[04:54] <JohnFlux> Hi all
[04:54] <JohnFlux> 9.10 allocates 10GB of swap for me, by default
[04:54] <JohnFlux> can we adjust that please?
[04:55] <JohnFlux> is there anyone that believes that the ordinary user needs 10GB of swap by default?
[05:00] <cody-somerville> Its a bug.
[05:00] <cody-somerville> The installer team is aware of it.
[05:01] <JohnFlux> cody-somerville: someone pointed out to me that you need lots of space for suspend to disk
[05:01] <JohnFlux> so it's not as silly as I first though
[05:01] <JohnFlux> thought
[05:01] <JohnFlux> I'm wondering if a combination of swapfile and swap partition could be used
[05:01] <JohnFlux> so have a swap partition _just_ for suspending
[05:01] <JohnFlux> and then a swap file for the actual swap
[05:02] <cody-somerville> You need as much swap as you do RAM and most people don't have 10GB of ram.
[05:03] <JohnFlux> right - I have 4GB
[05:03] <JohnFlux> so this way the installer could allocate the swap partition as big as the RAM
[05:03] <JohnFlux> then setup a growable swap file
[05:04] <JohnFlux> thus avoiding the problem of having to allocate 10GB of swap to be safe
[05:04] <cody-somerville> it really is a bug
[05:04] <cody-somerville> I heard cjwatson talking about it the other day.
[05:05] <JohnFlux> what is the value that it should be?
[05:05] <JohnFlux> perhaps about 6GB I guess?
[05:06] <cody-somerville> little over 4gb would probably be fine.
[05:06] <JohnFlux> cody-somerville: that would mean the swap space is only a little then
[05:06] <JohnFlux> cody-somerville: if you want suspend to work you have to reserve that 4gb
[09:43] <CIA-3> grub-installer: cjwatson * r847 ubuntu/ (debian/changelog grub-installer): (log message trimmed)
[09:43] <CIA-3> grub-installer: Ask grub-installer/only_debian at high priority again (LP: #47135).
[09:43] <CIA-3> grub-installer: Originally, in Warty, we ensured that this question wouldn't be visible
[09:43] <CIA-3> grub-installer: by default to support simple installations of Ubuntu with as few
[09:43] <CIA-3> grub-installer: questions as possible. However, since then, we wrote Ubiquity, which
[09:43] <CIA-3> grub-installer: has taken over responsibility for the simple installation case, and it's
[09:43] <CIA-3> grub-installer: become clear that automatically installing GRUB to the first disk's MBR
[09:50] <CIA-3> grub-installer: cjwatson * r848 ubuntu/debian/ (68 files in 2 dirs):
[09:50] <CIA-3> grub-installer: Update grub-installer/bootdev text to avoid GRUB device naming that
[09:50] <CIA-3> grub-installer: changed between GRUB Legacy and GRUB 2, and to use libata-style device
[09:50] <CIA-3> grub-installer: naming since that is more accurate for most people (LP: #391775)
[10:10] <ev> I wonder how many torches and pitchforks would be outside my door if I seeded proper vim back onto the live CD
[10:11] <StevenK> ev: Hell, I'd support that :-)
[10:14]  * cjwatson tries to figure out what to do about mdadm not working properly in the initramfs
[10:14] <cjwatson> I've worked out what the problem is
[10:14] <ev> oh?
[10:15] <cjwatson> the MD 0.90 metadata format operates by putting a superblock at the *end* of the physical volume
[10:15] <cjwatson> (this is arguably a little silly and the 1.x formats work differently, but GRUB doesn't support them yet so they're not an option)
[10:16] <cjwatson> this means that when mdadm starts up, it's told to just look at everything in /proc/partitions, and it finds that both /dev/sda and /dev/sda1 appear to be the same physical volume
[10:16] <cjwatson> and it's not smart enough to say "oh, I'll take the partition then"
[10:17] <cjwatson> the reason I suddenly ran into this is that I'm using exact-number-of-MiB disks as a result of using kvm (though maybe most disks are exact-MiB?  I'm not sure) and so the change to use MiB alignment by default means that the end of the partition lands exactly on the end of the disk
[10:17] <cjwatson> if there's a gap at the end, it's not a problem
[10:18] <ev> ah, yikes
[10:18] <ev> is there a bug number for this?
[10:18] <cjwatson> I'm very tempted to say "always allow a 4KiB gap at the end of the disk" to work around it
[10:18] <ev> that seems entirely reasonable for lucid
[10:18] <cjwatson> yeah, bug 527401
[10:18] <ev> though it wouldn't cover us for disks brought in from the outside
[10:19] <ev> outside our set of tools, that is
[10:19] <cjwatson> they'd already be failing, I think
[10:19] <cjwatson> the mdadm FAQ says you have to configure mdadm specially if you want to do that
[10:19] <cjwatson> personally I sort of think mdadm ought to be made smarter
[10:19] <cjwatson> but feels risky given available time
[10:20] <ev> indeed
[10:40] <cjwatson> I think this may end up wasting up to 1MiB at the end of every disk
[10:40] <cjwatson> but, well
[10:50] <ev> I'd be very surprised if anyone actually noticed, given how much additional space is taken by ext4.
[10:50] <ev> yay more webkitgtk bugs (bug 509838)
[11:14] <cjwatson>   * Make sure the user we're not processing events before we've set up
[11:14] <cjwatson>     the language list in the KDE frontend (LP 528678).
[11:14] <cjwatson> ev: ^- is the ": #" intentionally missing from that?
[11:15] <cjwatson> or did you just fix part of it?
[11:15] <ev> correct
[11:15] <cjwatson> ah, ok
[11:15] <ev> I haven't fixed the unicode to QString fun yet
[11:15] <cjwatson> I just noticed the different highlighting in my editor :)
[11:15] <ev> thanks for checking though
[11:26] <cjwatson> yay, that partman change did indeed fix raid installation
[11:26] <cjwatson> end-to-end test worked
[11:29] <ev> nice
[11:32] <CIA-3> partman-base: cjwatson * r198 ubuntu/ (debian/changelog parted_server.c):
[11:32] <CIA-3> partman-base: Always leave a small gap at the end of the disk (except on device-mapper
[11:32] <CIA-3> partman-base: devices), to avoid confusing mdadm by leaving the MD 0.90 superblock at
[11:32] <CIA-3> partman-base: the end of the disk as well as the end of a partition (LP: #527401).
[11:34] <CIA-3> partman-base: cjwatson * r199 ubuntu/debian/changelog: releasing version 139ubuntu2
[11:35] <CIA-3> grub-installer: cjwatson * r849 ubuntu/debian/po/ (66 files): avoid unnecessary delta
[11:36] <CIA-3> grub-installer: cjwatson * r850 ubuntu/debian/changelog: releasing version 1.49ubuntu9
[11:36] <ev> hrm, the locale isn't being generated in ubiquity before we get to console-setup
[11:37] <ev> so iconv gets fed a bogus value for $(locale charmap)
[11:37]  * ev digs further
[12:03] <davmor2> ev: don't dig too much further you'll be in china
[12:04] <ev> davmor2: :)
[12:24] <ev> cjwatson: if you have a moment, can I pick your brain on localization (for bug #542310)?  In console-setup, there's a call to iconv to do transliteration in the event that the post-install locale is non-UTF-8.  This breaks (creates question mark characters) in ubiquity because the value of $(locale charmap) is ANSI_X3.4-1968 because we haven't generated locales for el_GR.UTF-8 yet.
[12:24] <ev> d-i doesn't have iconv, so it avoids the call entirely.  I see two options.  We can either remove the transliteration from console-setup, given that we're only using UTF-8 locales (to my knowledge), or we can generate the locale before we get to console-setup.
[12:29] <cjwatson> ugh.  it seems to me that this is a debconf bug
[12:30] <ev> debconf?  This is long before text enters debconf.
[12:30] <cjwatson> isn't kbdnames substituted at some point?
[12:30] <cjwatson> what I mean is, SUBST encoding isn't defined, but it would be better to just be able to substitute UTF-8 text into debconf and let it worry about it, rather than having to manually recode to the user's locale
[12:30] <ev> ahhh
[12:31] <ev> indeed
[12:34] <cjwatson> I'm thinking of something like http://paste.ubuntu.com/401758/
[12:35] <cjwatson> but even without that, I think removing the transliteration would in practice be much better than leaving it in there
[12:35] <cjwatson> the number of people who'll set a non-UTF-8 locale and then later come round to reconfigure console-setup is going to be vanishingly small
[12:35] <cjwatson> as you say, we only use UTF-8 locales ourselves
[12:35] <cjwatson> with the possible exception of "No localization"
[12:38] <ev> definitely agreed
[12:41] <ev> by the way, Ivanka was just singing the praises of how fast the installer is now
[12:43] <ev> okay, I'll make the change to console-setup then
[12:44] <ev> cjwatson: did you want to commit that debconf change, or are you keen to leave that until Manic Monkey?
[12:49] <CIA-3> console-setup: evand * r142 console-setup/debian/ (changelog config.proto):
[12:49] <CIA-3> console-setup: Remove iconv transliteration. We don't use non-UTF-8 locales, and
[12:49] <CIA-3> console-setup: it breaks ubiquity as we haven't generated the configured locale by
[12:49] <CIA-3> console-setup: the time we hit console-setup (LP: #542310).
[12:50] <cjwatson> ev: fast> excellent :)
[12:50] <cjwatson> ev: I'll bring it up with Joey, but I don't think we need it for Lucid
[12:50] <ev> cjwatson: cool, thanks
[13:02] <CIA-3> console-setup: evand * r143 console-setup/debian/changelog: releasing version 1.34ubuntu13
[13:10] <xivulon> hi davmor2, any bug on your end I should be aware of, lp seems fairly quiet
[13:11] <davmor2> xivulon: not that I know too I'll have a play latter and see if there are any faults I can see
[13:11] <ev> xivulon: sounds like a call for testing on ubuntuforums is in order :)
[13:11] <ev> or whatever was done at beta that had lots of people in here asking to help test wubi
[13:11] <xivulon> ev we do not have a dedicated forum anymore :(
[13:11] <xivulon> so I do not follow the forums that much these days
[13:12] <ev> ask in a general forum and ask the moderators to sticky it?
[13:12] <xivulon> I still cannot boot using virtualbox (stops before loading initrd) so I am a bit in the dark
[13:13] <cjwatson> it would be worth a quick test to make sure the new ntfs-3g hasn't broken things
[13:13] <xivulon> and I need to load up the new translations
[13:14] <xivulon> I will see if kvm works out better
[13:16] <ev> cjwatson: per Jane's email, shouldn't DEBVERSION in debian-cd be "10.04 LTS", or will that break things?
[13:18] <xivulon> ps if there is any plan to improve the web visibility on ubuntu.com I will redirect wubi-installer.org, otherwise I could do with some of the new artwork
[13:19] <ev> xivulon: michaelforrest is looking into applying the new artwork to wubi
[13:19] <ev> apologies, I forgot to mention that earlier
[13:19] <xivulon> ah nice, will that also affect the app itself, the wubi website, ubuntu.com page?
[13:20] <ev> I'm not sure how they website layout is changing with respect to wubi, beyond what they've already done
[13:20] <ev> but it wont be the default for windows users until we can sort out migration
[13:20] <xivulon> It makes sense
[13:20] <xivulon> I have applied for mentorship to gsoc
[13:26] <xivulon> ev quick reminder about the logo in the certificate. That is actually the very first dialog many users see.
[13:26] <ev> xivulon: ah, good thinking
[13:26] <ev> michaelforrest: ^
[13:26] <ev> xivulon: mentorship> awesome
[13:27] <ev> right, lunch
[13:28] <xivulon> am already following up on #ubuntu-artwork
[13:45] <cjwatson> ev: which e-mail from Jane?  anyway, yes, that would break at least some things - among other things it's substituted into a URL for the release notes
[13:46] <cjwatson> there may be some centring issues on boot screens as well
[13:46] <cjwatson> not that it can't be changed, but it would need care
[13:51] <ev> cjwatson: warthogs, "Ubuntu 10.04 LTS"
[13:51]  * ev lunches
[13:52] <cjwatson> oh, yeah, I see it
[13:52] <cjwatson> we probably ought to make an effort to do something about that
[14:09] <xivulon> ev, davmor2 I have a call for testing on http://ubuntuforums.org/showthread.php?p=9030372
[14:09] <xivulon> you might want to subscribe
[14:17] <ev> xivulon: thanks
[14:18] <ev> cjwatson: we could add a ubiquity/lts template and subst it in wherever we use ${RELEASE}.
[14:18] <cjwatson> does LTS get translated?
[14:24] <ev> hrm
[14:24] <ev> I'm really not sure.  Sounds like a question for Jane.
[14:24] <ev> I would suspect not, but I am by no means authoritative :)
[14:49] <CIA-3> ubiquity: cjwatson * r3992 ubiquity/ (debian/changelog debian/control scripts/install.py):
[14:49] <CIA-3> ubiquity: Depend on uboot-mkimage and redboot-tools on armel, and ensure that the
[14:49] <CIA-3> ubiquity: appropriate one for the subarchitecture is kept installed, so that
[14:49] <CIA-3> ubiquity: flash-kernel can work properly without network access (LP: #537007).
[15:47] <CIA-3> ubiquity: evand * r3993 ubiquity/ (4 files in 3 dirs):
[15:47] <CIA-3> ubiquity: Handle window manager buttons properly in greeter mode
[15:47] <CIA-3> ubiquity: (LP 539827).
[16:33] <cjwatson> ev: are we going to make text non-selectable, per foundations-lucid-installer-design-improvements?
[16:33] <ev> cjwatson: yes
[16:34] <ev> just updating ubiquity for unexpected python-apt changes, then I'll do that
[16:34] <cjwatson> ok
[16:37] <CIA-3> ubiquity: evand * r3994 ubiquity/ (debian/changelog scripts/install.py):
[16:37] <CIA-3> ubiquity: Update DebconfInstallProgress for changes to python-apt
[16:37] <CIA-3> ubiquity: (LP: #548852).
[16:38] <CIA-3> ubiquity: evand * r3995 ubiquity/debian/changelog: Wrong LP bug number in previous commit.
[17:15] <CIA-3> ubiquity: evand * r3996 ubiquity/ (13 files in 3 dirs):
[17:15] <CIA-3> ubiquity: Make labels unselectable now that we're in beta, per foundations-
[17:15] <CIA-3> ubiquity: lucid-installer-design-improvements.
[17:31] <ev> cjwatson: From Jane> don't translate "LTS", not sure on Xubuntu being LTS
[17:31] <cjwatson> so no need for ubiquity/lts
[17:32] <ev> and we just change the strings that have LTS in them each release?
[17:33] <cjwatson> doesn't it pick up "10.04 LTS" from .disk/info?
[17:33] <cjwatson> at least assuming that we updated DEBVERSION somehow
[17:33] <ev> oh ok, I thought we weren't going to do that, on account of things breaking
[17:33] <ev> but yeah, if we update DEBVERSION there's no need to do much in ubiquity (save maybe update things that depend on spaces as delimiters)
[17:33] <cjwatson> I don't think it's very many things, it just isn't entirely trivial
[17:33] <cjwatson> I'd rather do it there than in ubiquity I think ...
[17:34] <ev> noted
[17:34] <ev> it would equally update it for things like usb-creator and wubi, which would be nice
[17:45] <CIA-3> debian-installer: ogra * r1263 ubuntu/ (6 files in 4 dirs): add support for armel+omap architecture
[17:57] <CIA-3> ubiquity: evand * r3997 ubiquity/ (debian/changelog gui/gtk/ubiquity.ui): Make the quit warning dialog transient for the main window (modal).
[17:57] <ev> right, I'm going to upload a new ubiquity
[17:59] <CIA-3> ubiquity: evand * r3998 ubiquity/ (d-i/manifest debian/changelog):
[17:59] <CIA-3> ubiquity: Automatic update of included source packages: console-setup
[17:59] <CIA-3> ubiquity: 1.34ubuntu13, grub-installer 1.49ubuntu9, partman-base 139ubuntu2.
[18:05] <CIA-3> ubiquity: evand * r3999 ubiquity/debian/changelog: releasing version 2.2.9
[20:38] <gellmar> hello! What's wrong with netboot kernels on lucid?
[20:38] <gellmar> the previous one worked like a charm
[20:39] <gellmar> and the latest one pops a list of drivers to recognize my HDD
[20:39] <gellmar> and I can not continue
[21:47] <CIA-3> ubiquity: superm1 * r4000 ubiquity/ (debian/changelog ubiquity/i18n.py):
[21:47] <CIA-3> ubiquity: Fix the default language selection for when only-show-installable-languages
[21:47] <CIA-3> ubiquity: is set to true.