[11:18] <xnox> Bug #1003890
[11:18] <ubot2> Launchpad bug 1003890 in ubiquity "unittest FAIL: test_pages_fit_on_a_netbook (test_gtkui.TestFrontend)" [Undecided,New] https://launchpad.net/bugs/1003890
[11:19] <CIA-62> ubiquity: cjwatson * r5481 trunk/ (bin/ubiquity debian/changelog):
[11:19] <CIA-62> ubiquity: If ubiquity is started up in a non-UTF-8 locale, force it into C.UTF-8
[11:19] <CIA-62> ubiquity: and fail immediately if that's unavailable (LP: #1003851).
[11:19] <cjwatson> xnox: Ah, that tends to be somewhat environment-specific unfortunately
[11:19] <cjwatson> xnox: I suspect you'll find it works in a clean chroot (e.g. an sbuild environment)
[11:20] <cjwatson> xnox: Feel free to try to figure out exactly how to improve the test so that it works regardless; but it doesn't block builds and the like
[11:20] <xnox> cjwatson: see bug, I had to create a symlink to '../pixmaps' to actually load the images -> calculate the real height of stuff -> fail the test
[11:21] <xnox> i see if I can fail it in the sbuild as well ;-)
[11:21] <cjwatson> Yes, pixmaps presence makes a difference
[11:22] <xnox> but do we care about that test? cause 500 seems small
[11:22] <cjwatson> Anyway, I tried to tackle that a while back and gave up, so I'm not going to try again, but don't let me stop you
[11:22] <cjwatson> Yes, we care, we get bugs when it's violated.
[11:22] <cjwatson> The test is there for a reason :-
[11:22] <cjwatson> )
[11:22] <xnox> ok
[11:22] <cjwatson> It's a very bad user experience when people can't see the Next button
[11:34] <CIA-62> ubiquity: evand * r5482 ubiquity/ (15 files in 9 dirs):
[11:34] <CIA-62> ubiquity: Remove migration-assistant following foundations-q-testing-
[11:34] <CIA-62> ubiquity: migration-assistant. Thanks Dmitrijs Ledkovs!
[11:34] <ev> ^ xnox thanks
[11:34] <xnox> ev: ok =) that was quick review ;-)
[11:35] <ev> I've landed it, but I'm still poking at it to be absolutely sure
[11:36] <xnox> ev: I'm not sure if anything needs doing in ubiquity/frontend/gtk_ui.py with respect to backing out, now that we do not have migration assistant
[11:47] <ev> if you're referring to the comment on gtk_ui:set_page:1142, then it still applies
[11:47] <ev> but I'll reword it
[11:47] <xnox> ok
[11:51] <CIA-62> ubiquity: evand * r5483 ubiquity/ubiquity/frontend/gtk_ui.py: Tidy up comment.
[12:10] <CIA-62> ubiquity: cjwatson * r5484 trunk/ (debian/changelog ubiquity/i18n.py):
[12:10] <CIA-62> ubiquity: KDE fails to round-trip strings containing U+FEFF ZERO WIDTH NO-BREAK
[12:10] <CIA-62> ubiquity: SPACE, such as the translations of a few language names. Strip these
[12:10] <CIA-62> ubiquity: from language name translations.
[12:13] <xnox> How do you do RAID1 with UEFI. Is /efi a separate partition on both drives or can /efi be on top of RAID?
[12:19] <cjwatson> I don't think the spec says.
[12:20] <cjwatson> I'd say you need a GPT partition entry with the appropriate GUID on any drive the firmware might attempt to boot from.  (You can probably get away with it on only one.)
[12:33] <xnox> (and if 'the one' is borked, I do want to boot of the other one =) that's the point of RAID1 right?!)
[12:34] <cjwatson> Mm, except that if you use an actual RAID partition for it, I'd have thought you'd get the wrong GUID
[12:34] <cjwatson> I don't remember how RAID on GPT works
[12:35] <cjwatson> Clearly you *want* it on all, but it might take some hacking, not sure
[12:35]  * xnox can't wait for quad housing to arrive to test this properly
[12:35] <xnox> cause we do currently support UEFI installations from 12.04 media, afaik
[12:35] <cjwatson> Easy enough to KVM up the GPT side of this to see what it looks like; you can probably do most of it without actual EFI firmware
[12:35] <cjwatson> Yes
[12:36] <cjwatson> You may find the partman-partitioning/default_label=gpt installer boot parameter helpful
[12:36] <xnox> my machine is lvm2 on top of luks and as it turned out kvm disk IO is painfully slow this way.
[12:36] <cjwatson> No separate non-luks/lvm partition?
[12:37] <cjwatson> I do all my KVM stuff on an external USB disk; it's not exactly speedy but it's usable for installer testing
[12:37]  * xnox is planning to reinstll to have some scratch non-luks space
[12:37]  * xnox ... and UEFI/GPT
[14:34] <tbrijeski> hi folks - can anyone tell me where ubiquity picks up its gtk3 theme from? On a non gtk3 desktop it is unreadable
[14:40] <stgraber> tbrijeski: ubiquity gtk is only tested with ubuntu-artwork installed (which depends on light-themes and sets the required gsettings keys to make it the gtk3 thmee)
[14:40] <stgraber> so ubiquity itself doesn't come with a gtk3 theme, though there are a few css overrides in the code itself that won't look good with anything but light-themes
[14:41] <tbrijeski> I know it doesn't come with it but I have tried a few different thigns to get it to work but it isn't setting it
[14:42] <tbrijeski> I have set the gtk-3.0 thjeme in /usr/share/themes/Default, /root/.config/gtk-3.0, the live user ~/.config/gtk-3.0 and it doesn't seem to help
[14:42] <stgraber> installing ubuntu-artwork should do the trick unless you have another theme overriding gsettings
[14:45] <xnox> ogra_: would you please look at bug 940025 =))))
[14:45] <ubot2> Launchpad bug 940025 in debian-installer-netboot-images "[public] armadaxp net installer fails to properly encrypt disks (dm-crypt)" [High,Confirmed] https://launchpad.net/bugs/940025
[14:45] <xnox> from comments it may be related to kernel settings / improprer console usage tty7
[14:46] <xnox> cjwatson: ^^^ you might know something about it as well
[14:46] <tbrijeski> stgraber - I'm running live now and figured it out - it will use a gtk3 theme as set by the live user but the theme has to have a gtk3 component in the theme directory so if the gtk2 desktop uses a gtk2 only theme it will look ugly.
[14:47] <cjwatson> xnox: unlikely to have time today I'm afraid
[14:48] <tbrijeski> i copied one of the gtk3 theme folders over into the selected gtk2 theme and made sure the index.theme reflected this theme.
[14:49] <tbrijeski> stgraber - thanks just the same - I'll post it up on my forum for the remastersys users that have run into this. Have a great day!
[14:55] <xnox> ogra_: bug 848782 as well?
[14:55] <ubot2> Launchpad bug 848782 in flash-kernel "Serial console not enabled for passphrase prompt when using LUKS" [Medium,Confirmed] https://launchpad.net/bugs/848782
[17:08] <ogra_> xnox, i plan to actually update to a completely new flash-kernel, i'll look into these bugs afterwards in case they are still there
[17:08] <xnox> ogra_: ok thank you. Hmm... the first bug
[17:08] <xnox> in description says "use full disk - guided LVM"
[17:09] <xnox> yet the log says: RAID, XFS, ext3, ext4 and mount/write errors.....
[17:09] <xnox> unless partman is looping across the devices, I got confused as to why the log mentions those
[17:09] <xnox> and I have asked the reported to retest on a clean drive
[20:31] <roadmr> Hey folks! I'm using d-i partman/early_command string debconf-set partman-auto/disk /dev/sda but I get "debconf-set: not found" in the logs. What to use instead?
[20:32] <roadmr> (my early_command is a bit more complex but the problem remains that debconf-set is not found - and I need to use some early_command magic, so I can't just d-i partman-auto/disk string /dev/sda)
[20:34] <xnox> roadmr: my guess is that debconf is optional dependency (?!) and not available yet. I might be very wrong though. wait for someone better to answer
[20:34] <xnox> or try mailing list from the topic.
[20:38] <roadmr> xnox: thanks :)
[20:39] <roadmr> xnox: what baffles me is that the example preseed has something that does just that, but doesn't seem to work on ubiquity (works fine on d-i, though)
[20:56] <xnox> roadmr that's interesting. can you pastebin what you are trying to do?