[00:26] superm1, I've not heard of any reports of that, and end up chatting with folks using those preinstalled images daily. How are you constructing the preinstalled? Are you using jasper? [00:27] persia, i'm reproducing it using a standard daily x86 image actually [00:27] i was asking ogra since that's the only place i know people are reliably testing oem-config (gtk) regularly [00:27] He may know more than I. Aren't there also folk in -testing that are running OEM installs as one of the test cases? [00:28] i tricked apport into reporting it (https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/641478) [00:28] Ubuntu bug 641478 in ubiquity (Ubuntu) "debconf-communicate crashed with SIGABRT in __kernel_vsyscall() (affects: 1) (heat: 10)" [Medium,New] [00:28] * persia has a suspicion that the various hack-arounds in jasper differ from those in casper sufficiently to cloud the issue [00:28] well the big difference there to here is whether there is a user 'oem' [00:29] which might actually be playing into this potentially [00:30] For jasper-initiated things, I think ogra forced running as root to avoid having to clean up after the creation of an oem user. [00:31] i've got a snapshot in a VM up to this point, i can see if it reproduces with the user 'root' for oem-config instead of oem easily [00:35] it looks like running as root prevents the crash [00:35] that does lead to other problems though, like the sound applet being broken [00:35] And networking, and ... [00:36] yeah and all that fun stuff [00:36] On the other hand, it becomes obvious where the paths differ, so the bug becomes more traceable (and it's known that jasper images can't help) [00:37] removing the oem user shouldn't be too much trouble on these jasper images, is there a reason it can't be done? [00:39] ogra was in a hurry that day? [00:40] assuming this other issue can get cleaned up, it would probably improve the experience on the jasper images to switch to oem if possible [00:40] Really, there's N things wrong with jasper, even if you ignore the architecture, coding style, and all implementation details. I've had firm promises that it will receive a complete rewrite. [00:41] That said, I doubt ogra will want to change it unless it's required to avoid regression at this point, and if other tools cause that regression, he's likely to want them reverted rather than change jasper. [00:42] (but I'm guessing: I don't really have a transcript of ogra's stream-of-conciousness printing out next to my desk) === cyphermox_ is now known as cyphermox [03:01] persia: jasper jasper livecd-rootfs arm I need a smoke [03:02] persia: there you go (sorry ogra ;-) ) [03:03] hm, insomnia does odd things to my sense of humour apparently [09:10] cjwatson, Ah, so you have the teletype :) [09:44] superm1, i have seen it once when i worked around the issue with creating an oem user so that might prove your theory [09:45] cjwatson, i smoke in my office :) [09:46] (fridays are just insane for me sorry if i give a weird impression on these :) ) [19:11] ubiquity: superm1 * r4400 ubiquity/ (bin/oem-config-wrapper debian/changelog): [19:11] ubiquity: Prevent a crash of debconf-communicate when removing oem-config. [19:11] ubiquity: (LP: #641478) [23:41] ubiquity: superm1 * r4402 ubiquity/ (debian/changelog ubiquity/i18n.py): [23:41] ubiquity: Raise privileges when running the language page in only-installable [23:41] ubiquity: mode. (LP: #647792) [23:54] ubiquity: superm1 * r4403 ubiquity/debian/changelog: add bug number