[09:32] <yann|work> question about a customized alternate intrepid cd - any idea what could cause the installer not to see tasks minimal/standard/xubuntu-desktop ?
[09:33] <yann|work> would the lack of a Release.gpg cause this ?
[09:47] <davmor2> evand, cjwatson bug 403021 seems to still be in effect on nv/nvidia drivers but not others.
[09:59] <cjwatson> davmor2: I think the X team will have to deal with that if it's hardware-specific
[10:03] <davmor2> cjwatson: do you think it would be best to write a new bug out for it then?
[10:06] <cjwatson> I'll just reassign it
[10:06] <cjwatson> evand: you're the assignee at the moment - is that OK with you?
[10:19] <evand> cjwatson: fine by me
[11:02] <evand> cool! For whatever reason, it didn't register with me that we'd get the slideshow and encrypted home in oem-config for free.  Hooray for mterry and merges
[11:09] <davmor2> evand: I saw that but I'm not going to test till post a5
[11:34] <davmor2> cjwatson: I thought that grub was meant to be hidden now?
[11:35] <cjwatson> only if you don't have other operating systems installed
[11:38] <davmor2> ah I got another one on the other disc in the machine I forgot about that :)
[12:25] <CIA-33> wubi: evand * r150 trunk/ (data/wubildr.cfg debian/changelog):
[12:25] <CIA-33> wubi: Use the correct syntax for grub2 (elif does not exist). Thanks to
[12:25] <CIA-33> wubi: Colin Watson for the tip.
[14:01] <evand> cjwatson: Have you seen https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/419194 ?  I just want to make sure it hasn't fallen into an email black hole.
[14:03] <cjwatson> oh, sorry, didn't realise that was waiting for me; I'll reply in the bug
[14:07] <evand> no worries and thanks
[14:20] <rgreening> heya evand
[14:23] <CIA-33> usb-creator: rgreening * r151 usb-creator/ (TODO debian/changelog usbcreator/frontends/gtk/frontend.py): * Fix exit button(s) which fail to exit in usb-creator-gtk (on fail or successful install dialog)
[14:24] <evand> rgreening: hi
[14:25] <rgreening> evand: I just made a change, but it doesn't completely work... bugger. exit button signals in usb-creator-gtk. I thought adding quit method signal was it... seems not...
[14:26] <ogra> cjwatson, is the diverting of update-initramfs from casper a new thing in karmic ? i cant remember that it was like that in jaunty
[14:26] <evand> where was the exit button failing to work for you?  it's been working fine here in all my testing.
[14:27] <cjwatson> ogra: bzr blame is your friend, it's what I'd have to do to answer your question anyway
[14:27] <ogra> heh, ok
[14:27] <rgreening> evand: on successfull install, gtk/python hung in background
[14:27] <cjwatson> or the changelog ...
[14:27] <rgreening> so, clicking exit did not close down the app correctly, just hid the main menu.
[14:28] <rgreening> dialog window I mean..
[14:28] <ogra> cjwatson, also the update-initramfs run at the end of the install, is that one new ? we already run one in-target a while before
[14:28] <ogra> cjwatson, looking at bug 422101
[14:28] <cjwatson> I have no idea whether it's new, sorry
[14:28]  * ogra doesnt see what makes the second one necessary
[14:29] <cjwatson> I doubt it's deliberate
[14:29] <cjwatson> the installer is a complex system, you sometimes get emergent behaviour
[14:29] <ogra> indeed
[14:30] <rgreening> evand: hmmm... maybe the problem is in def success in gtk/frontend.py. it calls sys.exit(0) and not gtk.main_quit().
[14:30] <cjwatson> it's not like there's a run_update_initramfs_again_to_annoy_ogra() method :)
[14:30] <cjwatson> in this case, it's run again due to a dpkg trigger while removing packages
[14:31] <cjwatson> according to the log, triggered while removing dmraid
[14:31] <cjwatson> perhaps /dev is not bind-mounted at that point or something?
[14:33] <evand> rgreening: looking into it now
[14:33] <rgreening> evand: ok.
[14:35] <rgreening> evand: fixed
[14:36] <evand> oh?
[14:36] <rgreening> evand: needs to call gtk.main_quit() in success
[14:36] <evand> ah, good deal
[14:36] <rgreening> not sys.exit
[14:36] <rgreening> :)
[14:36] <evand> thanks a bunch!
[14:36] <rgreening> I should have realized that
[14:36] <rgreening> lol
[14:36] <rgreening> I'll update
[14:38] <CIA-33> usb-creator: rgreening * r152 usb-creator/usbcreator/frontends/gtk/frontend.py:
[14:38] <CIA-33> usb-creator: * Use gtk.main_quit() in success method reather than sys.exit(0), otherwise app hangs.
[14:38] <CIA-33> usb-creator: * Revert previous attempt to fix issue above.
[14:38] <rgreening> ok, one less issue :)
[14:39] <rgreening> evand: how goes getting ISO images to work again?
[14:39] <evand> it goes.  I've been buried under ubiquity/oem-config bugs today
[14:39] <rgreening> ah.
[14:46] <CIA-33> usb-creator: evand * r153 usb-creator/ (debian/changelog usbcreator/frontends/gtk/frontend.py):
[14:46] <CIA-33> usb-creator: Fail more gracefully when an error occurs while setting up the install
[14:46] <CIA-33> usb-creator: routine.
[15:25] <CIA-33> usb-creator: rgreening * r154 usb-creator/usbcreator/frontends/kde/frontend.py: * Update kde/frontend.py to match gtk/frontend.py for handle graceful failed installations
[15:43] <rgreening> evand: did you notice that from trunk that the target list now contains a bunch of devices which should not appear. For one, any device with 0 capacity and 0 free should not show.I also see my CD-ROM show up in available targets. It's a read-only device, so it should be filtered too. SOmething broke this, but not sure what.
[15:50] <rgreening> evand: did you figure out how to get Format working yet?
[15:58] <superm1> cjwatson, i've been getting general protection crashes in libgobject-2.0.so.0.2105.0, usually when getting to the formating stage, but apport isn't catching them, any tips?  just gdb around the python process for ubiquity?
[16:00] <cjwatson> that, or maybe 'ulimit -c 0' inside the sudo first will convince it to drop a core file
[16:00] <superm1> k
[16:07] <superm1> must have intimidated it by threatening a debugger, now it doesn't even want to reproduce :(
[16:19] <yann|work> does anyone has an idea about what in a customized alternate intrepid cd could cause the installer not to see tasks minimal/standard/xubuntu-desktop ?
[16:20] <yann|work> the fun thing is, I have no problem with any customization, as long as I do not add a new component besides main and friends
[16:21] <yann|work> even removing the Release.gpg does not cause the problem, so the event triggering the problem seems to be the sole addition of this "extra" component
[16:21] <yann|work> I can't imagine *that* being the root problem
[16:22] <evand> rgreening: that's known.  I'm halfway through bringing back disk block devices in support of formatting and disk image writing.
[16:23] <evand> rgreening: re format button> https://bugs.freedesktop.org/show_bug.cgi?id=23541
[16:26] <rgreening> right, was just wondering progress :)
[16:28] <rgreening> evand: did you see my other note about the target list having devices listed it should not?
[16:29] <rgreening> evand: oh, and what was it that was required to get the ISO image to work again? Rather than having to burn a CDR first?
[16:39] <CIA-33> usb-creator: rgreening * r155 usb-creator/ (debian/changelog usbcreator/frontends/kde/frontend.py): * Remove unused get_solid_drive() from kde frontend
[16:58] <davmor2> cjwatson: as soon as this last install finishe on my vista box I'm going to reinstall vista and try and help figure out why it's not showing up in grub
[18:05] <CIA-33> usb-creator: evand * r156 usb-creator/ (5 files in 5 dirs):
[18:05] <CIA-33> usb-creator: * Fully re-enable ISO image support (LP: #422671)
[18:05] <CIA-33> usb-creator: * Add some debugging breadcrumbs.
[18:24] <CIA-33> usb-creator: evand * r157 trunk/TODO: Flesh out the TODO item on PolicyKit support.
[18:26] <CIA-33> usb-creator: rgreening * r157 usb-creator/usbcreator/frontends/kde/frontend.py: * Update KDE frontend to support ISO images again
[18:27] <rgreening> hmm.. evand diverged... r157
[18:28] <evand> rgreening: eh? You've diverged?
[18:28] <rgreening> we both committed r157 at same moment
[18:28] <evand> bzr log lp:usb-creator shows r157 as what I just committed
[18:28] <evand> ah
[18:28] <rgreening> let me merge
[18:28] <rgreening> 1 sec
[18:28] <evand> if I may give a tip, use bzr bind whenever possible
[18:29] <rgreening> never used bind?
[18:29] <CIA-33> usb-creator: rgreening * r158 usb-creator/TODO: Attempt to merge divergence
[18:30] <evand> it will check with lp:usb-creator to make sure you're up to date before committing, then lock lp:usb-creator, and then commit
[18:30] <rgreening> ah. I run bzr commit then bzr push
[18:30] <rgreening> so, bzr bind then push instead?
[18:31] <superm1> you just bind once, then your push is combined with the commit action
[18:31] <rgreening> oh
[18:31] <rgreening> nice
[18:31]  * rgreening is still green with bzr goodiness
[18:33] <rgreening> ok, I've done the bind. should be good from now on then.
[18:34] <rgreening> ty for the hint
[18:58] <rgreening> evand: Ok, so, testing with ISO image works now. However, at end of install, I get /dev/loop0 and my usb stick /dev/sdc1 still mounted. Shouldn't these get released in back-end when finished with them?
[18:59] <rgreening> e.g. unmounted...
[20:14] <davmor2> cjwatson: right got vista on and ubuntu installed along side and the only options visible on grub2 are those of Ubuntu's
[21:07] <CIA-33> ubiquity: cjwatson * r3428 ubiquity/ (4 files in 3 dirs):
[21:07] <CIA-33> ubiquity: Add support for the biosgrub partman method, to be added in
[21:07] <CIA-33> ubiquity: partman-partitioning 71.
[21:08] <cjwatson> davmor2: can you mount your Vista partition on /mnt please, then 'ls -al /mnt/; ls -al /mnt/boot/'
[21:09] <davmor2> cjwatson: is this the system bit you want me to mount with the boot info in?
[21:10] <cjwatson> yeah
[21:10] <cjwatson> presumably it's an NTFS partition?
[21:15] <davmor2> cjwatson: http://pastebin.com/f7e0322de
[21:16] <cjwatson> hmm! that ought to work then
[21:17] <cjwatson> installer syslog? (pointer to somewhere you already uploaded this works too, if applicable ...)
[21:17] <davmor2> cjwatson: it's not going to be thrown by the fact that the partition is called system and there is another called vista is it?
[21:18] <cjwatson> it shouldn't care what the partition is called
[21:18] <cjwatson> it just goes through all the partitions it can find and sees what it can detect
[21:19] <davmor2> cjwatson: the weird thing is, I get the pause in grub2 just no option to install
[21:19] <cjwatson> syslog and maybe also /boot/grub/grub.cfg would be good
[21:19] <davmor2> http://pastebin.com/f710768a
[21:19] <davmor2> that is syslog
[21:19] <cjwatson> you caught me just as I was about to go and watch some Angel though ...
[21:20] <cjwatson> "Unknown post id, it may have expired or been deleted"
[21:20] <davmor2> from /var/log/install
[21:20] <cjwatson> (your pastebin url)
[21:20] <davmor2> http://pastebin.com/f7107681a
[21:20] <davmor2> I missed the 1
[21:22] <davmor2> grub.cfg is http://pastebin.com/f4b127946
[21:22] <cjwatson> hmph. I'm going to have to come back to this later, sorry :(
[21:24] <CIA-33> clock-setup: cjwatson * r209 ubuntu/ (debian/changelog finish-install.d/10clock-setup):
[21:24] <CIA-33> clock-setup: Pass --utc or --localtime option to hwclock as appropriate, as well as
[21:24] <CIA-33> clock-setup: --noadjfile (LP: #423247).
[21:24] <davmor2> cjwatson: no probs the vista install is on backup and I always have cds of ubuntu :)
[21:25] <cjwatson> yeah, I hate wasting your time though
[21:25] <cjwatson> but I think my wife might forget who I am soon, so ...
[21:25] <CIA-33> clock-setup: cjwatson * r210 ubuntu/debian/changelog: releasing version 0.98ubuntu2
[21:25] <davmor2> snap I'm off in a second
[22:07] <tormod> how do I unpack an initrd.lz? lzma -dc fails
[22:10] <tormod> so is the normal initrd gonna stay gzip and the CD one use lzma?
[22:57] <tormod> heh, I came up with poor man's more (that works) in the initrd: mas() { i=1; while read a; do sed -n $i,`expr $i + 10`p < $1 ; i=`expr $i + 11`; done }
[23:17] <cjwatson> tormod: err, lzma -dc ought to do it ...
[23:18] <cjwatson> tormod: yeah, lzma is too slow and memory-hungry for the normal initrd, but in the live CD case the trade-offs flip round because space is so much more important
[23:19] <cjwatson> as in, that space lets us fit in another language or two
[23:22] <tormod> I get: initrd.lz:  unknown suffix -- unchanged
[23:24] <tormod> ok, works: cat ../initrd.lz |lzma -dc|cpio -id
[23:24] <tormod> I thought it would work as gzip did.
[23:25] <cjwatson> oh, lzma doesn't like the .lz suffix
[23:25] <cjwatson> but .lzma, which is what it wants, would create problems when writing the CD contents to a FAT filesystem, which is something people do do from time to time
[23:25] <cjwatson> so yeah, compromise
[23:26] <cjwatson> when xz is a bit more widely deployed we may be able to switch to that
[23:26] <tormod> hmm with -S lz it would work
[23:26] <cjwatson> yeah
[23:26] <cjwatson> sorry, I did deal with this problem a while back but forgot ...
[23:27] <tormod> btw, is there a small editor with few deps that is convenient to drop into the initrd? sed -i gets tiring...
[23:28] <cjwatson> nano is the usual candidate
[23:29] <tormod> have you tried it?
[23:29] <cjwatson> it's the only editor provided in d-i; you might say that ...
[23:29] <cjwatson> I'm not its biggest fan but it's serviceable
[23:30] <tormod> just wondered if ncurses etc would work fine
[23:30] <cjwatson> you might need to copy in terminfo as well
[23:30] <cjwatson> but if you don't want to include curses, then about your only option is going to be ed
[23:31] <cjwatson> which might not necessarily be a bad choice, at least it's interactive ...
[23:31] <cjwatson> (just)
[23:31] <tormod> it's better than sed :)
[23:32] <tormod> btw, is the broken more (=cat ?) a feature or bug?
[23:34] <cjwatson> it might be worth enabling CONFIG_FEATURE_USE_TERMIOS in the initramfs
[23:34] <cjwatson> er, in busybox
[23:34] <cjwatson> feel free to file a bug