[02:18] ev: I am now a kubuntu member :) [06:15] cjwatson: Ubuntu Server dailies failed to build for the last day or two [06:16] cjwatson: also, using the last daily I successfully have, there's an issue with the iscsi bits in d-i or partman [06:18] "This menu allows you to configure iSCSI volumes.\n iSCSI configuration actions\n Log into iSCSI targets\n Finish\n " [06:19] but none of these will let me get past this screen [09:29] kirkland: that's just bug 546929 - today's CD built successfully a few minutes after you said that, and should be happier [09:29] Launchpad bug 546929 in linux "most PATA/SATA modules missing in Lucid netboot" [Critical,Fix released] https://launchpad.net/bugs/546929 [09:44] installation-guide: cjwatson * r473 ubuntu/debian/changelog: releasing version 20081208ubuntu7 [09:48] heads up: freeze exception request for yesterday's changes to the slideshow and installer window is bug 551503 [09:48] Launchpad bug 551503 in ubiquity-slideshow-ubuntu "[FFe] Add LTS designator, make smaller, and new Ubuntu One icon." [Undecided,New] https://launchpad.net/bugs/551503 [09:59] usb-creator: evand * r286 usb-creator/debian/changelog: releasing version 0.2.19 [10:00] usb-creator: evand * r287 usb-creator/ (3 files in 3 dirs): [10:00] usb-creator: Fix format failing due to the device being busy (because we were [10:00] usb-creator: mounting it unnecessarily). [10:23] base-installer: cjwatson * r398 ubuntu/debian/changelog: releasing version 1.103ubuntu6 [11:34] cjwatson: do you recall why we default to "do not use" instead of the same filesystem as before without format set? I thought this was a change someone made around hardy, but it seems to go back as far as partman-target r1. [11:35] oh, maybe I'm reading this wrong [11:40] need to know where to mount it ... [11:40] and I think same filesystem as before without format set throws an error complaining about that [11:40] of course how to present this in the ubiquity UI is perhaps a different matter - this is pretty much a raw mapping from d-i [11:41] well, it throws a warning saying that it's going to remove system files [11:41] but surely we could find the mountpoint from fstab if set by UUID [11:42] (mind you, I'm not suggesting we do this in Lucid :) ) [11:42] no, we shouldn't be mounting things automagically [11:42] (which fstab, for starters?) [11:42] ah, chicken and egg [11:42] partman-basicfilesystems (56ubuntu4) hardy; urgency=low [11:42] * Disable automounting unless partman/automount is preseeded to true. This [11:42] makes LP #106209 much less likely to occur, since future installations [11:42] Launchpad bug 106209 in partman-basicfilesystems "fsck Unable to resolve UUID" [High,Fix released] https://launchpad.net/bugs/106209 [11:42] are less likely to format a partition whose UUID we have in /etc/fstab. [11:42] -- Colin Watson Wed, 09 Apr 2008 08:18:47 +0100 [11:43] there were other reasons to disable automounting, such as issues with filesystems from hibernated OS instances [11:43] well, at least my memory for dates isn't failing [11:44] indeed [11:44] ubiquity's UI is awkward though - it would be nice if it automatically switched "do not use" to the correct filesystem type if you specify a mount point [11:44] or something [11:44] yeah, that would definitely be a step in the right direction [12:20] ubiquity: evand * r4010 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): Translate the skip button in the new install window (LP: #551554). [12:54] cjwatson: thanks; testing now [12:55] it's working for me in kvm [12:55] damn, vgabios doesn't support DDC [12:55] I'm going to have to do this whole video mode detection thing either by zen, or by reboting [12:55] *rebooting [12:55] or I suppose I could rebuild vgabios but that's going to get into yak-shaving pretty quickly ... [12:57] cjwatson: i don't know what DDC is ... is that a feature, or something that's broken with our vgabios [12:58] it's a feature which doesn't normally make sense under emulation :-) [12:58] cjwatson: okay, I'm past the previous iscsi question; but my old preseed file isn't getting me past partitioning automatically any more [12:58] what question is it stopping at? [12:59] http://pastebin.ubuntu.com/406485/ [12:59] "If you choose guided partitioning for an entire disk, you will next be asked which disk should be used" [12:59] screenshot? [13:00] i'm on real hardware, camera phone screenshot okay? [13:00] weird, it's calling the disk sdb [13:00] never really did that before [13:00] my preseed hardcodes sda [13:00] camera phone is fine [13:00] i can change that easily enough [13:01] but yeah, disk name change would account for that [13:01] you can put a script in partman/early_command if you need to compute that dynamically [13:04] http://people.canonical.com/~kirkland/IMG_3427.JPG [13:06] right, for some reason /dev/sda must not exist, I have no idea why [13:06] you'll probably also need to add: [13:06] d-i partman/confirm_nooverwrite boolean true [13:07] to avoid problems later in some configurations [13:08] cjwatson: okay, i'll add that now [13:09] cjwatson: s/sda/sdb/ solved my immediate problem, while i go hack a better solution [13:20] arg, yeah, bummer... I have two slightly different machines i need to preseed 5 different installs against, and one is seeing its disk as sda and the other as sdb [13:22] kirkland: does any of them have more than one disk? [13:22] cjwatson: no, they don't [13:23] kirkland: then just leave out partman-auto/disk - you don't need to set it on a single-disk machine [13:23] cjwatson: i was just checking if i plugged a stray sd card in or something [13:23] no dice [13:23] cool [13:23] let me try that [13:23] if partman-auto/method is set and there's only one disk, then it will just autopartition the first disk, whatever that is [13:23] cjwatson: what about grub-installer/bootdev ? [13:24] you shouldn't normally need to set that either [13:24] * kirkland comments out those two and retests [13:25] you're already preseeding grub-installer/only_debian and grub-installer/with_other_os to true, which should mean that grub-installer/bootdev won't generally be asked [13:28] cjwatson: sweet, that works ;-) [13:31] ubiquity-slideshow-ubuntu: evand * r256 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 19 [14:02] ubiquity: evand * r4011 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): [14:02] ubiquity: Update the interface when the user presses one of the buttons on the [14:02] ubiquity: finished_dialog (LP: #551574). [14:02] ev: should bug 542310 be reassigned to console-setup and closed? [14:02] Launchpad bug 542310 in ubiquity "Corrupted layout/variant list on Step 3 in installer" [High,Fix committed] https://launchpad.net/bugs/542310 [14:04] cjwatson: done, thanks [14:05] ubiquity: evand * r4012 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): [14:05] ubiquity: Use the fancy new 'reboot required' panel notification [14:05] ubiquity: (LP: #551537). [14:06] so we have a few areas whereby one can trigger a crash by hitting next and then playing with the interface [14:07] this is possible on the console-setup page, for example, by pressing next and then clicking through the radio buttons [14:08] I'm tempted to guard against this by disabling any editable part of the interface when the user presses next, just like it already does with the back and next buttons [14:08] can we disable it without everything going to grey? [14:09] but to avoid that, which admittedly looks ugly and might not cover all cases of callbacks firing, perhaps we should just do hasattr checks in the callbacks [14:09] * ev wonders aloud if this could be written as a decorator that checks to make sure the dbfilter matches the ui component [14:09] that would be cute [14:09] s/callbacks/signals/ [14:10] I'll investigate that then [15:07] ah, this is actually really easy, given that we hold reference to the individual dbfilter on a Controller instance, which we set to None when we're done with it. [15:09] how's this: http://paste.ubuntu.com/406540/ [15:11] incidentally, this will also magically fix that scary AttributeError: 'NoneType' object has no attribute 'ok_handler' bug (bug 539827) [15:11] Launchpad bug 539827 in ubiquity "ubiquity crashes after clicking "try ubuntu"" [High,Triaged] https://launchpad.net/bugs/539827 [15:44] ev: it's certainly aesthetically pleasing; if it works, I say go for it [15:45] yay, I'll put it through its paces and then commit [15:45] thanks for the look-over [16:24] cjwatson: how can I tell what version of partman-base is being used for the lucid installer? I want to make sure I'm pulling the new one. [16:25] bipolar: isoinfo -lR -i foo.iso | grep partman-base_ [16:25] or you can check the .list file alongside the .iso on the web site [16:26] oh, you're netbooting aren't you? [16:26] cjwatson: hmm.. I'm using PXE to install [16:26] look at /var/lib/dpkg/status in your running image then, after the "downloading installer components" step [16:26] cjwatson: ok [16:27] 'more' sucks. :p [16:29] nano -v [16:29] is what I use [16:29] son of a.... [16:29] I didn't even know nano was in the installer... [16:30] 139ubuntu3 [16:31] I'm getting the 'No root file system' error. [16:31] I'm going to remove my preseeded crypto settings and select it by hand and see what it's doing. [16:39] ubiquity: evand * r4013 ubiquity/ (debian/changelog ubiquity/components/ubi-usersetup.py): [16:39] ubiquity: Move info_loop call from run to prepare so that it doesn't get [16:39] ubiquity: called with each debconf question asked in the user-setup component [16:39] ubiquity: (LP: #551742). [16:41] cjwatson: ok. this is strange. If I comment ALL partman*/* lines from my preseed, I get a window titled "[!!] partition disks" with one option "Configure iSCSI volumes", and "Undo changes to partitions" and "Finish ... write changes to disk". If I select "Undo changes to disk" the window disappears and I get just the blue background. If I select "Finish..." I get the error "No root file system". [16:41] bipolar: I'm in a meeting just now, but will get back to you [16:41] kk [16:41] I got one to go to as well. bbiab. [17:15] ubiquity: evand * r4014 ubiquity/ (6 files in 3 dirs): [17:15] ubiquity: Add only_this_page decorator to be used on functions that should [17:15] ubiquity: never be called once we've left their page. [17:29] back [17:35] usb-creator: evand * r288 usb-creator/debian/changelog: releasing version 0.2.20 [17:35] bipolar: so, I'll need to see log files (syslog and partman), preferably from an installation run with DEBCONF_DEBUG=developer on the kernel command line [17:35] usb-creator: evand * r289 usb-creator/ (debian/changelog usbcreator/install.py): [17:35] usb-creator: Change 'Flushing writes to the disk...' to 'Finishing...', as the [17:35] usb-creator: former is too technical. Thanks Iain Farrell! [17:37] cjwatson: does DEBCONF_DEBUG=developer mean the same as =5? thats what I have set now. [17:37] nevermind, I'll change it. it's not hard :P [17:43] =developer and =5 are equivalent, yes [17:43] cjwatson: I'm working on getting the logs... but here's something interesting. Running 'cat /proc/partitions' in the installer console returns no partitions. [17:45] bipolar: have you updated the pxe image you're using recently? [17:45] if not, you should ... [17:45] cjwatson: the only thing to update would be the initrd/kernel image. I'll do that. [17:45] the d-i netboot image was only updated last night for the fix for bug 546929 [17:45] bipolar: yes, that's the bit I mean [17:45] Launchpad bug 546929 in linux "most PATA/SATA modules missing in Lucid netboot" [Critical,Fix released] https://launchpad.net/bugs/546929 [17:45] I think I last updated them friday. [17:45] ah. too soon [17:45] update them, should go away [17:45] ok [17:46] will do [17:46] I should set up a cron job for this :P [17:55] cjwatson: It's asked me for my passphrase and is formating /. Thats a good sign. :) [18:17] ubiquity: evand * r4015 ubiquity/ (157 files in 3 dirs): Update translations from Launchpad. [18:18] ubiquity: evand * r4016 ubiquity/ (d-i/manifest debian/changelog): [18:18] ubiquity: Automatic update of included source packages: base-installer [18:18] ubiquity: 1.103ubuntu6, partman-base 139ubuntu3. [18:21] ubiquity: evand * r4017 ubiquity/debian/changelog: Freeze exception (LP: #551503). [18:23] ev: ping [18:23] rgreening: pong [18:23] how goes it? I see some changes to usb-creator.. seems the kde version is crashing atm... [18:24] ah, I haven't looked into the kde frontend in a bit. I've written that down as a task for tomorrow. [18:24] target = self.__backend.targets[udi] [18:24] KeyError: u'' [18:24] interesting [18:24] heh [18:25] If I get some time, Ill poke at it. right now it is useless. the start button remains greyed out cause of the crash [18:25] awesome (on the if you get some time, not that it's crashing) [18:25] lol [18:25] ya [18:25] :) [18:26] Im hoping to get to UDS again.. who knows. I was real busy with work last few months... [18:26] did you apply? [18:26] ya. [18:26] coolness [18:26] well, it is brussells.. mm chocolate [18:27] oh it's in Brussels, but it's not *in* Brussels [18:27] lol [18:28] from what I've heard, you'll need hiking equipment and a few days worth of food to make it into the city proper [18:28] yeah... [18:28] haha [18:28] but the hotel pics look AWESOME [18:29] haha [18:35] ubiquity: evand * r4018 ubiquity/debian/changelog: releasing version 2.2.11 [18:55] cjwatson: successful install is successful. Took a while for apt-cacher to download all the new stuff, but it's done and running now. [18:55] cjwatson: thanks! [18:57] bipolar: glad to hear it [19:18] cjwatson: do we still have the options for free software only on the live cd's now? [19:26] should do, don't remember removing those [20:29] ev: usb-creator-kde is foobar. dunno if its a combination of pyqt b0rk and/or changes to the backend. [21:24] ev: figured it out. QVariant() can be removed and anywhere I use toString() should be toPyObject() instead now. Changes in the PyQt bindings and how QVariant now works. I'll update later so we can push a new rev out which works. [21:24] apparantly it has been broken since recent pyqt update (4.7.1 possibly earlier) [21:47] usb-creator: rgreening * r290 trunk/ (debian/changelog usbcreator/frontends/kde/frontend.py): Update KDE frontend to work with PyQt in Lucid (QVariant changed). [21:49] ev: that should get us back to working for the moment. We will still need to review any changes which have been made to gtk frontend but not the kde one. I'll try and look at in next few days.