=== CarlFK1 is now known as CarlFK [09:39] evand: Gentle prod to check on the error message I reported with wubi [09:41] StevenK: -ENOCONTEXT [09:41] was this a bug report? [09:43] evand: This was testing the wubi-r144.exe you prepared for me [09:44] ah, where did you report the error message? If it was in here I probably lost it / didn't notice it. [09:45] evand: It was in here, yes. [09:45] You even replied :-) [09:45] apologies then [09:45] oh [09:45] then my memory is shot [09:45] can you post it again anyway? [09:45] evand: Huzzah, that gives me an Install inside Windows button, which when clicked says "'NoneType' object has no attribute 'name'" [09:45] cute [09:46] Yay for long running IRC sessions [09:46] I'll prod that just as soon as I'm done sorting out a new ubiquity-slideshow-ubuntu [09:46] yeah, I gave up on irssi + screen + server a while ago and just use pidgin now. [10:04] ubiquity-slideshow-ubuntu: evand * r128 ubiquity-slideshow-ubuntu/debian/ (changelog control copyright): [10:04] ubiquity-slideshow-ubuntu: * Move translate-toolkit and gimp to build-deps-indep. [10:04] ubiquity-slideshow-ubuntu: * Removing trailing comma from dependencies. [10:04] ubiquity-slideshow-ubuntu: * Drop coreutils dependency as it's in essential. [10:04] ubiquity-slideshow-ubuntu: * Use the proper Vcs-Bzr field, rather than the old XS-Vcs-Bzr. [10:04] ubiquity-slideshow-ubuntu: * Replace entire GPL-2 copyright with a pointer to common-licenses. [10:04] ubiquity-slideshow-ubuntu: * Bump standards version to 3.8.3. [10:10] ubiquity-slideshow-ubuntu: evand * r129 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 3 [11:01] wubi: evand * r145 trunk/ (data/isolist.ini debian/changelog): Kubuntu Netbook Edition is now Kubuntu Netbook. [11:09] wubi: evand * r146 trunk/data/images/ (3 files): Forgot to rename files in the previous commit. [11:58] ah, it's down to Wubi looking for an amd64 Kubuntu Netbook image, which doesn't exist [12:17] hmm, nice to see that people were doing routine d-i maintenance in my absence ;-) [12:17] debian-installer: cjwatson * r1129 ubuntu/ (10 files in 3 dirs): Move to 2.6.31-6 kernels. [12:24] libdebian-installer: cjwatson * r196 ubuntu/ (debian/changelog src/system/subarch-arm-linux.c): merge lp:~mcasadevall/libdebian-installer/dove-soc [12:26] libdebian-installer: cjwatson * r197 ubuntu/src/system/subarch-arm-linux.c: typo [12:27] libdebian-installer: cjwatson * r198 ubuntu/debian/changelog: releasing version 0.63ubuntu2 [12:30] base-installer: cjwatson * r375 ubuntu/ (4 files in 3 dirs): merge lp:~mcasadevall/base-installer/dove-soc [12:31] base-installer: cjwatson * r376 ubuntu/debian/changelog: releasing version 1.101ubuntu3 [12:34] debian-installer: cjwatson * r1130 ubuntu/ (6 files in 4 dirs): merge lp:~mcasadevall/debian-installer/dove-soc [12:34] my apologies, it didn't occur to me that it would be needed, but definitely noted for next time you're on holiday. [12:36] :-) just teasing [12:36] I'll give libd-i a bit to build before uploading that [12:43] cjwatson: Regarding https://bugs.edge.launchpad.net/dell/+bug/413135 , I dug around a bit, but ran out of time last week to further investigate it. Is this bug still on your radar? [12:43] Launchpad bug 413135 in oem-priority "Installer needs way to install PAE kernel on i386 9.10 DVD" [Undecided,New] [12:43] yes, it is [12:43] well, the general idea is, I hadn't seen the actual bug 'til now [12:44] okay, good deal [12:44] Jerone was sending nagging mails about that while I was on holiday [12:50] wubi: evand * r147 trunk/ (3 files in 3 dirs): [12:50] wubi: Use i386 if an amd64 version of the selected distribution does not [12:50] wubi: exist. [12:55] cjwatson hi, I am stacked with grub2, I get to the point where the linux kernel and initrd are loaded, but as soon as I run the `boot` command it jams [12:57] grub2 is generated here: http://bazaar.launchpad.net/%7Eubuntu-installer/wubi/trunk/annotate/head%3A/Makefile#L74 [13:04] sorry, I'm not very good at debugging that kind of thing yet [13:05] try to narrow down whether it's failing in grub or in early kernel startup [13:08] base-installer: cjwatson * r377 ubuntu/ (6 files in 5 dirs): Prefer PAE kernels on machines with >3GB of RAM (LP: #413135). [13:12] cjwatson do you know how to get some debug info? I turned on debug mode in grub but does not help [13:12] xivulon: wubi r147 got past grub for me. [13:12] woo [13:12] perhaps your issue is a grub bug, rather than a misconfiguration? [13:12] may be is because I use VB... [13:13] or maybe the version of grub I am using on my builid system... [13:13] perhaps [13:13] a few hours wasted... :( [13:14] good to hear that, how far do you go? [13:14] xivulon: well, you've possibly found a bug, so it's not a waste. [13:14] I would definitely take it to the grub mailing list though [13:14] * xivulon does not like bugs when there is no shell [13:14] if, as cjwatson says, you can determine it's in grub and not in the initramfs. [13:15] xivulon: to the kubuntu netbook desktop [13:16] ah cool [13:16] (which is just because ubiquity crashed) [13:17] debian-installer: cjwatson * r1131 ubuntu/ (4 files in 2 dirs): Add pcmcia-modules to hd-media images (LP: #415675). [13:17] evand is it #243105? [13:17] We'll need to do some changes in update-grub as well, hopefully it will be easier this time around [13:17] no, it appears to be a ubiquity-frontend-kde/kde issue [13:19] xivulon: I don't really, sorry [13:20] xivulon: if it might be in the initramfs, adding debug= to the boot parameters line might help [13:20] and remove quiet splash, of course [13:21] ah, python-kde4 is broken. Hilarious. [13:22] StevenK: So I fixed the issue you were hitting, and as noted above, I've gotten to ubiquity, but it's crashing due to a broken symlink in python-kde4. I imagine wubi kubuntu netbook installs will work once that is fixed. [13:23] (bug #417494 - seems to be fixed already) [13:23] Launchpad bug 417494 in kdebindings "Broken symlink in python-kde4" [Undecided,Fix released] https://launchpad.net/bugs/417494 [13:26] evand, OEM needs me this week. Which means I probably won't do much ubiquity plugin work. Thankfully, it's in a state where it could be pushed (read reviewed). The last thing I wanted to do this week is reorganize the plugins a bit to be more extendable. But I don't promise I can get to that. So I'm for-realz asking for a review sometime this week in anticipation of FF [13:26] cjwatson, looks more like grub, it also jams with no initramfs at all, only vmlinuz and seems to be jamming quite early on [13:27] mterry: okay, it's already on my todo list for today. I'll make sure I get to it next. [13:27] thanks for the heads up [13:27] evand, in addition to reading the code, look at https://wiki.ubuntu.com/Ubiquity/Plugins to see how it's supposed to work and an example plugin you could play with [13:27] okay [13:27] evand, cool, thanks [13:29] base-installer: cjwatson * r378 ubuntu/debian/changelog: releasing version 1.101ubuntu4 [13:47] xivulon: http://grub.enbug.org/HowToDebug may be helpful, or not [13:47] 'set debug=all' might be useful in any case ... [13:56] cjwatson thx will give it a try tonight [14:22] evand: UNR, even? [14:26] StevenK: yeah, that too :) [14:28] evand: So there's a new .exe published, or not yet? [14:29] StevenK: there is now [14:30] evand: Excellent, I'll test it tomorrow. Thanks! [14:30] sure thing [15:02] ubiquity-slideshow-ubuntu: evand * r130 ubiquity-slideshow-ubuntu/COPYING: Include the GPLv2 in COPYING. [16:06] cjwatson, thanks for looking at base-installer; I need to tweak the dove d-i branch though so please don't merge that one quite yet. (I wrote the branch on the assumption we would get a vmlinuz, not a uImage) [16:06] oh, I already did [16:06] but haven't uploaded [16:06] cjwatson, I can fix that quickly [16:06] cjwatson, hold on [16:10] cjwatson, checking out the branch; I'll apply my change and testbuild the ful larmel build so we can make sure we don't break imx51 this quickly (I apologize, I didn't except you to get on this as quickly as you did coming off vacation :-)) [16:25] evand: hiya, testing the latest iso's for encrypt-home [16:25] evand: is formatting ext4 supposed to take a long time at 5% ? [16:58] kirkland: it could. Large drive? Might want to check the output of ps auxf [16:58] evand: guest or host? [16:59] guest [16:59] to make sure it is in fact formatting [17:36] cjwatson: so we've run into a bit of a problem regarding encrypted swap as it's currently implemented. The password ends up in clear text on the swap device because the latter is not overwritten until its needed. The workaround that kirkland and I have come up with for this is to dd /dev/zero over all the swap devices at the end of the install. Do you think it's acceptable? It strikes me as being really hackish. [17:36] evand: only do that dd *if* encrypted-swap is flagged "on" [17:36] kirkland: indeed [17:36] evand: as this operation will take a few seconds [17:36] I'm wondering if it would just be better to create a cryptswap type in partman and handle things there. [17:36] evand: and we shouldn't penalize non-swap-encrypted installations [17:36] absolutely [17:37] evand: i know you're not asking for my opinion, but wiping the swap slate clean on shutdown doesn't seem *that* hackish to me [17:38] cjwatson: (for what it's worth, the current implementation is user-setup-apply calling ecryptfs-setup-swap, which configures the target filesystem for swap use, but does not actually enable it) [17:38] evand: and as i understand it, a cryptswap in partman won't help in the encrypted-home case, because we don't know that we need encrypted swap until *after* partman has run [17:39] hrm [17:40] my concern was more having partman-target (or whichever component) check user-setup/encrypt-home, but I guess that's not necessary [17:40] evand: but that runs *before* user-setup/encrypt-home is set (when preseeding is not used) [17:41] no I mean my concern with fixing this issue as we've proposed [17:41] ignore the cryptswap idea [17:41] * kirkland is lost :-) [17:42] sorry, I'm thinking this through as I type [17:42] probably not the best approach [17:42] the one thing partman gives us that this approach does not is some structure [17:42] * kirkland tries to keep up [17:42] the only way I can think of doing this is parsing crypttab for swap [17:42] which I guess isn't that bad [17:43] evand: right, that's what I'm thinking [17:43] evand: thinking about this as a paragraph in english ... -> [17:44] evand: "If you've indicated that you want us to prepare your swap devices for encryption, we should wipe them clean for you, as part of the installation/setup process." [17:46] I don't think we actually need to give the user the option. I'm concerned that more people understand what "encrypt my home directory" means than "prepare my swap devices for encryption" [17:46] evand: agreed [17:46] evand: i'm not proposing we expose "wipe the swap" as an option [17:47] evand: i'm suggesting that's implied by swap encryption, which is implied by homedir encryption ;-) [17:47] evand: bootstrapping security is fun! [17:47] sure [17:47] haha [17:47] I think a dd is ok [17:48] I don't think a cryptswap type would really be right - it's not cryptswap, it's swap on an encrypted volume [17:49] cjwatson: the method evand is using to create encrypted swap (ecryptfs-setup-swap), adds entries to /etc/crypttab that are cryptswap types [17:51] that's not the same as partman thinking that's the type though ... [17:51] cjwatson: noted; thanks === cody-somerville_ is now known as cody-somerville [18:14] cjwatson: is this a suitable place as far as you're concerned (http://pastebin.ubuntu.com/258790/) or should I move the code into finish-install.d/99wipe_swap and duplicate it in ubiquity? [18:20] evand: I think it's OK, just make it not fail if /etc/crypttab doesn't exist [18:21] ah, good call [18:22] swapoff would probably be a good idea too :-/ [18:28] user-setup: evand * r202 ubuntu/ (debian/changelog user-setup-apply): [18:28] user-setup: Zero out swap devices at the end of install when encryption is [18:28] user-setup: enabled. [18:38] user-setup: evand * r203 ubuntu/debian/changelog: releasing version 1.27ubuntu9 [19:04] evand: \o/ [19:04] evand: might i expect another iso today, or should i just pick it up tomorrow? [19:41] cjwatson, merged changes: lp:~mcasadevall/debian-installer/dove-soc [19:41] cjwatson, the imx51 targets also build, but I'm not sure they target the newest kernel [21:24] ubiquity: evand * r3395 ubiquity/ (d-i/manifest debian/changelog): [21:24] ubiquity: Automatic update of included source packages: base-installer [21:24] ubiquity: 1.101ubuntu4, user-setup 1.27ubuntu9. [21:36] ubiquity: evand * r3396 ubiquity/debian/changelog: releasing version 1.99.12 [22:50] NCommander: thanks [22:50] cjwatson, sorry on the delay for that === robbiew is now known as robbiew-afk [22:52] *shrug8 [22:52] * [23:11] debian-installer: cjwatson * r1132 ubuntu/build/config/armel/ (dove.cfg dove/netboot.cfg): merge lp:~mcasadevall/debian-installer/dove-soc [23:29] debian-installer: cjwatson * r1133 ubuntu/debian/changelog: releasing version 20081029ubuntu52