[00:57] <lifeless> wgrant: what techniques do you use to debug fails-in-ec2 branches?
[00:58] <wgrant> lifeless: A test or two that only fail in ec2?
[00:59] <lifeless> a fixture
[00:59] <lifeless> rabbitmq
[01:04] <wgrant> Ah, that one.
[01:04] <wgrant> I've previously used ec2 demo and run the relevant tests manually.
[01:26] <wgrant> lifeless: r12987
[01:26] <wgrant> It seems to now touch the config dir in twistedsftp, but it's twistedftp that uses GPGHandler.
[01:26] <wgrant> Am I missing something, or is that wrong?
[01:30] <lifeless> jml: if you reopen something, please triage it - avoids double-work
[01:32] <lifeless> wgrant: lib/lp/poppy/twistedsftp.py isn't the new poppy?
[01:34] <wgrant> lifeless: I guess it will still work, since they run in the same process.
[01:34] <wgrant> But twistedftp and twistedsftp are two fairly independent things that live in the same host.
[01:34] <wgrant> twistedsftp doesn't use GPGHandler.
[01:34] <wgrant> (yet)
[01:36] <LPCIBot> Project db-devel build #528: FAILURE in 5 hr 12 min: https://lpci.wedontsleep.org/job/db-devel/528/
[02:24] <wgrant> Hmm.
[02:24] <wgrant> So it's not good :(
[02:26] <wgrant> lifeless: I think this is going to leak. A SFTPServer is instantiated for each authenticated SFTP session, so: A) it doesn't touch until the first SFTP user authenticates, despite it being required for FTP, and B) each new session will create the LoopingCall, and they won't be destroyed until the end.
[02:26] <wgrant> Right?
[02:28] <lifeless> yes, thats faulty
[02:28] <wgrant> I think it should be moved into the tac.
[02:28] <wgrant> I guess wallyworld's not back yet.
[02:29] <lifeless> that might work better, yes
[02:34] <jtv> salutations fellow carbon-based lifeforms
[02:38] <spm> yo jtv
[02:39] <jtv> yo indeed, spm!
[02:40] <StevenK> jtv: Still in Cali?
[02:43] <jtv> StevenK: no, back but pushing the timezones hard.  Got home around 8 hours ago.
[02:43] <StevenK> jtv: Nice!
[02:43] <jtv> I'm not sure I'll think so in a few hours.  :)
[02:44] <StevenK> Haha. :-)
[02:44] <jtv> By the way, travel tip: if you're going to find out that your bookings have been all screwed up, checkin around 04:00 (yes, AM) on a super-security-conscious flight is a bad idea.
[02:44] <jtv> Sorry, a bad idea *for a time to find out* that you're not actually booked on your "confirmed" flight.
[02:45] <jtv> Also, if you do manage to get past that and out on a flight that actually leaves slightly earlier and you get to your hotel on the other side of the world, do not joke that your hotel reservation may be equally messed up because it just might turn true.
[02:46] <jtv> Speaking of super-security-conscious, both flights happened after the OBL thing yet I was only run through the give-us-a-look-under-those-clothes-then on the way _out_.
[02:47] <jtv> Which earned me a detailed patdown because my shorts' lower-left pocket was designed with slightly too much cloth, and thus showed up as a suspicious Something Potentially Hidden on the scan.  Come on.
[02:48] <jtv> And I stood in line for half an hour for _that_ ride?
[02:49] <jtv> Also, do not believe the helpful signs about the security checks when changing planes in Tokyo.
[02:56] <jtv> A quiet connection still dies regularly, but at least I seem to get better bandwidth than in Silicon Valley.
[03:08] <lifeless> jtv: so did I spot a bug then?
[03:10] <jtv> lifeless: I think so, yes.  Arguably a "soft" one in that it may still be nice to know about the TMs that aren't filtered out, but only arguably.
[03:10] <jtv> (Note by the way that, alack, we do not expect any useful selectivity from this extra filter)
[03:11] <lifeless> that was my next question :)
[03:11] <jtv> "No, sorry, you can't use the extra clause for optimization."  :-)
[03:28] <wgrant> jtv: Hm, why does Julian want them all generated separately?
[03:28] <wgrant> There is no reason to do that at present.
[03:28] <wgrant> It may perhaps be useful when we can add extra suites.
[03:28] <jtv> I suppose, yes.  Or when the script fails.
[03:29] <jtv> Doesn't _solve_ anything as such, but gives us a more fine-grained do-over.
[03:36] <jtv> wgrant: I'm slightly distracted atm (unpleasant family news) but let me just look up what exactly Julian said.
[04:07] <jtv> wgrant: I can't find any trace of the discussion about doing the indexes per suite.  May have been otp.
[04:10] <jtv> wgrant: btw have you been able to run the new (python) publish-ftpmaster through its paces recently?  It should be in pretty good shape now and I'm eager to sign off on the basic changeover.
[04:11] <wgrant> jtv: OK. Was just wondering if I was missing something, but I guess the extra flexibility is useful.
[04:11] <wgrant> I haven't run it in a couple of weeks, but it hasn't changed much lately, right?
[04:11] <jtv> Not really, apart from the built-in index creation.
[04:11] <wgrant> Right.
[04:11] <jtv> It all hinges on the magnitude of "a couple" of weeks.
[04:12] <wgrant> We are limited to partial fake runs now.
[04:12] <wgrant> Since DF only has 120GB free with the new DB.
[04:12] <wgrant> Which might just be enough to squeeze one series in.
[04:12] <wgrant> If we are lucky.
[04:13] <jtv> So what is it that happened to the DB?
[04:13] <wgrant> We restored it over the weekend.
[04:13] <wgrant> So it is now fresh and clean.
[04:13] <wgrant> But some dozens of gigabytes larger.
[04:13] <wgrant> We are considering purging branchrevision.
[04:13] <wgrant> To save a hundred gigabytes or so.
[04:14] <jtv> I wonder how stub's been getting on with the id-dropping.
[04:15] <wgrant> That will only save like 25GB, but at least it's something.
[04:15] <wgrant> Hmmm, plus indices.
[04:15] <wgrant> So maybe more.
[04:16] <wgrant> In fact it's involved in two indices.
[04:16] <wgrant> So it's not an unsubstantial saving.
[04:37] <lifeless> spm: yo
[04:37] <lifeless> spm: we have a spammer
[04:37] <spm> yo
[04:37] <spm> probably
[04:37] <lifeless> bugs 423094,5,6
[04:38] <lifeless> possibly more
[04:38] <wgrant> That's no longer LOSAish.
[04:39] <lifeless> oh
[04:39] <wgrant> Suspended.
[04:39] <lifeless> ok, so wgrant, we have a spammer ;>
[04:39] <lifeless> but some of the bugs are u1, you may not be able to see them all
[04:39] <wgrant> Indeed, there are eight tasks involved, but I can only see 7.
[04:40] <wgrant> We need a good solution to this.
[04:41] <lifeless> indeed
[04:41] <wgrant> I can see 3,4,5,6,7. 2 and 8 haven't been spammed.
[04:41] <wgrant> Maybe there's a deactivated project here somewhere.
[04:42] <wgrant> spm: How many tasks do you see at the URL in https://pastebin.canonical.com/47251/?
[04:43] <spm> 7
[04:43] <wgrant> Hmm. Why are there 8 karma entries, then :(
[04:43] <wgrant> Thanks.
[04:43]  * wgrant cleans them up.
[04:43] <lifeless> wgrant: 2 on one bug
[04:43] <wgrant> They're not normally that stupid.
[04:43] <lifeless> hey, I got the mails
[04:46] <LPCIBot> Project windmill-devel build #50: STILL FAILING in 1 hr 7 min: https://lpci.wedontsleep.org/job/windmill-devel/50/
[04:46] <wgrant> Anyway, all gone now.
[05:57] <wgrant> lifeless: Does Storm do DISTINCT ON?
[05:57] <wgrant> I can't see any code for it.
[06:01] <lifeless> wgrant: yes
[06:01] <lifeless> wgrant: well, we can make it do it.
[06:01] <lifeless> its -ffffugly- but works
[06:01] <lifeless> wgrant: grep for 'DISTINCT ON'
[06:02] <wgrant> No references. I'm thinking that SQL() in a Select's columns attribute should do it.
[06:02] <StevenK>                 # XXX: GavinPanella 2010-09-17 bug=374777: This SQL(...) is a
[06:02] <StevenK>                 # hack; it does not seem to be possible to express DISTINCT ON
[06:02] <StevenK>                 # with Storm.
[06:02] <wgrant> Oh, in our tree. RIght.
[06:03] <wgrant> Indeed, that's how it's done. That is indeed fugly, but it seems to work.
[06:03] <lifeless> wgrant: your grep foo is weak
[06:03] <wgrant> lifeless: I thought you meant storm.
[06:03] <lifeless> heh no sorry
[06:04] <wgrant> It seems like this is a fairly easy Storm fix, though.
[06:05] <wgrant> The syntax is even obvious: .config(distinct=(Foo.bar,))
[06:54] <lifeless> :( bug triage
[06:54] <wgrant> We're on it this week, but I haven't gone there yet.
[06:55] <wgrant> StevenK: Bug #779406 is yours.
[06:55] <_mup_> Bug #779406: Staging rollout for revno 10513 failed <canonical-losa-lp> <Launchpad itself:New> < https://launchpad.net/bugs/779406 >
[06:55] <StevenK> Oh damn it
[06:55] <wgrant> DSD parent stuff doesn't handle existing DSDs.
[06:55] <wgrant> Won't affect production yet, but will probably tomorrow.
[06:56] <StevenK> We don't have any DSDs in production?
[06:56] <wgrant> Julian wants some this week.
[06:56] <wgrant> For UDS.
[06:56] <wgrant> I believe.
[06:56] <StevenK> Hmmmmm, this is un-good
[06:57] <wgrant> Yes.
[06:57] <wgrant> Perhaps we can coerce him to use staging instead.
[07:31] <lifeless> jml: so when shall we talk ?
[07:31] <lifeless> poolie: -very- early draft, and incomplete because we started this call - https://dev.launchpad.net/ArchitectureGuide/Services
[07:41] <poolie> jml: ha ha
[08:46] <lifeless> jml: I am back, and at your service
[08:53] <bigjools> morning all
[08:54] <wgrant> Morning bigjools.
[08:54] <wgrant> Dogfood is alive!
[08:54] <bigjools> hmmn?
[08:54] <bigjools> I hope you didn't start it
[08:55] <bigjools> you've bloody started it
[08:55] <wgrant> I did, after fixing permissions. I needed it for QA to fix poppy. But it turned out it was bad so it didn't matter :/
[08:55] <wgrant> Should I not have?
[08:55] <bigjools> well my screen is full of errors
[08:55] <wgrant> Right.
[08:56] <wgrant> The slony fix script is fucked.
[08:56] <bigjools> and various bits of data need fixing
[08:56] <wgrant> But rerunning with -U postgres worked fine.
[08:56] <bigjools> like the builders
[08:56] <wgrant> Then ran security.py/upgrade.py and ran the builder fixer.
[08:56] <wgrant> Yeah.
[08:56] <bigjools> you fixed the b iuolders?
[08:56] <wgrant> Of course.
[08:56] <wgrant> I don't like destroying production much :)
[08:56] <wgrant> I followed the wiki page.
[08:56] <bigjools> ok now we need production chroots re-uploading to DF
[08:57] <wgrant> It's OK for now, but yes, we should do that at some point before lamont replaces them.
[08:57] <wgrant> I tested it and builds and uploads all work fine.
[08:57] <wgrant> And don't demolish prod.
[08:58] <bigjools> ok thanks
[08:58] <bigjools> did you happen to notice when it finished updating?
[08:58] <wgrant> It spent most of Sunday turning the foreign keys back on.
[08:58] <wgrant> (yes, that took more than 12 hours)
[08:58] <wgrant> But let's see...
[08:59] <StevenK> Sunday night, when wgrant pinged me with excitment
[08:59] <StevenK> Well, .au Sunday night
[08:59] <wgrant> Yeah. It finished some time Sunday afternoon.
[08:59] <wgrant> When I checked on Sunday morning it was very slowly ALTER TABLEing.
[09:00] <wgrant> And then it became clear that it hadn't gone further than pg_restore :/
[09:02] <LPCIBot> Project windmill-db-devel build #252: STILL FAILING in 1 hr 7 min: https://lpci.wedontsleep.org/job/windmill-db-devel/252/
[09:03]  * bigjools chuckles at jml's response on bug 778437
[09:03] <_mup_> Bug #778437: Recipe build success sends emails, please stop doing that <email> <recipe> <Launchpad itself:Triaged> < https://launchpad.net/bugs/778437 >
[09:03] <adeuring> good morning
[09:18] <mrevell> Morning
[09:22] <lifeless> jml: is now good? Its getting late is all
[09:27] <lifeless> bah, hes probably not actually connected... /me starts phone calls instead
[09:29] <beuno> lifeless, we're still in the plenary
[09:29] <lifeless> beuno: oh
[09:29] <beuno> and kiko just saked if you where here
[09:30] <beuno> *asked
[09:30] <lifeless> beuno: I thought it was over already
[09:30] <beuno> something about a pie in the face  :)
[09:30] <lifeless> beuno: hah :)
[09:30] <beuno> lifeless, yeah, it was suppose to
[09:30] <lifeless> beuno: he wants another? :P
[09:30] <beuno> lifeless, he just challenged everyone in Linaro
[09:30] <beuno> so eems so
[09:33] <wgrant> Which naysayer shall be pied this time?
[09:34] <lifeless> beuno: and what was the challenge?
[09:49] <stub> Ahh... so I suspect if a request creates a notification, and the request needs to be retried due to a conflict, the retried request will create a second notification and both will be displayed. Because the session machinery isn't tied to the transaction machinery (by design).
[09:52] <lifeless> stub: sounds plausible
[09:53] <lifeless> poolie: hi; I was trying to locate jml before it gets late late here
[10:04] <jtv> henninge, can I trouble you for a review?  https://code.launchpad.net/~jtv/launchpad/bug-779701/+merge/60346
[10:08] <henninge> jtv: Hi! I don't like being troubled ... ;-)
[10:08] <henninge> jtv:  ... but maybe it's fun?
[10:10] <lifeless> beuno: the keynote is over now, right ?
[10:11] <beuno> lifeless, it is
[10:11] <beuno> everybody is running around to sessions
[10:17] <henninge> jtv: just in case I was unclear: I am taking it. ;)
[10:18] <jtv> henninge: thanks!
[10:18]  * mwhudson blinks
[10:18] <mwhudson> is lib/contrib/oauth.py in the launchpad tree used ?
[10:19] <mwhudson> there is also oauth = 1.0 in versions.cfg
[10:20] <mwhudson> ok, the answer is yes
[10:20] <wgrant> Yes, contrib.oauth is used explicitly.
[10:20] <lifeless> poolie tried to remove it
[10:21] <wgrant> poolie tried to fix that.
[10:21] <lifeless> it went boom
[10:21] <wgrant> But there were test failures.
[10:21] <lifeless> shakalaka
[10:21] <wgrant> Curse you lifeless.
[10:21] <lifeless> boom
[10:21] <lifeless> shakalaka
[10:21] <lifeless> boom
[10:21] <lifeless> :)
[10:21] <mwhudson> wtf
[10:21] <bigjools> shake the room
[10:21] <mwhudson> this whole area is littered with unmaintained forks
[10:21] <mwhudson> afaicty
[10:21] <StevenK> lifeless: If you keep that up, I'm flying to NZ to silence you
[10:21] <wgrant> BeautifulSoup?
[10:21] <wgrant> mwhudson: Most of them are gone.
[10:22] <lifeless> StevenK: bring it :)
[10:22] <wgrant> It used to be much worse.
[10:22] <mwhudson> wgrant: do you mean in our stuff, or in general?
[10:22] <wgrant> Oh, BeautifulSoup isn't in contrib.
[10:22] <wgrant> It's straight in lib/
[10:22] <lifeless> -win-
[10:22] <wgrant> Our stuff.
[10:22] <mwhudson> wgrant: right, i meant in general
[10:25] <wgrant> mwhudson: I like to ignore the general Python situation.
[10:25] <wgrant> It is too upsetting.
[10:29] <stub> I'm just amazed all the people who know the one-hit wonder from 1987.
[10:30] <jml> lifeless: hi
[10:30] <mwhudson> wgrant: i don't think it's just python is it?  there are effectively unmaintained forks of the oauth *spec* aiui
[10:31] <wgrant> mwhudson: Hmm, I've not heard about that.
[10:31] <wgrant> And I mean Python in general, not just OAuth :)
[10:32] <mwhudson> maybe i'm exaggerating
[10:33] <mwhudson> certainly, lib/contrib/oauth.py does not implement http://tools.ietf.org/html/rfc5849
[10:36] <wgrant> mwhudson: It's not OAuth 2.0?
[10:36] <wgrant> I don't know.
[10:40] <mwhudson> wgrant: no, i don't think so, i think it's an older version of oauth 1
[10:41] <wgrant> Awesome.
[11:11] <henninge> jtv: why only FROZEN distroseries?
[11:12] <jtv> henninge: this is an initialization step, and initialization of distroseries happens in FROZEN state.
[11:12] <jtv> (It's a bit odd, and the series will also go into that state later on prior to release)
[11:12] <henninge> indeed
[11:12] <jtv> But this way we'll only have to fake those files for one Ubuntu release, and the rest will just continue as they were.
[11:15] <henninge> jtv: I'd probably replace "series = self.factory.makeDistroSeries(status=SeriesStatus.FROZEN)" with a small factory method on the test case which could then have a comment explaining why.
[11:17] <jtv> henninge: come to think of it, I may already have one... let me check
[11:17] <wgrant> What's with the status restriction?
[11:19] <henninge> jtv: is this removing duplicates?
[11:19] <henninge> self.assertEqual(sorted(markers), sorted(list(set(markers))))
[11:19] <jtv> Yes
[11:19] <henninge> jtv: how would that test fail?
[11:20] <jtv> Duplicates or missing items.
[11:20] <henninge> this test is to show that it "_uses_separate_files_per_suite".
[11:20] <henninge> jtv: I don't understand how it show that.
[11:21] <gmb> adeuring: I made the changes we talked about on Friday, could you take another look at https://code.launchpad.net/~gmb/launchpad/bug-772609/+merge/60176 for me?
[11:21] <adeuring> gmb: sure
[11:21] <gmb> Thanks
[11:21] <jtv> henninge: the test looks up the filenames for the respective suites, then shows that there are no duplicates.
[11:39] <adeuring> gmb: r=me
[11:40] <gmb> Thanks
[11:41] <adeuring> gmb: just one minor question: I assume that the "import BugNotificationRecipients" in personIsAlsoNotifiedSubscriber is necessary to avoid a circular import?
[11:42] <gmb> adeuring: Ah, yes. I'll add a comment to that effect.
[11:42] <adeuring> gmb: great., thanks!
[11:49] <bigjools> wgrant: can you remember when the fix went out to unembargo changelogs properly?
[11:50] <wgrant> bigjools: A while ago, but I've had production data fixed since then... why?
[11:50] <wgrant> I can dig up dates if it's important.
[11:50] <bigjools> wgrant: I've got one that was created 2011-04-21 but is marked restricted
[11:51] <bigjools> it sounds close to the date but ...
[11:53] <wgrant> bigjools: I last fixed the data on April 25, around this time.
[11:53] <wgrant> bigjools: Can you see when the pub was created?
[11:54] <bigjools> wgrant: https://launchpad.net/ubuntu/+source/gmp/2:5.0.1+dfsg-7ubuntu2
[11:54] <wgrant> Reading logs I now remember that I misremembered.
[11:54] <bigjools> so uploaded to the P3A on 21st, unembargoed on 28th
[11:54] <wgrant> 20:26 < wgrant> (I guess cocoplum is still running the bad code, so more bad data is being created :()
[11:55] <wgrant> So we need to fix it up one more time.
[11:55] <wgrant> LOSA ping.
[11:55] <bigjools> staging has the bad data too ...
[11:55] <bigjools> wgrant: wrong channel
[11:55] <bigjools> :)
[11:58] <wgrant> Oh good, staging unbroke.
[12:04] <jtv> bigjools: I should test index creation on staging now, not df, right?
[12:05] <bigjools> jtv: no, DF is still the place unfortunately.
[12:05] <jtv> Is it in a shape where I can create a series & run publish-ftpmaster?
[12:06] <bigjools> jtv: yes, should be ok
[12:06] <jtv> There I go then.  It's been too long since I wrecked a server.
[12:27] <lifeless> bigjools: wgrant: fyi - https://dev.launchpad.net/ArchitectureGuide/Services
[12:27] <lifeless> its about 2/3rds written up
[12:28] <wgrant> There are not enough \o/s in the universe.
[12:59] <bigjools> lifeless: sweet
[13:12] <LPCIBot> Yippie, build fixed!
[13:12] <LPCIBot> Project db-devel build #529: FIXED in 5 hr 30 min: https://lpci.wedontsleep.org/job/db-devel/529/
[13:13] <poolie> StevenK: hi
[13:26] <lifeless> gary_poster: stub: statik: https://dev.launchpad.net/ArchitectureGuide/Services may interest you
[13:27] <lifeless> gary_poster: I haven't discussed this with you at all as you were away last week :) - but if you can spare time near the end of your day (after I've slept) I've love to do so
[13:27] <poolie> bigjools: hi, thanks for commenting on the bug
[13:28] <gary_poster> lifeless, very cool!  I'd love to talk about it.  I'm actually taking today and tomorrow off as well--I'm just here to make sure my squad is all set for the next two days, then I'm disappearing.  I'll take a more thorough glance at the page later.  It would be an exciting initiative.
[13:28] <bigjools> poolie: np - I am responsible for a lot of the last changes to the email code, hence its hideousness but at the same time I understand what is supposed to happen :)
[13:29] <poolie> i'm not sure if you mean "so don't just cut it out"
[13:29] <poolie> or if it's just purely a reminder for the future
[13:30] <lifeless> gary_poster: kk; well - I am looking for critical analysis and for problems/benefits on both current and possible layout
[13:30] <gary_poster> lifeless, ok, I'll think in that direction.
[13:31] <lifeless> gary_poster: this is a synthesis of many ideas that have been floated by many folk : I want to get the best possible chance of fixing real issues and not adding terrible new ones :)
[13:31] <StevenK> poolie: Sorry, I was having dinner.
[13:31] <gary_poster> lifeless, I hear you :-)
[13:31] <lifeless> gary_poster: the page is only about 2/3 written - I have another page of notes and concepts to turn into prose
[13:32] <poolie> np
[13:32] <lifeless>  - details about how we might do it, possible migration strategy, making each stage pay for itself sort of thing
[13:33] <StevenK> poolie: Just doing a quick grep turns up test_handleStatusNotifies in lib/lp/code/model/tests/test_sourcepackagerecipebuild.py -- I can certainly see that one failing.
[13:37] <poolie> StevenK: that test looks a lot to me like it's asserting there are 0 notifications from ok builds
[13:39] <StevenK> poolie: Which is, in fact, backward. :-(
[13:44] <poolie> well, i don't understand why it was passing already
[13:44] <poolie>  i wondered if it was checking for some other kind of notification
[13:44] <poolie> is there any?
[13:46] <wgrant> I haven't read the test, but it's possibly checking for the upload notification.
[13:46] <wgrant> (you're suppressing the build notification)
[13:47] <poolie> that seems likely
[13:47] <poolie> makeSourcePackageRecipeBuildJob
[13:49] <wgrant> Hm, no, that's a build notification.
[13:49] <wgrant> Interesting.
[13:50] <poolie> perhaps it doesn't actually reach the point it would be fired
[13:50] <poolie> i can trace into it and try to get it to fail with the old code
[13:52] <wgrant> poolie: I guess it's too asynchronous now.
[13:52] <wgrant> poolie: handleStatus is meant to be called by Twisted.
[13:52] <wgrant> poolie: It returns a deferred.
[13:56] <deryck> Morning, all.
[13:56] <StevenK> wgrant, poolie: In either case, the test is wrong, and should be removed, no?
[14:00] <poolie> wgrant: oh, that could interact quite badly with a test that's looking for side effects
[14:00] <wgrant> poolie: Just a bit.
[14:00] <wgrant> Well, looking for absence of side-effects without checking that it actually finished.
[14:01] <poolie> stevenk if it's almost right, perhaps it should be updated to fail, then updated to pass again
[14:01] <deryck> henninge-lunch, ping for standup
[14:06] <LPCIBot> Project windmill-db-devel build #253: STILL FAILING in 1 hr 9 min: https://lpci.wedontsleep.org/job/windmill-db-devel/253/
[14:09] <henninge> deryck: sorry, forgot the time ... and my notebook
[14:09] <deryck> henninge, no worries.  we're done now though.
[14:09] <henninge> deryck: ok
[14:10] <deryck> henninge, the card for bug 196679 is still in progress on the board, but that's ready for deployment, yes?
[14:10] <_mup_> Bug #196679: ValueError on +language-packs <lp-translations> <oops> <qa-untestable> <Launchpad itself:Fix Committed by henninge> < https://launchpad.net/bugs/196679 >
[14:10] <henninge> deryck: I was actually able to reproduce the bug on Friday so I am able to provide an actual fix for the bug.
[14:11] <henninge> deryck: that's straight forward but I have not yet returned to it today.
[14:11] <henninge> deryck: will be finished today, though.
[14:11] <deryck> henninge, ok.  you're OCR today, so no expectations from me about finishing it.
[14:11] <deryck> henninge, just wanted to sort out that board
[14:21] <jml> do we support binonlynmu uploads?
[14:21] <jml> istr talking about it a while ago
[14:22] <bigjools> jml: no we don't
[14:22] <jml> bigjools: thanks. Do you know if there's a bug report for it?
[14:22] <bigjools> it's BinNMU if you want to get the terminology right :)
[14:22] <bigjools> ohhhh yes
[14:22] <bigjools> there's bugs all right
[14:23] <jml> bigjools: hmm, I'm sure I heard people hear say "only" :)
[14:23] <bigjools> https://bugs.launchpad.net/launchpad/+bug/245594
[14:23] <_mup_> Bug #245594: Rebuilds of binary packages without source changes <feature> <lp-soyuz> <motu> <Launchpad itself:Triaged> < https://launchpad.net/bugs/245594 >
[14:23] <jml> bigjools: thanks.
[14:25] <bigjools> jml: the crux is that people want to re-generate binaries for the already-uploaded source
[14:25] <bigjools> which implies an auto version bump for the binaries
[14:25] <jml> bigjools: makes sense.
[14:36] <sinzui> henninge: benji, will either of you have time to review https://code.launchpad.net/~sinzui/launchpad/anwers-api-0/+merge/60238
[14:37] <benji> sinzui: I can.
[14:38] <wgrant> sinzui: I have a couple of Answers API bugs that I plan to fix tomorrow, after finding them just now when hiding spam. Mostly fixing IQuestionSet['getByID']'s return type, exposing messages, and fixing setCommentVisibility's comment numbers so they match the ones shown in the UI.
[14:38] <wgrant> I'm not going to conflict with yours, am I>?
[14:38] <sinzui> fab
[14:39] <sinzui> wgrant: I do not think so. I exported answer contact management in this branch
[14:39] <wgrant> Great.
[14:39] <henninge> benji: Hi! I'll pick something else from activereviews.
[14:39] <wgrant> sinzui: I shall speak with you in 9 hours, I suppose.
[14:39] <wgrant> For the first time in a while.
[14:39] <benji> henninge: sounds good
[14:39] <henninge> rvba_: is either of your branches more urgent?
[14:40] <sinzui> wgrant: yes. I am sorry about last week. I fell ill late in my day and could not get off the sofa
[14:40] <henninge> rvba_: oh, one is a db patch. Not for me ;)
[14:40] <wgrant> sinzui: :( Better now?
[14:41] <sinzui> Yes, thanks
[14:41] <rvba_> henninge: the other is really not urgent ... but if you want to review it, thanks fine by me :)
[14:41] <rvba_> s/thanks/that's/
[15:21] <adeuring> deryck: could you please run these two queries on staging: https://pastebin.canonical.com/47268/ ?
[15:22] <deryck> adeuring, sure
[15:22] <adeuring> thanks!
[15:24] <deryck> first one still going .... ;)
[15:47] <benji> sinzui: I'm done with https://code.edge.launchpad.net/~sinzui/launchpad/anwers-api-0/+merge/60238
[15:49] <sinzui> thanks benji
[15:49] <benji> my pleasure
[16:26] <benji> henninge: are you doing any reviews right now?  I'm doing the unclamed ones and want to be sure I don't double-up.
[16:27] <henninge> benji: I was but I forgot to claim it ...:(
[16:27] <henninge> benji: pick anyone you like.
[16:27] <benji> k, thanks
[16:32] <rvba_> benji: thanks for the review!
[16:32] <benji> my pleasure
[16:38] <benji> henninge: an idea: to keep us from stepping on each other's toes (something I seem to be good at), should we have a gentleman's agreement to both claim the review and tell the other that we claimed it?
[16:44] <poolie> thanks for the review benji!
[16:44] <benji> my pleasure
[16:45] <henninge> benji: I am outta here.
[16:46] <poolie> review_s_
[16:46] <poolie> remarkable
[16:46] <poolie> bigjools: can you fix the syntax beore you land the flag?
[16:46] <poolie> or do a followup?
[16:47] <bigjools> poolie: ?
[16:47] <poolie> flags should have underscores not dashes, like python
[16:47] <poolie> names
[16:47] <bigjools> oh, didn;t know that.  I was basing it on an existing similar flag
[16:47] <poolie> at the moment they're inconsistent which seems kind of gross and likely to cause mistakes
[16:47] <poolie> :)
[16:48]  * bigjools fixorates
[16:48] <poolie> thanks
[16:48] <poolie> i guess i should fix them all to be one way
[16:48] <poolie> ust wanted to stop it getting worse
[16:48] <bigjools> I hope hitting ctrl-c terminates the ec2 instance :)
[16:49] <jml> bigjools: './utilities/ec2 list' should tell you
[16:49] <poolie> if it's already gone it's not a big deal
[16:49] <bigjools> yeah waiting for it to finish ...
[16:49] <poolie> you can kill it from the ec2 console
[16:49] <bigjools> yeah
[16:50] <bigjools> jeepers it's been going for 4 minutes now
[16:51] <bigjools> right, sorted, thanks for pointing that out poolie, I am now wiser than I was yesterday
[16:57] <poolie> np
[16:57] <poolie> generally i am not a big fan of pedantic conventions but here it seems worth while
[16:57] <poolie> since they'll be communicated to losas etc
[17:07] <bigjools> poolie: I am in total agreement with everything you just wrote :)
[17:26] <LPCIBot> Project windmill-devel build #51: STILL FAILING in 1 hr 14 min: https://lpci.wedontsleep.org/job/windmill-devel/51/
[17:59] <bigjools> night everyone
[18:14] <LPCIBot> Project windmill-devel build #52: STILL FAILING in 47 min: https://lpci.wedontsleep.org/job/windmill-devel/52/
[19:01] <sinzui> jcsackett: do you have time to mumble?
[19:02] <jcsackett> sure, give me one moment.
[19:05] <sinzui> jcsackett: I think I need to restart mumble again
[19:05] <jcsackett> sinzui: i am on mumble, i am not sure i have voice.
[19:05] <jcsackett> sinzui: righto.
[19:06] <sinzui> jcsackett: I do not see an indication you are speaking
[19:07] <jcsackett> sinzui: i have just restarted. i too see no indication of you speaking.
[19:08] <jcsackett> sinzui: i heard you.
[19:10] <jcsackett> sinzui: i think perhaps my mike is not working.
[19:10]  * jcsackett is checking audio settings.
[19:13] <jcsackett> sinzui: i think mumble just does not work for me today. i just had a full crash trying to fix it.
[19:13] <sinzui> ouch
[19:14] <jcsackett> sinzui: and now a unity crash trying to start mumble from terminal.
[19:15]  * jcsackett tries mumble one last time.
[19:16] <jcsackett> sinzui: i gather you still do not here me?
[19:16] <jcsackett> s/here/hear/
[19:16] <sinzui> jcsackett: No. Can I call you Patsy, my mute servant
[19:17] <jcsackett> sinzui: i would prefer not. :-P
[19:17] <jcsackett> skype? ekiga?
[19:17] <sinzui> I will start skype
[20:08] <LPCIBot> Project windmill-devel build #53: STILL FAILING in 1 hr 7 min: https://lpci.wedontsleep.org/job/windmill-devel/53/
[20:11] <sinzui> benji: I think your hide notification link script wires itself to every link in the notifcation: https://bugs.launchpad.net/launchpad/+bug/779538
[20:11] <_mup_> Bug #779538: Link from superseded blueprint to its successor disappears when you click it <javascript> <regression> <Launchpad itself:Triaged> < https://launchpad.net/bugs/779538 >
[20:13] <benji> sinzui: ooh, that's not nice; I'll take a look at it.  Do you have an example?
[20:13] <sinzui>  https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-o-auto-rotate
[20:13] <sinzui> benji: it will also affect an deactivated project or bug converted to a question
[20:19] <abentley> benji: Could you please review https://code.launchpad.net/~abentley/launchpad/person-product-merges/+merge/60417 ?
[20:19] <benji> abentley: sure
[20:21] <benji> sinzui: for me it hides the message box but doesn't prevent the browser from navigating to the link, is that the same behavior you see?
[20:21] <benji> if so, I think I can fix that
[20:22] <sinzui> benji: yes. many user want to open the link in a new tab and see the link they used in the previous tab
[20:23] <benji> absolutely, I wasn't arguing for not considering it a bug, just verifying what you were seeing
[20:23]  * benji hates web apps that break middle-click/control-click to open a new tab.
[20:35] <benji> abentley: I'm done with https://code.launchpad.net/~abentley/launchpad/person-product-merges/+merge/60417
[20:36] <abentley> benji: I prefer to write single-line imports as multi-line imports, so that we don't have to change the formatting later.
[20:39] <benji> abentley: I'm not stingy with vertical whitespace, but that's a bit too much even for me. ;)  It's a good thing we're trying to not be so up tight about style issues. :)
[20:42] <benji> sinzui: is there a bug for the overly agressive hiding?  If not, I'll file one.
[20:43] <sinzui> I pasted you the bug, so I thought
[20:43] <sinzui> https://bugs.launchpad.net/launchpad/+bug/779538
[20:43] <_mup_> Bug #779538: notifications disappear when you click *any* link in the text <javascript> <regression> <Launchpad itself:Triaged> < https://launchpad.net/bugs/779538 >
[20:43] <sinzui> ^ benji
[20:43] <benji> thanks
[20:45] <abentley> deryck: I'm looking at fixing https://bugs.launchpad.net/launchpad/+bug/777958 and I'd like to do a pre-imp call.
[20:45] <_mup_> Bug #777958: branch upgrade jobs keep transaction open <oops> <Launchpad itself:Triaged> < https://launchpad.net/bugs/777958 >
[20:49] <lifeless> sinzui: deryck: https://dev.launchpad.net/ArchitectureGuide/Services - incomplete, but the product of the discussions so far
[20:49] <sinzui> thanks
[20:54] <lifeless> jml: !! http://pypi.python.org/pypi/van.pg
[21:03] <lifeless> benji: hi; please triage-as-you-file, it saves double work (bug 779958)
[21:03] <_mup_> Bug #779958: lazr.restful.error.expose isn't documented <lazr.restful:New> < https://launchpad.net/bugs/779958 >
[21:04] <benji> lifeless: will do
[21:05] <lifeless> benji: thanks! I've triaged that particular one; Its just a kindness to the maintenance squads really :)
[21:05] <benji> thanks
[21:11] <lifeless> statik: hi
[21:14] <abentley> deryck: ping
[21:15] <deryck> hi abentley
[21:15] <deryck> abentley, i can chat here in just a sec....
[21:15] <deryck> just got back
[21:15] <abentley> deryck: ah, cool.
[21:15] <deryck> school run took much longer than I expected
[21:18] <deryck> lifeless, thanks for the link.  our call is top of this hour tomorrow, yes?
[21:18] <lifeless> deryck: or today if you like, statik has timeshifted for uds, and flacoste is @ uds
[21:19] <deryck> lifeless, abentley is waiting on a call now.  so tomorrow would be better.
[21:19] <lifeless> deryck: I'm easy
[21:19] <abentley> deryck: okay.
[21:19] <deryck> abentley, firing up mumble now
[21:20] <abentley> deryck: I'm looking at fixing https://bugs.launchpad.net/launchpad/+bug/777958 and I'd like to do a pre-imp call.
[21:20] <_mup_> Bug #777958: branch upgrade jobs keep transaction open <oops> <Launchpad itself:Triaged> < https://launchpad.net/bugs/777958 >
[21:37] <LPCIBot> Project windmill-db-devel build #254: STILL FAILING in 1 hr 7 min: https://lpci.wedontsleep.org/job/windmill-db-devel/254/
[22:12] <sinzui> Is there a method on storm objects that can be called to reinstantiate the object after I call reconnect_stores()?
[22:15] <lifeless> sinzui: *blink*
[22:16] <lifeless> sinzui: I tihnk you need to reload the objects
[22:16] <lifeless> sinzui: looking at existing tests
[22:17] <sinzui> lifeless: I was regetting the objects (IPersonSet.get()) but that clutters a lot of tests. review.reload() would be nicer to read
[22:17] <lifeless> sinzui: this is in a layer that can't use switchDbUser ?
[22:17] <sinzui> of well. I will write a helper. for the test harness
[22:21] <LPCIBot> Project windmill-devel build #54: STILL FAILING in 1 hr 7 min: https://lpci.wedontsleep.org/job/windmill-devel/54/
[22:27] <deryck> catch you all later on
[22:32] <lifeless> how do you get a js console on chromium?
[22:37] <benji> lifeless: shift-control-J on mine
[22:37] <lifeless> benji: thanks
[22:38] <benji> (or navigate into the deep, deep, dark menus)
[22:38] <benji> np
[22:55] <rvba_> Hey guys I have a traversal problem I can't wrap my head against.
[22:56] <rvba_> I want this kind of url to publish a DSD: /derived-distro/derived-series/+source/package-name/+difference/parent-distro/parent-series
[22:56] <rvba_> My problem is that I can't see how to use @stepthrough because I need *2* parameters (parent-distro *and* parent-series)
[22:56] <rvba_> Any advice?
[22:57] <lifeless> its fugly but you probably want an adapted parent distro as a place holder
[22:59] <rvba_> you mean an object that would be accessible at  /derived-distro/derived-series/+source/package-name/+difference/parent-distro ?
[23:01] <lifeless> yes
[23:01] <lifeless> representing the diff of derived-distro/derived-series/+source/package-name to parent-distro
[23:01] <lifeless> which is noddy
[23:02] <lifeless> so you wouldn't make a fat interface
[23:02] <lifeless> you'd only offer enough to traverse
[23:02] <rvba_> I see
[23:02] <lifeless> basically currying a factory for the thing you do want
[23:02] <rvba_> right
[23:03] <lifeless> which is the difff from  derived-distro/derived-series to parent-distro/parent-series of package-name
[23:03] <rvba_> makes sense ... I kinda hoped to get away with something simpler ... but I suppose I don't really have a choice here :)
[23:04] <LPCIBot> Project windmill-devel build #55: STILL FAILING in 43 min: https://lpci.wedontsleep.org/job/windmill-devel/55/
[23:05] <rvba_> lifeless: thanks a lot, I'll sleep on it and get this done tomorrow.
[23:05] <rvba_> 'night everyone.
[23:08] <lifeless> rvba-afk: gnight