[05:26] <persia> StevenK, evand: The changes for lpia == i686 are in ubiquity, and they are for a dependency on grub : no relation to grub-installer.  They will be included in the upload (already in trunk)
[05:28] <StevenK> Ah, kay
[06:12] <CIA-52> ubiquity: evand * r2892 ubiquity/ (debian/changelog ubiquity/components/summary.py):
[06:12] <CIA-52> ubiquity: If ubiquity is installing from a disk, such as a USB drive, then default
[06:12] <CIA-52> ubiquity: to writing grub to the disk that /boot or / is to be written to, as (hd0)
[06:12] <CIA-52> ubiquity: will most likely be the installation medium (LP: #282037).
[06:12] <StevenK> evand \o/
[06:15]  * evand still has a bad feeling in his stomach over the problem not having a correct solution.
[06:15] <evand> but as Colin notes, this appears to be the best option of the bunch.
[06:15] <StevenK> Dear Q1. Download language packs quicker. kthxbye
[06:15] <evand> haha
[06:15] <evand> Dear Q1, appear on my desk.  kthxbi.
[06:16] <StevenK> Haha
[06:16] <StevenK> Will a photo help? :_)
[06:16] <evand> haha, probably not
[06:16] <persia> evand, You don't want a Q1 : it's the least useful keyboard ever invented.
[06:16] <StevenK> Agreed.
[06:17] <StevenK> If you have to type more than four characters, you want to drop kick the thing through a wall or two
[06:17] <evand> I don't know, there's a Dell XPS sitting in front of me.
[06:17] <evand> My hands are not nearly massive enough to reach all of its keys.
[06:17] <evand> haha
[06:17] <persia> evand, moving one's arms is easier than shaving ones fingers into toothpicks.
[06:18] <evand> hahaha, noted
[06:18]  * TheMuso got to type on an EEE PC 10 inch and 7 inch last weekend. Both keyboards would give me massive cramps if I had to use them long term.
[06:18]  * StevenK digs up an photo
[06:18] <StevenK> evand: http://www.gottabemobile.com/content/binary/WindowsLiveWriter/SamsungQ1UltraUMPCSpecsandPricing_813B/Q1UltraFrontBcopy2.jpg
[06:18] <evand> oww
[06:19] <persia> TheMuso, those are even actual keyboards.  The Q1 has two little grids of tiny buttons on the upper left and right corners.
[06:19] <StevenK> evand: Hence the drop-kicking comment
[06:19] <evand> indeed!
[06:19] <TheMuso> persia: heh right, but I still think those EEE PC keybaords are too small for someone with big hands like myself.
[06:20] <StevenK> This is why I like my X40 keyboard
[06:20] <persia> The only problem with the X40 keyboard is that it gets fussy when you pour liquids in it.
[06:20] <superm1> don't most keyboards?
[06:21] <persia> Not well-designed ones.
[06:21] <persia> It's a common use case, but only a few manufacturers put in flow channels to avoid circuit damage.
[06:27] <StevenK> I fail to see how pouring liquids onto a keyboard is a common use-case.
[06:27] <StevenK> An accidental one, sure.
[06:27] <persia> common use cases aren't always intentional :)
[06:34] <CIA-52> ubiquity: evand * r2893 ubiquity/ (d-i/manifest debian/changelog):
[06:34] <CIA-52> ubiquity: Automatic update of included source packages: apt-setup
[06:34] <CIA-52> ubiquity: 1:0.37ubuntu5, partman-auto 78ubuntu3, partman-base 121ubuntu7,
[06:34] <CIA-52> ubiquity: partman-target 55ubuntu4, user-setup 1.20ubuntu9.
[06:47] <CIA-52> ubiquity: evand * r2894 ubiquity/debian/changelog: releasing version 1.10.5
[06:49] <CIA-52> ubiquity: evand * r2895 ubiquity/ (configure configure.ac m4/intltool.m4): Bump to 1.10.6
[06:55] <persia> Hrm?  1.10.5?  1.10.6?
[06:55]  * persia checks bzr
[06:58] <persia> Oh.  I get it.  the aotuconf change is *after* release.  Never mind.
[07:06] <StevenK> Bwahahaha!
[07:12] <persia> StevenK, ?
[07:13] <StevenK> http://people.ubuntu.com/~stevenk/IMG_2097.JPG
[07:14] <persia> StevenK, Erm.  640x480 isn't really a supported resolution :)
[07:14] <StevenK> :-)
[07:14] <persia> Now if someone was handy with glade...
[07:15] <evand> glade isn't going to fix the timezone map ;)
[07:15] <StevenK> And that's what I was chuckling over
[07:16] <StevenK> It's a mess
[07:16] <evand> haha
[07:16]  * StevenK pokes Ubiquity
[07:17] <persia> evand, I was thinking to select timezone differently at 640x480.  With luck, nobody wants to use that resolution for Jaunty.
[07:18] <StevenK> Yes, for Jaunty, you want to give them a dialog box that says "No timezone for you. Hook up an external monitor" :-P
[07:18] <evand> so I should note that we planned a (another) replacement timezone map for intrepid, but I ran out of time
[07:18] <StevenK> I think I broke Ubiquity
[07:18] <persia> Oh.  Hrm.  I was just hoping most people would just not use VGA.
[07:18] <persia> What did you do?
[07:19] <evand> It wouldn't solve the problem perfectly, but it would surely do better than the existing one at 640x480: https://wiki.ubuntu.com/UbiquityVisualRefresh
[07:19] <StevenK> Created a 1673MB / partition and hit Okay
[07:19] <persia> Right.  And then it should ask you for username and password, unless you're using my casper branch.
[07:20] <evand> perhaps at a sufficiently low resolution we should just show a drop down, given that there isn't much room for a proper timezone picker in that screenshot
[07:20] <StevenK> But it just has a watch cursor and nothing in the Ubiquity window
[07:20] <StevenK> evand: There isn't much room for a dropdown either
[07:20] <persia> The partman page refresh is a bit odd, but it shouldn't do that.
[07:20] <evand> StevenK: check for a backtrace in /var/log/installer/debug ?
[07:21] <StevenK> TypeError: int argument required
[07:21] <persia> evand, Could do, but surely there's a way to have a map at that resolution : I know I have a world map applet on my Zaurus.
[07:21] <StevenK> In /usr/lib/ubiquity/ubiquity/segmented_bar.py
[07:21] <persia> Where?
[07:21] <evand> persia: I'd love a screenshot of that as an example for future work, if you think it was a good exercise in UI design
[07:22] <StevenK> layout.set_text('%d%%' % round(self.segments[i].percent * 100))
[07:22] <persia> evand, I'll dig it up, but it's a full-screen map/clock applet, so perhaps not ideal for ubiquity unchanged.
[07:22] <evand> wow, clearly I'm an idiot
[07:24] <evand> persia: noted
[07:29] <CIA-52> ubiquity: evand * r2896 ubiquity/ (debian/changelog ubiquity/segmented_bar.py): Fix a crash in segmented_bar by providing the expected integer type.
[07:29] <CIA-52> ubiquity: evand * r2896 ubiquity/ (debian/changelog ubiquity/segmented_bar.py): Fix a crash in segmented_bar by providing the expected integer type.
[07:29] <evand> whoops
[07:30] <StevenK> Ah ha, so it was real bug :-)
[07:33]  * StevenK digs up the change
[07:43] <StevenK> evand: How does one unlock stuff?
[07:43]  * StevenK checks if the old ubiquity even died
[07:46] <persia> StevenK, You might also have to kill the partitioner.  Personally, I usually find it easier to reboot to redo an install.
[07:46] <StevenK> Yeah, the old ubiquity wasn't dead
[07:46]  * StevenK is starting to dislike vesa
[07:48] <evand> StevenK: sudo killall ubiquity parted_server; ps aux | grep part is what I generally use
[07:48] <evand> there's probably an easier way
[07:52]  * StevenK tries to break Ubiquity more
[08:02] <StevenK> evand: So, I managed to break my key
[08:03] <StevenK> evand: You were attempting to hide the installer device in partman?
[08:04] <evand> StevenK: break your key?
[08:05] <evand> and yes, that was one of the changes that went into partman as part of the ubiquity upload
[08:05] <StevenK> Oh 1.10.5
[08:05] <StevenK> evand: I was assuming /dev/sdb == the second disk in the device, not the install key.
[08:05] <StevenK> So I might have forcibly created a new parition table on it
[08:07] <evand> sdb should still be the second disk on that machine.  The change just hides any device that has mounted partitions
[08:08] <persia> Shouldn't that include the device that provides /cdrom ?
[08:09] <persia> (unless StevenK was using an outdated ubiquity)
[08:09] <StevenK> I am
[11:09] <baali> does preseed file work with ubiquity??
[11:11] <persia> baali, Yes.
[11:11] <baali> persia, yeah thanx just now got the link https://wiki.ubuntu.com/UbiquityAutomation
[11:13] <persia> baali, One thing not mentioned on that page is that preseeding also works in non-automatic mode, although it will still ask all the questions, it prepopulates the values with those from the preseed.
[11:14] <baali> aah
[11:51] <baali> here is mt problem http://pastebin.ca/1225784 any comments or directions
[11:56] <cjwatson> baali: none of the pkgsel/tasksel stuff works with ubiquity
[11:56] <cjwatson> baali: ubiquity copies the live filesystem on the CD
[11:57] <baali> cjwatson, so should i copy installer folder from DVD to cd and make changes accordingly
[11:57] <cjwatson> baali: if you're working with ubiquity, you probably ought to update the live filesystem with whatever extra packages you want; you should also update /casper/filesystem.manifest and /casper/filesystem.manifest-desktop on the CD
[11:57] <baali> or any other way out
[11:57] <baali> okay wait i guess i have a link for that
[11:58] <cjwatson> I'm pretty sure there's stuff on help.ubuntu.com/community for live CD customisation
[12:00] <baali> cjwatson, okay thank you will look into that
[12:01] <baali> is this one -> http://ubuntuforums.org/showthread.php?t=688872
[12:02] <cjwatson> baali: I was thinking of https://help.ubuntu.com/community/LiveCDCustomization
[15:01] <sugarmice> Hi chaps - is this the right place to ask about customising the Live CD, or is there a better place?
[15:02] <cjwatson> well, I do need to reply to your mail :) sorry, it's been a very busy couple of weeks
[15:02] <sugarmice> I can imagine.  :-)
[15:03] <cjwatson> for customising the boot screen it's a combination of gfxboot-theme-ubuntu and bits in the /isolinux/ directory on the CD
[15:03] <sugarmice> I'm having another go based on the beta of Ibex
[15:03] <cjwatson> what failures did you encounter when trying?
[15:03] <cjwatson> there's a bug about persistent which is release-critical and therefore we'll be sorting out before intrepid goes final
[15:04] <sugarmice> Aha - can you link me to the bug?  I should keep an eye on that
[15:04] <cjwatson> sugarmice: I think http://www.ubuntu.com/aboutus/trademarkpolicy should answer your branding questions
[15:05] <cjwatson> https://launchpad.net/bugs/274076
[15:07] <sugarmice> OK, so it looks as though I should be safe with calling my Live CD "Xubuntu, Wellcome Trust Remix"
[15:09] <cjwatson> yes, that sounds reasonable
[15:10] <sugarmice> But that I shouldn't add the Wellcome Trust's logo (which (a) is beyond my artistic abilities and (b) would imply a relationship that doesn't exist)
[15:13] <sugarmice> That only leaves me with one question, which is how to make modifications to the CD's boot menu, but I'm beginning to think I might not need to.  Since persistent is problematic, I'm thinking I might recommend the sites use wubi, in cases where they don't want a permanent installation.
[15:16] <cjwatson> most textual alterations should be doable by changing isolinux.cfg and possibly gfxboot.cfg
[15:23] <sugarmice> Aha, yes, thank you.  There's a text.cfg file.  Great.  That's exactly what I needed.
[16:27] <bdmurray> does bug 280900 sound familiar?
[16:29] <cjwatson> bdmurray: evand committed a fix for that earlier today
[16:29] <evand> bdmurray: already fixed in bzr, marking as such
[16:30] <bdmurray> alright, there are some possible duplicates  - Is the traceback what I'd want to look for to confirm any duplicates?
[16:30] <cjwatson> look for this bit:
[16:30] <cjwatson>     layout.set_text('%d%%' % round(self.segments[i].percent * 100))
[16:30] <cjwatson> TypeError: int argument required
[16:30] <CIA-52> ubiquity: evand * r2897 ubiquity/debian/changelog: Add LP bug reference
[16:30] <evand> thanks
[17:14] <baali> cjwatson, it worked thank you very much :)
[18:01] <cjwatson> evand: is 282756 due to 280900? I wasn't quite sure if partman getting SIGPIPE was due to the crash
[18:03] <evand> I suspect so
[19:05] <cr3> cjwatson: I'm using that source.list-setup file during the installation but it doesn't seem like its being apt-get update'd, so pkgsel can't find the given packages
[19:05] <cr3> cjwatson: is that a bug or is there a place that would be appropriate to run the update?
[19:10] <cr3> cjwatson: one workaround that comes to mind is not using pkgsel and relying on late_command instead
[19:10] <cr3> that would be a shame to not leverage existing preseed functionality though
[19:20] <cjwatson> cr3: sources.list.apt-setup is where the real sources.list gets moved aside until it's actually needed
[19:20] <cjwatson> cr3: it's moved back to sources.list just before processing pkgsel/include
[19:20] <cjwatson> cr3: if it isn't working, can I see your preseed file and the log?
[19:23] <cr3> cjwatson: it gets moved indeed, but does the apt database get updated as well?
[19:24] <cr3> cjwatson: I don't have the syslog, I'm trying another run with the pkgsel packages in the late_command. but I'll try again and get you the requested information
[19:24] <cjwatson> cr3: yes, should do
[19:24] <cjwatson> cr3: oh, are you adding new repositories by appending to sources.list.apt-setup?
[19:27] <cr3> cjwatson: I have a udeb installed at early_command which creates sources.list.apt-setup with new repositories
[19:32] <cjwatson> cr3: OK, that would have been broken by my recent change to fix bug 267884. Any reason you aren't using the available apt-setup preseeding, such as apt-setup/local0/repository?
[19:32] <cr3> cjwatson: aha, that's why my builds have stopped being tested from the 9th :)
[19:33] <cr3> cjwatson: the reason is that I'm testing ISO images as integrally as possible using a network install and using the packages from the image on a local web server. If I also provide archive.ubuntu.com during the installation process, more recent packages might be installed which are no longer integral to the image.
[19:34] <cjwatson> cr3: you can override the use of archive.ubuntu.com though
[19:34] <cjwatson> I'd rather extend the available apt-setup preseeding to cover what you need if necessary than revert the fix for bug 267884 (which I consider important)
[19:35] <cr3> cjwatson: how can I override? I mean, won't the installer pick the latest packages across all available repositories?
[19:36] <cjwatson> I mean that you can tell it to use a different mirror rather than archive.ubuntu.com
[19:36] <cjwatson> mirror/http/hostname and mirror/http/directory
[19:37] <cjwatson> I expect you'd need to set apt-setup/security_host and apt-setup/security_path too
[20:21] <ago_> I have issues mounting dmloop devices, at first I get "no such file or directory" that is probably due to /dev/mapper/loop0 taking a bit to come up, but after that, mounting manually /dev/mapper/loop0 to /root results in: "unable to read superblock"
[20:21] <cjwatson> when I tested dm-loop here I got kernel BUG messages in kern.log; haven't tracked them down yet
[20:21] <cjwatson> suggests to me it's a bit buggered
[20:21] <ago_> retrying sometimes succeed, but I have no explanation http://img98.imageshack.us/my.php?image=dmlosetupzd1.png
[20:22] <cjwatson> maybe I was driving it wrong though, this was just with a read-only iso9660 mount
[20:22] <ago_> I can use good old mount -o loop with the file and it works well
[20:33] <evand> cjwatson: could you elaborate on why by-path would fix the /cdrom issues when installing from a USB device?  I'm not sure I follow.
[20:35] <cjwatson> evand: I think by-path is a hardware identifier rather than a filesystem identifier
[20:35] <cjwatson> it was sort of a throwaway comment though, I didn't really expect us to do that for intrepid at this point ...
[20:35] <evand> ok
[20:36] <evand> I'll continue with the other approach then
[20:43] <CIA-52> netcfg: cjwatson * r630 ubuntu/ (debian/changelog netcfg-common.c):
[20:43] <CIA-52> netcfg: If we detect that the user entered a fully-qualified domain name as the
[20:43] <CIA-52> netcfg: hostname, then remove the domain part from the hostname as well as
[20:43] <CIA-52> netcfg: copying it to netcfg/get_domain (LP: #11019).
[20:49] <CIA-52> netcfg: cjwatson * r631 ubuntu/debian/changelog: releasing version 1.44ubuntu4
[21:15] <ago_> if I use dmlosetup manually I get kernel panic...
[21:52] <cr3> how can I set a proxy in apt-setup for my local? repositories?
[21:53] <cr3> it seems like the one specified under mirror/http/proxy is not being copied to the installed system
[22:03] <cr3> however, I see that the mirror/$protocol/proxy is being used for retrieving the key from the local? repositories
[22:04] <cr3> and the installation guide only contains the following reference to proxy: /usr/share/doc/installation-guide-i386/en/apbs04.html:d-i mirror/http/proxy string
[22:04] <cr3> actually, this might work. brb
[22:05] <cjwatson> mirror/http/proxy should be copied to /etc/apt/apt.conf in the target system
[22:48] <evand> Odd, my CIA commits for usb-creator are not showing up anymore.
[22:48]  * evand kicks CIA-52 
[22:48] <CIA-52> ow
[23:12] <cjwatson> evand: seemed to have fallen out of the filter rules, I don't know why. I've put it back now
[23:12] <cjwatson> the change history offers no hints
[23:12] <cjwatson> it was added, and then ... added again
[23:18] <evand> hah, fantastic.  Thanks