[04:08] micahg: what system has that limit? [04:08] lifeless: 1and1 [04:16] steven@mangled:~% find Maildir -type f | wc -l [04:16] 168334 [09:00] good morning [09:00] Hello adeuring. [09:00] hi rvba! [09:02] Morning adeuring, rvba. [09:02] Hi wgrant! === rvba changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: rvba | Firefighting: - | Critical bugtasks: 3*10^2 === Noldorin is now known as Qwerty === Qwerty is now known as Noldorin === rvba changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: - | Firefighting: - | Critical bugtasks: 3*10^2 [21:53] Morning all [22:01] Morning everyone. [22:02] wgrant: Morning [22:07] Morning [22:15] i'm trying to follow the code path for archive uploads to figure out what changes would be needed for https://lists.ubuntu.com/archives/technical-board/2011-November/001137.html - i see the poppy daemon moving uploads into /var/tmp/poppy/incoming once the upload completes, but what actually processes the uploads after that? [22:16] process-upload [22:17] broder: You probably want NascentUpload under lib/lp/archiveuploader [22:18] StevenK: ok, i'll look there - thx [22:19] * StevenK looks at the failed stable -> db-devel merge [22:21] StevenK: There shouldn't really be a conflict. [22:21] It's sinzui's test disablement. [22:21] Which is just a workaround for there being no webops. [22:21] I thought the fix was the same in both. [22:26] broder: I've seen people in Ubuntu mention the -backports permission differences over the years, but never seen any code to support the idea that there are any. [22:26] wgrant: ah, good. i was just starting to come to that conclusion, too :) [22:27] Isn't it the code that checks the pocket dependant on the series status? [22:27] That's unrelated. [22:27] That just affects when backports opens for a new series. [22:27] It should also be removed, but it's not a problem here yet. [22:28] wait, hmm? that sounds potentially related to my project [22:29] Ah, yes, so it is. [22:29] Not directly to this issue, though [22:29] Currently post-release pockets are only uploadable to when the series is frozen or release. [22:29] This seems pretty arbitrary. [22:30] Yeah, I'd want to turn that off for backports at least [22:30] See DistroSeries.canUploadToPocket [22:30] Seems like it should be turned off for everything, since they all require manual approval anyway. [22:30] And we do 0-day SRUs. [22:30] It just happens to work now because the series is usually frozen by the time the 0-day SRU is started. [22:33] So permissions on the backports pocket are enforced as they are for any other pocket, and they will always land in UNAPPROVED, but right now non-release uploads will be rejected until the series is frozen? That means that the last bit would be the only thing requiring change for pre-release backports [22:33] Yes [22:33] I believe so. [22:34] Sweet. I think I can pull that off [22:36] Oh, right. And I need to change the buildd configuration so that builds for backports run with the same component isolation as release/proposed/security [22:36] That's in lp.soyuz.adapters.archivedependencies [22:36] Should mostly be a matter of deleting 3 lines of code a few tests. [22:37] Let me see. [22:37] get_components_for_context [22:37] def get_components_for_context(component, pocket): [22:37] if pocket == PackagePublishingPocket.BACKPORTS: [22:37] return component_dependencies['multiverse'] [22:37] return component_dependencies[component.name] [22:37] Well, that looks easy... [22:52] hey launchpadders [23:00] Evening jelmer.