[00:19] <cjwatson> cody-somerville: cdrom-detect/try-usb is all that's available right now
[00:20] <cody-somerville> Are you familiar with the load-media installer component? what does that do?
[00:21] <cjwatson> not familiar enough to explain it, no
[00:22] <cjwatson> but it fits in the same slot as load-cdrom (i.e. the "Loading installer components" step), not in the same slot as cdrom-detect
[00:23] <cody-somerville> ok
[00:23] <cjwatson> mountmedia does mount fixed disks, but I have no experience with the whole assembly
[00:23] <cjwatson> it *might* be a reasonable thing to experiment with if you need this, but it might be a red herring too :)
[00:24] <cody-somerville> :)
[00:25] <cjwatson> load-media itself seems to be only on Debian's floppy images right now. its dependencies media-retriever and mountmedia are on a wider selection of images
[00:25] <cjwatson> although that makes sense because any given image can only really have one of {load-cdrom, download-installer, load-media}
[00:26] <cjwatson> anyway, meant to go to bed half an hour ago - night
[00:26] <cody-somerville> Night :)
[00:32] <cody-somerville> hmm... it says my preseed file is corrupt :/
[00:45] <CIA-28> casper: TheMuso * r601 trunk/ (3 files in 3 dirs):
[00:45] <CIA-28> casper: - Fix incorrect paths when chowning files.
[00:45] <CIA-28> casper: - Don't use /root for the ubiquity hook accessibility script, /target is
[00:45] <CIA-28> casper:  where the installed system is located.
[00:55] <CIA-28> casper: TheMuso * r602 trunk/ (3 files in 3 dirs):
[00:55] <CIA-28> casper: Remove laptop detect code. Its not currently working, and getting it to
[00:55] <CIA-28> casper: work requires invasive changes.
[01:01] <persia> cjwatson, Thanks for the update.  I lost track of that bug, and had removed those as part of bug 347207 (for which I'm just cleaning up).
[02:35] <CIA-28> casper: TheMuso * r603 trunk/ (3 files in 3 dirs):
[02:35] <CIA-28> casper: Only set pulseaudio settings and create the .orca directory if sed
[02:35] <CIA-28> casper: packages are installed on the live filesystem and installed system.
[02:47] <CIA-28> casper: TheMuso * r604 trunk/ (debian/changelog ubiquity-hooks/30accessibility): ubiquity-hooks/30accessibility: Copy orca settings to the installed system.
[03:39] <CIA-28> casper: TheMuso * r605 trunk/debian/changelog: releasing version 1.166
[10:08] <z1d4n3> Hello.
[10:08] <z1d4n3> Anyone around? I'm having an issue with the installer in the 9.04 beta. It won't pick up any of my HDD's partitions.
[10:11] <evand> z1d4n3: Are any of the partitions mounted?
[10:12] <z1d4n3> No
[10:13] <z1d4n3> I can mount them. Ubuntu picks them up fine.
[10:13] <z1d4n3> The installer sees my drive, just sda.
[10:13] <z1d4n3> No partition table.
[10:14] <evand> So just to clarify, the installer sees your disk and lets you install to it on the automatic partitioning page, but if you select advanced, it doesn't show any of the partitions that you know exist?
[10:14] <evand> z1d4n3: Does `sudo parted /dev/sda print all` show them?
[10:14] <z1d4n3> Yes, exactly
[10:15] <z1d4n3> Uhh, "Error: Can't have overlapping partitions."
[10:16] <z1d4n3> "sudo fdisk -l" lists them all.
[10:16] <z1d4n3> (Thanks for the help!)
[10:16] <evand> sounds like your partition table is a bit wonky.
[10:18] <z1d4n3> Any way to fix a wonky partition table, or do I need to reformat the disk?
[10:18] <evand> z1d4n3: first, I'd make some backups if there's anything you care about on there
[10:19] <evand> then I'd use fdisk to make sure the partition boundaries don't overlap and line up properly.  It might be more trouble than it's worth, honestly.
[10:20] <z1d4n3> All the data is backed up, however I just spent ~3 hours setting up an XP install. Shame to waste it.
[10:20] <z1d4n3> Meh, thanks for the advice.
[10:20] <z1d4n3> Have a good weekend.
[10:22] <evand> Wondering if we can handle that situation better by at least throwing up a dialog with that error message.  Might have a poke through the partman source code a bit later as I'm not sure how such errors are currently handled.
[10:22] <davmor2> evand: if he has installed xp over the linux install I bet it has something to do with the 8meg of blank space that xp install after it's partition :)
[10:23] <evand> surely Microsoft's partitioner knows how to properly write a partition table
[10:23] <evand> davmor2: link?
[10:25] <davmor2> evand: xp's partitioner is kinda screwy when it comes to linux like it sees it as an invalid partition
[10:25] <davmor2> vista's knows what is there
[10:26] <davmor2> xp sp3 I think knows what is there but sp2 and earlier definately have issues
[10:28] <davmor2> evand: you need to remember it's microsoft there is only one partition the one with windows on :D
[10:30] <evand> heh
[10:30] <davmor2> evand: I'll have a look for some info
[10:33] <cjwatson> exceptions from parted normally propagate upward and eventually do show up as dialogs
[10:34] <cjwatson> but it depends on exactly where they're issued
[10:52] <evand> ah
[10:53] <cjwatson> might be instructive to look into how it works anyway, it's quite a spectacular pile of software ;-)
[11:09] <evand> absolutely, just need to find the time
[11:44] <CIA-28> base-installer: cjwatson * r357 ubuntu/debian/changelog: releasing version 1.98ubuntu4
[11:49] <CIA-28> console-setup: cjwatson * r99 ubuntu/ (Keyboard/KeyboardNames.pl debian/changelog): releasing version 1.28ubuntu7
[11:51] <CIA-28> grub-installer: cjwatson * r777 ubuntu/debian/changelog: releasing version 1.36ubuntu5
[11:54] <CIA-28> hw-detect: cjwatson * r112 ubuntu/debian/changelog: releasing version 1.71ubuntu6
[11:55] <CIA-28> libdebian-installer: cjwatson * r190 ubuntu/debian/changelog: releasing version 0.61ubuntu2
[11:56] <CIA-28> lilo-installer: cjwatson * r434 ubuntu/debian/changelog: releasing version 1.29ubuntu2
[11:58] <CIA-28> partman-base: cjwatson * r145 ubuntu/debian/changelog: releasing version 129ubuntu4
[12:00] <CIA-28> partman-target: cjwatson * r756 ubuntu/debian/changelog: releasing version 58ubuntu6
[12:02] <CIA-28> pkgsel: cjwatson * r135 ubuntu/debian/changelog: releasing version 0.20ubuntu14
[12:04] <CIA-28> apt-setup: cjwatson * r165 ubuntu/debian/changelog: releasing version 1:0.37ubuntu11
[13:06] <ogra> cjwatson, if i have a pre-partitioned, pre-fromatted SD card the livesystem sits on, should i be able to install to the free partitions ?
[13:09] <cjwatson> ogra: yes in principle, though you won't be able to change them
[13:09] <ogra> i.e. i have mmcblk0p1 with the livesystem, mmcblk0p2 as swap and mmcblk0p3 as an empty 4G partition ... i get a warning that the content of mmcblk0p3 will be erased, but then ubiquity tries to unmount /cdrom and kicks me back to the partitioner, telling me it needs to make changes to the part. table (which isnt true)
[13:09] <cjwatson> ogra: but the implementation is wrong right now
[13:09] <ogra> ah
[13:09] <cjwatson> ogra: I'd like to debug that
[13:09] <cjwatson> ogra: please file a bug with syslog and partman attached
[13:09] <ogra> where do i find partman ? seems not to be in /var/log/installer
[13:10] <cjwatson> /var/log
[13:10] <cjwatson> actually, just use 'ubuntu-bug ubiquity' please?
[13:10] <ogra> ah, i looked to deep down :)
[13:10] <cjwatson> that'll save you having to find them and will automatically give me everything
[13:11]  * ogra wonders if that will work without him having his LP data handy 
[13:11] <ogra> hmm, seems to
[13:11] <cjwatson> you'll need to log in I'm sure
[13:12] <ogra> it didnt ask
[13:12] <ogra> ah, now it does ... was just slow
[13:15] <xivulon> evand any feedback on wubi in beta?
[13:19] <ogra> wow, ubuntu-bug knows that /usr/share/ubiquity/install.py was modified manually ... cool !
[13:19] <ogra> cjwatson, bug 349581
[13:20] <xivulon> unfortunately they have merged the wubi forum with the general ones and it is more difficult now to get feedback
[13:27] <cjwatson> ubuntu-bug is pretty hot
[13:27] <cjwatson> though I wish it wouldn't gzip the logs
[13:28] <evand> I need to start using it.  Muscle memory keeps preventing me from doing so.
[13:28] <evand> xivulon: none that I've seen.
[13:31] <cjwatson> ogra: the reason it thinks it's changed is that apparently you're changing mmcblk0p3 from ext4 to ext3
[13:32] <cjwatson> ogra: was that intentional?
[13:32] <cjwatson> (actually, as it happens, I don't think this should require changing the partition table; but it may indicate a different bug ...)
[13:33] <ogra> cjwatson, only in the the second try (wasnt intentional, i just selected the wrong fs)
[13:33]  * ogra has to rush out now ... back later
[13:34] <cjwatson> ogra: the second try is the only relevant one; in the first try it didn't think the partition table was changed
[13:58] <xivulon> evand: http://ubuntuforums.org/showthread.php?t=1107950
[14:07] <evand> xivulon: thanks!
[14:08] <evand> xivulon: I'll be away next week on vacation, but if you need an upload of wubi, just shoot me an email and I should be able to take care of it as I'll have my laptop with me
[14:14] <cody-somerville> cjwatson, does the diff ubiquity does between filesystem.manifest and filesystem.manifest-desktop remove dependencies?
[14:14] <CIA-28> ubiquity: evand * r3136 ubiquity/ (3 files in 3 dirs):
[14:14] <CIA-28> ubiquity: Move selecting a disk from radio buttons to a drop down box on the
[14:14] <CIA-28> ubiquity: automatic partitioning page. This saves a lot of space when
[14:14] <CIA-28> ubiquity: multiple disks are present and makes the code a little cleaner
[14:14] <CIA-28> ubiquity: (LP 325958).
[14:14] <cjwatson> filesystem.manifest and filesystem.manifest-desktop are both already dependency-expanded
[14:14] <cody-somerville> cjwatson, ie. if I just remove ubiquity from filesystem.manifest-desktop will it do the right thing?
[14:15] <cjwatson> cody-somerville: no
[14:15] <cody-somerville> okay
[14:15] <cjwatson> I don't think so
[14:15] <cjwatson> cody-somerville: the easiest way is to make sure to spit out filesystem.manifest-desktop at the appropriate point in the livefs build process, before installing ubiquity
[14:18] <cody-somerville> cjwatson, is the live-installer d-i component in Ubuntu?
[14:18] <cjwatson> it's in universe, but entirely unmaintained for Ubuntu
[14:18] <cjwatson> you could have answered this question using rmadison
[14:18] <cody-somerville> oh, neat
[14:35] <cody-somerville> cjwatson, does d-i expect all its components to be in main?
[14:35] <cjwatson> yes
[14:36] <cjwatson> cody-somerville: you should be able to override this by preseeding mirror/udeb/components to "main, restricted, universe, multiverse" (it's "main, restricted" by default)
[14:36] <cjwatson> although that could have other unintended consequences
[14:36] <cjwatson> if you want to do it as a boot parameter, mirror/udeb/components=main,restricted,universe,multiverse should work
[14:37] <cody-somerville> lh is hard coded to download the d-i components from main
[14:37] <cody-somerville> so I was thinking I could add support to it to download d-i components from all components enabled in the build and just put them in /main on the disk
[14:38] <cjwatson> seems plausible
[14:40] <persia> cjwatson, I'm not opposed to allow-password-empty, I just think MID can be fixed to not use it, and didn't want to maintain it.  I'm happy to won'tfix bug #347207 if you want to use it for other things.
[14:41] <cody-somerville> cjwatson, Also, I noticed that d-i tries to setup a PPOE connection. How do I preseed it to not do that?
[14:43] <cjwatson> cody-somerville: that's one of the unintended consequences of using udebs from universe
[14:43] <cjwatson> cody-somerville: you'll have to look into that yourself; it's not something we maintain
[14:46] <cody-somerville> so just excluding the udeb would work?
[14:50] <cjwatson> you'll have to look into that yourself
[14:50] <cjwatson> pppoe is not part of the questions asked by the core installer in main
[14:58] <xivulon> evand: noted, I'll be away from the 10th, will try to sort out the most annoying bugs this w/e, you are around right?
[15:29] <superm1> cjwatson, now that grub-pc is in main and showing up on live dvds, it should be possible to just preseed db_get grub-installer/grub2_instead_of_grub_legacy in situations that grub2 would be preferred, right?
[15:29] <cjwatson> yes, d-i grub-installer/grub2_instead_of_grub_legacy boolean true
[15:30] <cjwatson> we'd definitely appreciate testing of that
[15:30] <superm1> will most definitely.  it will help to lessen a very ugly delta in china installations that's currently there.
[17:02] <cjwatson> TheMuso: can you have a look at bug 311179?
[17:02] <cjwatson> it makes some sense, but I'm not sure how to detect this reliably ...
[17:22] <cjwatson> (InstallerDevelopment -> Installer/Development)
[17:31] <cody-somerville> cjwatson, How do I preseed partman not to ask to umount disks?
[17:33] <superm1>   d-i ubiquity/partman-skip-unmount boolean true
[17:34] <cjwatson> in general, to find out how to preseed something, first check the installation guide (although I don't think this is there) and then run the installer in debugging mode so that you can see what question it's asking
[17:34] <cjwatson> hmm, superm1's advice is valid but omits that the question is also in partman-base now
[17:35] <cjwatson> evand: if partman-base is handling this, should that script be removed from ubiquity?
[17:35] <cjwatson> (for the implementation in partman itself, the preseeding would be 'd-i partman/filter_mounted boolean false')
[17:39] <superm1> actually in our seed we have the combination of both of those as I look, i don't recall the reasoning for using both, but before removing the ubiquity one, it would be best to ensure that it's really not necessary anymore
[17:39] <cjwatson> of course
[18:44] <evand> indeed, I suspect that chuck of code can go away
[18:44] <evand> though, perhaps we should hal-lock disks in partman so we don't run into a situation where they mount a disk after the first partman check
[18:45] <cjwatson> I'm not too worried about that, as it strikes me as a "doctor, doctor, it hurts when I do this" kind of situation
[19:45] <evand> heh, fair enough
[20:15] <cody-somerville> When I try to mounting a squashfs, I get mount: mounting /dev/loop0 on /mnt failed: No such device
[20:16] <cody-somerville> What might be the problem?
[20:18] <cody-somerville> hmm... I just called a segmentation fault in mount
[20:22] <evand> cody-somerville: mount -t squashfs -o loop /mnt ?
[20:22] <evand> or is this in the depths of partman now
[20:22] <cody-somerville> Thats what I'm doing
[20:22] <cody-somerville> but I get mount: mounting /dev/loop0 on /mnt failed: No such device
[20:24] <cody-somerville> Then I did something and it caused "BUG: unable to handle kernel NULL pointer dereference at 00000080"
[20:32] <evand> Does /dev/loop0 actually exist?
[20:34] <cody-somerville> yes
[20:38] <evand> Not sure, I'd ask in -kernel.
[20:44] <btm> does beta installer for jaunty have new gpg keys in it?
[20:44] <btm> for the repositories?
[21:17] <cjwatson> btm: the archive's gpg key has not changed, so why would it need to?
[21:18] <btm> cjwatson: I was getting a GPG signature error using netboot=20081029ubuntu21. Moving to netboot=20081029ubuntu28 resolved the problem.
[21:21] <btm> during the base install. last time I tested the installer was a couple weeks ago and it was working. 'spose it doesn't matter since the beta release works.
[21:29] <cjwatson> I don't know why that would have been; it could have been due to Release and Release.gpg genuinely being out of sync on your mirror for a period, and it so happened that by the time you moved to ubuntu28 it was fixed
[21:29] <cjwatson> at any rate, the keys definitely haven't changed
[21:35] <btm> I had a different issue with my local mirror, which lead me to testing against us.archive.ubuntu.com, which lead to the upgrade of the installer, but anyhow, it looks like it'll work locally as soon as apt-mirror finishes pulling in packages now.
[21:42] <cody-somerville> cjwatson, the squashfs module isn't in Ubuntu's d-i's initrd
[21:43] <cody-somerville> cjwatson, Is there a work around or what not to be able to mount squashfs?
[21:50] <btm> cody-somerville: could you copy the module from linux-image-2.6.28-11-generic over scp?
[21:51] <cody-somerville> no
[21:55] <shtylman> evand: apprently that battery thing has popped up again...
[21:56] <shtylman> someone reported having the bug...
[21:56] <btm> cody-somerville: why not?
[21:59] <cody-somerville> because it doesn't fix the actual problem
[22:03] <btm> I thought you were looking for a work around.
[22:04] <cody-somerville> a work around as in including the udeb that provides the squashfs modules
[22:15] <btm> I don't believe there is one, but a wget/insmod in preseed/early_command should work.
[23:50] <evand> https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/349937 - that's the second case I've seen of hw-detect and X failing (I'm assuming the former as a result of the latter)