[00:01] persia: oh lord. probably. [00:02] ScottK: hate life [00:02] * persia checks the LP patch landing schedule [00:04] OK. There's a chance of landing that for 13th September, so it's possible to sort for release, but that's likely the earliest it can be done. [00:04] Oops, read that wrong. Landing 6th September. [00:05] (next window wouldbe 11th october) [00:06] strictly, it only needs LP if you need to build Task headers from it [00:06] if all you need is the metapackage, you can get by without [00:07] ScottK: nngh. somebody NBSed half the armel kernel udebs when they weren't NBS [00:07] livecd-rootfs uses the task. [00:08] slangasek: do you fancy figuring out how to get the rest of the armel kernel udebs back into the archive? [00:08] persia: but can be easily instructed not to [00:08] cjwatson: oh dear [00:09] cjwatson: is that my doing? the only NBS I did was copy'n'pasting of the command the NBS page gives out :/ [00:09] cjwatson, I guess, although I don't like that: I remember some missing packages when ubuntu-mid wasn't using a task. I suppose it can be swapped from task to meta, the seed moved, LP fiddled, and livecd-rootfs moved back. [00:09] it needs to be checked very carefully for kernel udebs [00:09] I think sometimes it outputs too much when the kernel is mid-build on one architecture, aside from the other failings [00:10] cjwatson: which accounted for about 80% of a several-hundred-package-long list, sigh [00:10] I basically apply a rule that I never NBS the current kernel version, which is perhaps a rule that should be hardcoded somewhere [00:10] * persia will try to sort out the set of steps required to transition to a new seed, and file a bug outlining them for later discussion. [00:10] regardless, we need to figure out how to get them back [00:10] (there were a lot of accumulated udebs in that list) [00:10] indeed [00:11] (I usually feed it through xargs -n1 | egrep ... | xargs [00:11] ) [00:11] you've done udeb recoveries in the past somehow, I think? Pinging of soyuz folks? [00:11] xargs> noted [00:11] it was years ago [00:11] not sure how many soyuz people will be around, but can try [00:12] should I just do a no-change upload of the kernel? [00:12] That's 16 hours for armel, and only covers the one kernel (there are several) [00:12] let's see if we can recover without that [00:12] only one kernel is affected afaik [00:13] which is the affected, linux? [00:13] yeah [00:14] I haven't exhaustively checked but I think just on armel [00:16] missing packages: block-modules-2.6.35-19-omap-di block-modules-2.6.35-19-versatile-di fat-modules-2.6.35-19-versatile-di firewire-core-modules-2.6.35-19-versatile-di fs-core-modules-2.6.35-19-omap-di fs-secondary-modules-2.6.35-19-omap-di input-modules-2.6.35-19-versatile-di irda-modules-2.6.35-19-omap-di md-modules-2.6.35-19-omap-di md-modules-2.6.35-19-versatile-di message-modules-2.6.35-19-versatile-di ... [00:16] ... mouse-modules-2.6.35-19-omap-di nfs-modules-2.6.35-19-omap-di nic-modules-2.6.35-19-omap-di nic-modules-2.6.35-19-versatile-di nic-usb-modules-2.6.35-19-omap-di pata-modules-2.6.35-19-versatile-di plip-modules-2.6.35-19-omap-di vlan-modules-2.6.35-19-omap-di vlan-modules-2.6.35-19-versatile-di [00:19] I have copies of all those files on my mirror, if someone wants to pull back and check the sigs. [00:19] (as armel udebs) [00:20] no, that is not a problem [00:20] they're all on the Launchpad librarian [00:20] Oh, right. [00:20] they need to have the publishing records restored, that's all [00:29] cjwatson: if we don't get a response from soyuz by evening (my time), I think it'll be simpler for me to just do the no-change rebuild that I can shepherd through over the weekend [00:29] the build queue appears to be completely empty right nowv [00:29] so while it would mean waiting for a rebuild, there at least wouldn't be any contention [00:48] slangasek: jpds was suggesting invoking the batphone, but perhaps you're right that that would be the simplest approach [00:49] slangasek: could you take over any coordination required? I need to go to bed [00:49] cjwatson: absolutely [00:49] thanks [04:40] in case someone who knows a bit about the seeding magic is still around, I'd appreciate some help understanding why oem-config-gtk that I added to the dvd seed for edubuntu doesn't get included on the DVD where all other packages in that seed get on it [04:42] a local run of germinate seems fine and the logs on people.canonical.com don't show any error (though they don't show the inclusion of oem-config-gtk). The build scripts seem to be using the right revision of the branch, so I really don't get what's happening there. [04:42] (it's probably a very simple mistake/misunderstanding but after spending an hour looking at it, I really have no more ideas) [08:07] stgraber, oem-config-gtk doesn't appear to be in any tasks right now. Maybe it needs a cron.germinate run to happen? === apache2logger is now known as apachelogger === bjf[afk] is now known as bjf[vacation] [17:36] persia: well, that'd be weird because I pushed that seed change on the 8th of August and then another seed change (including gnome-nanny) on the 16th and the change of the 16th worked properly (as in, that package is now on our dvd) [17:47] * lamont plays hardball with the amd64 ppa builders [17:55] all better [18:07] ubuntu-archive: please consider accepting eglibc and binutils this weekend (no changes for generated code), needed for the linaro cross toolchain [21:01] uploading a new LTSP now, ideally that should be in beta as otherwise the LTSP install on ubuntu alternate is simply going to fail [21:02] I'm running a test install just now, if the new package works I'll upload it [21:27] tested and uploaded [23:32] cjwatson: I see that the kernel failed to build on all architectures. [23:33] ogasawara: ^^ can you help me out? I tried to do a no-change upload of the kernel to restore some udebs that I had accidentally removed; apparently the kernel doesn't *allow* no-change rebuilds, it complains about missing ABI files