[08:29] <ev> infinity: I am now
[09:27] <ev> https://launchpadlibrarian.net/80641817/buildlog_ubuntu-oneiric-amd64.ubiquity_2.7.36_FAILEDTOBUILD.txt.gz - only happens in a pbuilder / on the buildds, but what. on. earth.
[09:27] <ev> looking into it
[09:28] <ev> lots of a11y bugs were filed overnight, but I'm hoping Mackenzie knows of a way to provide access to widgets via ATK without knowing their textual form beforehand
[09:28] <ev> so that we don't break the testing stuff by fixing them
[09:37] <cjwatson> it's not yet another dbus failure is it?  I think we need to stub out dbus.SystemBus and all its kin a bit more comprehensively, rather than trying to whack-a-mole all the call sites
[09:48] <ev> might be. Was just looking into whether gtk_ui's custom excephook was blatting over the one in tests/test_gtkwidgets
[09:48] <ev> but good point
[10:48] <ev> ah ha! It's something else in the stack breaking these tests
[10:48] <ev> presumably GtkBuilder, which doesn't seem to be calling __init__ on any of the custom widgets
[10:51] <ev> we really, really need integration tests. If a new copy of GTK gets uploaded, I want ubiquity's test suite run immediately with it before it's allowed in the archive.
[11:05] <ev> pygobject 3 looks like the culprit here
[13:07] <Peanut> cjwatson: these disks are 2000398934016 bytes, 2000.4 GB
[13:08] <cjwatson> right, so that will be covered by the fix I just made to partman-partitioning upstream, but as I say I don't think there's an easy workaround
[13:08] <Peanut> 'just', as in how long ago?
[13:08] <cjwatson> 14:05 <CIA-3> cjwatson * r2cd9f3f partman-partitioning/ (debian/changelog lib/disk-label.sh): Don't force GPT on disks between 2TB (decimal) and 2TiB (binary) inclusive.
[13:08] <cjwatson> three minutes
[13:10] <cjwatson> the grub-installer patch in lucid I was referring to was   * Fix GPT detection to account for changed fdisk output (LP: #538100).
[13:10] <cjwatson> if you can show me a syslog of the failure then I can check whether that atches
[13:10] <cjwatson> *matches
[13:10] <Peanut> Ah yes, that's very quick, thank you - maybe I'll have to try backporting that into karmic64. But I'm also going to try to convince the cluster owner to get onto LTS instead of karmic64 - I'm fine with GPT, as long as I can create a bootable system in the end. I'm re-running the install now but creating the 7GB partition *again* takes a bit of time before I can show you the changelog.
[13:11] <cjwatson> sure
=== kentb-out is now known as kentb
[13:22] <stgraber> ev: did I miss something or is bug 856669 not supposed to be possible? (.label is set in __init__) reminds me of that weird manual partitioning bug in the Qt frontend that was pretty similar (and that's probably still around)
[13:22] <ubot2> Launchpad bug 856669 in ubiquity "ubiquity crashed with AttributeError in do_set_property(): 'StateBox' object has no attribute 'label'" [Undecided,Confirmed] https://launchpad.net/bugs/856669
[13:22] <ev> stgraber: it's a bug in pygobject
[13:23] <stgraber> oh, that'd explain it then :)
[13:24] <ev> stgraber: http://paste.ubuntu.com/695628/ are my thoughts so far
[13:26] <Peanut> cjwatson: 'Unable to install GRUB in /dev/sda, executing 'grub-install /dev/sda' failed. This is a fatal error."
[13:27] <Peanut> "grub-installer: grub-setup: error: Embedding is not possible, bu
[13:27] <Peanut> t this is required when the root device is on a RAID array or LVM volume."
[13:29] <cjwatson> Peanut: ah, you probably don't have a BIOS Boot Partition created if you have a pre-existing DOS-based preseed file, OK
[13:29] <stgraber> ev: fun... I guess you already poked desktop/pitti to get that fixed?
[13:29] <cjwatson> tedious
[13:29] <ev> stgraber: indeed!
[13:30] <Peanut> cjwatson: does your anwser imply that this is something I can fix in the preseed recipe?
[13:31] <cjwatson> Peanut: probably, but it might be easier to apply a pickaxe
[13:31] <cjwatson> d-i partman/early_command string sed -i 's/2TB/4TB/' /lib/partman/lib/disk-label.sh
[13:31] <cjwatson> something like that (and I apologise for that being incredibly gross)
[13:31] <cjwatson> the 4TB number is fairly arbitrary, it's just bigger
[13:32] <cjwatson> once you have disks that are actually bigger than 2TiB then you probably will really need to switch to GPT
[13:32] <cjwatson> but you don't have to deal with that problem right now
[13:33] <Peanut> Oh, that is a nice pickaxe - but then that means I have to have to create our own partman package, as the machine-to-be-installed gets it as a package from the install server (who is proxying it from an Ubuntu installer).
[13:33] <cjwatson> no it doesn't
[13:33] <cjwatson> that's a line that can go in your preseed file
[13:34] <cjwatson> it hacks the code on the fly
[13:34] <Peanut> ooh, shiny
[13:40] <ev> stgraber: I've updated bug 856669 accordingly
[13:40] <ubot2> Launchpad bug 856669 in ubiquity "ubiquity crashed with AttributeError in do_set_property(): 'StateBox' object has no attribute 'label'" [Undecided,Confirmed] https://launchpad.net/bugs/856669
[13:44] <stgraber> ev: thanks
[14:05] <Peanut> cjwatson: Thanks very much for your help, the install just completed.
[14:07] <cjwatson> great
[14:31] <CIA-45> partman-base: cjwatson * r1353 ubuntu/ (debian/changelog lib/base.sh parted_devices.c):
[14:31] <CIA-45> partman-base: Handle /dev/zram* the same way as /dev/ramzswap* (exclude from
[14:31] <CIA-45> partman-base: partitioning, and don't disable swap devices there).
[14:33] <CIA-45> partman-base: cjwatson * r1354 ubuntu/debian/changelog: releasing version 151ubuntu2
[16:40]  * ev wonders aloud if tests or performance tests using complete plugins (so d-i module side and all) would be worthwhile and whether it would be better to do a sed to correct the paths or just unpack the built ubiquity and use fakechroot
[17:12] <infinity> ev: Have some time to discuss https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/820514 ?
[17:12] <ubot2> Ubuntu bug 820514 in ubiquity "oem-config-remove-gtk not found during preinstalled desktop initialization" [High,Confirmed]
[17:13] <infinity> ev: I apologise for my last comment being a bit on the grumpy side, I think I woke up on the wrong side of the release management bed yesterday. :P
[20:00] <CIA-45> grub-installer: cjwatson * r1211 ubuntu/ (debian/changelog grub-installer):
[20:00] <CIA-45> grub-installer: Simplify /proc and /sys mounting; make sure they're consistently mounted
[20:00] <CIA-45> grub-installer: for the entire life of grub-installer, and consistently unmounted on
[20:00] <CIA-45> grub-installer: exit (LP: #851704).
[20:25] <CIA-45> grub-installer: cjwatson * r1212 ubuntu/ (81 files in 3 dirs): merge from Debian 1.68
[20:34] <CIA-45> grub-installer: cjwatson * r1213 ubuntu/debian/changelog: releasing version 1.68ubuntu1
[20:44] <CIA-45> grub-installer: cjwatson * r1214 ubuntu/ (debian/changelog grub-installer): Really allow grub/grub2 choice for ext4 (lost in merge).
[20:45] <CIA-45> grub-installer: cjwatson * r1215 ubuntu/debian/changelog: releasing version 1.68ubuntu2
[20:46] <CIA-45> grub-installer: cjwatson * r1203 natty-proposed/ (3 files in 2 dirs):
[20:46] <CIA-45> grub-installer: Restore the ability to choose to install GRUB Legacy by preseeding
[20:46] <CIA-45> grub-installer: (grub-installer/grub2_instead_of_grub_legacy=false), removed in 1.60 as
[20:46] <CIA-45> grub-installer: a side-effect of removing a warning message. This is still useful e.g.
[20:46] <CIA-45> grub-installer: when installing inside Xen (LP: #857548).
[20:50] <CIA-45> grub-installer: cjwatson * r1204 natty-proposed/debian/changelog: releasing version 1.60ubuntu4
[22:13] <CIA-45> hw-detect: cjwatson * r1466 ubuntu/ (debian/changelog hw-detect.sh):
[22:13] <CIA-45> hw-detect: Load xenbus_probe_frontend if we're running under the Xen hypervisor
[22:13] <CIA-45> hw-detect: (LP: #857662).
[22:21] <CIA-45> hw-detect: cjwatson * r1467 ubuntu/hw-detect.sh: xenbus_probe_frontend probably needs to go in the initrd for xen-blkfront
[22:23] <CIA-45> hw-detect: cjwatson * r1468 ubuntu/debian/changelog: releasing version 1.81ubuntu3
[22:28] <CIA-45> hw-detect: cjwatson * r1466 natty-proposed/ (debian/changelog hw-detect.sh):
[22:28] <CIA-45> hw-detect: Load xenbus_probe_frontend if we're running under the Xen hypervisor
[22:28] <CIA-45> hw-detect: (LP: #857662).
[22:44] <CIA-45> hw-detect: cjwatson * r1467 natty-proposed/debian/changelog: releasing version 1.81ubuntu2.1
=== kentb is now known as kentb-out