=== Andre_Gondim-afk is now known as Andre_Gondim
[09:30] <davmor2> morning all
=== cking is now known as cking-afk
=== cking-afk is now known as cking
=== cking is now known as cking-afk
=== cking-afk is now known as cking
=== yofel_ is now known as yofel
=== cking is now known as cking-afk
=== Andre_Gondim is now known as Andre_Gondim-afk
=== McPeter is now known as Noob`s
[22:34] <persia> So, regarding SRU validation: There's a reporter in #ubuntu-bugs that says that upgrade for https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/508173 renders GPT hardware unbootable.
[22:34] <ubot4> Launchpad bug 508173 in grub2 (Ubuntu Lucid) (and 2 other projects) "postinst has errors with grub-probe that cause the system to stop booting (affects: 1) (heat: 16)" [Medium,Fix released]
[22:35] <persia> Could anyone remind me how we're supposed to respond?  This mostly affects Apple hardware, but there's bundles of that around.
[22:37] <sbeattie> persia: mark the bug as verification-failed, have the reporter open a new bug about the regression and refer to the regression bug report in a followup comment to 508173
[22:37] <persia> So, the reporter is having a bit of trouble opening a new bug using any automation.  Do we care?
[22:37] <persia> Should it be an unautomated bug?
[22:38] <persia> Also, do we know any regular testers with Apple hardware that could do verification testing?  I worry that we may want to take some significant action to avoid making Ubuntu not work there.
[22:39] <sbeattie> an automated bug would be preferred, but an unautomated bug is fine.
[22:40] <persia> OK, cool.  Unfortunately, this sort of hardware can't USB-boot, and the reporter doesn't have a blank CD about.
[22:40] <sbeattie> Bugger, looks like the update made it out of SRU, and into -updates?
[22:40] <persia> So there's no means to get back into Ubuntu until a shop can get visited, or there's some grub manipulation tool that can run inside OS X.
[22:41] <persia> Unfortunately.
[22:41] <persia> I'm not sure we test GPT heavily.
[22:42]  * sbeattie doesn't know what GPT is.
[22:43] <persia> It's "GUID Partition Table", used on many EFI-booting systems.
[22:43] <persia> EFI is the Extensible Firmware Interface.  GUID might be Globally Unique Identifier, but that doesn't make sense to me.
[22:44] <persia> Anyway, it's how Apple Intel machines get partitioned, rather than using a DOS disklabel.
[22:45] <persia> So testing would involve a (newer) Mac
[22:45] <persia> (but note that at least some newer Macs have issues getting installed, making this tricky)
[22:47] <sbeattie> Does our installer support EFI/GPT?
[22:48] <sbeattie> virtualbox-ose claims to support EFI booting, and the next version coming is supposedly going to support OSX installs.
[22:48] <sbeattie> (just wondering if virtualbox can be used as an accurate test environment)
[22:53] <persia> It kinda does, for some sorts of installs.
[22:54] <persia> At least for my hardware, default selections for desktop works and default selections for server fails.
[22:54] <persia> But I think mine differs from lots of stuff out there, or there would be more complaints, as default selections for desktop failed prior to Beta2 in lucid.
[22:55] <persia> (and my hardware is very new)
[23:10] <persia> sbeattie: I think I've gotten 508173 in the right state.  Could you confirm?
[23:13] <sbeattie> persia: yeah, looks okay. Probably best to get cjwatson looking at bug 576662
[23:13] <ubot4> Launchpad bug 576662 in grub2 (Ubuntu) "latest grub2 renders gpt based system unbootable (affects: 1)" [Undecided,New] https://launchpad.net/bugs/576662
[23:15] <persia> Bad time-of-day, and even time-of-year (elections/UDS), but yeah.
=== Noob`s is now known as McPeter