[00:40] <CIA-14> ubiquity: cjwatson * r2910 ubiquity/ (debian/changelog scripts/install.py): (log message trimmed)
[00:40] <CIA-14> ubiquity: * Fix excessive file-copy blacklisting, which had several problems
[00:40] <CIA-14> ubiquity:  (LP: #285238):
[00:40] <CIA-14> ubiquity:  - wasn't deciding which language packs to use before generating the
[00:40] <CIA-14> ubiquity:  blacklist;
[00:40] <CIA-14> ubiquity:  - was filtering out packages with prerms before doing ok-to-remove
[00:40] <CIA-14> ubiquity:  calculations;
[02:02] <cjwatson> evand: how are those bugs going?
[02:02] <cjwatson> I have to crash RSN ...
[02:03] <evand> working on them.  Had to take a call earlier.  I have a patch for grub-installer that I'm testing.
[02:04] <cjwatson> ok, I'm going to be away for most of the day tomorrow, but I can probably squeeze in a review first thing tomorrow if there's anything you're uncomfortable with just uploading
[02:04] <cjwatson> though it might be most expedient to upload and ask Steve to review
[02:05] <evand> ok, noted
[02:05] <cjwatson> good luck
[02:05] <evand> indeed, so long as I get another set of eyes on them
[02:05] <evand> thanks
[02:05] <evand> cheers
[04:14] <evand> Is it by design that the alternate CD syslinux does not have a countdown before it automatically boots?
[04:15] <evand> My concern is this: http://refit.sourceforge.net/doc/c4s3_keyboard.html
[04:16] <evand> Though admittedly not a great concern.
[04:23] <evand> hrm, USB disk d-i installs are failing in apt-setup / pkgsel
[04:48] <evand> Given that you cannot create symlinks on vfat, what should usb-creator do when it encounters one?  Recursively copy?
[05:29] <evand> Though that wont work for . -> ubuntu
[07:00] <evand> cjwatson: https://bugs.launchpad.net/ubuntu/+source/grub-installer/+bug/282037/comments/9 - I've tested that as best I could.
[07:01] <evand> I have not been able to get a proper d-i setup working yet, but I tested both state= paths on the desktop CD.
[08:39] <evand> cjwatson: see also (if you have time) https://bugs.edge.launchpad.net/ubuntu/+source/cdrom-detect/+bug/234185
[08:40] <evand> otherwise I'll just ask slangasek to look when he can.
[08:40] <evand> as suggested
[09:20] <evand> ok, so the problem is in apt-setup as neither /hd-media is present nor is OVERRIDE_LEAVE_CD_MOUNTED set.
[09:25] <evand> ah, and pkgsel is equally angry at life.
[09:41] <cjwatson> evand: by design although I forget exactly why :-)
[09:41] <cjwatson> evand: d-i isn't supposed to have any of the OVERRIDE_* things set
[09:42] <cjwatson> they're basically for ubiquity
[09:42] <cjwatson> so shouldn't /cdrom just be mounted in a USB install?
[09:42] <cjwatson> /hd-media is only if you're using iso-scan
[09:44] <cjwatson> grub-installer looks reasonable
[09:49] <cjwatson> evand: the only thing I notice about cdrom-detect is that it will fail to check .disk/info in the path where you had to select a CD-ROM driver module manually (rare though that is)
[09:50] <cjwatson> evand: it might be worth leaving the code outside the while loop there with some kind of guard
[10:38] <evand>  /cdrom is mounted
[10:38] <evand> apt-setup tries to remount it with iso9660 RO
[10:39] <evand> I mentioned hd-media and OVERRIDE as they avoid that code path
[10:39] <evand> (among others needed to be avoided in that file)
[10:39] <cjwatson> ah
[10:39] <cjwatson> maybe check whether it's iso9660 to start with as well
[10:40] <cjwatson> I know it's getting into check soup but that's probably the simplest fix
[10:40] <evand> indeed, that's what I did in the first pass, but then I realized there were other areas in that file that needed to be skipped
[10:40] <cjwatson> (and make sure to update the bit further down)
[10:40] <cjwatson> only two bits
[10:40] <cjwatson> you can just set a flag or something
[10:40] <evand> and then made a USB_INSTALL variable set by a check against cdrom-detect/try-usb, but that seems wrong
[10:41] <evand> as usb-creator will set that key for both desktop and alternate
[10:41] <cjwatson> no, I wouldn't do that
[10:41] <evand> indeed
[10:41] <cjwatson> add a check to the first [ -d /hd-media ] block for whether /cdrom is currently mounted as iso9660
[10:41] <evand> ok
[10:41] <cjwatson> then, if it actually does the mount, set a flag
[10:42] <evand> sounds reasonable
[10:42] <cjwatson> in the other block further down, replace the [ ! -d /hd-media ] && ... bit with checking that flag instead
[10:42] <cjwatson> seems most straightforward
[10:42] <evand> that wont fix pkgsel, whatever it's breaking on, but that's something seperate to tackle when I wake back up.
[10:42] <evand> separate*
[10:43] <cjwatson> right, it does it approximately unconditionally for CDs
[10:43] <cjwatson> again, I suspect an is-it-mounted-as-iso9660 check would do the job for now
[10:43] <evand> does what?
[10:43] <evand> ah
[10:44] <cjwatson> since load-install-cd will always mount as iso9660, so unmounting anything that isn't iso9660 is clearly not going to work
[10:44] <cjwatson> there, a justification :)
[10:44] <evand> heh
[10:44] <cjwatson> it does it> pkgsel and umount, I mean
[10:44] <cjwatson> sorry, in a rush so telegraphic
[10:44] <cjwatson> stop
[10:44] <evand> hahah
[10:44] <evand> no worries
[10:45] <evand> TheMuso_: before I forget, I do believe this will end up counting as a successful test of dm-raid.  Though I think I have a bug to file there with respect to resizing failing by trying to use the wrong device.