[01:15] <StevenK> wgrant: O hai, Mr. OCR. https://code.launchpad.net/~stevenk/launchpad/builder-description-mandatory/+merge/106729
[01:25] <wgrant> StevenK: Does code rely on the description being set?
[01:26] <wgrant> StevenK: The description shouldn't really exist at all; I'd prefer to fix the database to make it nullable.
[01:27] <StevenK> wgrant: DB patches are making me very cranky at the moment.
[01:27] <StevenK> wgrant: But if you think a DROP CONSTRAINT is the best way forward, I can do that.
[01:28] <wgrant> StevenK: Check what uses Builder.description
[01:28] <wgrant> I'm not sure infinity or lamont really want to keep coming up with awesome descriptions for builders.
[01:28] <StevenK> steven@undermined:~/launchpad/lp-branches/devel% bzr grep -i Builder.description | wc -l
[01:28] <StevenK> 0
[01:29] <wgrant> It's never going to be used like that.
[01:29] <wgrant> That'd only be if we were querying on it.
[01:29] <StevenK> <- Literal troll
[01:30] <bigjools> are there many bridges to hide under in western Sydney?
[01:30] <wgrant> He lives next to a river.
[01:30] <wgrant> So he should have one nearby.
[01:30] <wgrant> I'm not sure about goats, though
[01:30] <StevenK> There is one just outside, in fact.
[01:30] <wgrant> Bogans might have to do.
[01:30]  * StevenK kicks wgrant into Coburg.
[01:31] <wgrant> We don't talk about Coburg.
[01:31] <StevenK> I don't think Builder.description is actually used by anything at all.
[01:32] <wgrant> I think it's in the UI
[01:32] <bigjools> it should be nullable
[01:32] <wgrant> Most definitely.
[01:32] <wgrant> It's just a question of how easy it is
[01:32] <wallyworld_> wgrant: it appears the legacy access mirroring trigger for bugsubscription creates an artifact grant even if the bug is already visible via a policy grant
[01:32] <wgrant> I suspect the answer is "trivial"
[01:32] <cjwatson> RFC by Soyuz people: https://bugs.launchpad.net/launchpad/+bug/1002589
[01:32] <_mup_> Bug #1002589: Deploy new Ubuntu archive signing key <Launchpad itself:New> < https://launchpad.net/bugs/1002589 >
[01:32] <wgrant> wallyworld_: That's correct and very deliberate.
[01:32] <wallyworld_> why? :-(
[01:33] <wgrant> cjwatson: :( why didn't we do that before precise
[01:33]  * bigjools looks
[01:33] <cjwatson> Because we sucked at getting the necessary people together.
[01:33] <cjwatson> We finally managed it at UDS after lots of false starts.
[01:33] <StevenK> cjwatson: Why is my MP linked in that bug?
[01:33] <cjwatson> wuh
[01:33] <cjwatson> paste damage, will edit, sorry
[01:34] <bigjools> heh
[01:34] <wgrant> wallyworld_: Because the rules around when an APG should prevent creation of an AAG etc. weren't defined when they were written.
[01:34] <StevenK> I hate to say it, but maybe a new column on distroseries
[01:34] <cjwatson> better
[01:34] <wgrant> wallyworld_: And they would make it more complicated.
[01:35] <cjwatson> StevenK: troublesome with dogfood though?
[01:35] <bigjools> cjwatson: all the signing is done in ubuntu-specific shell scripts
[01:35] <wgrant> cjwatson: We can change the DB on DF
[01:35] <cjwatson> bigjools: I know, but it still needs to work on both prod and df
[01:35] <wgrant> wallyworld_: Is this problematic?
[01:35] <wgrant> wallyworld_: The triggers will be turned off before we use APGs
[01:35] <cjwatson> wgrant: on every restore?
[01:35] <wgrant> cjwatson: Yes
[01:35] <StevenK> I didn't think DF had the archive signing key?
[01:35] <bigjools> cjwatson: this is a great time to get the scripts out of the tree then
[01:35] <wgrant> cjwatson: They only happen every 6 months
[01:35] <cjwatson> Is there an automatic way to do that?
[01:35] <bigjools> StevenK: it doesn't
[01:36] <StevenK> bigjools: Good.
[01:36] <cjwatson> wgrant: Which is just long enough for people to forget manual procedures
[01:36] <wallyworld_> wgrant: so now i have a job which remove subscriptions for bugs which become invisible when info type changes. and my test was failing since the aag was providing visibility
[01:36] <wallyworld_> so i will have to disable triggers in my test
[01:36] <wgrant> cjwatson: People will remember them when the signing fails after they forget :)
[01:36] <cjwatson> bigjools: Well, uh, sure, but we still need to solve the problem :-)
[01:36] <bigjools> I am strongly of the opinion that there should be no LP changes here
[01:36] <wgrant> wallyworld_: Ah, that's awkward.
[01:36] <wallyworld_> wgrant: the aag was created by the trigger, not me
[01:37] <bigjools> cjwatson: can move the ubuntu run-parts stuff out of the LP tree and hard code what you need in your own scripts?
[01:37] <wgrant> cjwatson: The idea is that cronscripts/publishing/distro-parts/ubuntu shouldn't be part of the LP tree
[01:37] <cjwatson> It is kind of bizarre that these Ubuntu-specific scripts were created without any provision for Ubuntu-specific editing outside LP, I'll grant you
[01:37] <bigjools> huh?
[01:37] <wgrant> cjwatson: The Ubuntu primary archive is signed by those scripts that are meant to be managed by !LP, not the native signing stuff.
[01:37] <bigjools> there is provision
[01:37] <cjwatson> Uh, not right now
[01:38] <cjwatson> Except by us landing LP changes
[01:38] <wgrant> You tell us you want to edit them.
[01:38] <wallyworld_> wgrant:  i was creating an apg on info type 1 and then changing bug to info type 2 and expecting it would be invisible, but i had done a subscribe and that created the aag which kept the bug visible
[01:38] <wgrant> We change the config to point at your copy on cocoplum.
[01:38] <bigjools> it's always been waiting for you guys to move them out of tree
[01:38] <wgrant> Done
[01:38] <bigjools> we have done all we can
[01:38] <cjwatson> If you're waiting for us, you need to tell us :-)
[01:38] <cjwatson> So if it's config-able, fine, I can deal with that (maybe tomorrow)
[01:39] <wgrant> wallyworld_: Right, you either need to not subscribe, or disable the trigger if this is a key part of your test.
[01:39] <wgrant> cjwatson: wgrant@lplucid:~/launchpad/lp-production-configs/trunk$ bzr grep distro-parts
[01:39] <wgrant> ftpmaster-publish/launchpad-lazr.conf:run_parts_location: cronscripts/publishing/distro-parts
[01:39] <wgrant> cjwatson: We just need to change that to point at a location you can write to
[01:39] <wallyworld_> wgrant: but i need to subscribe because the whole poiint of the test is to see that the unsubscribe happens :=)
[01:39] <wallyworld_> so will disable triggers
[01:39] <wgrant> cjwatson: Then we can delete these horrors from our otherwise perfect, flawless, and high-quality tree :)
[01:39] <cjwatson> OK (and we'll need to set things up so that that's an ubuntu-archive owned branch that ops can pull, rather than direct editing on cocoplum, of course)
[01:40] <wgrant> Right :)
[01:40] <StevenK> *cough* cdimage
[01:40] <wgrant> wallyworld_: Ah, right.
[01:40] <wgrant> wallyworld_: Indeed
[01:40] <cjwatson> StevenK: ?
[01:41] <StevenK> cjwatson: Just thinking of cdimage-private, but that's special in it's own right, and not really relevant to this situation now that I think about it
[01:42] <StevenK> wgrant: I can't see description on https://launchpad.net/builders/palmer
[01:42] <cjwatson> Also (while it's not really relevant to this channel anyway) slangasek managed to get hold of sabdfl and get agreement that we don't need the private branch for its former reason any more, so that is cleanupable at some point
[01:42] <StevenK> wgrant: Or /builders for that reason
[01:42] <wgrant> StevenK: Maybe you can drop the column and nobody will notice.
[01:42] <cjwatson> bigjools,wgrant: OK, thanks, I know what to do then; I'd like to keep the bug open until we have stuff moved over to our control, and probably then reassign it to some relevant component owned by us
[01:43] <wgrant> cjwatson: sure.
[01:43] <bigjools> cjwatson: perfect
[01:43] <StevenK> wgrant: Two DB patches, I guess. DROP CONSTRAINT ; drop the model code; DROP COLUMN
[01:43] <wgrant> StevenK: DROP NOT NULL, but yes.
[01:43] <bigjools> cjwatson: out of interest, why are you guys wanting to use a new key?
[01:43] <StevenK> bigjools: 1024D versus 4096E
[01:43] <StevenK> 4096R
[01:44] <bigjools> fairy muff
[01:45] <cjwatson> Also that key was generated eight years ago
[01:45] <cjwatson> Also https://bugs.launchpad.net/launchpad/+bug/804252 unfixable until we change keys
[01:45] <_mup_> Bug #804252: Please support InRelease files <Launchpad itself:Triaged by cjwatson> < https://launchpad.net/bugs/804252 >
[01:46] <bigjools> cool
[01:46] <cjwatson> I think we should probably be cycling keys roughly once per LTS cycle
[01:46] <bigjools> seems eminently sensible
[01:47] <bigjools> isn't it kinda late there Colin? :)
[01:47] <StevenK> wgrant: :-( I was going to delete the branch/MP
[01:48] <cjwatson> bigjools: Yes
[01:48] <wgrant> StevenK: That works too
[01:48] <cjwatson> Haven't you learned that I have odd sleeping habits yet? :-)
[01:48] <bigjools> cjwatson: I have noticed lately, yes!  EBABY?
[01:48] <cjwatson> Sometimes that, sometimes EINSOMNIA
[01:49] <StevenK> Yeah, Soyuz keeps me awake at night too
[01:49] <bigjools> I get to say EBABIES soon...  /o\
[01:49] <StevenK> :-P
[01:57] <wallyworld_> bigjools: hahahahahahahahaha
[02:02] <StevenK> bigjools: It's next month?
[02:03] <bigjools> end of next month
[02:03] <bigjools> cometh the apocalypse
[02:03] <StevenK> Haha
[02:03] <StevenK> I was about to ask if you're looking forward to it ...
[02:03] <bigjools> I am
[02:03] <bigjools> despite the stress I know is coming :)
[02:04] <StevenK> But Uncle Ian will be a big help, won't he?
[02:05] <bigjools> he *sure* will :D
[02:30] <wgrant> lifeless: http://people.canonical.com/~wgrant/launchpad/soyuz-arch.png is roughly what I've envisaged for a few years. The Soyuz bit is largely non-LP-specific, can be used by LS/OEM/everyone/etc. And Launchpad the gigantic monolithic application, while it can display information using API calls, becomes largely unaware of the details of archives and packages -- they're all done through the separate (probably Launchpad-branded) package ...
[02:30] <wgrant> ... service frontend.
[02:45] <lifeless> so other than 'web frontend' there, that all makes a lot of sense to me
[02:45] <lifeless> if you want to talk more, I'm off my prior call, and have had foodish
[02:51] <wgrant> lifeless: What doesn't make sense about that?
[02:54] <lifeless> why would LP talk to a web frontend to the package management service?
[02:54] <lifeless> why wouldn't it talk to the service itself?
[02:55] <wgrant> Um, yes, that.
[02:55] <wgrant> Never mind my arrows.
[02:55] <lifeless> wgrant: I'd expect users to still come through the LP frontend facade too, per what I was saying before
[02:55] <lifeless> But I'd expect that to be very thin - just universal
[02:56] <lifeless> There are about a billion ways to construct something functional
[02:56] <lifeless> its hard to say what works best today
[02:56] <lifeless> I'm not sure how to turn gut feelings in this space into cheap experiments
[02:57] <wgrant> That interface would need to be really minimal for things to work well
[03:00] <lifeless> I'm not sure why you say that
[03:00] <lifeless> Are you worried about duplication in functionality?
[03:00] <lifeless> Or development friction?
[03:02] <wgrant> Development friction mostly.
[03:04] <lifeless> So, I think there would be some friction, but conversely, I think having all the presentation and UI away from the service will make changes to the service rarer, and keep its API focused on functional changes.
[03:04] <lifeless> So a bit of swings and roundabouts; I'd expect most changes that don't need schema evolution to be presentation only.
[03:05] <lifeless> one way to think of this is that you're splitting out one piece at a time
[03:05] <lifeless> and UI is a separate piece
[03:05] <lifeless> it can always be split out itself later.
[03:06] <wgrant> The presentation and UI is separate from the service in that diagram, and indeed would probably be the last bit to go.
[03:06] <lifeless> so, one of the things I think we may disagree on is whether it *should* go
[03:07] <lifeless> I don't think we know enough yet to know whether it should or shouldn't. I am arging that UI and presentation should split off from the data storing services (including lp:launchpad).
[03:07] <lifeless> But I'm also arguing that all the UI should stay together.
[03:08] <lifeless> I think it would be a nice environment to have a bunch of API only services (with no admin UI or anything), and a single UI+presentation layer
[03:08] <lifeless> stateless UI+Presentation
[03:09] <wgrant> For some bits, sure. I don't think it's particularly valuable to have Soyuz as one of those bits.
[03:09] <lifeless> What makes it special?
[03:11] <lifeless> also, in unrelated news. \o/. My belkin blocks multicast between LAN and Wifi AFAICT. Wiiiin. Not.
[03:11] <wgrant> Oh, that's very handy of it
[03:11] <lifeless> (Q: why does avahi not work in my network: A: Belkin)
[03:11] <wgrant> This is why we don't use crappy home routers :)
[03:17] <bigjools> Linksys FTW
[03:17] <lifeless> Yeah, I had one, but it don't do N.
[03:19] <bigjools> N schmen
[03:20] <bigjools> I don't see the attraction of it, I rarely need that kind of speed on mobile devices and when I do I plug in
[03:21] <bigjools> which is like once a year
[03:21] <lifeless> my ethernet hub is hard to get at
[03:21] <bigjools> when we get broadband speeds that regularly exceed G then I'll get N :)
[03:29] <wgrant> I think my WAP is from like 2003
[03:29] <wgrant> a/b/g
[03:29] <wgrant> Works fine
[03:29] <lifeless> wgrant: I don't want unnecessary friction in dev, definitely don't want.
[03:29] <lifeless> wgrant: what I'm worried about is adding friction by partitioning things we should be integrating.
[03:29] <wgrant> lifeless: Sure, that is a very real concern.
[03:30] <lifeless> wgrant: I view services as a great way to stop square-peg round hole *and* to stop thousand-implementations-of-same-thing syndromes
[03:30] <wgrant> But it is my feeling that there isn't a huge amount of beneficial integration that this will make awkward.
[03:31] <lifeless> keeping the UI and presentation together is a relatively cheap way to mitigate the risk, because it will help folk think about the user experience specifically, when they are working on any UI or presentation issue.
[03:31] <StevenK> bigjools: I just bought a Linksys WAP, it's been really good.
[03:31] <wgrant> WAPs are easy to get right
[03:32] <wgrant> Cheap routers aren't.
[03:33] <lifeless> Leenuchs
[03:34] <wgrant> Yeah
[03:34] <wgrant> I hope Marvell will release another useful SoC soon.
[03:34] <wgrant> ARMv5 Kirkwood is getting a bit old.
[03:39]  * wgrant testfixes
[03:59] <wgrant> Bah
[04:00] <wgrant> I'm still 10 deletions away from being LOC-debt free.
[04:00] <wgrant> (although I have lots of credit once DB patches are rolled into the main schema dump)
[04:30] <StevenK> wgrant: I'm at +142 or so, but I should hit -2000 once I can delete show_information_type_in_ui
[04:31] <wgrant> Yeah
[04:31] <StevenK> stub: Thanks for the db review
[04:34] <StevenK> wgrant: Hm, just thinking about it, I should drop into negative from dropping builder.description too
[04:35] <wgrant> StevenK: The two sets of DB patch boilerplate will probably prevent that
[04:35] <bigjools> StevenK: yeah I have the blue one that everyone has
[04:35] <wgrant> WRT54GL?
[04:37] <StevenK> I bought a E3200
[04:37] <StevenK> It's a router, but I'm running it in bridge mode
[04:37] <StevenK> wgrant: Oh, all +16 lines? :-)
[04:39] <bigjools> wgrant: y6es
[04:39] <StevenK> wgrant: I've grabbed a lock on LPS, I'll change your FDT to tonight
[04:39] <lifeless> wgrant: You have mail; hopefully I haven't misrepesented you :)
[04:39] <bigjools> though I am not using it any more since the modem has wifi which successfully covers the house
[04:40] <bigjools> and while it's tempting to blanket the neighbourhood in my wifi signals, I won't
[04:40] <StevenK> It's Queensland, like there are any other signals around.
[04:40] <wgrant> StevenK: Ah, thanks
[04:41] <bigjools> StevenK: I could say the same about Boganville where you live
[04:41] <StevenK> bigjools: Pft, I moved out of Blacktown.
[04:42] <wgrant> lifeless: I don't see mail.
[04:43] <wgrant> Hopefully it's not going via forster, or I'll never see it.
[04:43] <lifeless> that or I messed up
[04:43] <lifeless> argh
[04:43] <lifeless> will -> robbie.
[04:43] <lifeless> FAIL
[04:44] <wgrant> Heh
[04:45] <wgrant> That's more like me.
[04:46] <StevenK> 4287 AttributeError: 'NoneType' object has no attribute 'email'
[04:46] <StevenK>     Bug: https://launchpad.net/bugs/580461
[04:46] <_mup_> Bug #580461: Login with the email once associated to a merge account OOPS <lp-foundations> <oops> <openid> <qa-ok> <Canonical SSO provider:Invalid> <Launchpad itself:Fix Released by stub> < https://launchpad.net/bugs/580461 >
[04:46] <StevenK> Hnady
[04:46] <StevenK> *Handy
[04:47] <wgrant> That might be from the user I suspended
[04:47] <wgrant> Looking
[04:47] <wgrant> Yeah
[04:47] <wgrant> It is
[04:47] <wgrant> lifeless: So codehosting doesn't handle it, but the auth API crashes...
[04:47] <wgrant> 4285 * 20s == around a day
[04:47] <wgrant> So it fits
[04:47] <lifeless> win :>
[04:47] <lifeless> and \y/ for complex internal models
[04:55] <StevenK> wgrant: +18/-37 for the death of builder.description. I was hoping there was more.
[04:56] <wgrant> lifeless: The email starts a little incoherent, but gets better.
[04:56] <wgrant> I think you overstate the complexity in running a separate service, though :)
[04:57] <StevenK> steven@undermined:~/launchpad/lp-branches/drop-builder-description% bzr di | tail -n 7 | head -n 3
[04:57] <StevenK> -    ...     title="The Prosciuto Builder", description="Uhmmm", owner=cprov,
[04:57] <StevenK> -    ...     virtualized=False)
[04:57] <StevenK> +    ...     title="The Prosciuto Builder", owner=cprov, virtualized=False)
[04:57] <StevenK> wgrant: ^ Oh look, there's all the proof you need that the column is utterly bong.
[04:59] <wgrant> Oh
[04:59] <wgrant> I didn't think it would actually have tests
[04:59] <wgrant> That explains the diff.
[05:10]  * StevenK destroys the use of a_builder
[05:11] <StevenK> +27/-56 and the doctest sucks less
[05:45] <StevenK> wgrant: https://bugs.launchpad.net/launchpad/+bug/973001 WTF?
[05:45] <_mup_> Bug #973001: NotImplementedError BugTarget <security proxied lp.registry.model.distributionsourcepackage.DistributionSourcePackage instance at 0x12ef41d0> is not supported by . on +expirable-bugs page <oops> <Launchpad itself:Triaged> < https://launchpad.net/bugs/973001 >
[05:51] <wgrant> StevenK: Seems pretty self-explanatory to me.
[05:55] <StevenK> Hmmmm. Sure, but how the frack do you fix it
[06:08] <wgrant> Make it work or deregister the view.
[06:08] <wgrant> Making it work should be pretty trivial.
[06:08] <wgrant> I'd expect at most two lines of code.
[06:08] <wgrant> StevenK: ^^
[06:08] <StevenK> wgrant: deregister the view is hard, given it's registered against IBugTarget
[06:09] <wgrant> Alternatively delete the page
[06:09] <wgrant> Replace it with a bug search
[06:10] <wgrant> That's probably besty
[06:13] <StevenK> wgrant: So just redirect?
[06:15] <wgrant> The only difficulty is that the page currently has a block of text at the top
[06:56] <jelmer> g'morning
[06:57] <spm> o/ jelmer
[07:00] <jam> morning jelmer, mgz, vila, czajkowski
[07:00] <vila> morning all
[07:01] <czajkowski> morning
[07:13] <lifeless> jelmer: hi; subunit ?
[07:23] <jelmer> lifeless: hi
[07:23] <jelmer> lifeless: sorry, haven't had time to work on it yet
[07:24] <czajkowski> hmm trying to mark a bug duplicate of another one have the bug open but it keeps saying invalid
[07:24] <czajkowski> :/
[07:25] <jelmer> lifeless: my spare time is limited at the moment; if you or somebody else wants to take over that's okay, I don't want this to block on me.
[07:26] <czajkowski> nm engaged noggin
[07:40] <lifeless> jelmer: if it helps, you can count it as work, as gary's squad is impacted in their work by it
[07:43] <jelmer> czajkowski: :)
[07:43] <jelmer> lifeless: that does help, thanks
[07:48] <lifeless> czajkowski: how is bug 996537 a dupe of bug 68997 ?
[07:48] <_mup_> Bug #996537: Apport-Collect post many times and spams many times <apport-bug> <i386> <precise> <Launchpad itself:Confirmed> <apport (Ubuntu):Invalid> < https://launchpad.net/bugs/996537 >
[07:48] <_mup_> Bug #68997: kwalletmanager icon not in system tray, unstartable from menu <kdeutils (Ubuntu):Fix Released> < https://launchpad.net/bugs/68997 >
[07:49] <lifeless> czajkowski: I'm guessing you dropped a digit ..
[07:54] <adeuring> good morning
[07:57] <czajkowski> lifeless: just running into the tower bbiab
[07:57] <lifeless> czajkowski: don't hurt yourself :)
[08:21] <czajkowski> lifeless: I didn't cant say the same about the two cyclists outside and a caby
[08:21] <lifeless> czajkowski: ?!
[08:21] <czajkowski> I didnt hurt myself, there was a cycle crash outside millbank on the way in
[08:23] <lifeless> ah
[08:53] <StevenK> wgrant: Heh, and I would have expected you to revert.
[08:59] <wgrant> StevenK: The fix was trivial and restricted to the one test, so there was no reason to.
[08:59] <wgrant> I don't revert pointlessly :)
[09:00] <jamestunnicliffe> Hi, I have a buildbot fail (https://lpbuildbot.canonical.com/builders/lucid_lp/builds/2077) but I don't have permission to view the error. Can someone paste it somewhere for me?
[09:01] <wgrant> jamestunnicliffe: There was a trivial bug in a test, which I landed a fix for.
[09:02] <wgrant> -        item = self.MockWorkItem(False)
[09:02] <wgrant> +        item = self.MockWorkItem(False, False)
[09:02] <wgrant> Should be through buildbot in about 90 minutes
[09:02] <jamestunnicliffe> wgrant: Ah, cool. Thanks.
[09:06] <czajkowski> gmb: top of the morning to you
[09:07] <gmb> Morning czajkowski :)
[11:40] <cjwatson> How long should I expect it to take to mirror branches onto http://bazaar.launchpad.net?
[11:40] <czajkowski> jelmer: ^^
[11:40] <mgz> it should be instant, but it's lagging.
[11:40] <cjwatson> I've got one that I pushed some minutes ago that's available on bzr+ssh but not http yet.  Wondering if this is normal and I should just be more patient, or if it's a problem.
[11:40] <cjwatson> Right.  Anything to do with the replication lag?
[11:40] <jelmer> cjwatson, czajkowski: they should start pretty much instantly
[11:41] <cjwatson> It's lp:~ubuntu-archive/ubuntu-archive-publishing/trunk if that matters.
[11:41] <jelmer> https://code.launchpad.net/+code-imports/+machines
[11:41] <cjwatson> Oh, I didn't realise that went through the code import machinery.
[11:41] <jelmer> cjwatson: ah, that's just the scanning you're waiting for
[11:42] <jelmer> cjwatson: so it's indeed unrelated; there is no mirrorring involved for hosted branches
[11:42] <mgz> last count db replication was three hours behind
[11:42] <mgz> and a branch I pushed at 9:40 is still not visible over http
[11:46]  * cjwatson upgrades dogfood for some QA
[12:40] <cjwatson> OK, I've caused /srv/launchpad.net/publisher-parts/ubuntu/ to exist on both cocoplum and mawson.  Can somebody change run_parts_location to /srv/launchpad.net/publisher-parts in ftpmaster-publish/launchpad-lazr.conf and dogfood-publish/launchpad-lazr.conf?
[12:40] <cjwatson> I chose the path to match what's used in derived-distro-publish/launchpad-lazr.conf.
[12:43] <cjwatson> (Though, hmm, my import seems a bit buggy.  Fixing up by hand.)
[12:48] <wgrant> cjwatson: Did you bzr split it?
[12:48] <cjwatson> No, I used fastexport/import so I wouldn't have to carry LP's entire history around
[12:49] <wgrant> Ah
[12:49] <cjwatson> But it seemed to get confused and miss a few revisions at the end.  I recommitted them by hand.
[13:01] <jamestunnicliffe> Hi, so I have got two changes that have landed on qastaging. While they are both functionally correct, there is a bug when they run together such that something correct, but slightly unexpected happens.
[13:01] <jamestunnicliffe> The easiest thing to do would be land the changes and then submit a new bug fix (it is trivial)
[13:01] <jamestunnicliffe> The alternative is back out both changes and create 1 change for both.
[13:01] <jamestunnicliffe> Opinions?
[13:02] <wgrant> Does it prevent them from being released to production?
[13:02] <jamestunnicliffe> wgrant: I don't think so. https://qastaging.launchpad.net/~linaro-infrastructure/+upcomingwork shows the page that changed.
[13:03] <wgrant> Then it doesn't matter -- do whatever is easiest.
[13:03] <wgrant> All that matters is that everyone else is not blocked :)
[13:03] <jamestunnicliffe> wgrant: Cool. Will tag the changes as OK then get a new fix in ASAP.
[13:04] <wgrant> Great.
[14:27] <czajkowski> oh great fire alarm on
[14:48] <rick_h_> sinzui: great blog post!
[14:48]  * rick_h_ does a happy dance with the idea of all that cruft removed/cleaned up
[14:49] <sinzui> thank you rick_h_, the only developer to have added a sensible browser guard to the code base
[14:50] <sinzui> rick_h_, I am really glad it is over. I packed away the sd card with the vm image. I now have full use of my computer. I can read mail and hack
[14:51] <rick_h_> hah, yea. I was dreading that the job was heading my way and I'd have to reinstall the windows VM. Last time I had to create a usb install disk from the cd etc ugh!
[14:51] <rick_h_> my goal in life is that the next laptop I get will support esata booting, and I'll just toss the drive it comes with into an external sata enclosure for IE testing
[14:54] <sinzui> I really wish winetricks had created a true IE8 or IE9 browser. That was painless to install, but cookies and https rules are broken.
[14:54] <rick_h_> yuck, but yea guess I can see the https part at least. So much backend stuff to make that work
[14:55] <sinzui> The whole security model was broken. You could not log into Ubuntu SSO and state with cookies was out of the question.
[15:13] <rick_h_> gmb: just want to check if you're peeking at either of the rewviews to avoid duping?
[15:14] <gmb> rick_h_: I peeked at wallyworld's, saw the line count, thought better of it, and I don't have enough domain knowledge to safely review cjwatson's.
[15:14] <rick_h_> ok
[15:21] <nigelb> sinzui: Great glob post.
[15:21] <nigelb> err
[15:21] <nigelb> *blog post
[15:22] <mgz> glob is a much better word.
[15:22] <sinzui> glob is correct
[15:22] <nigelb> glob is also a person I talk to :P
[15:22] <nigelb> (on another network)
[15:22] <gmb> nigelb: Never fear; this is Launchpad. Speling is optional.
[15:23] <nigelb> gmb: Heh :)
[15:28] <allenap> Anyone fancy doing a review for postgresfixture? https://code.launchpad.net/~allenap/postgresfixture/run-with-command/+merge/106830
[15:29] <rick_h_> allenap: on the list. Getting backed up atm
[15:29] <allenap> rick_h_: Cool, thanks.
[15:31] <gmb> allenap, rick_h_ I'll take it and holler if I've any domain-specific questions.
[15:31] <allenap> gmb: Thank you :)
[15:32] <gmb> allenap: Although there should be a check on Merge Proposals: assert commit_text != description. ;)
[15:32] <allenap> gmb: Sorry, not a very good description I know. It's a fairly simple change though, but I'll add more if you want?
[15:33] <gmb> allenap: No, it's fine. I'll ask questions as I go if necessary :)
[15:35] <jamestunnicliffe> gmb / rick_h_: Very quick review request (5 lines) if you want a not very involved one to tick off: https://code.launchpad.net/~dooferlad/launchpad/postponed-is-done-wi-auto-open-fix/+merge/106836
[15:35] <gmb> jamestunnicliffe: I'll take a look when I've finished with allenap's.
[15:35] <jamestunnicliffe> gmb: Thanks.
[15:37] <allenap> gmb: Ah, to try it out: make && bin/python -m postgresfixture --help
[15:37] <gmb> allenap: Ta.
[15:42] <gmb> allenap: Scarily, it makes perfect sense to me. r=me.
[15:43] <allenap> gmb: \o/ thanks
[15:43] <gmb> jamestunnicliffe: Are you offsetting your +6 lines against danilos's megabranch (we should track this somewhere)?
[15:44] <danilos> gmb, it's all tracked already :)
[15:44] <gmb> Ah, cool.
[15:44] <gmb> Where?
[15:44] <danilos> gmb, https://docs.google.com/a/linaro.org/spreadsheet/ccc?key=0AvOsYPy8e7yUdGkyRmx2WGFwT3NnSjdHVW04Q1pvSmc
[15:45] <gmb> danilos: Ah, great. Now, what we really need is to have that tracked in dev.launchpad.net - or at least linked to therefrom.
[15:46] <danilos> gmb, I thought I added it to https://dev.launchpad.net/Projects/WorkItems/, but I guess not
[15:48] <danilos> gmb, it's there now
[15:48] <gmb> Thanks.
[15:54] <gmb> jamestunnicliffe: r=me
[16:13] <jamestunnicliffe> gmb: I don't have merge privs, so that is another job for you :-|
[16:14] <rick_h_> thanks for the catch up gmb
[16:14] <rick_h_> jamestunnicliffe: which branch is this? I'll submit to ec2 if needed
[16:15] <jamestunnicliffe> rick_h_: lp:~dooferlad/launchpad/postponed-is-done-wi-auto-open-fix
[16:15] <rick_h_> jamestunnicliffe: ok, pulling down to submit
[16:15] <jamestunnicliffe> rick_h_: Thanks!
[16:16] <sinzui> abentley, ping
[16:16] <rick_h_> jamestunnicliffe: can you set a commit msg in there please?
[16:17] <jamestunnicliffe> rick_h_: The descriptive merge message was at rev 15283. Am I missing something?
[16:19] <rick_h_> jamestunnicliffe: in your MP there's a field for setting a commit message which gets used when things are merged together
[16:19] <sinzui> jamestunnicliffe,  visit https://code.launchpad.net/~dooferlad/launchpad/postponed-is-done-wi-auto-open-fix/+merge/106836
[16:19] <gmb> rick_h_: Thanks for taking care of that merge.
[16:20] <rick_h_> gmb: np
[16:20] <jamestunnicliffe> rick_h_, sinzui: Got it.
[16:20] <gmb> Why did I put that hyphen in there?
[16:20] <rick_h_> jamestunnicliffe: ty much
[16:23] <jamestunnicliffe> rick_h_: Is this supposed to include [r=??]
[16:23] <rick_h_> jamestunnicliffe: no. It's just the commit message about the change that will show in the bzr log of trunk
[16:23] <jamestunnicliffe> rick_h_: OK
[16:23] <rick_h_> and should be more of what changed vs the bug report
[16:24] <rick_h_> So something like "auto expand work items list considers postponed as incomplete work"
[16:24] <jamestunnicliffe> rick_h_: Done.
[16:25] <rick_h_> awesome
[16:28] <lifeless> cjwatson: the post-restore fixup code is in losa scripts, AFAIK
[16:33] <rick_h_> jcsackett sinzui either of you avail for sanity check of https://code.launchpad.net/~cjwatson/launchpad/archive-build-score-api/+merge/106804
[16:33] <rick_h_> it seems ok, but mentions some security sanity checking and I start to get nervous
[16:34]  * sinzui looks
[16:35] <rick_h_> jamestunnicliffe: so your branch is off to the land of ec2. Should get an email in 6hrs-ish if things failed at all/passed
[16:35] <jamestunnicliffe> rick_h_: thanks
[16:37] <cjwatson> lifeless: anything that's mortal-visible?
[16:40] <czajkowski> sinzui: great blog post
[16:42] <sinzui> rick_h_, the branch is good.
[16:42] <rick_h_> sinzui: ok, thanks for the confirmation
[16:42] <sinzui> I am going to comment about another bug that I think Collin to tell us if it is fixed, if if he knows how to fix it
[16:44] <rick_h_> sounds like a plan to me
[16:51] <cjwatson> done
[16:58] <cjwatson> sinzui: think I'm OK to go ahead and land?  I don't think I can do better at that description myself (I'm fine at explaining to people who know the field, less so to people who don't)
[16:59] <cjwatson> and it basically looks like wgrant sorted it out a while back but didn't see that bug
[16:59] <sinzui> cjwatson, I agreed. thank you
[17:00] <lifeless> cjwatson: possibly; nag someone in -ops perhaps
[17:16] <abentley> sinzui: pong
[17:17] <sinzui> hi abentley. I just sent an email to the launchpad-developers list asking if celery is ready in production. maybe you want to reply to it since I had a few questions
[17:23] <abentley> sinzui: I don't see it yet, but I'll reply when I do.
[17:23] <sinzui> abentley, thank you. I curse our slow mailman
[22:14] <sinzui> wallyworld_, lib/lp/bugs/tests/test_doc.py
[22:58] <sinzui> wallyworld_, wgrant, StevenK: this is one of my favorite bugs https://bugs.launchpad.net/launchpad/+bug/568768
[22:58] <_mup_> Bug #568768: Bug UI Inconsistancy <confusing-ui> <javascript> <lp-bugs> <ui> <Launchpad itself:Triaged> < https://launchpad.net/bugs/568768 >
[23:59] <rick_h_> sinzui: that's a good one