[07:30] <jibel> ev, cjwatson could you update wubi isolist to 10.04.3  ?
[09:14] <cjwatson> jibel: I committed the change for that a few days ago and asked ev for a new build
[09:15] <cjwatson> huh, I'm confused now
[09:15] <cjwatson> http://people.canonical.com/~evand/wubi/lucid/ has wubi-r191.exe dating from 21 Jan
[09:15] <cjwatson> my commit was r191, dated 15 Jul
[09:15] <cjwatson> ev: did you forget to push your 10.04.2 wubi commit, perhaps?
[09:15] <ev> trying to figure out how to update my lucid chroot with a 3.0 kernel
[09:16] <ev> probably :-/
[09:20] <cjwatson> use a 3.0.0 kernel instead, it'll be less painful
[09:20] <cjwatson> maybe you could use a checkout to avoid that problem?
[10:36] <ev> cjwatson: new wubi up
[10:36] <ev> thanks for the tips
[10:44] <cjwatson> jibel: is there any hope of validating the installer changes in lucid-proposed in time?
[10:47] <cjwatson> jibel: given that Peter checked my grub-installer patch for bug 695842 when applied out of band, I think regression-testing some other system with apt-setup/proposed=true passed as an installer boot parameter (alternate or server install) would be enough to let that into lucid-updates.  What do you think?
[10:47] <ubot4> Launchpad bug 695842 in grub-installer (Ubuntu Maverick) (and 2 other projects) "Cannot install on HP Proliant DL385 G7 - dual RAID controllers (affects: 1) (heat: 26)" [High,Fix committed] https://launchpad.net/bugs/695842
[10:48] <cjwatson> jibel: bug 442941 should be easy to validate - the permissions check would be enough, given that the update-manager test will be hard to do in this case
[10:48] <cjwatson> s/would/should/
[10:48] <ubot4> Launchpad bug 442941 in ubiquity (Ubuntu Oneiric) (and 14 other projects) "debconf failed to upgrade from 1.5.27ubuntu1 to 1.5.27ubuntu2: exit status 128 - Use of uninitialized value $reply in scalar chomp at /usr/share/perl5/Debconf/FrontEnd/Passthrough.pm line 66 (affects: 452) (dups: 300) (heat: 2461)" [High,Fix released] https://launchpad.net/bugs/442941
[10:49] <cjwatson> pitti: I think we need to waive the waiting period for base-files and debian-installer to get them into lucid-updates in time for 10.04.3
[10:50] <cjwatson> is anything else in lucid-proposed needed for 10.04.3?
[10:51] <cjwatson> oh, bug 777759 too, that should be easy to validate as well (per my comment 1 indicating that we can probably just do regression-testing there)
[10:51] <ubot4> Launchpad bug 777759 in casper (Ubuntu Natty) (and 3 other projects) "10.10 live CD prompts to upgrade to 11.04 (affects: 1) (heat: 57)" [High,Fix committed] https://launchpad.net/bugs/777759
[10:51] <jibel> cjwatson, 442941 if tha's the only check then I verified it on many installs.
[10:51] <jibel> 777759 is verified, i'll update the status
[10:52] <cjwatson> jibel: 442941> the permissions check?
[10:52] <jibel> right, permissions check. Anything else I can verify ?
[10:53] <cjwatson> it's tricky with an up-to-date image, is the problem
[10:53] <cjwatson> since there's naturally nothing to update
[10:53] <cjwatson> and doing any preparatory operation that touches the debconf database in any way will invalidate the test
[10:54] <cjwatson> perhaps check that you can install a package on top with software-center or whatever
[10:54] <cjwatson> just as an extra regression test
[10:55] <cjwatson> other than that I don't think there's much we can check - we'll just have to revisit if the problem recurs later
[10:55] <jibel> that's one of the post-install test I run. installing something with software-center.
[10:56] <jibel> 695842, I can only do regression-testing, since I don't have the hardware.
[10:57] <jibel> cjwatson, I get the problem again with oem install, I can't pass the second stage of the installation.
[10:58] <cjwatson> 695842> I think we can accept that given Peter's comment 17; but since I wrote the patch, I'd like a second opinion from pitti
[10:58] <jibel> how can I debug this ? it is not reproducible. It happens on less than 10% of the install I did.
[10:58] <cjwatson> can I have logs, to start with?
[10:59] <cjwatson> I can at least see if there's anything obvious in them
[10:59] <jibel> ok
[11:00] <cjwatson> it must mean that oem-config has failed part-way through
[11:00] <cjwatson> I think you'll find that /var/lib/oem-config/run still exists?
[11:00] <cjwatson> (it'll be zero-length)
[11:00] <jibel> yes
[11:01] <cjwatson> this happens if oem-config exits non-zero
[11:02] <jibel> right, that's the only error I see in the logs "oem-config main process terminated with status 1"
[11:05] <cjwatson> did you run the second stage with the 'debug-oem-config' boot parameter
[11:05] <cjwatson> ?
[11:05] <pitti> jibel, cjwatson: I'm fine with regression-testing on bug 695842, indeed
[11:05] <ubot4> Launchpad bug 695842 in grub-installer (Ubuntu Maverick) (and 2 other projects) "Cannot install on HP Proliant DL385 G7 - dual RAID controllers (affects: 1) (heat: 26)" [High,Fix committed] https://launchpad.net/bugs/695842
[11:06] <pitti> cjwatson: want me to move to -updates?
[11:06] <cjwatson> let's wait 'til regression-testing has actually been done
[11:06] <pitti> cjwatson: the current lucid dailies use -proposed?
[11:07] <cjwatson> oh, yes, that grub-installer version should be what's on the CD
[11:07] <cjwatson> I'll just double-check that
[11:07] <cjwatson> yeah, it is
[11:07] <pitti> i. e. the iso testing, and the following certification testing pretty much serves as regression testing
[11:08] <cjwatson> should do, yes
[11:08] <cjwatson> jibel: I'd like to see the logs just in case there's something that doesn't look like an error but that might give me a clue, though
[11:12] <jibel> cjwatson, logs attached to bug 812230
[11:12] <ubot4> Launchpad bug 812230 in ubiquity (Ubuntu) "OEM install always boots to 'prepare for shipping' (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/812230
[11:14] <cjwatson> oh, I misread your original report
[11:14] <cjwatson> you never even get oem-config running, do you
[11:16] <cjwatson> jibel: can you boot with the '--verbose' boot parameter, and then attach /var/log/syslog?
[11:19] <jibel> cjwatson, I add it to the kernel boot command ?
[11:21] <jibel> interesting, I also get logged out after few minutes.
[11:21] <cjwatson> yes
[11:24] <jibel> https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/812230/+attachment/2214765/+files/syslog
[11:24] <ubot4> Launchpad bug 812230 in ubiquity (Ubuntu) "OEM install always boots to 'prepare for shipping' (affects: 1) (heat: 6)" [Undecided,New]
[11:24] <jibel> I think ev already fixed something similar with upstart.
[11:26] <cjwatson> if you have details, that would be helpful
[11:28] <cjwatson> it apparently at least tries to run
[11:28] <cjwatson> then immediately afterwards, gdm starts
[11:28] <cjwatson> I wonder if they're both running on different VTs
[11:29] <cjwatson> this looks like a straightforward race between 'starting gdm' and 'stopping rc'
[11:30] <cjwatson> it was fixed in oneiric
[11:30] <cjwatson> actually natty, at least in part
[11:30] <cjwatson> however, it took a few goes to get it right; I'm not comfortable with backporting it in a hurry
[11:31] <cjwatson> bug 650703
[11:31] <ubot4> Launchpad bug 650703 in ubiquity (Ubuntu Natty) (and 4 other projects) "oem-config-prepare works, but oem-config fails to start after reboot (affects: 15) (dups: 2) (heat: 45)" [Critical,Fix released] https://launchpad.net/bugs/650703
[11:31] <cjwatson> hah, d'oh, it has a 10.04.3 task
[11:31] <cjwatson> well, I don't know.  what do other people think?
[11:31] <jibel> yep, that's the one
[11:33] <cjwatson> I think the fix is http://paste.ubuntu.com/646431/
[11:34] <cjwatson> what we could do is move all the current installer bits to -updates, and then I do a fresh upload for this bug
[11:34] <cjwatson> then we build new images against -proposed
[11:35] <cjwatson> if they pass, we waive the waiting period and let the new one in; if not, we leave it in -proposed
[11:35] <cjwatson> then either way, we build 10.04.3 images against -updates
[11:50] <cjwatson> anyone?  Bueller?
[12:14] <cjwatson> pitti: do you think I can go ahead and copy base-files and debian-installer to -updates despite the short wait, on the basis of iso testing?
[12:14] <pitti> cjwatson: pease do; base-files only was the .3 bump anyway, right?
[12:14] <cjwatson> yes
[12:16] <pitti> cjwatson: want me to move ubiquity, too? was verified now
[12:17] <jibel> re 650703, sounds good to me.
[12:18] <cjwatson> done base-files and debian-installer
[12:18] <jibel> do you think of a way to reproduce it ?
[12:18] <cjwatson> pitti: grub-installer needs to go before ubiquity
[12:19] <pitti> ah
[12:22] <cjwatson> jibel: well, it's a race between 'stopping rc' and 'starting gdm', so to reproduce it, you could try to force 'starting gdm' to happen later to widen the window
[12:22] <cjwatson> jibel: here's one possible hack that might do it:
[12:22] <cjwatson> jibel: edit /etc/init/dbus.conf, and find the 'post-start' line; in oneiric that reads 'post-start exec kill -USR1 1'
[12:22] <cjwatson> change that to:
[12:22] <cjwatson> post-start script
[12:22] <cjwatson>     kill -USR1 1
[12:23] <cjwatson>     sleep 10
[12:23] <cjwatson> end script
[12:23] <jibel> ok, trying
[12:24] <cjwatson> ah, grub-installer verified now; I'll go ahead and move the installer bits, then
[12:24] <jibel> sorry for the delay
[12:26] <cjwatson> that's ok
[12:35] <cjwatson> urgh, 'sru-release lucid ubiquity' is timing out
[12:35] <cjwatson> I bet it's the bazillion duplicates of bug 442941
[12:35] <ubot4> Launchpad bug 442941 in ubiquity (Ubuntu Oneiric) (and 14 other projects) "debconf failed to upgrade from 1.5.27ubuntu1 to 1.5.27ubuntu2: exit status 128 - Use of uninitialized value $reply in scalar chomp at /usr/share/perl5/Debconf/FrontEnd/Passthrough.pm line 66 (affects: 452) (dups: 300) (heat: 2506)" [High,Fix released] https://launchpad.net/bugs/442941
[12:35]  * cjwatson wonders whether doing it on cocoplum will work better
[12:38] <cjwatson> right, there we go.  15 seconds, too long for the API I guess :-/
[12:39] <jibel> cjwatson, I can move part of the duplicates to another report if that helps.
[12:40] <cjwatson> it's ok, it's done now
[12:42] <jibel> I confirm that I can reproduce 650703 with the additional sleep in dbus.conf
[12:45] <cjwatson> good, preparing an upload now
[12:58] <cjwatson> pitti (or other ubuntu-sru member): could you review ubiquity in {lucid,maverick}-proposed, please?
[12:58] <cjwatson> well, the maverick version might not be there yet.  somebody in the bug report is asking for it urgently though so I thought I should do both
[12:59] <pitti> cjwatson: can do; still waiting a couple of minutes for the diff to arrive
[13:02] <lamont> fyi, chroot tarballs freshened for natty ('twas pre-release) and oneiric
[13:11] <pitti> cjwatson: so, admittedly I don't fully understand the correctness of the upstart condition changes, but as this was a backport, it looks fine to me
[13:11] <pitti> (accepted)
[13:11] <pitti> doing maverick a little later, when the queue diff arrives
[13:12] <cjwatson> the quick explanation is that this moves the 'stopping rc' off into a separate package that isn't installed except on server systems
[13:12] <pitti> ah, we don't install this; that makes it clearer indeed
[13:12] <cjwatson> so you get either 'stopping rc' or 'starting gdm' in your job graph but not both
[13:12] <cjwatson> typically, anyway
[13:13] <cjwatson> I don't think it's *ideal*, but I think it should be sufficient
[13:13] <cjwatson> so, once that builds and publishes, I'll do another build with -proposed, which should amount to fixing that ubiquity bug plus sorting out wubi isolist versions
[13:14] <cjwatson> once 650703 is validated, we should then push it to -updates and do a 10.04.3-candidate build with -updates only
[13:14] <cjwatson> and verify the hell out of that
[14:59] <jibel> I tried ubiquity 2.2.27 in lucid-proposed and that fixes 650703.
[15:01] <cjwatson> whee
[15:01] <cjwatson> with the dbus.conf hack to widen the race?
[15:01] <cjwatson> did you just upgrade it in place?
[15:02] <jibel> yes I upgraded in place
[15:02] <jibel> but I need to try on a real image to validate the fix because the conditions of the race are a bit different
[15:03] <jibel> for example with the sleep in dbus there is a process that locks debconf config.dat that is not running in the original bug.
[15:04] <cjwatson> OK, not entirely surprising that it would permute some things
[15:04] <cjwatson> I was waiting for the powerpc build to finish before starting images, since that will be less hassle
[16:14] <Ursinha> bug 675396
[16:14] <ubot4> Launchpad bug 675396 in nmap (Ubuntu) "Icon missing in Kubuntu Remix menu (affects: 3) (heat: 14)" [Low,New] https://launchpad.net/bugs/675396
[16:23] <cjwatson> Ursinha: is that the bug you meant to quote?  it doesn't seem urgent
[16:24] <Ursinha> cjwatson: I was just abusing the bot, sorry
[16:25] <cjwatson> ah, ok
[16:39] <micahg> !msgthebot > Ursinha
[16:39] <ubot4> Ursinha, please see my private message
[17:58] <GrueMaster> skaet: Why is ubuntu server armel+omap daily on the tracker for 10.04.3 release testing?  Is this something I need to look at?
[18:07] <skaet> GrueMaster, hmm,  that's not an image I was expecting to see...  and am not seeing the netbook ones on it.   hmm...
[18:07] <GrueMaster> There should be no armel images in 10.04.3,  We have no LTS releases at this time.
[18:08] <skaet> GrueMaster, ack.
[18:08] <GrueMaster> I do SRU testing for 10.04, but that support ends with 11.10 release (yea!).
[18:10] <skaet> cjwatson, ^^ I'm going ahead and disabling ubuntu server armel+omap,  since it wasn't on the 10.04.2 manifest.
[18:54] <skaet> cjwatson, have started off DIST=lucid buildlive ubuntu-netbook daily-preinstalled && (for-project ubuntu-netbook cron.daily-preinstalled &)
[20:46] <cjwatson> building another round of lucid images now that ubiquity 2.2.27 is in place