[00:30] <CIA-3> debian-installer: ogra * r1272 ubuntu/debian/changelog: releasing version 20081029ubuntu95
[06:20] <Vadtec> how do you get the graphical installer for 9.10 into "expert" mode? ive tried expert and priority=low in the boot options, but it doesnt seem to work
[09:44] <ev> gah, thanks for that fix superm1.
[09:47] <ara> ev, morning
[09:47] <ev> good morning ara
[09:48] <ara> ev, do you happen to remember a bug in ubiquity that prevented people to change the destination of grub from the drop down?
[09:48] <ev> yes, one second
[09:48] <ev> bug 539204
[09:49] <ara> ev, thanks, there are a lot of duplicates. I will mark them as dup
[09:49] <ev> ara: cool, thanks
[09:50] <ev> if they're using a version of ubiquity newer than 2.2.6, then it's a new bug.
[09:51] <ev> but I suspect they're all just using old CDs
[10:23] <CIA-3> ubiquity: evand * r4039 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py):
[10:23] <CIA-3> ubiquity: Move the install window color setting code closer to when the window
[10:23] <CIA-3> ubiquity: is shown, so the colors don't get reset before then.
[11:16] <CIA-3> ubiquity: evand * r4040 ubiquity/ (debian/changelog ubiquity/components/ubi-language.py):
[11:16] <CIA-3> ubiquity: Fix alpha warning being shown all the time with recent changes to
[11:16] <CIA-3> ubiquity: the KDE frontend.
[11:39] <ev> shtylman: I've filed bug 556376 per our discussion last week.
[11:42] <shtylman> ev: k.. I will check that out ... any ideas on the matter?
[11:42] <shtylman> how does the gtk frontend do the event loop without eating 100% cpu?
[11:44] <ev> shtylman: it runs a gtk mainloop without doing anything fancy (just calls gtk.main)
[11:44] <shtylman> so the main loop runs in a separate thread?
[11:48] <ev> no, see ubiquity/frontend/gtk_ui.py:499
[11:48] <ev> if you're noticing the threads_init call at the top of gtk_ui, that's for webkit
[11:55] <cjwatson> any sane event loop (and surely Qt's is in this category, I'd be astonished if it weren't) only uses the CPU when an event is actually received
[11:55] <cjwatson> the alternative is called "busy-waiting" but I wouldn't expect Qt to do that
[11:59] <cjwatson> if it did, then every process on a KDE system would be sitting there spinning all the time; performance would go through the floor, and power consumption would go through the ceiling
[12:38] <ara> ev, cjwatson: is anybody having a look to bug 539463?
[12:49] <cjwatson> ara: in future could you file new reports rather than attaching information to existing ones, please?
[12:50] <cjwatson> the problem with attaching information to existing ones is that often the same symptom is caused by multiple different problems
[12:50] <ara> cjwatson, sure, will do
[12:50] <cjwatson> ara: indeed, your problem is definitely different from that of the original report
[12:51] <cjwatson> er
[12:51] <cjwatson> ara: so if you could file a new bug now for your problem, I'd appreciate it
[12:51] <ara> cjwatson, ok, I'll get back to you
[12:51] <CIA-3> tzsetup: evand * r515 tzsetup/ (debian/changelog tzsetup):
[12:51] <CIA-3> tzsetup: Use LC_ALL=C when running sed over the geoip data, otherwise it
[12:51] <CIA-3> tzsetup: fails on UTF-8 text.
[12:51] <ev> ^ thanks ara, noticed that from your debug log
[12:52] <cjwatson> ara: this also means you can use apport to attach all the log files :)
[12:53] <ara> cjwatson, does ubuntu-bug ubiquity attach the correct logs?
[12:53] <cjwatson> yes
[12:58] <ara> cjwatson, bug 556436
[13:19] <CIA-3> tzsetup: evand * r516 tzsetup/debian/changelog: releasing version 1:0.26ubuntu6
[13:25] <CIA-3> grub-installer: cjwatson * r853 ubuntu/ (debian/changelog grub-installer):
[13:25] <CIA-3> grub-installer: 'nosplash' doesn't work because it isn't passed through by user-params.
[13:25] <CIA-3> grub-installer: Use debian-installer/splash=false for this behaviour instead
[13:25] <CIA-3> grub-installer: (LP: #548954).
[13:26] <CIA-3> grub-installer: cjwatson * r854 ubuntu/debian/changelog: releasing version 1.49ubuntu11
[14:16] <ev> (14:15:44) kwwii: just wanted you to know that I am working on the time zone map this afternoon
[14:16] <ev> (14:15:54) kwwii: so it will be ready tonight or tomorrow morning
[14:16] <ev> ^ for what it's worth
[14:33] <rgreening> hey ev
[14:33] <ev> hi
[14:34] <rgreening> ev: wrt bug 484252, can we disable format button in UI if a partition is selected and enable if it's not a partition? Or is there something which would preclude this case
[14:35] <ev> rgreening: well, that would horrendously break if they had written just a filesystem to the disk
[14:35] <ev> rgreening: I'm going to try to get some design team consultation on this
[14:36] <rgreening> yeah, the "select" a partition/disk and expecting it to be the only thing formatted seems pretty reasonable ot expect in the way the UI works today.. but its not the behaviour we have programmed. so we either need to change the UI or the behaviour.
[14:37] <ev> which is why I suggested the button label being changed to "wipe disk" which is pretty clear, in my opinion
[14:38] <ev> as much as possible, I really don't want to expose partitioning to users, especially in a tool with such a simple set of use cases as usb-creator
[14:38] <persia> ev: Do you see any reason not to process bug #460298 as an SRU?
[14:39] <ev> persia: indeed, I pulled up the backports wiki page this morning and then got caught up in other things
[14:39] <rgreening> hey persia
[14:39] <persia> Cool.  It's currently being discussed in the ubuntu-ja team meeting, and I just wanted to make sure you knew about it (as the nomination hadn't been approved).
[14:39] <ev> I'd like to shove an entire new version of usb-creator in karmic, given that it's mostly bug fixes
[14:39] <persia> hey rgreening.
[14:40] <rgreening> ev: we'd have to revert the UDisks for karmic right?
[14:41] <ev> yeah, I failed to consider that
[14:41] <ev> we'd have to create a branch that did s/DeviceKit/UDisks/
[14:42] <rgreening> ya. that should work...
[14:42] <rgreening> either that or we simply backport individual fixes/patches based on specific commits
[14:43] <rgreening> I leave that for you to decide. just let me know if the kde UI needs changes in the process.
[14:48] <cr3> I have a preseed to install on /dev/sda, the problem is that when there's a card reader the desktop image detects it as /dev/sdb but the alternate image detects it as /dev/sda
[14:49] <cjwatson> you need to make it deal with it dynamically
[14:49] <cjwatson> partman/early_command may be helpful
[14:51] <cr3> cjwatson: if I understand correctly, from the early_command script, I need to preseed the value of partman-auto/disk. if so, what's the command to set a d-i preseed value?
[14:59] <cjwatson> cr3: debconf-set
[14:59] <cjwatson> cr3: or maybe better, echo 'd-i ...' | debconf-set-selections
[15:00] <cjwatson> oh, wait, damn, I bet that doesn't work in that context
[15:00] <cjwatson> sorry, doing three things at once
[15:00] <cjwatson> . /usr/share/debconf/confmodule; ...; db_set partman-auto/disk ...
[15:00] <cjwatson> more likely to work
[15:04] <davmor2> ev: what happened to the web pic on the slides again?
[15:05] <cr3> cjwatson: thanks, I'm glad I asked :)
[15:05] <CIA-3> usb-creator: evand * r300 usb-creator/ (5 files in 4 dirs):
[15:05] <CIA-3> usb-creator: Change the format button's label to 'Erase Disk' as it wipes the
[15:05] <CIA-3> usb-creator: entire disk, rather than an individual partition.
[15:05] <ev> davmor2: eh?
[15:06] <davmor2> ev: on this mornings iso on wubi install in i386 there is no graphic on the web slide just an image marker
[15:06] <CIA-3> usb-creator: evand * r301 usb-creator/debian/changelog: Add LP bug reference to previous commit.
[15:07] <ev> davmor2: what distribution?
[15:07] <davmor2> Ubuntu
[15:07] <davmor2> 10.04
[15:07] <ev> not netbook edition or anything crazy like that?
[15:07] <davmor2> no
[15:08] <ev> very odd, I definitely didn't notice that with today's CDs
[15:09] <davmor2> ev: I'll do a side-by-side install after and double check for you
[15:09] <ev> thanks
[15:10] <superm1> i noticed the same thing on a une disk yesterday
[15:11] <ev> weird, it just points at /usr/share/pixmaps/firefox.png
[15:12] <superm1> er does it? <img class="icon" src="/usr/lib/firefox-3.6/icons/mozicon128.png" alt="" />
[15:12] <ev> ah, I'm looking at trunk
[15:12] <ev> right, I see why this is failing then :)
[15:13] <davmor2> :D
[15:14] <ev> davmor2: bug 554976 covers this as a UI freeze exception
[15:15] <superm1> ev, some time a while back we discussed some usability improvements to usb-creator to not be offering the full disk device when using an ISO.  any chance you think you're still going to make lucid with that?
[15:16] <ev> superm1: indeed, I have our discussion saved as a document, but it's not going to make lucid.  I've disabled the format warning entirely anticipating this would happen.
[15:16] <superm1> ah too bad :(
[15:22] <ev> yeah, definitely.  I'll endeavor to get a work item for fixing that in Maverick.
[15:23] <superm1> ok. one more question on usb-creator then.  are the windows builds actually happening?  in any sort of discoverable way?
[15:25] <ev> manually
[15:25] <ev> I have a TODO item for making sure the windows frontend is working, but it's growing cobwebs
[15:27] <superm1> when you get around to checking it, will the builds get published somewhere easy to find?
[15:28] <ev> well, they currently go in http://people.canonical.com/~evand/usb-creator/lucid/ - is that not discoverable?
[15:28] <ev> or do you mean for end users?
[15:31] <superm1> yeah i mean end users
[15:31] <davmor2> ev: why does resizing partition not move off 0% until it's completed?
[15:32] <superm1> i had no idea on that URL myself either and was having a hard time w/ google, i'll bookmark that
[15:35] <cjwatson> davmor2: because the underlying tool doesn't provide progress feedback
[15:35] <cjwatson> or at least not in a form that we can easily deal with
[15:36] <davmor2> cjwatson: would it not be better to just display the text "this might take a while" and remove the bar?
[15:40] <cjwatson> I'd rather fix it to deal with the resize feedback somehow; but neither answer is actually trivial
[15:40] <cjwatson> there is *no such thing* in the debconf protocol as an indeterminate progress bar
[15:40] <cjwatson> so right now it is not actually possible to remove the bar in any sensible way
[15:49] <davmor2> ev, cjwatson: https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/556373  have you guys seen this?
[15:49] <ev> yes, I changed the title
[15:49] <ev> haven't gotten to it yet though
[15:50] <ev> trying to figure out why the kde frontend is having difficulties on the partitioning page
[15:51] <davmor2> ev: cool
[16:00] <shtylman> ev: is the partitioning page returning the right result?
[16:01] <ev> shtylman: can you elaborate on what you mean?  The bug I'm investigating appears to be fallout from r4023
[16:03]  * shtylman tackes a look
[16:04] <ev> ah, think I've got it
[16:04] <shtylman> what was the problem?
[16:20] <ev> shtylman: missing definitions
[16:23] <CIA-3> ubiquity: evand * r4041 ubiquity/ (debian/changelog ubiquity/frontend/kde_ui.py):
[16:23] <CIA-3> ubiquity: Define allowed_change_step in the KDE frontend to fix the manual
[16:23] <CIA-3> ubiquity: partitioner buttons not working (LP: #556180).
[16:28] <cjwatson> hah, good catch
[16:28] <cjwatson> sorry about that, I was unwisely assuming some degree of parity ...
[16:49] <ev> quite okay, I had assumed the same parity which ate up quite a bit of time while I wasn't checking kde_ui.py :)
[16:52] <CIA-3> ubiquity: evand * r4043 ubiquity/ (debian/changelog ubiquity/components/ubi-language.py):
[16:52] <CIA-3> ubiquity: Do not show the 'ready to install' or alpha warning text in the KDE
[16:52] <CIA-3> ubiquity: frontend when in oem-config mode (LP: #556577).
[17:23] <CIA-3> ubiquity: evand * r4044 ubiquity/ (10 files in 2 dirs):
[17:23] <CIA-3> ubiquity: Update timezone images to cover changes to the Russian timezones and
[17:23] <CIA-3> ubiquity: make further adjustments around UTC+5, UTC+5.5, and UTC+7
[17:23] <CIA-3> ubiquity: (LP: #368060).
[17:29] <CIA-3> ubiquity: evand * r4045 ubiquity/ (d-i/manifest debian/changelog):
[17:29] <CIA-3> ubiquity: Automatic update of included source packages: grub-installer
[17:29] <CIA-3> ubiquity: 1.49ubuntu11.
[17:45] <CIA-3> ubiquity: evand * r4046 ubiquity/debian/changelog: releasing version 2.2.15
[18:43] <cjwatson> ev: I've sent a message to ~amichai2 regarding his ubiquity branch
[18:45] <cjwatson> copyright assignment and such
[20:15] <bipolar> cjwatson: thanks for adding the crypt passphrase to preseeding!
[20:16] <cjwatson> that's ok
[21:13] <shtylman> cjwatson: my only concern about lp:~amichai2/ubiquity/fixes is that the progress bar width is unbounded, which means the progress bar might expand/contract with different text ... which looks weird from a ui perspective
[21:13] <shtylman> other than that.. it seem ok
[21:14] <shtylman> the ui can be tweaked later if needed
[21:25] <NCommander> cjwatson: ping, are you planning another d-i upload before B2?
[21:48] <cr3> not that I'm complaining or anything, but is there any particular reason why the alternate image doesn't log any disk information (as found in media/.disk) anywhere under /var/log/installer?
[21:49] <cr3> I mean, .disk/info
[22:06] <cjwatson> NCommander: not currently
[22:07] <NCommander> cjwatson: alright, then I don't need to rush to fix the Dove netboot files as long as we'll have an upload for RC :-)
[22:07] <cjwatson> cr3: I thought it was saved in media-info - if it's not it's a bug
[22:07] <cjwatson> NCommander: certainly will
[22:07] <NCommander> cjwatson: thanks
[22:08] <cr3> cjwatson: I just finished an alternate install and this is the content of /var/log/installer: cdebconf/ hardware-summary initial-status.gz lsb-release partman status syslog
[22:10] <cjwatson> bug on cdrom-detect with that syslog, please?  assign me
[22:12] <cr3> cjwatson: done, reported bug #556842
[22:15] <cr3> I wonder if the problem might be related to:
[22:15] <cr3> Apr  6 18:31:16 debconf: Adding [CHOICES] -> [cdrom-checker: Verify the cd contents, cdrom-detect: Detect CDROM devices and mount the CD, eject-udeb: ejects CDs from d-i menu, file-preseed: load debconf preseed file, iso-scan: Scan hard drives for an
[22:15] <cr3> Apr  6 18:31:16 debconf: --> INPUT medium anna/choose_modules
[22:15] <cr3> Apr  6 18:31:16 debconf: <-- 30 question skipped
[22:16]  * cr3 tries to figure it out
[22:17] <cjwatson> no
[22:18] <cjwatson> that just means "not bothering you with boring question about which installer modules you want"
[22:19] <cr3> cjwatson: is there a trick to interrupt the installer at various places to debug stuff from the console?
[22:21] <cr3> cjwatson: by the way, I'm an idiot: this is a network install, so it makes sense that the media info is not being logged. I added a comment to the bug, but should I mark it as invalid?
[22:25] <cjwatson> cr3: yes, if it's a network install then there's no media info (the mutant netinst-from-CD trick is just treating it as an apt archive)
[22:25] <cjwatson> cr3: alt-f2 or alt-f3 to get a console
[22:25] <cr3> cjwatson: right, but to interrupt the installation assuming a preseeded install
[22:26] <cjwatson> drop the priority
[22:26] <cjwatson> if you make it medium, it'll stop at each menu entry transition
[22:26] <cr3> gotcha, thanks
[23:10] <bladernr> Is there a way to tell the alternate installer to load the restricted drivers (e.g. for the Broadcom B43 wireless card)