[00:07] <CIA-33> ubiquity: cjwatson * r3525 ubiquity/debian/changelog: releasing version 2.0.0
[02:06] <NCommander> hey cjwatson I'm working with partman-uboot, and it doesn't seem to let me specify the mountpoint of that partition in ubiquity (the box is greyed out). Any ideas?
[02:06] <NCommander> s/greyed out/greyed out showing /boot in it/g
[07:15] <cjwatson> NCommander: perhaps you need to add cases for partman-uboot/mountpoint wherever you see cases for partman-basicfilesystems/fat_mountpoint
[07:16] <cjwatson> NCommander: also note some unfortunate hardcoding of known filesystem types in the frontends
[07:16] <cjwatson> search for fat16 or something
[07:18] <NCommander> cjwatson, well, its doing the right thing
[07:18] <NCommander> cjwatson, it always makes the mountpoint /boot so ...
[07:18] <NCommander> I'm kinda inclined to leave it
[07:20] <NCommander> cjwatson, who can give me an exception to the translation string freeze. I'd like to improve the message in partman-uboot when it doesn't like the partition layout; obviously not critical, but it would be nice
[07:20] <cjwatson> eh, it's not like anyone had translated that
[07:20] <cjwatson> "kinda inclined to leave it" - I'm only really prepared to merge a patch for this if it actually handles the partman interaction properly :)
[07:21] <cjwatson> which I think does mean accounting for the different questions it asks
[07:25] <NCommander> cjwatson, come again?
[07:26] <NCommander> cjwatson, the messages about this layout not working shows up just file. Its just not clear on what the correct layout it
[07:28] <NCommander> cjwatson, anyway, just adding the package to the packaging lists did the trick. No special mucking around needed
[07:59] <NCommander> cjwatson, I'll cook up branches for you with the template change, the partman-auto change to go from ext2 -> uboot, and for ubiquity
[08:21] <cjwatson> NCommander: handling for partman-uboot/mountpoint needs to be added or else ubiquity won't offer the correct set of choices for it
[08:21] <cjwatson> at least not in all cases
[08:22] <cjwatson> I'm not just imposing random capricious requirements for no reason :)
[08:28] <NCommander> cjwatson, added where?
[08:29]  * NCommander misread what you said earlier
[08:29] <NCommander> sorry, my brain is in neutral at the moment
[08:34] <cjwatson> best to search for anywhere in ubiquity that mentions partman-basicfilesystems/fat_mountpoint or fat16
[08:34] <cjwatson> those are handy markers for places you probably also need to handle uboot
[08:35] <NCommander> cjwatson, woo, fun :-/. *sigh*
[08:36] <NCommander> cjwatson, and we enter final freeze today
[08:58] <evand> cjwatson: you're a star.  Thanks for your help, the new Webkit appears to fix the bug: https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/445756/comments/5 https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/440922/comments/7
[09:19] <cjwatson> evand: yay, sometimes hunches pay off
[09:41] <xivulon1> cjwatson, evand sent the strace logs via email
[09:41] <xivulon1> with strace on the jams happens somewhat earlier, not sure if I did anything wrong
[11:15] <CIA-33> ubiquity-slideshow-ubuntu: evand * r170 ubiquity-slideshow-ubuntu/debian/ (changelog copyright): Updated Ubuntu and Ubuntu One logo license (LP: #451263).
[11:49] <CIA-33> ubiquity-slideshow-ubuntu: evand * r171 ubiquity-slideshow-ubuntu/ (565 files in 13 dirs): Updated translations from Launchpad (source package).
[11:49] <evand> gah, I sure hope ubuntu-release arms themselves with filterdiff.
[11:54] <cjwatson> xivulon: hmm, that's odd, this log seems to suggest that it gets successfully to the point where you've confirmed partitioning, but then simply does nothing else - does that match your observations?
[11:56] <cjwatson> I'm wondering if we should stop sreadahead while starting ubiquity, to eliminate a source of random weirdness ...
[11:58] <davmor2> cjwatson: If it's alright to do so feel free to respin for tests ;)
[11:59] <cjwatson> well, it would be possible to test this without a respin
[11:59] <cjwatson> boot with break=casper-bottom, rm /root/etc/init/sreadahead.conf
[12:02] <davmor2> cjwatson: I'll have a go at that after then
[12:04] <CIA-33> usb-creator: evand * r235 trunk/debian/changelog: releasing version 0.2.11
[12:04] <CIA-33> ubiquity-slideshow-ubuntu: evand * r172 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 9
[12:08] <xivulon> cjwatson, the behaviour is slightly different if I have strace on, without I get to see the slides
[12:09] <xivulon> also as mentioned, I had this msg about sreadahead beign killed
[12:09] <xivulon> evand, by the way, small commit to enable ext4 by default
[12:10] <evand> xivulon: indeed, I saw that yesterday, just forgot to process it
[12:10] <evand> on it now
[12:10] <cjwatson> xivulon: is that wise? do we know that ext4 behaves better?
[12:10]  * evand stops :)
[12:11] <xivulon> was hoping it would improve things, but since it didn't we can probably avoid the commit
[12:11] <cjwatson> we can try, I suppose
[12:11] <cjwatson> we just don't have a lot of iterations lefet
[12:11] <xivulon> well not sure it didn't the partman jam is different from the filesystem hang (I assume)
[12:12] <cjwatson> I suggest using $default_filesystem{ } rather than filesystem{ ext4 }
[12:12] <cjwatson> that way it can be controlled from the grub menu
[12:12] <cjwatson> (if necessary)
[12:12] <cjwatson> xivulon: I'm not sure that assumption is warranted yet ...
[12:12] <cjwatson> I've gone over the code in partman that's deadlocking very carefully, and can't see any intrinsic flaws in it assuming that other things are working correctly
[12:14] <xivulon> will do more testing tonight, if you can think of anything I could do to help debugging send me an email
[12:15] <xivulon> evand please feel free to change wubi preseeding as suggested by colin
[12:38] <CIA-33> wubi: evand * r158 trunk/ (debian/changelog src/wubi/backends/common/backend.py): Use $default_filesystem in place of ext4.
[12:54] <evand> xivulon: done, built with grub2 1.97~beta4-1ubuntu1 and uploaded
[13:06] <CIA-33> debian-installer: lool * r1196 ubuntu/ (5 files in 4 dirs): Drop dove-z0 armel flavour as we dropped the kernel flavour.
[13:06] <lool> cjwatson: Ok to upload? ^
[13:06] <lool> (I didn't test build it; I expect it's relatively safe)
[13:08] <cjwatson> I'd like to check why powerpc is failing to build first
[13:08] <lool> Ok
[13:09] <cjwatson> oh my, genisoimage is segfaulting, that can't be goood
[13:09] <cjwatson> good
[13:11] <evand> lool: *cough* https://edge.launchpad.net/~ubuntu-cowboy ;)
[13:19] <tormod> cjwatson, installing on dmraid now: looks great, except "GrubInstaller failed with code 10"
[13:39] <tormod> cjwatson, I filed bug 452118, but will leave it untouched for a while in case you'd like some investigation
[13:45] <evand> hilarious.  On crash in only-ubiquity, ubiquity is running the noninteractive frontend, which happily plows through and installs to / on the liveCD, then gdm kicks in, and you're brought to a desktop with a full drive and half the contents missing.
[13:45] <evand> as far as I can tell, anyway
[13:45]  * evand digs
[13:49] <cjwatson> tormod: hmm - don't suppose you can repeat the installation test, but running 'ubiquity -d' from a terminal to start the installer? (I'd recommend rebooting first)
[13:50] <tormod> cjwatson, I can repeat the whole thing if needed.
[13:51] <cjwatson> with -d, we should get a useful debconf trace in /var/log/installer/debug
[13:51] <tormod> the whole install is rather quick now that partman has less to scan through
[13:51] <tormod> why is rebooting preferred? debconf cruft?
[13:52] <cjwatson> error code 10 means "bad parameters" to some debconf command, usually "question doesn't exist", but I don't know which question
[13:52] <cjwatson> es
[13:52] <cjwatson> yes
[13:55] <lool> evand: Sounds cool; I applied even if I dont know quite why  :)
[13:55] <tormod> okdok, brb
[13:55] <evand> :)
[13:57] <lool> evand: Gosh you're not in https://launchpad.net/~we-love-pitti
[13:57]  * evand joins
[14:14] <tormod> cjwatson, ok uploaded the debug log
[14:27] <Nivex> cjwatson: confirmed 441690 is fixed.  Thanks!
[14:27] <cjwatson> excellent
[14:35] <evand> cjwatson: just saw your changelog entry for 2.0.0.  Wonderfully put.  Thank you for all of your hard work in making that happen.
[14:37] <CIA-33> ubiquity: evand * r3526 ubiquity/debian/ (changelog ubiquity.ubiquity.upstart):
[14:37] <CIA-33> ubiquity: Check to see whether X crashed or ubiquity crashed before attempting
[14:37] <CIA-33> ubiquity: to bail into the noninteractive frontend from only-ubiquity mode
[14:37] <CIA-33> ubiquity: (LP: #444901).
[14:38] <cjwatson> evand: I figured we'd better not forget to bump the version, and it seemed like a good opportunity for nostalgia. :-)
[14:38] <evand> :)
[14:42] <tormod> cjwatson, I have only 1+ h left on this machine today, so please tell if I can test something
[14:50] <cjwatson> sorry, I was waiting for my browser to reboot ...
[14:57] <cjwatson> tormod: thanks, I have all I need now
[14:58] <tormod> cjwatson, great, should I try running grub-installer by hand?
[14:58] <tormod> or I can test again tomorrow (I keep Jaunty on this machine, so I can retest Karmic installations for some time)
[14:59] <tormod> btw, has the broken "Examples" link icon been reported?
[15:06] <cjwatson> tormod: I'll have it fixed by tomorrow; just trying to figure out the best fix at the moment
[15:06] <cjwatson> running grub-installer by hand is unlikely to be very useful ...
[15:07] <tormod> just to get things booting, should I use install-grub then?
[15:08] <tormod> err grub-install
[15:10] <CIA-33> grub-installer: cjwatson * r819 ubuntu/ (debian/changelog grub-installer): Always use GRUB Legacy when /boot is on a multipath device.
[15:10] <cjwatson> the installation will probably be broken if you do that
[15:10] <cjwatson> ubiquity hadn't finished yet
[15:10] <cjwatson> let me see if I can get you a patch
[15:14] <CIA-33> grub-installer: cjwatson * r820 ubuntu/ (debian/changelog grub-installer):
[15:14] <CIA-33> grub-installer: If /boot is on SATA RAID or multipath, don't ask a question to which
[15:14] <CIA-33> grub-installer: there's only one acceptable answer and the other causes grub-installer
[15:14] <CIA-33> grub-installer: to fall over (LP: #452118).
[15:15] <cjwatson> tormod: apply http://bazaar.launchpad.net/~ubuntu-core-dev/grub-installer/ubuntu/revision/820 to /usr/share/grub-installer/grub-installer before starting installation
[15:16] <CIA-33> grub-installer: cjwatson * r821 ubuntu/debian/changelog: releasing version 1.43ubuntu6
[15:16] <evand> *sigh* bug 156177
[15:18] <tormod> cjwatson, ok I thought boot loader was the very last thing to do before umounting target
[15:18] <tormod> thanks will try patch
[15:18] <cjwatson> it isn't - installing/removing extra packages comes after that, as well as copying log files over
[15:18] <tormod> can I try it without rebooting?
[15:19] <cjwatson> probably
[15:19] <tormod> nice
[15:19] <cjwatson> you'll have to run the install itself from scratch though
[15:20] <cjwatson> hmm, this d-i powerpc build failure is awkward
[15:20] <cjwatson> it appears to be a near-the-fencepost condition in genisoimage
[15:21] <cjwatson> I think if you're something like within five sectors of some boundary, it falls over
[15:21] <cjwatson> but I have no hardware to test a fix on
[15:26] <cjwatson> workaround: random padding file (ugh!)
[15:34] <tormod> cjwatson, "Unable to configure GRUB" Executing 'update-grub' failed. This is a fatal error
[15:35] <tormod> devkit-disks-daemon also died, no idea if it's related
[15:39] <cjwatson> tormod: try reproducing after a reboot again?
[15:39] <cjwatson> I don't know what might be wrong just from that error message alone
[15:39] <cjwatson> actually, get me the log before you reboot?
[15:39] <tormod> I will upload the log to the bug
[15:39] <tormod> done
[15:42] <tormod> this made a good menu.lst: sudo chroot /target update-grub
[15:43] <tormod> (that is, ignoring my other OS'es though, same with Migration-assistant)
[15:47] <cjwatson> urgh
[15:47] <cjwatson> dear me, that'll require some thought
[15:47] <tormod> I have a /target/var/lib/dpkg/info/grub.templates now though
[15:47] <cjwatson> I think we need a special implementation of X_LOADTEMPLATEFILE in debconffilter to account for the chroot
[15:54] <tormod> cjwatson, ok I gonna reboot for smoketest now
[15:56] <CIA-33> ubiquity: cjwatson * r3527 ubiquity/ (debian/changelog ubiquity/debconffilter.py):
[15:56] <CIA-33> ubiquity: Rewrite X_LOADTEMPLATEFILE commands in case they refer to template files
[15:56] <CIA-33> ubiquity: in the /target chroot (LP: #452118).
[16:03] <tormod> cjwatson: failed to boot because dmraid was not installed. what you said about remove/add packages _after_ grub installation :)
[16:04] <tormod> also it got the root wrong in menu.lst, hd0,0 instead of hd0,2
[16:12] <cjwatson> right, that's also due to your hacking
[16:12] <cjwatson> grub-installer fixes that up
[16:12] <cjwatson> we'll do it properly when we get dmraid implemented in grub2 :)
[16:13] <tormod> funny enough dmraid was installed according to apt-get, but e.g./sbin/dmraid was not there
[16:14] <tormod> I did a --reinstall in chroot, will try booting again
[16:18] <davmor2> cjwatson: tried your sreadahead line on kubuntu = fail.  Installer stopped at 100% Computing the new partitions
[16:19] <cjwatson> ok
[16:20] <CIA-33> debian-installer: cjwatson * r1197 ubuntu/ (5 files in 3 dirs):
[16:20] <CIA-33> debian-installer: Horrible workaround for LP #452212: add a 512*50-byte padding file to
[16:20] <CIA-33> debian-installer: the powerpc64_netboot mini.iso image tree, to avoid a genisoimage
[16:20] <CIA-33> debian-installer: segfault at certain tree sizes.
[16:26] <cjwatson> evand: have you got anywhere with that partman-target excessive-deletion bug?
[16:31] <evand> cjwatson: not yet, I'll try to sort out the blacklist generated from Contents tomorrow.  I'll make the whitelist for /usr/local, etc now.
[16:42] <davmor2> evand: so will that wubi change make it onto the iso tomorrow?
[16:42] <evand> yarp
[16:42] <davmor2> Okay cool I'll try it again with that in place :)
[17:11] <tormod> sports were canceled so I got to exercise dmraid and ubiquity some more...
[17:13] <tormod> cjwatson, I rebooted the live CD again, applied both the ubiquity 3527 and the grub-installer 820 and ... it just worked \o/
[17:13] <cjwatson> perfect
[17:13] <tormod> you rox
[17:13] <cjwatson> (phew)
[17:14] <tormod> the only trouble (serious enough) is that the raw devices are not hidden (they are in the live session)
[17:14] <tormod> but that must be a dmraid/udev issue and not the installer's fault
[17:25] <CIA-33> debian-installer: cjwatson * r1198 ubuntu/ (38 files in 2 dirs): Update help text translations from Launchpad.
[17:30] <tormod> the "devices not hidden" thing is intermittent, good old bug 362768 I guess
[17:31] <CIA-33> debian-installer: cjwatson * r1199 ubuntu/debian/changelog: releasing version 20081029ubuntu69
[18:04] <LaserJock> cjwatson: I've got a d-i related question for Edubuntu
[18:04] <LaserJock> cjwatson: the Edubuntu DVD doesn't have the LTSP option in the F4 boot menu
[18:04] <LaserJock> but it seems that LTSP is always installed when using d-i
[18:04] <LaserJock> ogra suggested we needed to have the ltsp.preseed
[18:06] <LaserJock> that is the ltsp.preseed from the ubuntu preseeds folder
[19:49] <cjwatson> LaserJock: that at least used to be included in the default installation option for Edubuntu
[19:49] <cjwatson> LaserJock: which certainly still includes most of the LTSP preseeding; the only thing it doesn't do is add ltsp-server-standalone and openssh-server to pkgsel/include
[19:51] <LaserJock> cjwatson: so right now it seems that LTSP gets installed no matter what task you pick
[19:52] <LaserJock> but there's no option either in the main boot selections or the F4 menu for it
[19:52] <LaserJock> so do the .preseed files determine that?
[19:53] <LaserJock> I'm sort of confused as to what the boot UI hooks up to
[19:56] <cjwatson> I believe it depends on what *boot menu option* you pick, not what task
[19:56] <cjwatson> this is almost certainly carried over from pre-DVD days; I need you guys to maintain it by sending meus patches, really :)
[19:56] <cjwatson> s/meus/us/
[19:57] <LaserJock> cjwatson: right, I'm just trying to figure out what stuff to muck around with
[19:58] <LaserJock> I found the .preseed files, but I'm not sure where the actual menu items come from
[19:58] <cjwatson> menu items are written out by the files in tools/boot/karmic/, in code
[19:59] <cjwatson> they generally include preseed/file=<something> which links up to the .seed files
[19:59] <cjwatson> the menu items are translated in the gfxboot-theme-ubuntu package
[19:59] <LaserJock> ok, that helps
[21:20] <LaserJock> cjwatson: do you know offhand what happens when the boot parameter line declares a preseed, but the preseed is not on the .iso?
[21:23] <cjwatson> you get an error screen in d-i, and otherwise it's as if you didn't preseed anything
[21:24] <LaserJock> ok
[21:25] <LaserJock> looks like that's happening on the Edubuntu DVD currently
[21:25] <LaserJock> wait, not it's not
[21:25] <LaserJock> nvm, looking at the wrong file listing
[21:41] <LaserJock> cjwatson: well, I figured out the problem with LTSP not showing up
[21:41] <LaserJock> cjwatson: does a merge request work OK?
[21:42] <cjwatson> should do
[22:03] <bdmurray> cjwatson: is bug 451872 in the right place?
[22:08] <cjwatson> bdmurray: same cause as bug 381037
[22:09] <bdmurray> cjwatson: okay, that's why I asked!
[22:09] <cjwatson> I checked, all the bits for e100 are in the initramfs already, so it's just a matter of adding the firmware loader
[22:12] <LaserJock> cjwatson: merge request sent, thanks for pointing the right direction.
[22:13] <bdmurray> and bug 445522 is probably the same
[22:13] <cjwatson> before duping, check whether the relevant module is actually in the initramfs
[22:14] <cjwatson> zcat /boot/initrd.img-whatever | cpio -it | grep modulename
[22:14] <cjwatson> actually, 445522 is clearly different
[22:14] <cjwatson> oh, wait, LACK OF SLEEP
[22:15] <cjwatson> 451872 is talking about the *installer*, not the installed-system initrd!
[22:15] <cjwatson> so that's a separate bug - similar cause, and belongs on udev, but it needs to be fixed independently
[22:15] <cjwatson> 445522 is the same as 451872
[22:16] <bdmurray> okay, I hadn't thought there dups of the udev one but similar
[22:16] <bdmurray> so I'll move them to udev and dup one to the other
[22:16] <cjwatson> double-wait
[22:16] <cjwatson> argh, I am so crap today
[22:18] <cjwatson> third try: they're still dups, but they belong on linux
[22:18] <cjwatson> on further investigation, udev-udeb *is* shipping the firmware loader as it should
[22:18] <cjwatson> the problem is that nothing is delivering those firmware blobs to the installer
[22:19] <cjwatson> not a general problem, as *some* of the e100 firmware is delivered
[22:20] <cjwatson> bdmurray: can you please make sure it's tagged as a regression and the kernel team is informed? I'll get a patch to them
[22:20] <cjwatson> because this can't feasibly be fixed post-release, and so we'll need to run it through post-hasste
[22:21] <bdmurray> cjwatson: yes, no problem
[22:25] <cjwatson> I think qlogic may have a similar issue
[22:26] <cjwatson> though I can't find a report of that
[22:29] <bdmurray> bug 328550 is regarding qlogic and firmware loading
[22:33] <cjwatson> sounds unrelated
[22:33] <cjwatson> that's post-install rather than in d-i
[22:37] <cjwatson> kernel-team@ has a pair of patches now