
[12:30] <lifeless> ddaa: ping
[12:30] <ddaa> lifeless: pouet
[12:30] <lifeless> morning
[12:30] <jamesh> lifeless: no pqm mail yet
[12:30] <cprov> lifeless: ok
[12:30] <lifeless> and, to save replying to a bunch of emails, I wanted to chat quickly about private branches
[12:31] <lifeless> we (Mark and I and various others) explicity considered private branches throughout Sao Carlos
[12:31] <lifeless> in that we wanted graceful degradation when branches aren't accessible, such as being behind a firewall
[12:32] <lifeless> but we did not want a 'checkbox to make it private' as that really suggests the wrong thing to users
[12:32] <lifeless> thats not explicitly stated as we spent a huge amount of time talking branches over 5 different ways, all the way through to the last week
[12:33] <lifeless> so only the end result really got written up - sorry.
[12:34] <ddaa> lifeless: I'm listening
[12:34] <lifeless> ddaa: thats the key
[12:35] <lifeless> I'm sure that raises questions and corollaries for you
[12:35] <mpt> Goooooooooooooooooooooooooooooooooood morning Launchpadders
[12:35] <lifeless> so lets talk those through
[12:37] <lifeless> stub: jamesh: Znarl is kicking the mail queue
[12:37] <lifeless> as soon as you get a pqm error please forward it to me
[12:38] <ddaa> lifeless: I'm still listening.
[12:39] <lifeless> ddaa: I've said my bit
[12:39] <salgado> lifeless, do you want my failure message too?
[12:39] <lifeless> salgado: that would be great
[12:39] <salgado> sent
[12:39] <lifeless> thanks
[12:40] <ddaa> okay, first thing (in order in the mail) is branch pulling
[12:41] <lifeless> Znarl: ping
[12:41] <Znarl> lifeless : Yep?
[12:41] <lifeless> stub: * Installing PL/PythonU
[12:41] <lifeless> createlang: language installation failed: ERROR:  could not access file "$libdir/plpython": No such file or directory
[12:41] <ddaa> there are some provisions for error reporting, though I consider them incomplete (I would like something like access to error logs through the librarian)
[12:42] <lifeless> Znarl: thats a missing launchpad dependency on balleny
[12:42] <ddaa> but in the short term they are just not implemented
[12:42] <lifeless> Znarl: probably I missed saying 'and the DatabaseSetup wiki pages requirements
[12:42] <ddaa> short term, I think any error in jblack cronscript should send error spam to launchpad-errors
[12:42] <lifeless> Znarl: can we get that pages pre-reqs in the base os, and in both chroots ?
[12:42] <lifeless> the base is of immediate importance
[12:42] <ddaa> then, switch to something more "compact" when it's implemented
[12:43] <Znarl> lifeless : ok.
[12:43] <ddaa> lifeless: is there a disagreement on that?
[12:43] <lifeless> Znarl: now if possible, as I'm trying to get pqm moved:)
[12:43] <lifeless> ddaa: thats fine with me, but that does not prevent using lp for our branches w/ james's script.
[12:44] <lifeless> it just means spam.
[12:44] <ddaa> that's not right for me as I'm concerned that this spam could make the error reporting unusable. But let's worry about this issue when it becomes a practical issue.
[12:45] <lifeless> right
[12:45] <lifeless> if anything I think that that is just motivation to get the error reporting implemented
[12:45] <ddaa> motivation is not the critical resource, time is
[12:46] <ddaa> Let's move on.
[12:47] <carlos> grrr the poor I/O performance is killing my laptop
[12:47] <ddaa> I do not think it would appropriate to have a branch page that reads "0 revisions" for private branches. Instead it should read "private branch" or something that clearly means that no data is expected to appear on that page, ever.
[12:47] <ddaa> But I guess you would call that "polish".
[12:48] <carlos> lifeless, I got an error from PQM that I think is related to its setup, do you want it to debug the new installation?
[12:48] <ddaa> lifeless: do we have agreement on that issue?
[12:48] <carlos> it comes from an old request I sent yesterday
[12:48] <lifeless> one sec, EADMIN
[12:50] <lifeless> ddaa: I think that that is a separate issue.
[12:51] <lifeless> I think that that revolves around whether launchpad should 'understand' having private branches, or just 'accomodate' them
[12:51] <lifeless> we took 'accomodation' as the strategy in brazil.
[12:51] <lifeless> carlos: no
[12:51] <carlos> ok
[12:51] <ddaa> I think that will lead to confused people. But I guess it's not considered important.
[12:52] <lifeless> ddaa: I think there are other ways of showing it
[12:52] <lifeless> i.e. 'no reading succeeded in the last year'
[12:52] <ddaa> that can means essentially one of two things:
[12:52] <ddaa> 1. the branch is private, nothing to do
[12:53] <jamesh> lifeless: mailed you the PQM failure message: createlang: language installation failed: ERROR:  could not access file "$libdir/plpython": No such file or directory
[12:53] <ddaa> 2. there's a typo in the URL, the owner should be notified
[12:53] <ddaa> 3. there's a bug somewhere, the admin should be notified
[12:53] <lifeless> jamesh: thanks
[12:53] <lifeless> ddaa: exceptions that we do not understand -> admin.
[12:53] <jamesh> lifeless: should I resubmit, or wait til things are up and running again?
[12:54] <lifeless> failures that we report to the system (see previous discussion) -> user via the branch message thread & error status
[12:54] <ddaa> I'm happy get admin notifications for all errors, as the amount of typoed URL should stay quite low.
[12:54] <lifeless> ddaa: you are trying to force this towards explicit statements about private branches
[12:54] <ddaa> I'm not trying to force anything.
[12:55] <ddaa> I'm exposing the issues I see.
[12:55] <ddaa> And I have not noticed a proposal that would address them.
[12:55] <lifeless> I think directly mailing you on every branch that fails to mirror would be counterproductive
[12:56] <lifeless> I'm not trying to address them as you present them, I'm trying to put them into the context we looked at them in s.c., where they matter less.
[12:56] <lifeless> jamesh: waut
[12:56] <lifeless> jamesh: wait
[12:57] <ddaa> lifeless: does that mean that those issue will to be considered under another perspective? That would muchly reduce the scope of this discussion.
[12:58] <lifeless> ddaa: so, case 1 if lp accomodates it is simply that users solve connection failed errors themselves, and admins only deal with exceptions - new forms of errors
[01:00] <ddaa> so, everybody that registers a private branch should be nagged to death about expected pull failures?
[01:00] <ddaa> I guess that's not what you mean.
[01:00] <lifeless> no
[01:00] <lifeless> you only need to nag on transitions
[01:00] <lifeless> pass -> fail
[01:00] <lifeless> fail -> pass
[01:01] <lifeless> so right at the start, they will get 'mirroring failed' once, and never get a 'service restored'
[01:03] <ddaa> Does that mean that pull attempts will be performed, and fail, after the initial "mirrorring failed", using e.g. the exponential delay strategy, without producing any feedback?
[01:04] <ddaa> and without requiring any user intervention
[01:04] <lifeless> that was the concept
[01:04] <ddaa> Mh... okay, that bit is starting to make sense.
[01:04] <lifeless> there will be feedback on the error status on the branch object
[01:04] <lifeless> but no emails need be sent
[01:05] <ddaa> So, on the broader issue:
[01:05] <stub> Anyone know how to switch on my tv out?
[01:05] <ddaa> the choice is to provide exactly as much private branc support that is needed for Launchpad development?
[01:06] <ddaa> And not be bothered about the fact that disclosure policy of Launchpad is quite unique, at least in my experience.
[01:07] <lifeless> ddaa: right
[01:07] <lifeless> ddaa: we can only usefully design for the use cases direct clients of the programmers have
[01:07] <lifeless> anything else is speculation
[01:07] <lifeless> smart speculation, but speculation
[01:07] <ddaa> that's why I suggested getting requirements from the business people, now that we have some.
[01:08] <lifeless> did that, 'mark', in sao carlos
[01:08] <ddaa> okay
[01:09] <ddaa> so, the final thing is that I think your argument about "private sftp is fine" is a bit at odd with reality...
[01:09] <ddaa> but I guess that would be classified as pedantry.
[01:09] <lifeless> anonymous sftp
[01:09] <ddaa> yes, anonymous sftp
[01:09] <lifeless> not private sftp
[01:09] <lifeless> anonymous sftp is sane and meaningful
[01:09] <ddaa> is that actually used out there?
[01:09] <lifeless> same as anonymous ssh is useful for some cvs servers these days
[01:10] <lifeless> and yes, pedantry
[01:11] <lifeless> ok the last 7 failed branches are retrying now
[01:11] <lifeless> that should take use through to my morning
[01:11] <lifeless> I'm going to try and stay up for this
[01:12] <salgado> cprov, who's supposed to know what are the titles and descriptions of the dbschema items that are needed by MirrorManagement
[01:12] <salgado> ?
[01:13] <lifeless> tomorrow I'll be working on performance improvements.
[01:13] <ddaa> lifeless: do we have a spec out there about the error reporting logic we just discussed?
[01:13] <lifeless> but I'm hitting 'beat' now.
[01:13] <lifeless> ddaa: I thought ...Telemetry was about it
[01:13] <lifeless> ddaa: but we can do one up.
[01:13] <cprov> salgado: Kinnison, me, you ;) are you about to start the implementation ?
[01:14] <ddaa> lifeless: my immediate concern is that I think that deserves at least one (probably several) entries in in the BazaarTaskList.
[01:14] <lifeless> ddaa: will it work with the current stuff being rolled out? however verbosely and non clearly ?
[01:14] <ddaa> The immediate issue (cannot use sftp URLs) does not, but the future new and improved error reporting does.
[01:15] <ddaa> lifeless: I have no visibility on the branch pulling code.
[01:15] <salgado> cprov, yes, I already started. would you list all dbschema classes with their respective items (including the title and description for each item)?
[01:15] <lifeless> ddaa: yes, but you do on the api it needs (exposed via xml-rpc in theory. We need a service for it soon!)
[01:16] <ddaa> lifeless: I do not see any reason for it not to work once the urls are in the database.
[01:16] <lifeless> ddaa: ok
[01:16] <lifeless> ddaa: so immediate thing, is to add a 'support sftp://' registration task.
[01:17] <ddaa> But getting them there is going to require a small, but non-trivial amount of work.
[01:17] <ddaa> Not something I can comfortably promise before the end of the week.
[01:17] <lifeless> ddaa: and there should already have been one on the error reporting in the registry from the supermirror, so there should be no task list changes there
[01:17] <lifeless> ddaa: thats ok, perhaps I or jamesh can do that. surely its easy to s/http/http|sftp/
[01:18] <ddaa> not that easy
[01:18] <ddaa> I expect that will requier a new pl-python procedure and new vocabulary. But maybe the required bits are already there.
[01:19] <ddaa> I'm not sure if vocabulary is the right term though...
[01:19] <ddaa> anyway, not a big task for somebody familiar with launchpad internals.
[01:22] <cprov> salgado: adding them to the spec
[01:23] <lifeless> ddaa: well, that should be added to - because regardless of form, the SM needs to be able to tell LP that its succeeded or failed to mirror
[01:23] <salgado> cprov, great. thank you!
[01:23] <lifeless> night all
[01:24] <ddaa> sure, but the task list is only for well specced task that I'm confident can be implemented in the order of one week.
[01:24] <ddaa> so far, this error reporting thing is quite fuzzy to me
[01:26] <lifeless> ok
[01:26] <lifeless> well its not immediate any which way
[01:41] <daf> mpt: did you get my mail?
[01:50] <daf> jamesh: still around?
[01:57] <salgado> cprov, did you finish editting the spec? I have some questions that I want to add there
[01:58] <salgado> lifeless, should I send my merge request again or pqm isn't ready yet?
[01:58] <cprov> salgado: not yet, but i can save now 
[01:59] <salgado> cprov, no need to, I can wait
[02:00] <cprov> salgado: too late, it's already saved 
[02:01] <mpt> daf, I got your question that is answered by <https://wiki.launchpad.canonical.com/LaunchpadCapitalization>, which I'm implementing right now, but I did not get your list of headings you want changed
[02:01] <jamesh> daf: yeah
[02:01] <jamesh> for a bit
[02:03] <daf> jamesh: ok, I just want to know what you think about my comment to bug 5461
[02:03] <Ubugtu> Malone bug #5461: redirects should canonicalise URLs In: launchpad (upstream), Severity: Normal, Assigned to: James Henstridge, Status: Fixed https://launchpad.net/bugs/5461
[02:03] <daf> jamesh: if you think it's a good idea, I might hold off on landing my bug-2230 branch
[02:05] <SteveA> daf: i think we should leave the status as it is
[02:05] <SteveA> daf: if you're using the status field, then you need to understand what you're doing
[02:06] <jamesh> daf: I don't think it's that big a deal.  There are about 3 codes that might be appropriate: 301 (Moved Permanently), 302 (Found) and 307 (Temporary Redirect)
[02:06] <daf> ok, in that case I'll land it
[02:06] <daf> thanks
[02:11] <carlos> see you later
[02:12] <salgado> Kinnison, cprov, I added some questions to the MirrorManagement spec. would one of you guys answer then for me?
[02:16] <Kinnison> salgado: I can do my best
[02:19] <mpt> bradb, ping
[02:20] <Nafallo> -ENOKIKO :-P
[02:20] <Nafallo> can someone add a sourcepackage to dapper for me? :-)
[02:20] <Nafallo> linuxdcpp
[02:22] <daf> mpt: thanks, that page answers my immediate question
[02:22] <daf> mpt: I haven't compiled that list yet
[02:22] <daf> mpt: I'll send it to you as soon as I have
[02:23] <salgado> Kinnison, great. thank you. :)
[02:23] <Kinnison> salgado: cprov will try and help too
[02:23] <cprov> salgado: I'm on it 
[02:24] <LarstiQ> I see the rollout has made it, very nice!
[02:24] <Kinnison> cprov: star
[02:25] <mpt> daf: good good
[02:28] <mpt> jamesh, what can I do with an OOPS-CODE?
[02:28] <mpt> where do I look it up?
[02:34] <bradb> mpt: pong
[02:35] <mpt> bradb, about bug 977
[02:35] <Ubugtu> Malone bug #977: Commenting on bug should add you to the CC list In: malone (upstream), Severity: Normal, Assigned to: Brad Bollenbach, Status: Accepted https://launchpad.net/bugs/977
[02:36] <mpt> would you be ok with that being adjusted to "... unless you're already subscribed to the product/package/whatever"?
[02:36] <SteveA> mpt: for production, chinstrap:/srv/gangotri-logs/2005-12-14
[02:36] <SteveA> that's today's oopses
[02:36] <mpt> bradb, because I don't want to get two copies of every bugmail just from being subscribed to the launchpad project (for example)
[02:36] <mpt> thanks SteveA 
[02:36] <SteveA> the oops filename has a number, related to the time it occured
[02:36] <SteveA> then a dot
[02:37] <SteveA> and then the OOPS code
[02:37] <SteveA> for staging, there is a parallel directory structure
[02:37] <SteveA> there is an rsync job running every 10 minutes or so
[02:37] <SteveA> that copies recent oops reports onto chinstrap
[02:38] <bradb> mpt: That seems pretty reasonable. I'm not in the right mode to think about how to clearly communicate that in the UI, but suggestions welcome.
[02:41] <SteveA> mpt, brad: why should this matter?
[02:42] <SteveA> surely we're doing a set() operation on recipients?
[02:42] <SteveA> so we send one mail out to each
[02:42] <mpt> oh, hmm
[02:42] <mpt> that's true
[02:43] <bradb> SteveA: It matters when the team has an email address.
[02:44] <bradb> If that's not the case, then yes, it'll be set()ified into just one email being sent to that person
[02:44] <SteveA> you mean when the team has an email address that is a list?
[02:44] <bradb> yeah
[02:45] <mpt> that's a catch-21
[02:45] <SteveA> that's not our problem, really
[02:45] <SteveA> if someone is subscribed to a list, and also individually
[02:45] <mpt> because once WhateverSubscriptions are implemented, teams won't *need* to have e-mail addresses that are lists
[02:45] <SteveA> that's the same as with anything else
[02:45] <SteveA> so, no problem
[02:46] <bradb> SteveA: It could be our problem. A few people at UBZ mentioned not wanting to get duplicate bugmail in this scenario.
[02:46] <bradb> The UI could be smart enough to say "you're already getting notifications about changes to this bug"
[02:47] <SteveA> let's wait until all these changes land
[02:47] <SteveA> and then see
[02:47] <SteveA> what do you think mpt?
[02:48] <bradb> lifeless: Can you add:
[02:48] <bradb>   sourcecode/bzrtools
[02:48] <bradb>   sourcecode/testresources
[02:48] <bradb> to .bzrignore in rocketfuel?
[02:48] <lifeless> no
[02:48] <lifeless> we should not be ignoring subtrees
[02:48] <lifeless> that will interact badly as subtree support comes onine
[02:49] <mpt> SteveA, I agree that it's not our problem, but we're causing the problem as long as WhateverSubscriptions aren't implemented
[02:49] <mpt> (and last I checked it wasn't specced yet)
[02:49] <SteveA> what exactly is WhateverSubscriptions?
[02:49] <lifeless> this is because the behaviour for current revision will change with upcoming bzrs
[02:50] <bradb> lifeless: hm, ok
[02:50] <SteveA> lifeless: is there any sense in ignoring now, and unignoring when we bzr improves to get subtree support?
[02:50] <mpt> SteveA, being able to subscribe to a product, or a package, or a project, or a distro, or a distro release
[02:51] <bradb> mpt: InitialBugContacts is step one of that.
[02:51] <mpt> indeed it is
[02:51] <SteveA> then i think we're fine.  IBC will land soon.
[02:52] <SteveA> there is one outstanding issue for kiko and me and brad to sort out
[02:52] <lifeless> SteveA: I really dont think thats a good idea
[02:53] <mpt> SteveA, for example, I'm subscribed to launchpad-bugs@ and get two copies of bugs that I'm explicitly subscribed to. If I could subscribe to the Launchpad project instead, launchpad-bugs@ would no longer be necessary, and I would no longer get duplicate bugmails.
[02:53] <mpt> (two copies of bugmails for bugs, rather)
[02:53] <salgado> cprov, dude, variable names can't start with a number, can you fix the items of MirrorSpeed while you're answering those questions?
[02:54] <cprov> salgado: ok ... crying  child right ONE instead on 1 ...
[02:55] <cprov> becomes funny for 128 ;)
[02:55] <lifeless> stub: around ?
[02:55] <stub> lifeless: yes
[02:55] <lifeless> error with newpqm
[02:55] <Kinnison> cprov: S1K S1M S256K
[02:56] <lifeless> forwarded to you, I think its pgsql related  - but its definately an area you have fiddled heaavily with
[02:56] <lifeless> can you check it out ?
[02:56] <salgado> cprov, did you rename DistributionMirror.mirrortype to DistributionMirror.content?
[02:56] <cprov> Kinnison: ok  and clarity is somewhere far from here 
[02:57] <cprov> salgado: thinking about it, just now 
[02:57] <Kinnison> MirrorSpeed.ms128k
[02:57] <Kinnison> I know hungarian sucks, but it solves the issue
[02:57] <stub> lifeless: OperationalError: FATAL:  Ident authentication failed for user "importd"
[02:58] <stub> I thought my wildcard entries would be fine :-(
[02:58] <stub> I'll fix it in a mo
[02:58] <lifeless> ah, I missed that line. that makes it clear though :)
[02:58] <salgado> cprov, well, the spec doesn't have mirrortype anymore, and it had 30 minutes ago
[02:58] <cprov> salgado: did you already create the tables ?
[03:00] <salgado> cprov, yes
[03:00] <salgado> tables, database classes, interfaces
[03:01] <cprov> salgado: if yes, I can change they back, but I think 'content' is better concept then 'type' ... you can rename it now or leave it to the future :(
[03:01] <salgado> so, please let me know if you change anything in the spec, as it's approved and I already started implementing
[03:01] <salgado> I agree with you content is a better name, I just want to know about the changes you make
[03:02] <salgado> so I can keep the code consistent with the spec
[03:04] <salgado> cprov, I have no clue as to what the items in MirrorStatus should be, and the spec doesn't mention anything related to possible statuses
[03:05] <salgado> cprov, same for the dbschema used in MirrorDistroArchRelease.pocket
[03:06] <cprov> the changes are kind of last minutes, and I've just saved because you need to add your questions
[03:06] <salgado> oh, okay. sorry then
[03:07] <cprov> salgado: pocket is PackagePublishingPocket 
[03:08] <cprov> salgado: MirrorStatus is comming, as your unresolved issues,  just one min
[03:10] <salgado> cprov, another question: why is there a status column in both MirrorDistroArchRelease and MirrorDistroReleaseSource, if we already have that column in DistributionMirror?
[03:12] <cprov> salgado: DRM has no status IIRC.
[03:12] <cprov> salgado: DM, sorry
[03:13] <salgado> that's right. my mistake
[03:14] <cprov> salgado: MDAR and MDRS have status because we control them separately
[03:15] <salgado> cprov, MirrorProbeRecord.datetime is not a very good name. do you have any suggestion?
[03:15] <cprov> datecreated, I think
[03:15] <cprov> salgado: if you agree I can fix it from my draft
[03:16] <salgado> datecreated is okay, I guess
[03:16] <cprov> fine
[03:20] <Kinnison> datecreated is the launchpad-style :-)
[03:21] <SteveA> although, it is usually a date and time
[03:21] <SteveA> so "whencreated" would be more accurate.  nonetheless, we use "datecreated"
[03:23] <stub> lifeless: postgres auth on balleny should now be fine
[03:23] <cprov> simply "created" sounds better, but as you mentioned LP-style is "datecreated", which is ok too
[03:23] <SteveA> "created" sounds like a boolean
[03:24] <Kinnison> SteveA: If only we'd bothered putting ?s on the end of boolean column names, etc
[03:24] <cprov> SteveA: xiiii, yes .. 
[03:24] <SteveA> created_question_mark ?
[03:25] <SteveA> like that band, therapy_question_mark
[03:25] <Kinnison> SteveA: More like FooBarRecord."created?"
[03:26] <cprov> sorry, MM spec is up
[03:28] <SteveA> Kinnison: is_created or was_created is a common pattern
[03:33] <Kinnison> SteveA: boo hiss
[03:33] <Kinnison> FooBarRecord.sqlbCreated
[04:32] <SteveA> key
[04:32] <kiko> hey there
[04:45] <carlos> hmm
[04:45] <carlos> the new pqm machine is missing the unzip command
[04:45] <carlos> lifeless, ?
[04:46] <carlos> my merge failed (again) because that
[04:47] <carlos> grr
[04:48] <lifeless> Znarl: ^^^
[04:48] <SteveA> carlos: oh?
[04:48] <SteveA> should unzip be in the launchpad dependencies package then?
[04:48] <lifeless> that should be documented somewhere yes
[04:48] <SteveA> we should be using the launchpad dependencies package for this stuff
[04:48] <lifeless> right.
[04:48] <SteveA> not relying on docs
[04:48] <SteveA> but on the package
[04:48] <lifeless> anyhoo, it seems its all a matter of attracting admin attention now
[04:49] <SteveA> what about the package
[04:49] <SteveA> ?
[04:49] <Znarl> lifeless : OK, adding.
[04:49] <lifeless> SteveA: I'm worried about pqm now the package per se
[04:49] <lifeless> Znarl: I'm crashing, its 3am
[04:49] <kiko> balleny huh?
[04:49] <lifeless> I've baby sat enough I think
[04:49] <kiko> congratulations lifeless 
[04:49] <SteveA> the package should be in malone, for bugtracking it
[04:49] <kiko> indeed it should
[04:49] <SteveA> it should also be in the launchpad project in launchpad
[04:49] <SteveA> so that we can file bugs on it
[04:50] <kiko> I'll do that later today.
[04:50] <carlos> SteveA, https://chinstrap.ubuntu.com/~dsilvers/paste/fileFsioey.html
[04:50] <SteveA> thanks kiko
[04:50] <lifeless> Znarl: can you please honour any like requests that come up tonight?
[04:50] <Znarl> lifeless : Ok, I'll keep an eye out for more complaints.
[04:50] <lifeless> Znarl: stevea or kiko should vet anything
[04:50] <kiko> lifeless, I vet you going to bed!
[04:50] <lifeless> but if they say 'aye' then stash it in the chroots and base too please.
[04:50] <lifeless> Znarl: thanks a million.
[04:50] <lifeless> kiko: :)
[04:50] <lifeless> kiko: its happening, ELYNNE for starters.
[04:50] <carlos> SteveA, is not launchpad directly but hct
[04:51] <Znarl> lifeless : OK, I'll run requests past stevea or kiko if I'm unsure of anything.
[04:51] <kiko> okay now that you smiled I'll take that back
[04:51] <SteveA> we need linux-31337rutekit on balleny please
[04:51] <lifeless> Znarl: sweet, thanks
[04:51] <kiko> LOL
[04:51] <kiko> SteveA, who owns launchpad-dependencies?
[04:51] <SteveA> mdz
[04:51] <Znarl> ...or maybe just kiko, not stevea.
[04:52] <kiko> yeah, who is that impostor SteveA anyway
[04:52] <kiko> I only trust it's him if he says "root kit" in lithuanian!
[05:00] <daf> strange: I submitted a request to PQM, and it seems the mail was delivered, but it hasn't landed and it's not still in the queue
[05:01] <kiko> daf, I think lifeless has moved it to a new box -- can you retry?
[05:01] <kiko> I don't know if pqm.ubuntu.com has been updates
[05:01] <daf> ah
[05:02] <kiko> updated
[05:02] <daf> my next door neighbour wants to talk to me about using Rosetta for PledgeBank, so I'm going to go over for a cup of tea
[05:19] <carlos> daf, dude, which kind of neighbours do you have? 
[05:19] <mpt> In what way, cyberix?
[05:19] <carlos> ;-)
[05:19] <cyberix> I'd like to write a feature request for Metacity. Can I do this somewhere in Launchpad?
[05:20] <mpt> cyberix, not really
[05:20] <cyberix> Ok. So I'm not really blind. Just in the wrong palce.
[05:20] <kiko> mpt, he could write a distro spec on that.
[05:20] <mpt> cyberix, Launchpad doesn't host feature specifications itself. That's usually done on a wiki.
[05:20] <carlos> I need to go out for two - three hours, see you later!
[05:21] <cyberix> Well, I could use 1EUR and create a bounty ;-P
[05:21] <mpt> cyberix: Once you have the spec written up somewhere you could link it to the Metacity product (or the Ubuntu distribution, as kiko suggested)
[05:21] <kiko> right.
[05:21] <mpt> but the former won't be useful if the Metacity maintainers aren't using Launchpad :-)
[05:22] <carlos> mpt, well, kiko pointed to a distro spec
[05:22] <kiko> right
[05:22] <kiko> so he could use the Ubuntu wiki.
[05:22] <cyberix> There is no "social machine" to throw the bugs forward?
[05:22] <carlos> I suppose the metacity maintainer would point upstream to that spec ;-)
[05:22] <kiko> cyberix, oh, there is!
[05:22] <kiko> cyberix, you can point bugs to an upstream product as well
[05:22] <kiko> and you can link bug reports
[05:24] <mpt> cyberix, there's no way of automatically reporting the bug in the upstream bug tracker, if that's what you mean
[05:24] <mpt> and bugs are not the same thing as specifications :-)
[05:24] <kiko> social machine can mean many things :)
[05:25] <mpt> It bothers me that specifications and bug reports are separate things
[05:25] <cyberix> "Pay for a feature/bugfix"-option in help menu. Along "translate"
[05:26] <mpt> and that bug reports and bounties are separate things, even
[05:26] <kiko> I see
[05:26] <cyberix> the translate-option is a great step towards community work.
[05:27] <mpt> hmm, bugtrackers-portlet-details.pt is smelly
[05:27] <kiko> it smells of cheese
[05:27] <mpt> yes
[05:27] <cyberix> In future I'm hoping to see similar ways of supporting, by giving money. For those who don't have the time or knownledge.
[05:28] <LarstiQ> cyberix: didn't Znarl already fix that?
[05:29] <cyberix> Znarl?
[05:30] <LarstiQ> argh
[05:30] <LarstiQ> s/cyberix/cprov/
[05:30] <LarstiQ> cyberix: tabcompletion snafu
[05:30] <cyberix> ok
[05:31] <cyberix> The pay system should not be only easy, but almost "too easy".
[05:31] <cprov> LarstiQ: AFAICS, not yet :(
[05:31] <cyberix> Like the software that recommends you stuff at Amazon.
[05:31] <LarstiQ> cprov: hmm, thought he did
[05:31] <Znarl> cprov : I thought I did too.
[05:31] <cyberix> So that every teenager starts to support Ubuntu :-)
[05:31] <LarstiQ> cyberix: like, money should roll out of your pocket by itself? ;)
[05:32] <cyberix> LarstiQ: You got the idea.
[05:33] <cyberix> The International money transfers are already a very big problem.
[05:33] <mpt> Deposit $20 in your Launchpad account, then offer 50 cents for this bug, 50 cents for that one, 50 cents for the other ...
[05:33] <cyberix> mpt: Something like that.
[05:33] <cprov> Znarl: ehe, It could be only effective in the n+1 job ...  let's see what jamesh's job get
[05:33] <SteveA> 50 cent bugs?  lends a whole new meaning to drive by debugging.
[05:33] <cyberix> SteveA: Ofcourse multiple users could donate to a bug they think is important.
[05:35] <cyberix> I might pay 5 or even 10 euros (if it was very easy) to someone who fixed the one bug in Gnome-terminal that irritates me most.
[05:35] <Hieronymus> xkeyboard-config has no translations, but there are translations in the source. How can I upload them?
[05:35] <cyberix> mpt: Ofcourse the ease of deposit is also a problem.
[05:36] <cyberix> mpt: Teens don't have credit cards.
[05:36] <kiko> cyberix, why don't you offer the bounty on irc.gnome.org, on #gnome or some other related channel?
[05:36] <mpt> I didn't get a credit card until I was 27
[05:36] <cyberix> kiko: Because it is not this one case that matters
[05:36] <kiko> (mpt, that's not something you need to go about saying in public)
[05:36] <cyberix> kiko: It is the general ease of Supporting Ubuntu by paying.
[05:37] <cyberix> kiko: And free software in general
[05:37] <kiko> cyberix, I understand
[05:37] <salgado> cprov, what do you think of renaming DistributionMirror.primary* to DistributionMirror.official*?
[05:37] <cprov> can someone help me to sort the last PQM failute -> https://chinstrap.ubuntu.com/~dsilvers/paste/fileHXlOYa.html
[05:38] <cyberix> kiko: I might still ask about the bug there. So thanks for the address. Didn't know they have own irc-network
[05:38] <cprov> salgado: not so good, because these names do not reflect what those fields are 
[05:39] <kiko> cyberix, yeah, hey all hang out in here
[05:39] <salgado> cprov, really? what those fields are then? 
[05:39] <cprov> salgado: "primarycandidate" explicit means "This mirror is official CANDIDATE and it can be tweaked by the onwer" ... doesn't it
[05:40] <kiko> cprov, the error is known, just retry -- Znarl has install zip now.
[05:41] <cprov> salgado: "primaryapproved" mean some LP-ENTITY has approved it 
[05:41] <cprov> kiko: ok, thx
[05:41] <cprov> salgado: "official" state is the AND combination of both
[05:41] <cprov> salgado: is it clear ?
[05:43] <cprov> salgado: even a good/complete mirror could not be "official", if the onwer doesn't agree with it  
[05:43] <kiko> Znarl, could you add deb-src lines for all releases and pockets in drescher's sources.list, and re-run apt-get update?
[05:43] <salgado> cprov, yes, I understood how to check if a mirror is official or not. my concern was that the names primarycandidate and primaryapproved are far from good
[05:44] <SteveA> salgado: is there a way to edit wiki names in launchpad?
[05:44] <mpt> SteveA, what was the better name you came with for "Translation Perms"?
[05:44] <salgado> SteveA, yes, /people/$person/+editwikinames
[05:44] <SteveA> salgado: is it in a menu?
[05:44] <SteveA> i couldn't find it
[05:44] <Znarl> kiko : Can you create an RT request please?
[05:45] <cprov> salgado: maybe official[candidate,approved]  would make more sense associated with "official" property
[05:45] <SteveA> mpt: remind me what they mean... it's about whether the translations need to be done by some team, or whether anyone can do them?
[05:45] <kiko> Znarl, I needed that right away or not at all :)
[05:45] <mpt> SteveA, yes
[05:45] <SteveA> so, it is the policy for accepting translations.
[05:45] <Znarl> kiko : Right away is fine, but I still need you to create an RT request.
[05:46] <kiko> okay, cool.
[05:46] <kiko> I just don't want to bother you if you are too immediately busy 
[05:46] <salgado> SteveA, no, it's in the portlet below the menu
[05:46] <SteveA> i see
[05:46] <Znarl> kiko : Elmo makes me dust the data centre racks if I do work without RT requests being created.
[05:46] <SteveA> it's "below the fold"
[05:47] <SteveA> and i didn't notice it
[05:47] <SteveA> salgado: with the new layout, it would be better on the LHS
[05:47] <SteveA> because the actions menu is so long now
[05:48] <kiko> Znarl, and I can't disagree with him. rt issue filed.
[05:49] <SteveA> salgado: want me to file a bug for that, or can you just do it?
[05:49] <SteveA> or, brad can do it on his vilnius-fixes branch
[05:50] <daf> Launchpad doesn't see mto be generating tracebacks
[05:50] <SteveA> salgado: i want to encourage you to ask either me, or mpt, about UI issues
[05:50] <daf> how long until James' error reporting stuff lands?
[05:50] <SteveA> salgado: mpt and i should be talking enough so that we're consistent
[05:50] <SteveA> daf: landed already
[05:51] <kiko> daf, can I hear a recording of your fiddling?
[05:51] <kiko> I didn't think you had a musical bent
[05:51] <SteveA> see chinstrap:/srv/gangotri-logs
[05:51] <daf> !?
[05:51] <SteveA> then by date
[05:51] <SteveA> filenames inside there are time-in-day.oopsid
[05:51] <stub> daf: I think it landed today
[05:52] <SteveA> but, developers should still be seeing TBs when logged in
[05:52] <SteveA> so, there's a bug there
[05:52] <daf> indeed
[05:52] <daf> I am logged in
[05:52] <salgado> SteveA, right, so what should I do with all those links that are on that portlet? I think placing then on the overview menu won't be good, as it'd make us have almost 20 links there
[05:52] <SteveA> salgado: we should just put that "edit" portlet on the left, not on the right
[05:53] <kiko> it needs balance
[05:53] <SteveA> then it would be much more visible
[05:53] <salgado> SteveA, but then, shouldn't we move the 'Edit GPG Keys' and all others to that portlet?
[05:53] <salgado> I mean, all others that are related to edit something related to that person
[05:54] <SteveA> that would add 7 items to the menu box
[05:54] <SteveA> giving a total of 15
[05:54] <SteveA> that may be too many.  try it out with the box on the right, and see how that works for you
[05:54] <SteveA> that's a simple change, and easy to make
[05:55] <SteveA> maybe show that screen to someone like johan, and ask him to change the wiki name.  see if he sees how. 
[05:55] <salgado> actually it'll add only 6
[05:55] <SteveA> if that isn't obvious to him, then we should move it into the menus
[05:56] <SteveA> salgado: i see... gpg and ssh keys are repeated
[05:56] <SteveA> so maybe only 5
[05:56] <SteveA> yeah, it adds 5
[05:57] <daf> kiko: fiddling?
[05:57] <salgado> yes, only 5
[05:57] <kiko> daf, I'm just chuckling at your activity report ;)
[05:57] <SteveA> salgado: okay, do it.  into the menu!
[05:58] <daf> kiko: heh, oops
[06:02] <kiko> BjornT_, excellent work on the gpg debugging, way to go
[06:06] <salgado> does pqm's email address changed?
[06:07] <kiko> that's not english!
[06:07] <kiko> cprov, I think we should try and fix the publisher ourselves.
[06:08] <salgado> the email address I should use to request a merge changed?
[06:09] <cprov> kiko: you're looking very optimistic today
[06:09] <cprov> kiko: I'm won't be afraid to face it if you have a plan ;)
[06:10] <kiko> cprov, can we try first just adding a friggin DAR?
[06:10] <kiko> otherwise I'll look at the code with you
[06:11] <cprov> kiko: ok, creating an empty DAR for breezy-autotest 
[06:20] <kiko> matsubara, bug 5788 is a dupe of the wrap-urls-bug
[06:20] <Ubugtu> Malone bug #5788: Branch URLs can easily be too wide for new layout In: launchpad (upstream), Severity: Normal, Assigned to: Nobody, Status: New https://launchpad.net/bugs/5788
[06:23] <mpt> kiko, marked it
[06:24] <Kinnison> I forgot to unsuspend workrave this morning, didn't I?
[06:25] <SteveA> stub: can i get that whiteboard report tomorrow sometime?
[06:30] <salgado> SteveA, any idea why the "Overview" menu is collapsed when you go to the "+edithackergotchi page? (the same happens for the +editemblem, in case of teams)
[06:35] <Ubugtu> Malone bug #5787: wrong menus on edit hackergotchi page In: launchpad (upstream), Severity: Normal, Assigned to: Dafydd Harries, Status: New https://launchpad.net/bugs/5787
[06:36] <SteveA> right
[06:36] <SteveA> see that bug
[06:36] <SteveA> probably a facet zcml thing now registered
[06:37] <salgado> apparently the generalform don't expect a facet in its declaration
[06:39] <daf> ooh, https://launchpad.net/errors gives me an oops
[06:40] <SteveA> daf: we don't want that any more
[06:40] <SteveA> it should give 404
[06:40] <daf> ah
[06:43] <SteveA> we use chinstrap etc. for that now
[06:43] <SteveA> salgado: oh?  that's interesting.  it's also a bug on generalform.
[06:44] <SteveA> salgado: best assign the bug to me then...
[06:48] <daf> bug 5789
[06:48] <Ubugtu> Malone bug #5789: /errors page should be removed In: launchpad (upstream), Severity: Normal, Assigned to: Nobody, Status: New https://launchpad.net/bugs/5789
[06:49] <salgado> SteveA, did you have a look at that patch I sent you thursday (and re-sent yesterday)? 
[06:54] <SteveA> salgado: you chose to set menu.request in _nearest_menu but in the other place that returns an application menu
[06:54] <SteveA> this should be more symmetrical
[06:54] <SteveA> iow, set it in the place that returns a facet menu, not in _nearest_menu
[06:55] <SteveA> this aspect of the menus system hasn't been tested
[06:55] <SteveA> i don't see a 'request' attribute in the appropriate menu interface
[06:57] <salgado> SteveA, right. I think I can fix this. I was most concerned with that XXX I left there, which I have no idea how to properly solve it
[06:58] <SteveA> that XXX has no date or owner
[06:59] <salgado> well, that's because I don't plan to merge this code with it there
[06:59] <SteveA> +            # XXX: Need to stick the view's predefined params here too, so we 
[06:59] <SteveA> +            # get the advanced form with the right values selected.
[06:59] <SteveA> 
[06:59] <SteveA> that's the one?
[06:59] <salgado> yes
[07:01] <SteveA> why are you using a view class to do the redirect, and not a redirection() in the navigation?
[07:01] <SteveA> hmm
[07:01] <SteveA> i guess
[07:01] <SteveA> because you only want to do so sometimes
[07:01] <SteveA> salgado: i can't really help here.  i don't know what the issues are, and i haven't seen the UI in action.
[07:01] <salgado> exactly. only when I get an advanced query parameter
[07:02] <SteveA> can you get a sample server set up at async, so i can try it out?
[07:02] <salgado> sure
[07:02] <SteveA> or give me a branch to look at (although i won't look until next week)
[07:05] <salgado> SteveA, http://async.com.br:8082/products/firefox/+bugs
[07:05] <salgado> SteveA, actually, http://async.com.br:8082/products/firefox/+bugs-critical
[07:06] <salgado> SteveA, that's a pre-defined search, which shows only the critical bugs
[07:07] <salgado> I want that when you click on "Advanced...", that you get to the advanced search, and "Critical" is the only value selected in the Severity widget
[07:09] <SteveA> salgado: mail these details to me and i'll look this evening, when brad has gone
[07:09] <salgado> okay
[07:12] <salgado> SteveA, I have a patch that moves those "hidden" links into the menu. it's almost trivial; would you review it?
[07:17] <SteveA> yes
[07:18] <salgado> SteveA, mailed it to you
[07:22] <SteveA> salgado: it is good.  one comment.
[07:22] <SteveA> leave in the empty metal slot div for the right portlets
[07:22] <SteveA> mark and i were going through a phase of i'd remove the empty ones, and he'd re-add them and so on
[07:22] <SteveA> without realizing it
[07:23] <SteveA> so we talked it over at UBZ, and decided to leave them in as a placeholder in general
[07:23] <salgado> I see. I'll remember that from now on
[07:23] <SteveA> salgado: maybe you'd like to add the 'facet' attr to the zcml for generalform?
[07:23] <SteveA> it isn't too hard.
[07:24] <SteveA> you can see the tests in the system doctests, and the code to do it in the webapp/metazcml.py
[07:24] <SteveA> there is a mix-in interface for specifying the zcml schema
[07:24] <SteveA> and then you just need to handle it, couple of lines of code there
[07:25] <salgado> okay. /me looks
[07:26] <salgado> SteveA, one question, though. I can only see a facet in the zcml declaration of pages that are not in the overview facet
[07:26] <salgado> IOW, I don't see why only the edithackergotchi and editemblem pages need this
[07:27] <mpt> those two pages should be merged anyhow
[07:27] <mpt> a person's hackergotchi is equivalent to a team's emblem
[07:28] <SteveA> salgado: the page-setup code doesn't even look for a facet higher up in the zcml
[07:28] <SteveA> as a workaround, you can set the __launchpad_facet__ attribute (i think that's it)
[07:28] <SteveA> in the init method
[07:28] <SteveA> to the appropriate facet
[07:28] <SteveA> maybe it is __launchpad_facetname__
[07:28] <SteveA> you'll find it in the facets code
[07:29] <SteveA> yes, that's it
[07:29] <SteveA>  __launchpad_facetname__
[07:29] <SteveA> that would be a quick workaround to fix the bug
[07:29] <SteveA> without having to fix generalform too
[07:30] <salgado> is it a problem if I do that, SteveA?
[07:30] <SteveA> no problem
[07:30] <SteveA> please file a bug on general form, though
[07:30] <salgado> will do
[07:41] <niemeyer> Heh.. GvR has been googled..
[07:45] <SteveA> as in?
[07:45] <SteveA> hired?
[07:46] <niemeyer> Yep
[07:46] <niemeyer> Good for us.. looks like he'll have more time for Python
[07:46] <SteveA> interesting
[07:51] <SteveA> niemeyer: got a URL about that?
[07:53] <salgado> SteveA, I added the rightportlets placeholders and did that workaround to fix the issue with generalform. want me to send the patch again?
[07:54] <SteveA> salgado: is kiko around?
[07:54] <SteveA> salgado: no need.  r=me
[07:54] <niemeyer> SteveA: Nope.. it's in the PSF list, which has no open archive IIRC
[07:56] <salgado> SteveA, kiko went out with cprov, I think
[07:56] <salgado> he's not here, that I'm sure
[07:57] <kiko> yep
[07:57] <kiko> I'm here
[08:43] <SteveA> hey mpt 
[08:44] <stickman> is account registration working?
[08:45] <mpt> ho SteveA 
[08:45] <SteveA> i just had a call with mark
[08:45] <SteveA> and kiko (for some of it)
[08:46] <SteveA> and mark said that launchpad is looking pretty nice in the UI, and commended the work you've done of late
[08:46] <mpt> What does Mark think of the new layout?
[08:46] <mpt> oh, good
[08:47] <salgado> stickman, it should be. did you have any problems?
[08:47] <stickman> i never receive a confirmation email.  not sure what's going on.
[08:50] <salgado> stickman, the confirmation emails can take some time to arrive, in some cases. when you tried to sign up?
[08:50] <stickman> i've tried multiple times over a period of several weeks. and with two different email addresses.
[08:52] <salgado> stickman, is it possible that the emails got caught in a spam filter?
[08:54] <stickman> nope.  checked there.  one account has not filtering at all.
[08:56] <salgado> kiko, we don't have access to the bounced emails yet, do we?
[08:56] <LarstiQ> stickman: hmm, should surely have had it by now then
[08:57] <SteveA> kiko: ?
[08:57] <stickman> wait.  tried another email and got something.
[09:00] <stickman> ok. registered now.  thanks all.
[09:02] <salgado> stickman, you're welcome
[09:04] <carlos> daf, hi
[09:05] <stickman> i'm relatively new to ubuntu so i figured i'd check it out.  
[09:12] <salgado> cprov, I added another question to the MirrorManagement spec:  What is DistributionMirror.pulsesource and who's allowed to change it?
[09:35] <mpt> home time for mpt
[10:25] <kiko> SteveA, I received no answer on that yet. :-(
[10:25] <kiko> I think I'm going to msg Znarl 
[10:32] <sivang> daf: when you have time, I'd appriciate comments on my remaining remarks inline RFS
[10:33] <sivang> jblack: what's needed in order to fininsh the rocketfuel-get script?
[11:38] <LarstiQ> sounds important :)
[11:39] <jordi> openoffice considering rosetta officially
[11:39] <LarstiQ> ooh
[11:39] <LarstiQ> still considering though?
[11:40] <jordi> they want two features
[11:40] <jordi> not too difficult
[11:40] <mdke> pretty hot
[11:40] <mdke> don't you have firefox and mailman too now?
[11:41] <jordi> mailman yes, not firefox
[11:41] <jordi> unless I'm totally out of date :)
[11:42] <jordi> carlos: are you sure this is talking about the global l10n project, or just the British team?
[11:42] <jordi> or whatever team, I mean
[11:42] <carlos> jordi, he says is the coordinator of the global l10n team
[11:42] <carlos> jordi, I don't see any reference to a concrete team...
[11:43] <carlos> see you tomorrow
[11:43] <carlos> jordi, do you need anything from me?
[11:44] <jordi> carlos: I am fearing this is not the whole openoffice
[11:45] <jordi> but Esperanto.
[11:45] <mdke> lol
[11:45] <mdke> that would be a catch anyhow :)
[11:45] <carlos> jordi, just ask him directly ;-)
[11:45] <jordi> mdke: we'd have Kurdish and Esperanto. :)
[11:45] <mdke> :)
[11:46] <mdke> carlos, if you have time tomorrow can you have a look at my mail to rosetta-users about the po's we download from rosetta and send back through xml2po?
[11:46] <jordi> from the first read it sounds to me like he's just a translator
[11:46] <jordi> no coordinator
[11:46] <jordi> well
[11:46] <jordi> the coordinator for just a translation team
[11:46] <carlos> mdke, sure, I have it in my queue, sorry for not answering it earlier...
[11:47] <mdke> carlos, no problem, i wasn't sure how busy you were :)
[11:47] <mdke> carlos, no massive rush
[11:47] <carlos> mdke, as an advance, I think those are Rosetta bugs
[11:47] <carlos> at least the duplicate problem is a know problem
[11:47] <carlos> see you tomorrow!!
[11:48] <mdke> cool
[11:48] <mdke> night