[00:04] <infinity> wxl: I'm going to spin up new RCs for everyone tonight after I get my last livecd-rootfs tweaks tested and in place.
[00:05] <wxl> infinity: kthx
[00:05] <elfy> infinity: awesome sauce
[00:06] <wxl> vivid's being funky too it seems, but we'll gross that bridge later ☺
[00:06] <elfy> I'll be up and about earlier UTC and will be able to see how our's look asap
[00:07] <elfy> wxl: I'll cross it instead if you don't mind - my days of gross are long gone :p
[00:07] <wxl> elfy: argh, you know what i meant. i was typing with one hand
[00:08] <wxl> infinity: tonight = what in utc?
[00:09] <elfy> wxl: tomorrow for you I guess - why not just tell your people to look for change in image name on tracker if you're going to be asleep?
[00:09] <elfy> because the time is likely to be when it's done :)
[00:11] <wxl> so evening utc tuesday then elfy ?
[00:13] <elfy> wxl: dude I don't know
[00:14] <elfy> if I have to mark it not ready then that's what I'll do - but as long as I can smoketest the image I suspect I won't
[00:14] <elfy> I'll just make it plain in release notes
[02:34] <infinity> wxl: Tonight meaning late for me, in North American time.  Still need to work the last kinks out of myth, studio, and kylin, and then push the changes for everyone.
[12:33] <Mirv> archive admin consultance (preNEW accept) needed for Compiz that adds compiz-mate package, https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-021 - summary of packaging changes https://ci-train.ubuntu.com/job/ubuntu-landing-021-2-publish/32/artifact/packaging_changes_compiz_1%3A0.9.12.1+15.04.20150213-0ubuntu1.diff
[15:57] <elfy> anyone got an idea why lubuntu/gnome/kylin/studio and xubuntu have no dailies for vivid today?
[16:02] <infinity> elfy: Picking GNOME at random from your list, it looks like it built fine.
[16:02] <infinity> elfy: Unless today isn't the 17th...
[16:06] <ogra_> yesterday the 17th was tomorrow
[16:06] <infinity> ogra_: Ow.
[16:08] <elfy> infinity: lol - so that built in the last short while :)
[16:08] <elfy> mine is still showing the 16th on the tracker
[16:09] <infinity> Speaking of builds, I *think* I finally sorted out everyone's task->meta messes for trusty.
[16:09] <infinity> And am rebuilding the world right now.
[16:09] <infinity> I'll check manifests when they're all done and see if I missed anything, but hopefully I got it all in this pass.
[16:11] <elfy> infinity: thanks :)
[16:15] <elfy> rest still showing as 16th - has the time these things build been changed?
[16:16] <ogra_> lubuntu failed it seems
[16:16]  * ogra_ notices a build failure mail
[17:11] <elfy> infinity: booted 64bit - that looks more like it :) thanks
[17:12] <infinity> elfy: Nice.  I'll go diffing things in a bit to make sure I got all the bits right in my own estimation, but if it looks sane, that's a good start.
[17:12] <elfy> indeed :)
[17:13] <elfy> still not sure what's up with vivid - but one thing at a time - more interested in trying to get trusty tested again
[17:21] <elfy> Riddell: you got a minute?
[17:48] <wxl> infinity: didn't see your email but i assume we're good? did you have any luck fixing the lubuntu alternate image?
[17:49] <infinity> wxl: I haven't sent the email yet, cause I want to diff all the manifests, but we should be good for desktop images.
[17:49] <wxl> infinity: so alternate is an unknown?
[17:49] <infinity> wxl: I'm less inclined to try to fix the alternate, to be honest.
[17:50] <infinity> wxl: But I can look at it later and see if the fix is simple enough.
[17:50] <wxl> infinity: k thanks, hope you didn't have to pull an all nighter :)
[17:50] <infinity> wxl: I might have...
[17:50] <wxl> :(
[17:50] <infinity> It happens. :P
[17:50] <wxl> well thank you for your hard work regardless, infinity
[18:07] <wxl> infinity: so server is going to suggest netboot/mini.iso then?
[18:16] <elfy> infinity - so I've quickly done all 10 of our tests with trusty on vb and am pleased to have 10 passes instead of fails - so thanks awfully :)
[18:27] <infinity> wxl: No, server doesn't have the same issue.
[18:27] <wxl> infinity: well that's lucky :)
[18:27] <infinity> wxl: Since server doesn't ship X.
[18:27] <wxl> ah, i didn't realize it was x-specific
[18:28] <infinity> wxl: Yeah, this is all about the X HWE stack.
[18:28] <infinity> elfy: That's encouraging.
[18:28] <wxl> infinity: so it's about the change from tasks to metapackages specifically for the x hwe stack?
[18:28] <infinity> wxl: Right.
[18:29] <wxl> infinity: oohhhhhhhhh, well now that clears things up quite a bit. thanks, then.
[18:29] <infinity> wxl: Because tasks don't change post-release, so we can't make the *-desktop tasks suddenly pull in the new X stack.
[18:29] <ogra_> will be interesting what happens to this in Mirland
[18:29] <infinity> ogra_: I suspect Mir will still need a fair bit of the X framework for a long while, unless "switch to Mir" also implies "and only support 3 video cards", so... We'll see.
[18:29] <wxl> infinity: wow, how did we ever survive using tasks before?
[18:30] <infinity> wxl: We did this same thing in precise with the HWE point releases there.
[18:30] <ogra_> infinity, right, so we might end up with two hwe stacks for a while
[18:30] <ogra_> for the graphical bits
[18:30] <wxl> infinity: so will we be metapackage only for 16.04 and on?
[18:31] <infinity> wxl: I'm pondering that question.  Doing the switch at .2 hurts a lot, so doing .0 with metapackages makes some sense.  But tasks are a bit more convenient for $reasons.
[18:31] <infinity> wxl: There are some ways we could make tasks post-release manglable, so thinking about that too.
[18:32] <wxl> infinity: yes, there has been some concern about not using tasks. still, i see this whole transition to be a royal pain in the you-know-what for the release team (i.e. you in this case!)
[18:35] <infinity> Hrm, and lubuntu powerpc is FTBFS.  Fun.
[18:37]  * infinity scratches his head at that one.
[18:39] <wxl> infinity: that is of major sucking, as lubuntu has changed to really only supporting ppc on lts
[18:42] <infinity> wxl: Well, people who install with 14.04 and upgrade are fine.  It's just the 14.04.2 livefs that's hating me.
[18:42] <infinity> wxl: But I'll look at it after I've rested a bit.
[19:01] <wxl> infinity: we're seeing the lack of xserver-xorg-input-synaptics still in the live session. is it possible that the live session would fail but the install would work fine?
[19:02] <infinity> wxl: Erm...
[19:02] <infinity> wxl: Which image?
[19:02] <wxl> infinity: desktop i386
[19:02] <infinity> wxl: And are you just looking for the package name, or the functionality? :P
[19:03] <wxl> infinity: is the package not required for functionality? :)
[19:04] <infinity> wxl: Oh, hrm.  Okay, I did all my testing on amd64...  i386 does seem confused. :/
[19:04] <wxl> <more expletives>
[19:05] <wxl> infinity: i hate to ask this, but i assume you can resolve this?
[19:06] <infinity> wxl: I'm sure it can be resolved.
[19:06] <infinity> wxl: I first need to sort out why it's broken.
[19:06] <wxl> infinity: thank you. please keep me updated.
[19:08] <infinity> I might need to explicitly install xserver-xorg-input-all-lts-utopic, that's all.
[19:08] <wxl> infinity: if that's what it is. i did do a diff on old versus new if that's helpful
[19:09] <infinity> wxl: Have the diff handy?
[19:10] <wxl> infinity: several available in the bug description of 1417784
[19:10] <wxl> oops bug 1417784
[19:12] <infinity> Yeah, it looks like, for some insane reason, it's picking evdev instead of all.
[19:12] <infinity> Implicitly installing all should fix it.
[19:12] <wxl> infinity: okie dokie. none of the other differences are an issue?
[19:14] <infinity> wxl: The others should have been fixed already in my previous pass, I believe.
[19:15] <wxl> okie dokie
[19:16] <wxl> infinity: let me know when we should go for a rebuild
[19:16] <infinity> wxl: I'm going to do a pass of all the manfiests for all arches and flavours, I guess, and see where we stand, so I'm not doing a bunch of fixes in a row.
[19:17] <infinity> wxl: But if you can test amd64, and assume I'll fix i386 soon, that'd be cool. :P
[19:17] <wxl> infinity: on it
[19:18] <infinity> wxl: Since it seems napping isn't going to be an option, I'm going to eat and take a TV/lunch break before I get to that, but should be sorted with new images spit out in, say, 4h or so.
[19:21] <wxl> cool, thanks infinity
[19:29] <jderose> found an issue with UEFI mode desktop installs - https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1422864
[20:33] <cyphermox> jderose: looking
[20:41] <jderose> cyphermox: thanks! i've never been clear on why the uefi-mode installs download the newest kernel but bios-mode doesn't. either way, it's not downloading the correct one :)
[20:45] <cyphermox> jderose: well, the "newest" is the correct one and is on the CD anyway, but something is convincing ubiquity to get an older one.
[20:46] <jderose> yeah, seems it's on  the 14.04.1 track still.
[20:49]  * infinity boggles at lubuntu missing a dep of lubuntu-desktop, but having lubuntu-desktop installed...
[20:50] <cyphermox> o.o;
[20:51] <infinity> Oh, because it's provided by gnome-system-tools as well.  Brilliant.
[20:52] <infinity> And, gnome-time-admin is a transitional package that shouldn't even be seeded.  Okay, that's one mystery I can stop caring about.
[20:53] <cyphermox> infinity: there is much yuckiness in gnome-system-tools :)
[20:57] <infinity> Meh, still a few other manifest diff bugs too.  *sigh*
[21:14] <cyphermox> infinity: so, it's base-installer explicitly getting you to install linux-signed-generic, linux-signed-image-generic, that gets queries in ubiquity's scripts/check-kernels
[21:14] <cyphermox> looks like base-installer/kernel/altmeta should allow for tweaking this, but perhaps not enough for signed kernels
[21:16] <cyphermox> or any really. maybe it's not meant for that
[21:17] <infinity> altmeta is meant to deal with that, yes.  But we don't preseed it on desktop installs because we don't have a kernel to select.
[21:17] <infinity> And therein lies the problem, I imagine.
[21:17] <cyphermox> hm, true
[21:18] <cyphermox> the sed looks unsufficient anyway
[21:18] <infinity> I wonder if we had a preseed on precise, and I'm just missing that bit somewhere.
[21:18] <infinity> It does?  altmeta works great on the server ISOs.  Or, should do.
[21:18] <cyphermox> trying to look that up but I'm waiting for the install to finish
[21:19] <cyphermox> infinity: sed 's/$/-$RET/', it won't change -generic into xyz-generic, but I may be misunderstanding the intent
[21:19] <cyphermox> download to finish, I mean
[21:20] <infinity> cyphermox: No, it changes linux-generic into linux-generic-lts-utopic
[21:20] <cyphermox> ah
[21:20] <cyphermox> magic!
[21:20] <infinity> cyphermox: $ (end of line) becomes -$RET (-lts-utopic)
[21:20] <cyphermox> yes, I can read regex
[21:20] <infinity> See, I assumed you could, but you seemed to be reading it backwards.  Maybe. :)
[21:20] <cyphermox> no
[21:20] <infinity> Or French.
[21:20] <infinity> Stop reading in French.
[21:20] <cyphermox> I just didn't know we had a linux-generic-lts-utopic
[21:21] <infinity> Ahh. :)
[21:21] <cyphermox> how about you start learning French? two official languages and all of that ;)
[21:21] <infinity> I was fully bilingual as a child.
[21:22] <cyphermox> d-i	base-installer/kernel/altmeta	string lts-saucy
[21:22] <infinity> I've lost most of my vocabulary since I stopped using it.
[21:22] <cyphermox> magic!
[21:22] <infinity> Magic, and stale magic at that.
[21:22] <infinity> Since that probably should have been trusty, if that was a 12.04.5 image. :/
[21:22] <cyphermox> nah, it was .4
[21:23] <infinity> Where is that preseed?
[21:23] <cyphermox> you mean on the CD or in livecd-rootfs/ubuntu-cdimage?
[21:23] <infinity> On the CD.
[21:24] <infinity> From debian-cd, it looks like it should be there for trusty too...
[21:24] <cyphermox> preseed/ubuntu.seed
[21:24]  * infinity downloads a trusty ISO to look.
[21:29] <cyphermox> ehhh
[21:29] <cyphermox> it's there already
[21:29] <infinity> That's what I assumed. :/
[21:30] <infinity> So, either the preseed isn't being read correctly, or it's not base-installer's fault we're getting that kernel installed.
[21:30] <cyphermox> I'd like to assume the preseed works
[21:30] <infinity> I know it works with d-i installs (server CD and netinst), but ubiuity and I have never been on the best of terms.
[21:31] <infinity> Perhaps a log from your install implies what it was thinking when it happened? :/
[21:32] <cyphermox> working on it
[21:32] <wxl> infinity: lubuntu amd64 desktop is a go, so at least that's good. alternate is hosed, but i'm not surprised.
[21:32] <infinity> wxl: amd64 is still a bit broken too, but glad you didn't notice. :P
[21:33] <wxl> aw great, infinity
[21:35] <melodie> hello
[21:37] <melodie> I have had a few talks with wxl about the issues met in Lubuntu. wxl said on the list and on another chan that infinity is in charge of the hard job : fix the issues.
[21:37] <infinity> melodie: infinity is working on it currently.
[21:38] <melodie> how can the fact that the packages linux-libc-dev and two depends of this package are not mentioned in the lubuntu trusty daily filesystem.manifest?
[21:38] <melodie> I have seen them in Ubuntu, in Xubuntu one of them. Not in Lubuntu
[21:38] <infinity> melodie: They never were in lubuntu's manifest.
[21:38] <melodie> can that affect the behavior of the installer or other important parts?
[21:38] <melodie> they were in 2012 when I did a search for a project of mine.
[21:38] <melodie> so that's unrelated?
[21:39] <infinity> melodie: Please compare to the 14.04 lubuntu manifest, not to another flavour.
[21:39] <melodie> infinity ok. will do.
[21:39] <melodie> thanks.
[21:42] <infinity> cyphermox: So, the preseed is definitely being found at least.  Whether it's influencing ubiquity's abuse of base-installer is another question entirely.
[21:42] <elfy> infinity: just quick mention that I'm sure you'll see - we appear to have xscreensaver and mate-notification-daemon in our newest image
[21:43] <infinity> elfy: Yeah, noticed already and committing fixes.
[21:43] <elfy> awesome :)
[21:43] <infinity> Well, fixes committed for mate-notification-daemon ... Need to sort out xscreensaver.
[21:43] <elfy> ok - ty
[21:48] <infinity> elfy: Okay, and xscreensaver sorted too, I think.
[21:49] <elfy> excellent
[21:49] <elfy> rebuilds coming sometime by tomorrow? or request myself ?
[21:49] <infinity> elfy: Well, I need to upload first. :P
[21:50] <infinity> elfy: Rebuilds will be done by me for everyone, I think.
[21:50] <elfy> lol - I didn't mean I'll go do that now :)
[21:50] <infinity> elfy: Hoping to hunt down this efi/signed thing too.
[21:50] <elfy> aah - super - I'll just wait for notification of it then :)
[21:51] <infinity> elfy: OOI, do you test amd64 on UEFI systems?
[21:51] <elfy> not personally
[21:51] <infinity> elfy: I assume LP: #1422864 is a universal bug (can't see why it wouldn't be).
[21:51] <elfy> not sure if anyone does specifically for us
[21:52] <elfy> tbh I've never needed to worry about UEFI
[21:52] <infinity> elfy: The time will come soon, I'm sure. :P
[21:53] <infinity> elfy: Hard to buy a new machine that doesn't force you to care.
[21:53] <infinity> Or, getting harder anyway.
[21:54] <elfy> well this machine has uefi pretty sure - but don't install - boot the usb to the other option
[21:54] <elfy> certainly no EFI in /boot which I thought is where it would be
[21:57] <infinity> cyphermox: If you make sense of the kernel thing, or need a sounding board, let me know.  I'm going to veg for a bit and try to not fall asleep.
[21:58] <cyphermox> ugh, doing the best I can
[21:58] <cyphermox> I'm testing this here but I can't seem to reproduce it just yet by hacking my way about it
[21:58] <cyphermox> now trying a proper install with efi
[21:58] <cyphermox> but really, efi or not you should get the same preseed behavior and the same kernel name mangling from altmeta and stuffs
[21:59] <infinity> cyphermox: Yes, but without efi, it never tries to install/keep the signed kernel, so that whole path might be skipped.
[21:59] <infinity> cyphermox: At least, this is my guess.
[22:00] <cyphermox> yeah, but that's exactly what I was trying to reproduce just looking at /var/lib/ubiquity/install-kernel and remove-kernel
[22:00] <cyphermox> they looked fine either way, if I just went running check-kernel for efi or non-efi
[22:02] <infinity>         preferred_kernels="$(echo "$preferred_kernels" | sed "s/$/-$RET/"; \
[22:02] <infinity>                              echo "$preferred_kernels")"
[22:02] <cyphermox> yep
[22:02] <infinity> for kernel in $preferred_kernels; do
[22:03] <infinity> I read that second block as "I'm going to try to install the 'signed' variant of everything!"
[22:03] <infinity> That's scripts/check-kernels
[22:03] <cyphermox> yes
[22:04] <infinity> At least, that's a naive stab in the dark on my part, I don't know how it all hooks back into reality.
[22:05] <infinity> But that would install linux-signed-generic-lts-utopic (already there) *and* linux-signed-generic
[22:05] <cyphermox> ah, here it's getting into remove-kernels... not good
[22:06] <infinity> ie: I think the second "echo $preferred_kernels" is probably what's biting us.
[22:25] <jderose> cyphermox: please let me know if you'd like me to help test anything... i have a qemu VM setup for uefi, so easy to do quick tests
[22:26] <infinity> jderose: I think we've logicked it out.  Easiest testing method will be live-patching a VM before running ubiquity, when there's a patch, someone might paste it to you for sanity checking.
[22:26] <jderose> infinity: cool, thanks
[22:46] <melodie> good night
[22:55] <infinity> jderose: http://paste.ubuntu.com/10280647/ might do the trick.
[22:55] <infinity> jderose: Not entirely sure where scripts/check-kernels lives on the live filesystem, but "dpkg -S check-scripts" should find it.
[22:55] <infinity> Err, check-kernels.
[22:56] <jderose> infinity: okay, i'll give it a test. so just modify this file in the live enviroment before starting the installer, right?
[22:56] <infinity> jderose: Right.
[22:58] <bluesabre> would any archive admins be interested in releasing this to the archive? :-) New source: lightdm-gtk-greeter-settings (vivid-proposed/primary) [1.1.0-0ubuntu1]
[23:12] <jderose> infinity: fix seems to work in that 3.13 is not getting installed. although that doesn't prove it would correctly download a newer 3.16 kernel were it available :)
[23:13] <infinity> jderose: Well, if the system reboots and has linux-signed-lts-utopic installed, that should be proof it worked correctly.
[23:13] <infinity> jderose: Since those packages are removed if not "needed".
[23:14] <jderose> infinity: but linux-signed-lts-utopic was installed before.... the problem was that 3.13 was also needlessly installed
[23:15] <infinity> jderose: Right.  I'm saying that it won't have regressed if it's still installing it. :P
[23:15] <jderose> ah, gotcha
[23:15] <infinity> jderose: So, if 3.16's signed is there and no 3.13 is there, we win.
[23:20] <infinity> jderose: As soon as you tell me the rebooted system looks sane, kernel-wise, I'll commit and upload this.
[23:22] <jderose> infinity: yup, rebooted, looked good
[23:23] <cyphermox> awesome
[23:32] <infinity> jderose: Thanks for the testing!
[23:32] <jderose> infinity: cyphermox: np. thanks for the quick fix!