[00:01] <wgrant> spm: Can you please run 'SELECT dsc_format, COUNT(*) FROM sourcepackagerelease GROUP BY dsc_format;' on prod for me?
[00:02] <spm>  dsc_format | count
[00:02] <spm> ------------+--------
[00:02] <spm>             |  70050
[00:02] <spm>  1.0        | 388445
[00:02] <spm> wgrant: ^^
[00:02] <wgrant> Good lord.
[00:02] <cody-somerville> ugh
[00:02] <wgrant> Hmmm.
[00:04] <wgrant> Thanks spm.
[00:05] <spm> np. I'm delighted with the response from yourself and cody-somerville tho :-D
[00:07] <wgrant> Doesn't *really* matter, since we'll just clobber all the null ones. But it's a bit strange that there are so many.
[00:09] <maxb> I have https://code.edge.launchpad.net/~maxb/pygettextpo/makefile-fix-version/+merge/14256 pending, who do I nudge about making it happen?
[00:20]  * thumper stabs launchpad.dev
[00:21] <wgrant> thumper: What is it doing to you today?
[00:21] <thumper> wgrant: stabbing launchpad
[00:21] <thumper> wgrant: some small tweaks, some bigger ones
[00:21] <thumper> wgrant: mostly ui stuff
[00:22] <wgrant> Ah.
[00:46] <thumper> sinzui: still around?
[00:52] <thumper> wgrant: perhaps you know...
[00:52] <thumper> if someone is starting a new project
[00:52] <thumper> what is the minimum they need to do to upload files?
[00:53] <cody-somerville> thumper, can you provide more context to your question?
[00:53] <thumper> someone cornered me in the weekend after moving a project to launchpad
[00:53] <thumper> and asked "how do I add download files"
[00:53] <thumper> I couldn't find the link
[00:53] <wgrant> thumper: Add a release.
[00:54] <wgrant> Files are added to releases.
[00:54] <thumper> how do you add a release?
[00:54] <thumper> is there a friendly button?
[00:54] <thumper> ah
[00:54] <wgrant> (releases are created on top of milestones within a series, but I think there is a button which adds a milestone and immediately turns it into a release)
[00:54] <thumper> create release
[01:02] <thumper> that should at least be mentioned somewhere on the main page
[01:02] <thumper> it isn't intuitive to a new user
[01:03] <wgrant> It's Registry. Of course it's not intuitive.
[01:04] <wgrant> Although it's much better than it used to be, the whole series thing is very obscure.
[01:05] <wgrant> thumper: Can the codehosting-related stuff be dropped from #launchpad's topic yet?
[01:06] <thumper> I'd say so
[01:06] <thumper> I'm not aware of any issues right now
[01:09] <spm> thumper: actually related to that - that emergency codehost sync logs we setup back when - I assume we can remove that as well now?
[01:10] <thumper> spm: what was set up exactly?
[01:11] <spm> thumper: huh. nm. it's been removed. la la. from memory, it was sync the codehost *logs every 3 mins
[01:11] <thumper> spm: yes we can stop that :)
[01:11]  * spm does nothing and pretends he does - all gone!
[01:12] <thumper> spm: gee, thanks
[01:12] <spm> haahahahaha
[01:12] <spm> I recall seeing it the other day, and assumed it was still there; but alas, one of the other losas was probably on a clean up phase :-)
[01:59]  * thumper smacks head on the table
[01:59] <thumper> refactoring the wrong test
[01:59] <thumper> d'oh
[02:05]  * wgrant tries to work out how to add breadcrumbs.
[02:20] <thumper> wgrant: automagically
[02:20] <wgrant> thumper: I find that unlikely.
[02:21] <thumper> wgrant: it kinda is
[02:21] <thumper> wgrant: by default it uses the canonical_url_iterator
[02:21] <thumper> wgrant: you can override the heirarchy it uses
[02:21] <wgrant> thumper: I know how that stuff works.
[02:22] <thumper> wgrant: take a look at lp.code.browser.branch.BranchHeirarchy
[02:22] <wgrant> But some objects don't have a breadcrumb.
[02:22] <wgrant> eg. builds.
[02:22]  * thumper tries to remember how branches did it
[02:22] <thumper> BranchMergeProposalBreadcrumb
[02:23] <thumper> You need an IBreadcrumb adapter
[02:23] <wgrant> Ah, looks like r9842 does the sort of stuff I want.
[02:23] <wgrant> Thanks.
[02:23] <thumper> np
[02:38] <wgrant> We should destroy custom titles whereever possible, right?
[02:38] <thumper> maybe
[02:38] <thumper> not sure on the edict
[02:42]  * wgrant trusts the old 3.0 heading rules for now, since they are at least better than what is implemented now.
[02:44] <thumper> wgrant: what is your time split between real study and LP right now?
[02:47] <wgrant> thumper: It depends how you calculate it. Last week was pretty much fully studying, due to the first batch of exams. There are no non-trivial exams remaining, so not much study is happening now.
[03:16]  * wgrant glances at pagetitles.py and screams.
[03:17] <thumper> wgrant: it is deprecated thank god
[03:17] <wgrant> It still has calendar stuff, I see.
[03:17] <wgrant> How useful and modern.
[03:18] <thumper> heh
[03:18] <ajmitch> I take it that's a bad thing? :)
[03:18] <wgrant> Well, the calendar system was closed aaages ago, and actually removed more than two years ago.
[03:20] <ajmitch> a great opportunity to clean up old code
[03:21] <spm> heh. get quadruple code karma for removing old code
[03:22] <ajmitch> does it have to be old code?
[03:22] <spm> that was a pun, not serious. :-)
[03:22]  * ajmitch can think of a few things that could be removed
[03:23] <spm> soyuz... malone... :-P
[03:24] <wgrant> There are lots of <browser:pages> with completely different views using the same view class. If I want to specify a title without using pagetitles.py, does that mean I have to create a subclass and rip it out into its own <browser:page>, or is there some less inconvenient method that I'm missing?
[03:32] <thumper> wgrant: make the page_title a property that is smart?
[03:32] <thumper> wgrant: or yes, subclass
[03:32] <wgrant> Subclassing seems less revolting.
[03:33] <wgrant> Just makes the ZCML suck a bit more.
[03:33] <wgrant> Neeeed Grooook.
[03:36] <wgrant> thumper: Has Martian been considered at all for LP?
[03:36] <thumper> yes
[03:37] <thumper> but that is about as much as I know
[03:37] <thumper> flacoste knows more
[03:37]  * ajmitch even attended a talk in the weekend that mentioned that
[03:42] <wgrant> How did kiwipycon go?
[03:43] <thumper> pretty good actually
[03:43] <thumper> I enjoyed myself
[03:44] <thumper> which is always a good measurement
[03:45] <wgrant> How many were there?
[03:48] <ajmitch> it was full, so about 150
[03:48] <wgrant> Not bad.
[03:48] <ajmitch> everything ran smoothly, talks were good
[06:12]  * wgrant notices the team leads mailing list, and discussion on it about it being private -- it isn't.
[06:12] <poolie> hee hee
[06:12] <poolie> srsly?
[06:13] <wgrant> Sadly.
[06:13] <ajmitch> let's hope that there's not too much interesting gossip on there
[06:13] <citrus> We know the list is public, and what we put in it is public. We will state if we removed something that was private
[06:14] <wgrant> I used LP to make the private list, but I do have one concern.  It appears
[06:14] <wgrant> that some people I added through LP just have an email address outside of
[06:14] <wgrant> Canonical.  I was hoping to only use our Canonical addresses for this list.
[06:14] <wgrant> That indicates that at least Brian believes it's private.
[06:15] <citrus> Brian is mistaken
[06:16] <citrus> karls emails are the best. His notes might include a Dalek invasion during an interlude in the meeting
[06:18] <wgrant> I see.
[06:18] <wgrant> Anyway, thought it best to mention it, since you guys don't have an awfully good track record of keeping private mailing lists private.
[06:19] <citrus> No we dont. The owner who demand privacy seem to want public lists.
[06:20] <citrus> sé la vié
[06:37] <wgrant> https://launchpad.net/builders/rothera/+history has a back link at the bottom. Were it not for that link, that template could go away. Now that we have good breadcrumbs, is there any reason to keep it?
[06:41] <citrus> I agree that the link is a distraction. we are also removing the links that use >> and <<
[06:44] <wgrant> citrus: What is the label policy for pages with broken context?
[06:44] <wgrant> (ie. anything under a distro)
[06:44] <wgrant> Should they include the object, since the main heading is not correct?
[06:44] <wgrant> s/broken context/broken root or watermark thingy/
[06:48] <citrus> wgrant: I do not understand. label is a loaded and very confusing term in launchpad now that it is used for forms, more headings, and surprisingly to fill the pages primary heading slot
[06:49]  * citrus starts throwing things when he has to think about what the consequences are setting a view's label and page_title
[06:49] <wgrant> citrus: I mean the label on the view, which becomes an h1.
[06:49]  * citrus starts throwing things
[06:50] <citrus> The label should be the title of the context object, or the heading of the context's subordinate information, or the action the user is performing on the context
[06:50] <wgrant> citrus: That's fine for projects.
[06:50] <citrus> I wish I had good examples for this
[06:50] <wgrant> But not for anything else.
[06:51] <wgrant> Well, projects and distros.
[06:51] <wgrant> But not project series, distro series, distro arch series, distro source packages, or disto series source packages.
[06:51] <wgrant> Or builders.
[06:51] <wgrant> Or most of the rest of the LP objects that could conceivably have actions.
[06:51] <citrus> it works for https://edge.launchpad.net/builders/rothera/+history
[06:52] <wgrant> citrus: That label isn't one of the three possibilities.
[06:52] <citrus> "the heading of the context's subordinate information"
[06:53] <citrus> The context is a builder, the information is that it is history
[06:53] <wgrant> citrus: But this includes the context's identity, whereas most other views do not.
[06:53] <citrus> indeed
[06:53] <wgrant> I am wondering whether it should or not.
[06:53] <wgrant> Since it is neither specified nor consistent.
[06:53] <citrus> We are not consistent. I got angry and stopped working on heading for 4 weeks
[06:53] <citrus> what colour is your bike shed?
[06:54] <citrus> The history example includes the smarty quotes which I think are vestigial to 1.0
[06:54] <wgrant> I'm mainly trying to correct the braindead titles produced for queue and build pages, but labels got dragged along too.
[06:57] <citrus> There are many headings that can be improved, many well thought heading were made bad the day all the page titles changed
[06:57] <wgrant> Yes.
[06:57] <wgrant> The actual titles are easy to fix.
[06:57] <citrus> wgrant, I have learned that I have to argue for each heading. I lost a lot of arguments
[06:57] <wgrant> Because they always include context.
[06:57] <wgrant> But labels have no such luxury.
[06:58] <citrus> The page_title is documented as not being required. Yet I have had to add that to more than dozen views last month to fix breadcrumbs. I am sure it is required.
[06:59] <citrus> I think a good label focuses your attention and task.
[06:59] <wgrant> Won't breadcrumbs include '+somethingstupid' unless you specify page_title?
[06:59] <citrus> Oh.now I see how that is screwed up
[07:00] <wgrant> ?
[07:01] <citrus> https://edge.launchpad.net/builders/rothera/+history
[07:01] <citrus> ^ I think the <h1> is in the tempalate's content. The view does not provide a label
[07:01] <wgrant> I know.
[07:01] <wgrant> My new version of the build listing template fixes that.
[07:01] <wgrant> The page_title is obviously 'Build history'
[07:01] <wgrant> The correct label is not at all obvious.
[07:02] <citrus> The correct page_title is "build history" because rothera (i386) is already in the breadcrumbs/title
[07:03] <wgrant> Right.
[07:04] <citrus> The page_title we are seeing could be the label if we extend beuno's rule that the heading for a page that is alternate view of an object should extend the object's title...
[07:04] <wgrant> I suspect that the label should be "Build history for “rothera (i386)”", but I really do not know.
[07:05] <citrus> eg: Lucid: Edit Lucid
[07:05] <citrus> eg: My team: My team members
[07:06] <citrus> wgrant: +1
[07:08] <wgrant> citrus: Inconveniently, all of the rest of the build pages live within a distroseries, and nobody seems to know how to refer to those.
[07:09] <citrus> I would not be surprised if each page was updated by a different developer
[07:10] <citrus> adding a post-3-ui-tag to a bug improves the chances of getting it scheduled. We are still allocating 50% of our capacity to oopses and UI bugs
[07:13] <citrus> wgrant: I just came across several of your bugs for foundations. They should have been triaged and tagged with post-3-ui-cleanup. I updated the bug so that they should up in jml and flacoste's reports
[07:17] <wgrant> citrus: Thanks.
[07:18]  * wgrant is tempted to email the list about the label issue.
[07:19] <citrus> go ahead. A few weeks have passed and I think the rules are a little clearer.
[07:19] <citrus> We might get a codified rules out of this.
[07:20] <citrus> oh, but a lot of engineers are distracted by the lazr-js sprint and UDS
[07:20]  * wgrant shall compose the email later after work stuff.
[07:21] <wgrant> Thanks for your help.
[07:45] <thumper> citrus: what's with the name change?
[07:45] <thumper> citrus: and what are you doing up?
[07:45] <citrus> I learned that I have today off
[07:46] <citrus> So I am working according to my own agenda
[07:46] <citrus> Except I am too tired to fix the answers pages, remove the style sheet, or kill off main-template
[07:46] <thumper> citrus: plz add wiki, kthxbye
[07:47]  * citrus install wikimedia
[07:47] <thumper> citrus: I'm about to use locals().extend(setup_func()) to a pagetest :)
[07:47] <thumper> locals().update that is, not extend
[07:48] <mwhudson> citrus: fix blueprints!!1
[07:48]  * mwhudson hides
[07:48] <citrus> mwhudson: please give me a UI review so that I can fix the worst parts
[07:49] <citrus> https://edge.launchpad.net/blueprint/+milestone/3.1.11
[07:50] <citrus> The problem with fixing a menu or removing the 1.0 button means I need to find ui reviewers who are all sprinting
[07:50] <mwhudson> citrus: argh, can't really help you there :(
[07:50] <citrus> I really need an alternate personality so that I could review my self
[08:57] <adeuring> good morning
[08:57] <al-maisan> moin adeuring
[08:57] <adeuring> hi al-maisan!
[09:23] <maxb> Who do I talk to about getting stuff landed to sourcecode/ branches?
[09:30] <mrevell> Morning
[09:35] <pgquiles> what source should I get for setting up a launchpad in production? branch 'stable'? branch 'db-stable'? branch 'devel'? branch 'db-devel'? is there a tarball anywhere? what are those 'db-something' branches?
[09:44] <maxb> pgquiles: You are aware of the image licensing constraint, right?
[09:45] <maxb> a.k.a. "Canonical does not grant you a licence to use the existing images in a production instance"
[09:45] <pgquiles> maxb: yes
[09:45] <pgquiles> I read that, in the FAQ IIRC
[09:46] <pgquiles> atm I'm trying to set it up, we'll replace the images, etc when/if I get it to work
[09:46] <pgquiles> we are only going to use launchad internally, though, not public site
[09:47] <maxb> I'd love to do the same, but I don't have the time to replace all the images :-/
[09:47] <maxb> Anyway, https://dev.launchpad.net/Trunk explains how the four branches relate
[09:48] <pgquiles> thanks
[10:49] <pgquiles> how can I download launchpad 3.1.10? I can't find a branch/tag/tarball/whatever by that name :-?
[10:59] <wgrant> bigjools: Want to review my branch at some point, or shall I use OCR?
[11:00] <wgrant> bigjools: I'd also like to get the production data unbroken.
[11:00] <bigjools> wgrant: I can, I'm just not sure *when* I can :/
[11:00] <bigjools> go for OCR and I'll try and tag along
[11:25] <bigjools> can someone do me a favour please: du -sh /var/lib/postgresql
[11:28] <wgrant> bigjools: Only 400MiB at the moment, but I've seen it get to a couple of gigabytes during tests.
[11:28] <bigjools> wgrant: ok ta - that's annoying, it won't fit in a ramdisk any more :(
[11:35] <maxb> pgquiles: You're going to find that the open-sourced bits of Launchpad are deliberately only the bits that are developer oriented. Anything that only pertains to production ... like the VCS branches pertaining to what it is exactly ... isn't public.
[11:43] <pgquiles> maxb: huh? so launchpad has been opensourced in a pretty useless way, hasn't it? :-/
[11:43] <maxb> It's been opensourced in a way which is perfect for allowing community contributors to enhance launchpad.net
[11:44] <maxb> That clarifying clause.... was perhaps conveniently omitted from the initial publicity
[11:45] <maxb> Don't get me wrong, I'm really happy that Canonical went as far as they did, when they didn't have to opensource it at all, but there are still strings attached
[11:48] <maxb> AFAIK, the best us community people can do to see the sourcecode of a particular production version is to look for the last revision in db-stable before stable->db-devel auto-merging re-commences
[11:49] <pgquiles> that's what I've done, I've modified the rocketfuel-setup script to install the db-stable branch
[11:49] <pgquiles> but I was also interested in the buildd configuration branch, which is proprietary :-/
[11:49] <wgrant> buildd configuration branch?
[11:49] <maxb> buildd configuration?
[11:51] <stub> The production branches are private so security fixes can land without exposure. Nothing lands on the production branches that doesn't also land on launchpad/devel or launchpad/db-devel. I don't know about the buildd config branch (including if it exists) - it probably contains keys that can't be exposed but I'm guessing.
[11:53] <maxb> Ultimately this boils down to : Canonical doesn't want people to run their own internal Launchpad instances, but people see "Launchpad is Open Source", and want to, and are then disappointed when they find out they can't.
[11:54] <maxb> s/can't/can't without lots of rebranding work and devising their own production-esque configs/
[11:54] <wgrant> One doesn't have to be too creative to get hold of productionish configs.
[11:55] <wgrant> But the configs aren't that interesting.
[11:55] <maxb> I've not really thought much about the second, because the first is a big blocker.
[11:55] <wgrant> pgquiles: What did you mean by "buildd configuration"?
[11:58] <pgquiles> wgrant: I want to install PPAs here. There is this project ( http://launchpad.net/lpbuildbot ) which seems to contain interesting stuff, but it's closed source
[11:58] <wgrant> pgquiles: That's not related.
[11:58] <wgrant> pgquiles: That's the configuration for the buildbot which runs the Launchpad test suite.
[11:58] <pgquiles> oh, I see
[11:58] <pgquiles> thanks
[11:59] <pgquiles> wgrant: there are no tarballs for releases, are there?
[11:59] <wgrant> pgquiles: No. It's all deployed through bzr internally.
[12:00] <pgquiles> so how do you upgrade to the next version (say, 3.1.11) ? bzr pull ?
[12:00] <wgrant> I don't know.
[12:03] <stub> You want 3.1.11 at the start of the month or at the end of the month? You generally want the latest lp:~launchpad-pqm/launchpad/db-stable
[12:09] <maxb> Public bug 368889 references private bug 280578 - could someone have a look at it and determine if it needs to be private, or if it's just a leftover from pre-opensourcing? Thanks.
[12:09] <mup> Bug #368889: https://code.edge.launchpad.net/~twisted-dev/twisted/trunk reports wrong branch format <branch-puller> <branch-scanner> <Launchpad Bazaar Integration:Triaged> <https://launchpad.net/bugs/368889>
[12:29] <pgquiles> ok, I got it working using db-stable (devel seems to have some problem with loom.branch not being found. But db-stable (3.1.10 r8660) works like a charm!
[12:29] <pgquiles> now reading about soyuz, let's see if I can set up the autobuilders
[15:04] <Ursinha> allenap, hi
[15:04] <allenap> Ursinha: hi
[15:05] <allenap> Ursinha: Whassup?
[15:05] <Ursinha> allenap, I'm sorting out a problem of bug mail spam a user reported in feedback@
[15:05] <Ursinha> allenap, he says he's getting email of a bug that's not related to him in launchpad
[16:35] <pgquiles> it seems that rocketfuel-setup does not install the keyserver ( keyserver.launchpad.dev goes to "page not found" and gpg --send-keys also fails). Do I need to install/configure anything else? (db-stable r8661 here)
[17:05] <maxb> pgquiles: "make run" only starts a variety of more essential dev services. Investigate "make run_all"
[17:07] <maxb> NB that the dev keyserver is a minimal shim to make the tests work, not the same software that runs the actual keyserver
[17:08] <pgquiles> maxb: make run_all doesn't start the keyserver either
[17:08] <pgquiles> I only need the keyserver for the PPAs
[17:12] <maxb> pgquiles: Have you seen https://dev.launchpad.net/Soyuz/HowToUseSoyuzLocally ?
[17:15] <pgquiles> maxb: that's exactly what I'm trying to follow
[17:16] <maxb> pgquiles: Have you actually done start-soyuz.sh?
[17:16] <maxb> The keyserver is called zeca
[17:16] <pgquiles> maxb: but it does not work for me. Even make-ubuntu-sane.py is broken, it looks like ubuntu.seriases is an empty list :-? odd
[17:19] <maxb> Have you run "make schema" ?
[17:19] <pgquiles> maxb: this is the output of make-ubuntu-sane.py: http://pastie.org/693926
[17:21] <pgquiles> maxb: yes, after rocketfuel-setup
[17:22] <pgquiles> the only thing I did different from the instructions at http://dev.launchpad.net is I'm using db-stable instead of devel, and it's because devel fails due to some error with bzr.loom
[17:22] <maxb> pgquiles: I believe that attribute got renamed recently
[17:25] <pgquiles> maxb: you don't know the new name, do you?
[17:25] <maxb> s/serieses/series/, I *think*
[17:31] <pgquiles> maxb: you think *right* :-)
[17:31] <pgquiles> thanks!
[17:34] <pgquiles> I've updated make-ubuntu-sane.py in the wiki
[17:38] <Ursinha> allenap, hi :) is it possible to unsubscribe and stop receiving bug mail from a bug I've filed?
[17:39] <allenap> Ursinha: Yes, I think/hope so. You're just added as a direct subscriber I think.
[17:40] <Ursinha> thanks allenap
[17:51] <pgquiles> ugh, keyserver still not working :-(
[18:10] <mrevell> night all
[18:23] <pgquiles> ah, I found the problem: /var/tmp/zeca did not exist. I created it with proper permissions and now the keyserver works fine.
[19:50] <pgquiles> I followed https://dev.launchpad.net/Soyuz/HowToUseSoyuzLocally to add builders but now I want to remove them. I've tried with "remove" instead of "add" ( scripts/ftpmaster-tools/manage-chroot.py -s lucid -a i386 remove -f chroot-ubuntu-lucid-i386.tar.bz2 ) but it does nothing. What am I doing wrong?
[19:56] <thumper> morning
[20:36] <jkakar> jml: In case you hadn't noticed, I've started some work to make it possible/pleasant to write unit tests for launchpadlib-using applications: https://code.edge.launchpad.net/~jkakar/launchpadlib/testing-support/+merge/14444
[20:37] <thumper> jkakar: awesome
[20:37] <jkakar> thumper: It's not quite ready to land, but I thnk it's the right way forward.  Would love to get some reviews/feedback and push it forward (hint, hint). :)
[20:38] <thumper> :)
[21:38] <wgrant> Hm. So has it really been decided that assertEqual(expected, observed) is the Right Way?
[21:39] <thumper> wgrant: that is how the code team does it
[21:40] <wgrant> How odd, but OK.
[21:40] <thumper> wgrant: but I don't think there is a "LP approved" way