[00:34] how does one set OVERRIDE_ALLOW_PRESEEDING? [00:34] can you just pass it on the kernel command line? will it get picked up? [00:50] never mind, it works on the kernel command-line [11:22] cjwatson, ev, partman and migration assistant both die on omap ubiquity http://pastebin.com/Q5qUwvB9 any ideas ? [11:24] Bus error? Goodness knows [11:24] i was wondering about /target/var/cache/debconf/config.dat [11:24] thats something i can ignore ? [11:24] irrelevant [11:24] (it seems to try to access it before partman even started [11:25] ok [11:25] it's just stuff run on ubiquity exit that only works if the system actually managed to get installed [11:25] note that i only have 256M and use 50% compcache [11:25] but i dont see any OOM in dmesg or anywhere so i think it should be fine even with that value [11:25] probably irrelevant too. bus error suggests a bug in some C program, or even miscompilation [11:25] gah [11:26] why would you suspect OOM? Bus error doesn't mean that [11:26] because i get OOM if i run with the default of 25% compcache [11:26] it's a bad memory access of some kind [11:26] so first step is to find out what program is being killed with that signal [11:26] getting only ~300M [11:26] strace may help [11:26] follow the evidence, don't guess wildly :) [11:27] * ogra tries to ... i'm a bit desparate since even the kernel isnt 100% yet [11:27] so it can be anything really [11:28] SIGBUS could certainly easily be a kernel bug [11:28] [ 1094.317077] Unhandled fault: external abort on non-linefetch (0x1018) at 0x401f9000 [11:28] [ 1094.317169] Process 5295(parted_server) has RLIMIT_CORE set to 0 [11:28] [ 1094.317169] Aborting core [11:28] dmesg^^^ [11:29] setting RTLIMIT_CORE to 0 is desired ? [11:29] ok, so then try attaching gdb to it before it SIGBUSses, and find out where it falls over [11:29] RLIMIT_CORE is 0 by default [11:35] seems the fix for bug 498525 is missing in the omap kernel [11:35] Launchpad bug 498525 in linux "[lucid] breaks apport: core dumps get aborted even if core_pattern is a pipe" [High,Fix released] https://launchpad.net/bugs/498525 [11:35] (omap is .33) [11:50] partman-auto: cjwatson * r312 ubuntu/ (2 files in 2 dirs): Don't offer resize_use_free on armel (LP: #550963). [11:50] Launchpad bug 550963 in partman-auto "On armel platforms, side-by-side installation should not be a standard option due to lack of user interface at boot." [Low,Fix committed] https://launchpad.net/bugs/550963 [12:22] talking to dbus through fakeroot is proving to be quite the pain [12:22] (KDE explodes if it can't talk to the session bus) [12:42] partman-auto: cjwatson * r313 ubuntu/debian/changelog: releasing version 89ubuntu5 [15:00] ubiquity: evand * r3923 testing/tests/ (run run-frontend): [15:00] ubiquity: * Build osextras if necessary. [15:00] ubiquity: * Set SUDO_UID and SUDO_GID. [15:00] ubiquity: * Add a note on why the KDE and noninteractive tests are disabled. [16:21] ogra: are you going to implement these new omap installer bugs you filed? I know I'm not going to have time [16:21] cjwatson, for flash-kernel ? sure [16:21] thats why i assigned them to me [16:21] i just want to give slangasek something [16:22] since they are new features [16:22] partman-uboot [16:22] bug 559301 is unassigned at the moment [16:22] Launchpad bug 559301 in partman-uboot "partman-uboot needs to handle omap installs" [High,New] https://launchpad.net/bugs/559301 [16:22] NCommander was offering help here [16:22] and he created the package [16:22] I'll assign it to your team, and you can take it from there [16:23] so i assume we're fine here, i just dont want to assign it to hijm right now since i dont want to distract him from OO.o fixing [16:23] cjwatson, ubuntu-armel please [16:23] oh [16:23] * cjwatson quickly changes it [16:23] we try to keep canonical out of the bugs :) [16:24] well, I don't see why that's necessary, but OK [16:24] in case an enthusiastic community dev comes and wants to help :) [16:24] cjwatson, i guess persia could explain that to you :) [16:25] If nothing else, the team has had significant historical involvement from InDT [16:39] cjwatson: did you make any block alignment improvments in 10.04? [16:40] in the partitioner, you mean? [16:41] yep [16:41] yes, significant [16:42] we try to apply optimal alignment by default now [16:42] great, did anyone do any benchmarks? [16:42] under kvm etc [16:43] no [16:43] correctness first :) [16:43] anyway, I wouldn't expect kvm to show much of a difference [16:44] the differences would be on real disks that care about alignment [16:44] great, i see [17:33] hmm.. i see the installer makes /boot xvda1 under xen, but sda5 normally(when using lvm) [19:27] ubiquity: cjwatson * r4062 ubiquity/ (157 files in 3 dirs): Update translations from Launchpad. [19:29] ubiquity: cjwatson * r4063 ubiquity/ (d-i/manifest debian/changelog): [19:29] ubiquity: Automatic update of included source packages: partman-auto 89ubuntu5, [19:29] ubiquity: tzsetup 1:0.26ubuntu7. [19:45] ev: that tzsetup change - how were you going to handle the fact that busybox's wget doesn't have --timeout? I commented on this in bug 519756, which is the same bug [19:45] Launchpad bug 519756 in tzsetup "default geoip wget timeout is very long" [Medium,Triaged] https://launchpad.net/bugs/519756 [19:48] ubiquity: cjwatson * r4064 ubiquity/debian/changelog: releasing version 2.2.16 [21:53] partman-base: cjwatson * r202 ubuntu/ (debian/changelog parted_server.c): [21:53] partman-base: Ignore free space smaller than the grain size of the partition creation [21:53] partman-base: constraint, rather than only free space smaller than a cylinder.