[00:05] <Blinkiz> Hi. I can't get the new option about encrypting my home dir with ecryptfs to work in latest jaunty iso. I have downloaded todays live-iso and at startup, I put in user-setup/encrypt-home=true. Problem is that no new options have come up under the installation. I followed the guide here: http://blog.dustinkirkland.com/2009/02/jaunty-encrypted-home-directories.html . What can be wrong?
[00:17] <Blinkiz> Found it. Its kind of a bug. Before you have chosen any of the two options, "log in automatically" or "require a password to log in", the third (new) option is not visible. Even when clicking on any of these two visible on, the third one does not appear. But a side rolling bar comes up that allows you to roll down and now the new option is visible. Clearly a bug. Sad that launchpad is down at the moment.. jeje. Letting you guys know in th
[00:17] <Blinkiz> is channel. My screen size is 1680x1050.
[00:19] <Blinkiz> Oh, btw, nice feature! Installing it on my girlfriends computer to try it out. Me? No, she has to test it first :P
[11:38] <CIA-28> partman-base: cjwatson * r147 ubuntu/ (debian/changelog debian/partman-base.templates init.d/parted): (log message trimmed)
[11:38] <CIA-28> partman-base: * Rewrite handling of mounted partitions. If the only thing mounted on a
[11:38] <CIA-28> partman-base:  disk is the installation medium and it uses more or less the whole disk,
[11:38] <CIA-28> partman-base:  then silently exclude that disk (LP: #347916); if the installation
[11:38] <CIA-28> partman-base:  medium is mounted but doesn't use the whole disk, issue a warning that
[11:38] <CIA-28> partman-base:  partitioning may be difficult; if anything else is mounted, offer to
[11:38] <CIA-28> partman-base:  unmount it.
[11:39] <ogra> wordy commit :)
[11:44] <cjwatson> and CIA truncated it too
[11:44] <cjwatson> superm1: FYI, partman/filter_mounted is going away; I hope you won't actually need it any more, but see the full changelog for partman-base r147
[11:49] <CIA-28> partman-auto: cjwatson * r283 ubuntu/ (debian/changelog lib/auto-shared.sh):
[11:49] <CIA-28> partman-auto: Exclude devices containing the installation medium from automatic
[11:49] <CIA-28> partman-auto: partitioning (LP: #347916). Requires partman-base 129ubuntu5 for this
[11:49] <CIA-28> partman-auto: change to work properly.
[11:55] <CIA-28> ubiquity: cjwatson * r3165 ubiquity/ (7 files in 5 dirs):
[11:55] <CIA-28> ubiquity: Display a warning on the advanced partitioning screen if the
[11:55] <CIA-28> ubiquity: installation medium is mounted but not filling the whole disk, so that
[11:55] <CIA-28> ubiquity: the disk containing it is still offered for partitioning (LP: #347916).
[12:05] <CIA-28> partman-base: cjwatson * r148 ubuntu/debian/po/ (64 files): debconf-updatepo
[12:10] <ogra> damned .. ubiquity goes to the configure_bootloader step just fine now, but flash-kernel fails (well, isnt installed at all)
[12:10] <cjwatson> I thought I made it a dependency
[12:11] <ogra> i thought it calls apt-install
[12:11] <ogra> (without any evidence indeed, just an assumption)
[12:11] <cjwatson> ubiquity's supposed to depend on it so that it doesn't have to actually install it
[12:11] <ogra> ah
[12:12] <ogra> its not in .manifest
[12:12] <cjwatson> ogra: what does 'dpkg-architecture -qDEB_HOST_ARCH_CPU' say on this machine?
[12:13]  * ogra installs dpkg-dev
[12:14] <ogra> arm
[12:14] <cjwatson> grah
[12:14] <ogra> and i dont see flash-kernel in the deps btw
[12:14] <cjwatson> can you show me the output of dpkg-architecture without arguments?
[12:14] <cjwatson> yes yes, it's not in the dependencies because DEB_HOST_ARCH_CPU wasn't what I expected
[12:15] <ogra> http://paste.ubuntu.com/142724/
[12:16] <cjwatson> how confusing. ok, will fix
[12:17] <ogra> yeah
[12:17] <ogra> i dont know why the "arm" is set here
[12:17] <ogra> doesnt make much sense
[12:18] <cjwatson> the CPU is indeed arm
[12:18] <ogra> well
[12:18] <cjwatson> so I think it's correct, it was just unexpected
[12:18] <ogra> its says armel everywhere else
[12:18] <cjwatson> it's fine as it is
[12:18] <cjwatson> can we not have a bikeshed about this now?
[12:19]  * ogra resets the board and retries with flash-kernel installed in advance
[12:19]  * StevenK resists the urge to comment "No, blue! Ohh, purple!"
[12:19] <StevenK> Oh, wait.
[12:20] <CIA-28> ubiquity: cjwatson * r3166 ubiquity/debian/ (changelog rules):
[12:20] <CIA-28> ubiquity: On armel, DEB_HOST_ARCH=armel but DEB_HOST_ARCH_CPU=arm. Check
[12:20] <CIA-28> ubiquity: DEB_HOST_ARCH rather than DEB_HOST_ARCH_CPU.
[12:20] <ogra> technically purple is a shade of blue though
[12:20] <ogra> just some red in it :)
[12:20] <StevenK> ogra: It's all bikeshedding :-P
[12:20] <ogra> heh
[12:20] <cjwatson> ogra: ^- should fix it, anyway
[12:21] <ogra> yep
[12:25] <ogra> hmm, that wont get me flash-kernel-installer indeed ...
[12:25] <cjwatson> err
[12:25] <cjwatson> flash-kernel-installer is built into ubiquity
[12:25] <ogra> already ?
[12:25] <cjwatson> ubiquity can't depend on udebs, so it incorporates them instead
[12:25] <cjwatson> yes
[12:26] <ogra> ah, good
[12:26] <ogra> i thought that depends on the dep
[12:31] <ogra> cjwatson, the two radio buttons not having text on the user credentials page in ubiquity is a known fact i assume ?
[12:31] <ogra> (the two ones at the bottom)
[12:32]  * ogra actually scrolled down for the first time on that page
[12:38] <cjwatson> ogra: I don't think that's known, no
[12:38] <cjwatson> they should be "Log in automatically" and "Require password", or some such
[12:38] <ogra> yeah, might be a missing translation though, i'll try an english install later
[13:17] <ogra> cjwatson, hmm, not a translation issue, its missing in english as well
[13:17] <ogra> aha, changing it makes the text appear
[13:32] <cjwatson> superm1: could you elaborate on your comment the other day that my ubiquity diff to fix bug 337748 didn't work (presumably, broke bug 290398 again)? I'm trying it out here and it seems to be doing the right thing - it filters language-support-aa out of self.langpacks, for example
[13:46] <CIA-28> ubiquity: cjwatson * r3167 ubiquity/ (debian/changelog gui/glade/stepLocation.glade): Fix spacing on timezone page (LP: #336745).
[14:05] <ogra> cjwatson, does ubiquity put an apt-install binary (or wrapper) anywhere ?
[14:05] <ogra> seems flash-kernel-installer calls it, if i run it manually it fails on it
[14:22] <cjwatson> ogra: yes, in /usr/lib/ubiquity/compat/. But don't try to run bits of ubiquity manually, really, it's a recipe for confusion
[14:22] <ogra> well, i wanted to see if the script works
[14:23] <ogra> i avent tested it in context yet
[14:23] <cjwatson> note that flash-kernel-installer will be running after /var/lib/ubiquity/apt-install-direct has been created
[14:23] <ogra> commenting out the apt-install bits (i installed the packages manually anyway) and installing flash-kernel in /target works though
[14:24] <cjwatson> ok, then I'd expect that to be fine
[14:24] <ogra> but ubiquity itself still fails in configure_bootloader()
[14:24] <cjwatson> this use of apt-install will effectively be a no-op in ubiquity anyway
[14:24] <cjwatson> look at logs before messing about doing stuff by hand
[14:24] <ogra> apt-install doesnt apply to /target, right
[14:24] <cjwatson> also, use ubiquity --debug and look at the debug log before messing about doing stuff by hnad
[14:24] <cjwatson> hand
[14:25] <cjwatson> ogra: wrong
[14:25] <ogra> well, the debug log only says flash-kernel failed with exit 1
[14:25] <cjwatson> ogra: the next step is to put set -x in the flash-kernel script
[14:25] <ogra> which made me look at flash-kernel first and check that it actually can run and there are no typos
[14:26] <ogra> (we cross checked it ten times, but you never know)
[14:26] <cjwatson> make sure to echo $? after testing it by hand, to see if it exited 1 despite producing no error messages
[14:26] <ogra> oh, indeed that i didnt do
[14:27] <ogra> hmm, shouldnt there be /var/log/installer in my installed system ?
[14:27] <ogra> ah, crap i guess it died before it copied logs
[14:30] <cjwatson> correct
[14:30] <cjwatson> but all the information is in /var/log/syslog in the livefs
[14:30] <ogra> which is gone after i tested the reboot :)
[14:46] <lool> cjwatson: Do you whether a Recommends of linux-image-imx51 on flash-kernel is required to have flash-kernel installed on the target system?
[14:47] <lool> cjwatson: Trying to decide whether 348382 is a blocker or not
[14:47] <lool> +know
[14:48] <lool> cjwatson: ogra mentionned that ubiquity would remove flash-kernel at the end of install just like it removes ubiquity and its deps
[14:51] <cjwatson> ogra: please stop being confused ;-)
[14:51] <cjwatson> lool: ubiquity ensures that flash-kernel remains installed on the relevant systems
[14:51] <cjwatson> explicitly
[14:52] <lool> cjwatson: Ok, so the bug is not a blocker; the dep doesn't prevent installation or anything, it's just nice to have it
[14:52] <lool> cjwatson: thanks
[14:52] <cjwatson> lool: correct
[15:02] <cjwatson> ogra: typo in your build-babbage-img script?
[15:02] <cjwatson>         cp $BUILDDIR/iso/install/vmlinuz $BUILDDIR/vmlinuz
[15:02] <cjwatson>         cp $BUILDIR/iso/install/initrd.gz $BUILDDIR/initrd.gz
[15:13] <ogra> cjwatson, no, thats from persia, he wants to builf alternate too from that script
[15:13] <ogra> cjwatson, look some lines up
[15:14] <ogra> cjwatson, sorry i'm trying not to be confused ... but its hard sometimes :)
[15:14] <cjwatson> ogra: that wasn't my point
[15:14] <cjwatson> ogra: BUILDDIR vs. BUILDIR
[15:15] <cjwatson> there's a D missing on the second line
[15:15] <ogra> oh, right
[15:15] <persia> ogra, It's still a typo (I do that a lot).
[15:15] <ogra> well, that code isnt used anywhere yet, thats why i missed it
[15:18] <ogra> ... typo fixed ...
[15:35] <superm1> cjwatson, re that diff for bug 337748, ubiquity was hung with that diff applied
[15:35] <superm1> did you test it with a dvd?
[15:36] <cjwatson> hung? bah
[15:36] <cjwatson> no, DVD testing is really difficult for me logistically
[15:36] <cjwatson> I'm behind 800Kbps ADSL
[15:36] <superm1> ah yuck
[15:36] <cjwatson> so basically I can't, except when I have absolutely no alternative
[15:37] <cjwatson> anyway I doubt it's testing on a DVD that's the determining factor here - it'll be testing with your huge locale preseeding thing :)
[15:37] <superm1> later this year it may become hard for me too when not in the office, my cable co is instituting monthly caps in my area at 5,10, 20, or 40 GB depending on how much I want to pay
[15:37] <cjwatson> since the added code is O(number of langpacks requested), not O(number of langpacks available)
[15:37] <superm1> right
[15:38] <cjwatson> I have a monthly cap of 50GB (though can pay extra for more ad-hoc), but usually it's the bandwidth that's the problem
[15:38] <cjwatson> basically it means a working day of not being able to use the network effectively to get a DVD
[15:39] <cjwatson> ok, well if it's poor performance, that gives me somewhere to start
[15:39] <superm1> i can try in a few hours again with it applied to get you a syslog in debug mode if it'd help
[15:42] <cjwatson> if you can, although I suspect it'll just have a big time delay in it :)
[15:43] <superm1> yeah, at least getting the area that it's got the delay will probably be helpful
[15:46] <cjwatson> hopefully in the added code ;-)
[15:47] <cjwatson> ah, that's better, decent-sized timezone map again
[15:53] <cjwatson> ogra: ISWYM, they're blank in English too. Argh
[15:56] <CIA-28> ubiquity: cjwatson * r3168 ubiquity/ (debian/changelog ubiquity/timezone_map.py):
[15:56] <CIA-28> ubiquity: Use the full size of the underlying images for the timezone map if the
[15:56] <CIA-28> ubiquity: screen is big enough.
[15:57] <cody-somerville> cjwatson, Just to confirm, debian-cd uses debian-cd task lists generated from germinate output and not the actual task meta-data in the mirror, right?
[15:58] <cody-somerville> If I remember correctly, you said the only reason we have that information in the repository is to support the net-based installs, right?
[15:58] <superm1> cjwatson, so reading that partman-base commit you refered to, how does that interact with the existing ubiquity option for filtering disks? does the ubiquity preseed do anything still?
[15:58] <cjwatson> superm1: I added an explicit comment to the ubiquity source for this
[15:58] <cjwatson> cody-somerville: right, and to support people running tasksel post-install
[15:59] <superm1> cjwatson, okay then later when i can retest that patch, i'll pull a new ubiquity and partman-base and see if i can test the interactions
[15:59] <cody-somerville> cjwatson, and a debian-cd task list is just a file with all the seeded packages and their dependencies expanded, one package per line?
[15:59] <cjwatson> superm1: the remaining purpose of the ubiquity code is insurance in case somebody mounts a partition during partitioning
[15:59] <cjwatson> cody-somerville: more or less, it also has architecture #ifdefs (processed through cpp), although you should never actually have to care what the file looks like
[16:00] <cjwatson> superm1: I wouldn't expect that you'll still need the ubiquity preseed, normally
[17:13] <cody-somerville> cjwatson, offhand do you know if livecd-rootfs uses the metapackages or taskel?
[17:14] <ogra> meta
[17:14] <ogra> look at the code :)
[17:14] <cody-somerville> ogra, my computer is very bogged at the moment as my panels are running in valgrind
[17:23] <cjwatson> cody-somerville: it uses the tasks
[17:23] <cjwatson> cody-somerville: not via tasksel, but via apt-get's foo^ syntax for installing tasks
[17:24] <cody-somerville> if it were to install ubuntu-standard and ubuntu-desktop, would the result be the same?
[17:24]  * ogra decides to keep away from answering any other questions today
[17:26] <cjwatson> cody-somerville: in a stable release, yes; in a development release, ideally but there can be temporary desynchronisation. But we like to have something that runs automatically and makes sure the tasks in the archive work ...
[17:28] <cody-somerville> Is minimal pulled in via debootstrap or would I have to install ubuntu-minimal as well to ensure parity?
[17:28] <ogra> cjwatson, hmm, so setting flash-kernel-installer "set -ex" i see ubiquity: "in-target flash-kernel" /usr/lib/ubiquity/flash-kernel/flash-kernel-installer: 1: in-target: not found
[17:29] <ogra> and i see apt-install working fine above ...
[17:30] <cjwatson> cody-somerville: livecd-rootfs does it just in case, but debootstrap will install minimal
[17:32] <cjwatson> ogra: ok, I'll look into that
[17:32] <cjwatson> ubiquity doesn't have in-target right now, but probably needs a compatibility shim for it
[17:32] <ogra> well, i can as well special case it in flash-kernel-installer
[17:32] <cjwatson> no, please don't
[17:33] <ogra> though that would end up in a croot call
[17:33] <ogra> *chroot
[17:33] <ogra> which in-target tries to prevent
[17:33] <ogra> do you want/need a bug  ?
[17:33] <cjwatson> yes please
[17:33] <cjwatson> in-target tries to prevent> WTF? no it doesn't
[17:34] <cjwatson> in-target is a wrapper around chroot which does some extra stuff like connecting up debconf properly and mounting virtual filesystems
[17:34] <ogra> yes, it tries to prevent me from making my own chroot calls :)
[17:35] <cjwatson> I'll look at it later. Dinner now
[17:48] <ogra> cjwatson, Bug #353941
[18:58] <CIA-28> debian-installer: cjwatson * r1079 ubuntu/debian/changelog: releasing version 20081029ubuntu31
[19:01] <CIA-28> partman-base: cjwatson * r149 ubuntu/debian/changelog: releasing version 129ubuntu5
[19:02] <CIA-28> partman-auto: cjwatson * r284 ubuntu/debian/changelog: releasing version 84ubuntu2
[19:04] <CIA-28> ubiquity: cjwatson * r3169 ubiquity/debian/po/ (80 files): debconf-updatepo
[19:07] <CIA-28> ubiquity: cjwatson * r3170 ubiquity/debian/ (changelog ubiquity.install-any):
[19:07] <CIA-28> ubiquity: Install chroot-setup.sh and in-target from debian-installer-utils
[19:07] <CIA-28> ubiquity: (LP: #353941).
[19:10] <ogra> sigh ... still no success
[19:11] <cjwatson> superm1: bug 353273: is this a DVD?
[19:12] <cjwatson> oh, grub-pc Conflicts: grub. Bah.
[19:14] <superm1> cjwatson, yeah that's a DVD
[19:15] <superm1> er wait, no that wasn't
[19:15] <superm1> that was a CD with a network connection plugged in.  dvd's weren't ready yet
[19:15] <cjwatson> yeah, the DVD broke due to my attempt to fix your previous bug ;-)
[19:15] <cjwatson> the problem here is that grub and grub-pc conflict
[19:16] <cjwatson> so I think the best answer is indeed to make ubiquity recommend grub rather than depending on it
[19:16] <cjwatson> that'll let us remove it temporarily
[19:16] <CIA-28> ubiquity: cjwatson * r3171 ubiquity/debian/ (changelog control rules):
[19:16] <CIA-28> ubiquity: Move bootloader dependencies to recommendations, to make it easier to
[19:16] <CIA-28> ubiquity: cope with grub and grub-pc conflicting (LP: #353273).
[19:16] <cjwatson> superm1: I've committed a change which ought to fix this, but please accept my apologies if it doesn't - as previously mentioned I have a hard time testing DVDs :-(
[19:16] <cjwatson> if it breaks again, please reopen that bug and I'll try to look in more detail
[19:17] <superm1> cjwatson, yeah I understand.
[19:17] <superm1> if you can upload a ubiquity again with these changes tonight before the next set of rebuilds, I can at least check from a CD
[19:17] <superm1> to be able sort out the grub-installer bit of it
[19:17] <cjwatson> aye, planning to
[19:20] <cjwatson> bug 353090 is worrying me though
[19:22] <ogra> yeah, its ver yweird
[19:22] <ogra> *very weird
[19:23] <superm1> compiz redraw problem maybe though?
[19:23]  * ogra wishes he would have to debug something thats at the start of the install, not at the end 
[19:23] <ogra> superm1, nope
[19:24] <ogra> happens on non compiz setups as well
[19:24] <ogra> either gtk or the way gtk is used i would suspect and it wasnt there before the scrollbar showed up on that page
[19:48] <ogra> ARGH!
[19:48] <ogra> mv /target/usr/sbin/flash-kernel /target/tmp/flash-kernel.$$
[19:48] <ogra> *sigh*
[22:59] <cjwatson> ogra: mm, it's possible it's a scrollbar thing, though there were several other changes at around the same time
[22:59] <cody-somerville> cjwatson, Currently to extend the Debian Installer you just need to include a udeb. Does ubiquity have the same sort of support?
[23:00] <cjwatson> cody-somerville: it's not as straightforward I'm afraid
[23:00] <cjwatson> cody-somerville: but I'd prefer to explain it at some time that isn't 11pm local :)
[23:00]  * cody-somerville nods.
[23:00] <cjwatson> cody-somerville: there's some explanation in doc/README in ubiquity's source
[23:00] <cody-somerville> I'll take a look
[23:01] <cjwatson> I suggest you read the materials available there, and come back if there's anything you don't understand
[23:01] <cjwatson> d-i/source/ and ubiquity/components/ may also be worth looking at
[23:13] <CIA-28> ubiquity: cjwatson * r3172 ubiquity/ (d-i/manifest debian/changelog):
[23:13] <CIA-28> ubiquity: Automatic update of included source packages: partman-auto 84ubuntu2,
[23:13] <CIA-28> ubiquity: partman-base 129ubuntu5, partman-ext3 56ubuntu1, user-setup
[23:13] <CIA-28> ubiquity: 1.23ubuntu16.
[23:29] <CIA-28> ubiquity: cjwatson * r3173 ubiquity/debian/changelog: releasing version 1.12.2