[00:07] <shtylman> cjwatson: lets say I have a package which depends on other packages (not unheard of I hear :) ... and lets say that the packages I depend on like to ask annoying questions? during their install ... but my package provides configuration files so those questions are irrelevant... can I supress them with my package during the install?
[00:08] <cjwatson> nope
[00:08] <shtylman> ;(
[00:09] <cjwatson> debconf questions are usually asked by .config scripts, which are way early on, before any packages are even unpacked
[00:09] <shtylman> I see
[00:09] <cjwatson> you'll need to work with the relevant package maintainers to improve things somehow
[00:09] <cjwatson> or preseed things in a non-packaged way
[00:10] <shtylman> k
=== robbiew is now known as robbiew_
[00:38] <CIA-3> partman-base: cjwatson * r186 ubuntu/debian/ (changelog control): Build against parted 2.2.
[00:41] <CIA-3> partman-base: cjwatson * r187 ubuntu/debian/changelog: releasing version 138ubuntu2
[00:46] <CIA-3> partconf: cjwatson * r837 ubuntu/debian/ (changelog control): Build against parted 2.2.
[00:47] <CIA-3> partconf: cjwatson * r838 ubuntu/debian/changelog: releasing version 1.32ubuntu2
[00:49] <StevenK> cjwatson: You'll do the -15 -> -16 update for d-i, or shall I?
[00:51] <cody-somerville> What involved there?
[00:51] <cody-somerville> *Whats
[00:51] <cjwatson> StevenK: I can do it now
[00:52] <cjwatson> cody-somerville: sed over some files in build/config/, and a seed change
[00:56] <CIA-3> debian-installer: cjwatson * r1254 ubuntu/ (8 files in 2 dirs): Move to 2.6.32-16 kernels.
[00:58] <CIA-3> debian-installer: cjwatson * r1255 ubuntu/debian/changelog: releasing version 20081029ubuntu89
[00:58] <superm1> i've been curious, how come you have always explicitly had to change that in the seeds?  is it not possible to just seed the meta and let the meta handle grabbing the newest?
[00:59] <cjwatson> no it isn't, but I'm too tired to explain properly :)
[01:00] <cjwatson> basically metapackages aren't applicable in this position
[01:00] <superm1> okay :)
[01:00] <cjwatson> think udebs
[01:00] <superm1> it just seems so tedious to have to always be updating, so that's a shame
[01:00] <cjwatson> it's trivial :)
[01:00] <cjwatson> there are like a million more tedious things ahead of that one
[02:47] <CIA-3> usb-creator: superm1 * r273 usb-creator/ (debian/changelog usbcreator/frontends/gtk/frontend.py):
[02:47] <CIA-3> usb-creator: * Don't force select a source unless the list is empty.
[02:47] <CIA-3> usb-creator: * Reverse the order of the populating ISOs and command line option --iso
[02:47] <CIA-3> usb-creator:  again so that --iso always trumps others.
[02:47] <CIA-3> usb-creator: * Hide the source_vbox again when --iso is used.
[06:54] <Anzenketh> Hi all working on the bug day For ubiquity is there anything that you would like me to put in special notes besides what is already there?
=== dpm-afk is now known as dpm
[09:31] <CIA-3> ubiquity: evand * r3900 ubiquity/ (7 files in 5 dirs): Merge with Roman's Kubuntu branch.
[09:44] <CIA-3> ubiquity: evand * r3901 ubiquity/debian/changelog: Add a changelog entry on behalf of Roman.
[09:46] <CIA-3> ubiquity: evand * r3902 ubiquity/debian/ (5 files):
[09:46] <CIA-3> ubiquity: Remove the intro plugin. It's now presented as part of the language
[09:46] <CIA-3> ubiquity: page when the greeter option (-g) is selected.
[10:19] <CIA-3> ubiquity: evand * r3903 ubiquity/ (debian/changelog ubiquity/components/ubi-language.py):
[10:19] <CIA-3> ubiquity: Support setting the language for the live environment from the
[10:19] <CIA-3> ubiquity: language page in greeter mode by finishing localechooser when 'Try
[10:19] <CIA-3> ubiquity: Ubuntu' is clicked.
[10:29] <CIA-3> ubiquity: evand * r3904 ubiquity/ubiquity/frontend/gtk_ui.py: Add missing ui reference.
[10:29] <ev> pychecker saves the day again
[10:41] <CIA-3> ubiquity: evand * r3905 ubiquity/ (3 files in 3 dirs):
[10:41] <CIA-3> ubiquity: Fix missing references in the noninteractive portion of the
[10:41] <CIA-3> ubiquity: usersetup plugin.
[11:05] <CIA-3> grub-installer: cjwatson * r841 ubuntu/ (debian/changelog grub-installer): Fix partition offset handling for GRUB Legacy.
[11:24] <CIA-3> ubiquity: evand * r3906 ubiquity/ubiquity/ (3 files in 2 dirs):
[11:24] <CIA-3> ubiquity: * Fix incorrect reference in KDE greeter code.
[11:24] <CIA-3> ubiquity: * Fix missing reference to the progress title in the GTK+ partman progress
[11:24] <CIA-3> ubiquity:  bar handling.
[11:24] <CIA-3> ubiquity: * Removed dead code from the noninteractive frontend.
[11:28] <ev> cjwatson: do you have any more changes to grub-installer in the pipe?  I'd like to release a new ubiquity and respin the CDs.
[11:29] <cjwatson> no, uploading now
[11:30] <CIA-3> grub-installer: cjwatson * r842 ubuntu/debian/changelog: releasing version 1.49ubuntu6
[11:31] <michaelforrest1> cjwatson: can't see the new branding on that boot splash screen in the installer yet - did it go in?
[11:32] <cjwatson> days ago ...
[11:32]  * cjwatson rsyncs to check but I'm pretty certain it was there last I looked
[11:33] <cjwatson> you sure you're checking a daily, not alpha-3?
[11:36] <ev> cjwatson: thanks!
[11:38] <michaelforrest1> cjwatson: I'm on the daily-live image yes
[11:39] <michaelforrest1> it's got the first version I sent you, but I sent you that new stuff on Thursday after the brand rollout with the new logo and glowy icons
[11:39] <michaelforrest1> on purple
[11:40] <michaelforrest1> sorry: "aubergine"
[11:44] <cjwatson> michaelforrest1: oh, was that the mail subject "boot image" with no body?
[11:44] <michaelforrest1> quite probably
[11:44] <cjwatson> no, that's got the old logo on black
[11:44] <cjwatson> michaelforrest1: I do not have this stuff from you
[11:45] <michaelforrest1> cjwatson: The email is called "New Brand for Boot Screen"
[11:45] <michaelforrest1> you replied to it
[11:45] <cjwatson> and I integrated the image from the followup you sent me to that
[11:45] <cjwatson> the 640x350 one
[11:46] <cjwatson> or at least I certainly thought I did
[11:46] <michaelforrest1> so it's supposed to look like this: http://assets.grimaceworks.com/boot_experience_640_wide.png
[11:47] <michaelforrest1> but when I look at the daily build, it still looks like this: http://assets.grimaceworks.com/boot-screen.png
[11:47] <cjwatson> right, that's what I thought I'd done.  I must have made some trivial mistake in the build scripts
[11:47] <cjwatson> thanks, I'll fix that
[11:47] <michaelforrest1> ok cool.
[11:47] <michaelforrest1> thanks
[11:47] <cjwatson> I probably just forgot to switch over the image filename or something
[11:48] <cjwatson> hmm, or the .png file is getting used rather than the .pcx.  I wonder why!
[12:16] <CIA-3> ubiquity: evand * r3907 ubiquity/debian/real-po/ (81 files): debconf-updatepo
[13:11] <CIA-3> ubiquity: evand * r3908 ubiquity/ (d-i/manifest debian/changelog debian/control):
[13:11] <CIA-3> ubiquity: Automatic update of included source packages: bterm-unifont
[13:11] <CIA-3> ubiquity: 1.1build1, flash-kernel 2.13ubuntu14, grub-installer 1.49ubuntu6,
[13:11] <CIA-3> ubiquity: migration-assistant 0.6.4build1, partconf 1.32ubuntu2, partman-base
[13:11] <CIA-3> ubiquity: 138ubuntu2.
[13:31] <CIA-3> ubiquity: evand * r3909 ubiquity/debian/changelog: releasing version 2.1.33
[13:44] <cjwatson> michaelforrest1: ha!  I bet it would have helped if I'd run 'bzr pull' on the machine that does the CD builds. ;-)
[13:45] <michaelforrest1> cjwatson: :)
[14:15] <cjwatson> michaelforrest1: can you please tell me how I can reproduce this problem about boot option text being missing, and there just being "live", "install", etc. instead?
[14:16] <michaelforrest1> we ran the installer from yesterday's daily
[14:16] <michaelforrest1> cjwatson: so I guess you should see it that way
[14:16] <cjwatson> sorry, can you please be more specific
[14:16] <cjwatson> what exactly did you do
[14:17] <michaelforrest1> put in the usb key and pressed a key during boot
[14:17] <cjwatson> and where exactly did you see this text?
[14:17] <michaelforrest1> in the gfxboot screen
[14:17] <cjwatson> you see, I tried this with today's daily, but there have been no relevant changes since yesterday
[14:17] <cjwatson> so obviously you and I are doing something different
[14:17] <cjwatson> perhaps it's specific to the use of USB keys?
[14:17] <michaelforrest1> this was Mark and Iain
[14:18] <cjwatson> what I am trying to do is turn this into an actual bug report
[14:18] <michaelforrest1> can you try TheDoctor ?
[14:18] <cjwatson> since it's clearly just a bug
[14:18] <cjwatson> ok
[14:18] <michaelforrest1> I came into the session late
=== robbiew_ is now known as robbiew
[14:43] <ev> cjwatson: I suspect it might be a usb-creator bug
[14:43] <ev> that kind of thing happened the last time configuration files were moved around in isolinux
[14:44] <cjwatson> yeah, it seems possible doesn't it.  It does seem to copy all of the isolinux directory though?
[14:44] <ev> it does, but then it mucks with it
[14:44] <ev> to add try-usb and such
[14:44] <ev> and that's generally where things burst into flames
[14:44] <cjwatson> the other obvious possibility is gfxboot-theme-ubuntu failing to find stuff under /syslinux/, but the filenames aren't complicated or anything
[14:45] <ev> indeed
[15:03] <shtylman> ev: the kubuntu branch got merged... I was under the impression it was waiting on me to add the changelog entries and take out the stale code?
[15:05] <ev> shtylman: I added changelog entries on your behalf.  I wanted to get an upload in as we're quickly approaching beta freeze.
[15:05] <ev> by all means take out the stale code from trunk
[15:05] <shtylman> ev: noted :)
[15:05]  * ev postoffice
[15:50] <michaelforrest1> ev: any idea why dd is so damn slow??
[15:51] <ev> michaelforrest1: ...it isn't?
[15:52] <ev> keep in mind you're writing to a USB disk.  Not exactly the fastest write speeds on the planet.
[15:52] <michaelforrest1> yeah but I'm talking like 2k/s
[15:52] <michaelforrest1> I have it set to bs=4096
[15:52] <michaelforrest1> it's crazy slow
[15:52] <ev> I was just going to ask
[15:52] <ev> bs=1M is what usb-creator uses
[15:53] <michaelforrest1> which would be bs=1024 right?
[15:53] <michaelforrest1> how long does usb-creator take?
[15:55] <ev> no, 1024 would be 1K
[15:55] <ev> michaelforrest1: I don't time it from fear of knowing the answer to that question
[15:56] <michaelforrest1> oh dear.
[15:57] <persia> Note that for fastest write, you'll want to get information about the geometry of the flash in the USB drive, and set bs to match the flash eraseblock size (or an integer multiple thereof).
[15:58] <cjwatson> I think the kernel exposes that through /sys now ...
[15:58] <persia> bs=1M usually works because most drives have 2K or 4K eraseblocks.
[15:58] <ev> ooo
[17:36] <michaelforrest> ev: does some special magic have to happen to make a boot loader?
[17:36] <michaelforrest> is it even possible to do this usb key with just dd ??
[17:47] <CIA-3> ubiquity: superm1 * r3910 ubiquity/ (bin/oem-config-firstboot debian/changelog): Remove extraneous '--', restoring oem-config-firstboot.
[17:49] <CIA-3> ubiquity: cjwatson * r3911 ubiquity/ (debian/changelog ubiquity/frontend/base.py): Add not_me to the base interface definition of Controller.translate.
[18:47] <mcasadevall> Can someone who is more experienced at reading udeb logs explain why the ia64 image failed to build (failed to find udebs)
[18:47] <mcasadevall> alternate image that is
=== mcasadevall is now known as NCommander
[19:21] <cr3> in the ubiquity success command script, I tried calling in-target update-grub and chroot /target update-grub but this didn't seem to have worked based on the content of /proc/cmdline
[20:47] <cr3> ev: if I modify /target/etc/default/grub in ubiquity/success_command, can I call update-grub directly or should I call the grub-installer script from ubiquity?
[22:07] <exlt> is there an update planned soon for the lucid network installer?  it is currently broken.. :(  kernel-image-2.6.32-16-generic-di etc were released to the archive, and a current netbooting client cannot find a kernel or modules..
[22:08] <cjwatson> exlt: I did that update last night
[22:09] <exlt> right - 20081029ubuntu88 is 2.6.32-14
[22:09] <cjwatson> cr3: I would advise calling update-grub directly
[22:09] <cjwatson> so use 20081029ubuntu89 :-)
[22:10] <exlt> http://archive.ubuntu.com/ubuntu/dists/lucid/main/installer-amd64/ shows 88
[22:10] <cr3> cjwatson: the problem was that I wasn't mount --binding lots of directories, seems to work now
[22:10] <exlt> have a location to pull from?
[22:10] <cjwatson> I assume it failed to build or something
[22:10] <cjwatson> there is no location more primary than that
[22:11] <cjwatson> NCommander: which log?  in http://people.canonical.com/~ubuntu-archive/cd-build-logs/ubuntu/lucid/ports_daily-live-20100309.log, ia64 succeeded
[22:11] <cjwatson> that's odd, the amd64 build succeeded ...
[22:12] <cjwatson> not amd64-specific either, it's missing on all architectures!
[22:13]  * cjwatson pokes around on the master archive machine
[22:13] <cjwatson> 2010-03-09 20:04:26 ERROR   Queue item ignored: Tarfile /tmp/tmpToUskw/debian-installer-images_20081029ubuntu89_i386.tar.gz has a file ./ which is illegal
[22:13] <cjwatson> for.  god's.  sake.
[22:14] <cjwatson> exlt: thanks, this is a regression in Launchpad which I am bringing up with its developers
[22:15] <exlt> sweet - I have data center techs holding for me  ;)
[22:17] <cjwatson> exlt: err, tell them to hang up
[22:17] <cjwatson> exlt: this probably isn't going to be resolved tonight, but hopefully within a day or so
[22:18] <exlt> not on the phone - just trying to install servers and wondering if I could fix
[22:18] <exlt> would be killer to not purge all the older kernel versions
[22:19] <cjwatson> we don't
[22:19] <cjwatson> somebody got trigger-happy about *this* one evidently, but it *is* a manual process
[22:19] <cjwatson> StevenK: it's usually you :-)
[22:19] <exlt> ah, the packages are there, but pruned from repository, so not "findable" by a client machine
[22:20] <cjwatson> err, no
[22:20] <cjwatson> the way it's supposed to work is that the old kernel is not removed from the repository until the installer has been updated
[22:20] <cjwatson> but removal is a manual process, so this is vulnerable to human error
[22:20] <exlt> hmm..
[22:29] <mcasadevall> cjwatson: alternate builds, live builds successed AFAIK
[22:30] <cjwatson> mcasadevall: but that's the URL I'm looking at - URL to the log you're looking at, please?
[22:30] <cjwatson> oh
[22:30]  * cjwatson can't read apparently!
[22:31] <cjwatson> mcasadevall: ok, that's the soyuz bug discussed a moment ago with exlt
[22:32] <cjwatson> I uploaded a d-i for the new kernel ABI, but it never got published
[22:32] <mcasadevall> cjwatson: but the kernel ABI in the log and whats on LP agree with each other
[22:33] <cjwatson> but the kernel ABI of the debian-installer binaries currently published does not
=== mcasadevall is now known as NCommander
[22:33]  * NCommander seems to have nic issues
[22:33] <NCommander> cjwatson: ah, so the problem will fix itself once LP is fixed?
[22:33] <cjwatson> yes
[22:33] <cjwatson> well, I might have to reupload d-i, but that's OK
[22:33]  * NCommander is going to attempt to bend ia64 images back into working order, and may take another whack at SPARC
[22:34] <cjwatson> and actually, the kernel ABI in the log does not agree with the kernel ABI in the archive
[22:34] <cjwatson> the log says 2.6.32-14, but the archive has 2.6.32-16
[22:36]  * NCommander guesses he can't read either