[01:11] <sergio_br22> every week I get some broken package in xenial :/
[01:11] <sergio_br22> stuff is really moving fast
[01:11] <sergio_br22> i just fixed a package last week, and it's broken again
[12:37] <Odd_Bloke> We have a script that generates a summary of merge proposals that we're interested in, and we want to get a list of all the merge proposals that have a review requested from our team.  However, team.getRequestedReviews excludes MPs that someone in the team opened (as documented at https://launchpad.net/+apidoc/devel.html#team-getRequestedReviews), making this difficult.
[12:37] <Odd_Bloke> I've had a poke around in a debugger, but I didn't see a good way of getting around this; is it possible to get what we want from the team object?
[12:38] <Odd_Bloke> (If not, we could presumably iterate over things submitted by members of the team to gather the list, but I didn't want to resort to that if I'm missing an easier way of doing it)
[12:38] <wgrant> Odd_Bloke: Unfortunately there's no review requested of the team once someone claims it.
[12:38] <wgrant> It is completely transferred.
[12:39] <wgrant> Oh.
[12:39] <Odd_Bloke> wgrant: That's fine; but it doesn't even return unclaimed reviews.
[12:39] <wgrant> I miread.
[12:39] <Odd_Bloke> :)
[12:43] <wgrant> Odd_Bloke: It doesn't exclude MPs that someone in the team opened. That bit of the docstring just means that it only returns MPs that have a direct review requested of the team or person on which you invoke the method; it doesn't include reviews requested of teams that they participate in, so it doesn't return the full set of reviews that they can do.
[12:43] <wgrant> That doesn't seem like it would be a problem for you. What's the exact issue you're seeing?
[12:51] <Odd_Bloke> wgrant: Ah, OK; let me try to get a minimal test case for you (after lunch ^_^).
[12:51] <wgrant> Odd_Bloke: I must regrettably sleep imminently, but someone else might be around then.
[12:55] <Saviq> hey guys, it started happening yesterday - s390x builds for unity8 get stuck each build https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-010/+build/9031715
[12:55] <Saviq> we've been cancelling them, but I thought it might be interesting for you
[12:56] <wgrant> Ah, is that why all the s390x builders keep dropping out...
[12:56]  * wgrant pokes.
[13:04] <wgrant> Saviq: Looks like an sbuild change, only hitting s390x because the buildds are xenial due to the new port. Keep cancelling the builds for now, but we're investigating a fix.
[13:08] <Saviq> wgrant, ack, tx
[13:35] <Odd_Bloke> wgrant: cjwatson: We've seen our arm64 builds get stuck at https://launchpad.net/~cloudware/+livefs/ubuntu/xenial/cpc/+build/52672 a couple of times now; any ideas what's going on?  (The last one which I cancelled did not end up with a build log)
[13:35] <Odd_Bloke> (That last one being https://launchpad.net/~cloudware/+livefs/ubuntu/xenial/cpc/+build/52586)
[13:37] <cjwatson> Odd_Bloke: Possibly the same as the s390x issue above.
[13:37] <cjwatson> I bet it depends on the apt in the chroot.
[13:39] <Odd_Bloke> Our s390x builds are passing (and hit that same codepath).
[13:39] <Odd_Bloke> (Which isn't to say it isn't the same issue manifesting itself differently, of course :p)
[13:40] <cjwatson> Well, wait, livefs builds can't be the same thing.
[13:40] <cjwatson> Odd_Bloke: I can't see this build, though.  What does it look like from your end?
[13:41] <Odd_Bloke> cjwatson: It's "Currently building" with http://paste.ubuntu.com/15107531/ as Buildlog.
[13:41] <cjwatson> Um, not sure
[13:41] <Odd_Bloke> cjwatson: Which is what the last build looked like before I cancelled it (which then didn't get a build log attached to it).
[13:41] <cjwatson> wgrant: ^- could that be the out-of-loop-devices thing?
[13:50] <wgrant> Doesn't really look like it, but I suppose it's possible...
[13:50] <wgrant> I cancelled one just like that this morning.
[13:51] <wgrant> Could also just be a boring nbd hang.
[14:39] <cjwatson> Saviq: fixes on their way
[15:13] <Saviq> cjwatson, there you go https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-010/+build/9032689
[15:14] <cjwatson> Saviq: thanks, just awaiting proposed-migration of sbuild now
[15:15] <cjwatson> at least for the s390x case.  this does affect some other architectures too, but the exact circumstances that trigger it are build-dependent
[16:08] <Odd_Bloke> cjwatson: Has anything changed for the arm64 builders that might have caused us to start seeing this?  (We've never seen it before yesterday, and it's now happening most runs)
[16:18] <cjwatson> Odd_Bloke: We're gradually moving things into scalingstack, but you seem to have had successful runs on scalingstack builders before.
[16:18] <cjwatson> Odd_Bloke: Which builder is this on?
[16:19] <cjwatson> Odd_Bloke: bos01-arm64-009, right?
[16:19] <Odd_Bloke> cjwatson: I can't see the builder which the failures happened on; our current build (which hasn't reached the potential hang point yet) is on -004.
[16:24] <cjwatson> Odd_Bloke: You've definitely had a mix before today.  I'm sorting out some other upgrades right now, but as soon as I get the chance let me try resetting the bos01-arm64-{001..010} guests and seeing if that helps.
[16:25] <Odd_Bloke> cjwatson: Thanks. <3
[16:32] <Odd_Bloke> (So it looks like it hangs when we try to use a loop device for not-the-first-time; don't have the full build log so I don't know if I missed one in between refreshes, but it might be the second time)
[17:10] <cjwatson> Saviq: s390x is all sorted now.  as I say the same thing will affect other architectures in some situations - I've filed a ticket for the upgrade
[17:11] <Saviq> cjwatson, thanks
[22:12] <qengho> Hi hi. I'm stuck by the bzr-git bug that pukes when a Git commit in a foreign import is GPG signed. https://bugs.launchpad.net/bzr-git/+bug/1084403
[22:13] <qengho> Given first-class-ish Git branches on LP, can the importer import straight to a git branch, instead of trying to bzr-ize it?
[22:21] <wgrant> qengho: Direct git->git imports are not currently supported, but I was actually going to start investigating that today.
[22:22] <wgrant> It's easy to mirror manually with a cronjob in the interim.
[23:38] <qengho> I rebooted, didn't see reply. Is there a way to import a Git repo to Launchpad, not Bzr-izing it along the way?
[23:41] <wgrant> 09:21:54      wgrant | qengho: Direct git->git imports are not currently supported, but I was actually going to start investigating that today.
[23:41] <wgrant> 09:22:05      wgrant | It's easy to mirror manually with a cronjob in the interim.
[23:42] <qengho> wgrant: right. Thanks.