[04:24] <evand> ffm: Shouldn't be necessary.  I actually pointed out that same issue to xivulon a while back, I just didn't file a bug for it.  It might already be fixed, I'll give a try with wubi HEAD in the morning, otherwise I'll remind him about that bug number.
[13:32] <ffm> evand: Thanks.
[15:34] <saispo> hi
[15:34] <saispo> anyone know if it's possible to include *-security and *-updates packages in a custom cd ?
[15:37] <grrrreg> I think so
[15:38] <saispo> i try but some error in the installation process :-|
[15:38] <saispo> grrrreg: must modify germinate and some shell scripts ?
[15:52] <cjwatson> grrrreg: getting anna to fetch from universe> not in gutsy, but in hardy (er, at least after the next debian-installer upload) you'll be able to preseed mirror/udeb/components=main,restricted,universe
[16:14] <grrrreg> cjwatson: thanks a lot !
[16:17] <cjwatson> oops, I'd disabled my CIA plugin
[16:17] <CIA-25> console-setup: cjwatson * r54 ubuntu/ (Keyboard/ckbcomp debian/changelog):
[16:17] <CIA-25> console-setup: * Silence error messages for keycodes 214, 215, 216, and 217, which are
[16:17] <CIA-25> console-setup:  various special keys in the default X.org PC keyboard map that we don't
[16:17] <CIA-25> console-setup:  need to handle.
[16:17] <CIA-25> console-setup: cjwatson * r55 ubuntu/ (Keyboard/ckbcomp debian/changelog): * Map dead_stroke to U+002D HYPHEN-MINUS.
[16:17] <CIA-25> console-setup: cjwatson * r56 ubuntu/ (4 files in 3 dirs):
[16:17] <CIA-25> console-setup: * Add a "Do not configure keyboard; keep kernel keymap" model option,
[16:17] <CIA-25> console-setup:  preseedable by console-setup/modelcode=SKIP (LP: #59889).
[16:17] <CIA-25> console-setup: cjwatson * r57 ubuntu/debian/ (changelog config.proto):
[16:17] <CIA-25> console-setup: * Drop the layout/variant priority to high on fresh installs
[16:17] <CIA-25> console-setup:  (LP: #205959).
[16:17] <CIA-25> console-setup: cjwatson * r58 ubuntu/ (Keyboard/ckbcomp debian/changelog):
[16:17] <CIA-25> console-setup: * Support rules with multiple consequents (thanks, Adam Simpkins;
[16:18] <CIA-25> console-setup:  LP: #139710).
[16:18] <CIA-25> console-setup: cjwatson * r59 ubuntu/ (debian/changelog setupcon):
[16:18] <CIA-25> console-setup: * Allow specifying alternate rules using XKBRULES (thanks, Adam Simpkins;
[16:18] <CIA-25> console-setup:  LP: #139712).
[16:18] <CIA-25> console-setup: cjwatson * r60 ubuntu/Keyboard/KeyboardNames.pl: regenerate
[16:18] <CIA-25> console-setup: cjwatson * r61 ubuntu/debian/console-setup.postinst: empty BOOTTIME_KMAP_MD5 if skipping keyboard configuration
[16:18] <CIA-25> console-setup: cjwatson * r62 ubuntu/debian/changelog: releasing version 1.21ubuntu5
[16:18] <CIA-25> net-retriever: cjwatson * r343 ubuntu/ (3 files in 2 dirs):
[16:18] <CIA-25> net-retriever: * Fetch udebs from components listed in mirror/udeb/components
[16:18] <CIA-25> net-retriever:  (LP: #191262).
[16:18] <CIA-25> net-retriever: cjwatson * r344 ubuntu/debian/changelog: releasing version 1.18ubuntu3
[16:18] <CIA-25> net-retriever: cjwatson * r345 ubuntu/ (debian/changelog net-retriever):
[16:18] <CIA-25> net-retriever: * Fetch security updates path from apt-setup/security_path if possible
[16:18] <CIA-25> net-retriever:  (LP: #51470).
[16:18] <CIA-25> net-retriever: cjwatson * r346 ubuntu/debian/changelog: releasing version 1.18ubuntu4
[16:18] <CIA-25> installation-guide: cjwatson * r415 ubuntu/ (debian/changelog en/appendix/preseed.xml): * Document mirror/udeb/components (LP: #191262).
[16:18]  * evand discovers that cabp escape can be used on question names, rejoices.
[16:18] <CIA-25> installation-guide: cjwatson * r416 ubuntu/ (debian/changelog en/appendix/preseed.xml): * Document apt-setup/security_path (LP: #51470).
[16:18] <CIA-25> installation-guide: cjwatson * r417 ubuntu/ (debian/changelog en/appendix/preseed.xml): * Document console-setup/modelcode=SKIP (LP: #59889).
[16:18] <CIA-25> installation-guide: cjwatson * r418 ubuntu/debian/changelog: releasing version 20080211ubuntu3
[16:19] <CIA-25> clock-setup: cjwatson * r190 clock-setup/ (3 files in 2 dirs):
[16:19] <CIA-25> clock-setup: * Add clock-setup/utc-auto template; if preseeded to true and we are the
[16:19] <CIA-25> clock-setup:  only OS on the disk, clock-setup/utc will not be asked.
[16:19] <CIA-25> clock-setup: cjwatson * r191 clock-setup/finish-install.d/10clock-setup: check clock-setup/utc-auto in a more constrained case
[16:19] <CIA-25> clock-setup: cjwatson * r192 clock-setup/debian/changelog: releasing version 0.92ubuntu3
[16:19] <mario_limonciell> cjwatson, in just running oem-config, I still see an oem user post run, isn't the solution used for bug 153311 then a security vulnerability?
[16:21] <CIA-25> ubiquity: cjwatson * r2595 ubiquity/ (3 files in 2 dirs):
[16:21] <CIA-25> ubiquity: * Fix crash if partitioning failed when partitioning was automated (see LP
[16:21] <CIA-25> ubiquity:  #206113).
[16:21] <cjwatson> mario_limonciell: hmm, it's supposed to be removed. damnit.
[16:21] <cjwatson>                 # Don't run again.
[16:21] <cjwatson>                 update-rc.d -f oem-config remove >/dev/null 2>&1
[16:21] <cjwatson>                 # Remove the temporary OEM configuration user, if possible.
[16:21] <cjwatson>                 if getent passwd oem >/dev/null; then
[16:21] <cjwatson>                         deluser --quiet --remove-home oem || true
[16:22] <cjwatson>                 fi
[16:22] <cjwatson>                 exit 0
[16:22] <cjwatson> what went wrong?
[16:22] <mario_limonciell> hm, is there a log?
[16:22] <mario_limonciell> i'll take a look
[16:22] <cjwatson> mario_limonciell: did oem-config fail? try /var/log/oem-config.log
[16:23] <mario_limonciell> it doesn't look like it did.  I'll rerun it this afternoon in debug mode
[16:23] <mario_limonciell> and see if I can narrow this down further
[16:24] <cjwatson> might be worth sticking set -x at the top of oem-config-firstboot
[16:24] <mario_limonciell> good idea
[16:24] <cjwatson> and probably 'exec 2>/var/log/oem-config-firstboot.log' too since I don't think it sends stderr anywhere sensible
[17:14] <mario_limonciell> cjwatson, ah, disregard my previous concerns with oem-config.  It was because of doing some test on another VT while waiting for it to run (Can't delete an in use user :))
[17:15] <cjwatson> ah :)
[17:15] <cjwatson> probably wouldn't hurt to be loud about the failure somewhere
[17:15] <mario_limonciell> well I found that on a different VT
[17:15] <mario_limonciell> so it was loud somewhere
[17:54] <mario_limonciell> cjwatson, as my understanding is, there is no way to enforce casper to load from a particular drive in its current state, correct?
[17:57] <cjwatson> you should be able to give it a UUID
[17:58] <cjwatson> /conf/uuid.conf in the initramfs, if set, must match one of /.disk/casper-uuid* on the image
[17:58] <mario_limonciell> but what if the uuid is identical on both images?
[17:58] <cjwatson> we do this by default for Ubuntu CDs
[17:58] <cjwatson> don't do that ;-)
[17:58] <mario_limonciell> this particular use case is for recovery disks
[17:58] <cjwatson> can you give me more context?
[17:58] <cjwatson> right, shouldn't recovery disks have a different UUID?
[17:58] <cjwatson> hmm, I guess not if you're just copying them, bah
[17:59] <mario_limonciell> well not when they are made from the original recovery disk
[17:59] <mario_limonciell> yeah
[17:59] <cjwatson> the point of this was kind of to help with recovery disks
[17:59] <cjwatson> how much control do you have over the recovery disk creation process?
[17:59] <mario_limonciell> total
[18:00] <mario_limonciell> so it would be feasible to create new UUID's I suppose when the disks are made
[18:00] <cjwatson> un-cpio the initramfs, whack in a new UUID, cpio it back up, stick the UUID in /.disk on the recovery image?
[18:00] <cjwatson> wouldn't hurt to have a tool to automate that
[18:00] <cjwatson> we could put it in casper
[18:01] <mario_limonciell> how is the UUID originally made, eg can just a random string of characters be used?
[18:04] <cjwatson> uuidgen -r
[18:04] <cjwatson> (/usr/share/initramfs-tools/hooks/casper)
[18:04] <mario_limonciell> ah very good. I'll go this route then.
[18:05] <mario_limonciell> i'll put all the steps in their own script that you can then add to casper
[18:05] <mario_limonciell> thanks
[18:05] <cjwatson> great
[18:09] <CIA-25> casper: cjwatson * r491 casper/ (2 files in 2 dirs):
[18:09] <CIA-25> casper: * Make scripts/casper-bottom/38disable_restricted_manager executable
[18:09] <CIA-25> casper:  again.
[18:11] <CIA-25> casper: cjwatson * r492 casper/debian/ (changelog control): * Depend on uuid-runtime for uuidgen.
[18:11] <cjwatson> mario_limonciell: ^-- might not have been helping
[18:12] <mario_limonciell> hr, um?  uuidgen is no longer in e2fsprogs
[18:12] <mario_limonciell> oh there we go uuid-runtime
[18:12] <mario_limonciell> okay thanks :)
[18:13] <mario_limonciell> could you add that package to the seeds of debs on the DVD then too?
[18:14] <CIA-25> casper: cjwatson * r493 casper/debian/changelog: releasing version 1.128
[18:14] <cjwatson> shouldn't be needed, casper is in supported-installer already so it'll be pulled in by dependencies
[18:15] <mario_limonciell> ah okay
[19:07] <cjwatson> is it just me, or is it really quite hard to target map points in the new timezone widget? you seem to have to click right on it
[19:07] <cjwatson> I'm sure it wasn't quite so picky in an earlier iteration of the new widget
[19:08] <evand> it's not just you, I very much need to find a solution to that problem before the release.
[19:08] <cjwatson> ok, good, as long as it's on somebody's plate :)
[19:09] <evand> heh, indeed.  My current thought is to either keep a static map with different levels of zoom (like the old widget) or scroll as the mouse approaches an edge.
[19:10] <evand> But I'll try to find a way to make having a ridiculously large surface that you can scroll across efficient enough to use before I resort to that.
[19:11] <evand> oh, you're not complaining about that, are you?
[19:11] <evand> hrm.
[19:11] <evand> I'll look into that
[19:28] <mario_limonciell> cjwatson, http://paste.ubuntu.com/6272/ should suffice if you want to add that to casper's source for whenever it gets uploaded next
[19:35] <cjwatson> mario_limonciell: looks plausible at a quick glance - trying to finish up for the day here though
[19:35] <mario_limonciell> cjwatson, take your time.  it will live in our recovery disk source package for now.  I can put it on a bug for you if you want
[19:35] <cjwatson> I've dumped it into my working tree
[19:35] <cjwatson> yes please
[19:36] <cjwatson> casper-uuid a plausible name?
[19:36] <cjwatson> oh, copyright statement wouldn't hurt
[19:36] <mario_limonciell> oh yeah.  should put that in
[19:36] <mario_limonciell> it'll be on the bug then
[19:36] <mario_limonciell> (with that version)
[20:15] <cr3> how can I install packages from busybox to reproduce a problem which might be occuring with packages defined in pkgsel/include?
[20:16] <cr3> I don't think that in-target apt-get install package_name is quite the same, but I might be wrong
[22:46] <TheMuso> evand: Any ideas on why the ubiquity GUI in only-ubiquity is running as root? I say this because while orca comes up speaking now, Orca/at-spi cannot see any widgets in the ubiquity window.
[23:06] <mario_limonciell> TheMuso, are all the other processes coming up as a user in ubiquity-dm?
[23:13] <TheMuso> mario_limonciell: Yep they are.
[23:13] <TheMuso> gnome-settings-daemon, at-spi-registryd, orca, metacity, the lot.
[23:21] <mario_limonciell> TheMuso, glancing over ubiquity-dm, all of those processes are preexec_fn with self.drop_privileges, whereas the ubiquity process isn't
[23:21] <xivulon> bdmurray, did you happen to do any other test for bug #204133 ?
[23:21] <mario_limonciell> so is ubiquity supposed to drop them when it first starts then?
[23:21] <xivulon> I cannot reproduce mysel and am no wiser than the last time we talked :(
[23:22] <mario_limonciell> TheMuso, oh nvm, frontend/base.py, supposed to drop there.
[23:23] <TheMuso> mario_limonciell: Right.
[23:23] <mario_limonciell> TheMuso, well here's the issue then
[23:23] <mario_limonciell> in frontend/base.py, it uses SUDO_UID and SUDO_GID to drop them
[23:23] <mario_limonciell> are those valid in ubiquity-dm ?
[23:23] <bdmurray> xivulon: I've been on holiday and just got back
[23:23] <TheMuso> hang on, let me have a look.
[23:23] <mario_limonciell> ubiquity-dm's drop_privileges uses         self.uid, self.gid = pwd.getpwnam(self.username)[2:4]
[23:24] <xivulon> bdmurray, np
[23:26] <TheMuso> mario_limonciell: right
[23:26] <TheMuso> ah yeah I see what you mean.
[23:26] <mario_limonciell> TheMuso, so two possible solutions then: 1) revamp the method used elsewhere that uses SUDO_UID/GID, 2) set SUDO_UID SUDO_GID before ubiquity starts
[23:27] <mario_limonciell> 2) feels more like a hack, but should be functional and less coding and verification necessary
[23:27] <TheMuso> I think 1 is a better bet possibly. I'll have a dig today.