[09:04] ev, good morning [09:05] ev, one question, what packages are installed if I select "Install 3rd party sw" during the installation? [09:08] * persia suspects ubuntu-restricted-addons [09:10] good morning ara [09:10] ara: persia is correct [09:10] ev, persia: OK, thanks both! [09:10] well, that and everything that jockey installs for the broadcom driver, if it finds the hardware [09:11] bcmwl-kernel-source, being the main one [09:11] ev, is there any documentation about this change? (i.e. a spec in the wiki) [09:12] just the giant installer redesign specification: https://blueprints.edge.launchpad.net/ubuntu/+spec/foundations-m-installer-redesign [09:12] ev, Does jockey only get invoked with install-3rd-party? [09:13] persia: correct [09:13] unless you use it in the live CD dekstop [09:13] desktop* [09:13] then it will copy the packages over to the installed system [09:13] packages> packages it needed to install [09:18] Just out of curiosity, from where did the name "Third Party Software" come? [09:22] mpt [09:22] Thanks :) [09:22] with support from Andrew and Amanda [09:23] Oh, I trust it's safe to say that, I just have reservations about calling it "external" from our repositories, and wanted to know who to chat with about emerging semantic concepts :) [09:24] ah, of course :) [09:24] he's your man then [09:24] Yep :) [09:41] * mpt hides [09:45] mpt, heh. Just something to add to the agenda next time we sit and chat. My experience is generally that I hear more good things from you than I have things to complain about :) [10:18] party !!! [10:19] party can never be bad :) [10:20] and external means that its not you who has to clean up afterwards ... thats a good thing ! :) [11:54] ev, so i notice that it takes about 20sec between oem-config-gtk vanishing and the aptdaemon UI piece showing up on omap4, during that time (and during the whole aptdaemon run) the mouse cursor doesnt show the busy corsor, is there a possiblity we can enable that for this part as well ? [11:59] hmm [12:00] * ogra wont file a wishlist bug if its tecnically complicated, thats why i'm asking [12:12] ogra: does this fix it? http://paste.ubuntu.com/498397/ [12:16] ev, oh, that requires some tinkering and a new install, will take some time [12:17] * ogra is just doing tests on the installed system [13:29] base-installer: cjwatson * r403 ubuntu/ (debian/changelog library.sh): [13:29] base-installer: * Backport from trunk: [13:29] base-installer: - Set Dir::Media::MountPath to /media/cdrom as well as [13:29] base-installer: Acquire::cdrom::mount, as otherwise apt doesn't consistently read from [13:29] base-installer: the right one. [13:30] base-installer: cjwatson * r404 ubuntu/debian/changelog: releasing version 1.107ubuntu2 [13:46] ev, hmm, if we set the hostname during oem-config, we should actually restart sevices that make use of it (syslog and avahi-daemon at least) [13:51] i still see the old hostname in the logs and can only connect to .local after manually restarting avahi [16:03] ev: are you planning on uploading ubiquity-slideshow-kubuntu at some point? [16:04] ah, I hadn't seen that you made changes [16:04] and that I forgot to debcommit -r [16:04] yes [16:04] after our team meeting [16:05] sounds like I got in just in time :) [16:12] :) [16:14] WIN [16:14] ev, so testing your busy cursor patch from above makes aptdaemon not start at all ... [16:15] lucid wubi install, upgrade grub-pc to maverick, IT STILL BOOTS [16:15] * cjwatson tries a full dist-upgrade [16:16] well whoops [16:16] cjwatson: so I'm quite keen to pick your brain on how we might automatically test Wubi [16:16] i see a busy cursor for a second though :) [16:16] before it switches to gdm [16:16] ogra: it's probably some variation on that, I've just clearly failed in my copy and paste madness [16:17] heh [16:17] ev: wow, er, not sure how to handle the Windows side of it and the multiple reboots ... [16:17] cjwatson: exactly [16:17] especially given the already complex interaction with grub [16:17] ev, but its not that important, more a cosmetical thing, what i find more important are the syslog and avahi issues i pointed out above [16:17] yes, my main concern is identifying grub failing as an actual failure and not wedging the test machine [16:18] esepcially avahi ... [16:18] I think I can roughly accomplish it using a machine that supports boot on power and a timeout, but that seems a bit dirty [16:18] (assuming we can cheat a little and change menu defaults) [16:20] I can't think of any other way to do it [16:20] I mean, if grub fails, it's not like you can run code [16:21] yeah, indeed [16:21] hm, maybe kvm-autotest? But that would mean no hope of doing it on real hardware [16:21] something to ponder [16:21] I'm hoping for a bit less of this provided nobody asks for this spare laptop back [16:22] (this> wubi grub failures for extended periods of time in future) [16:22] cjwatson: also, do you think it's worth it to try to suck out enough individual bits of dependencies of d-i components to test the interactions between them and ubiquity plugins, or should I just go full bore into a chroot? [16:22] suck out in what way? [16:23] The latter seems slow and unreliable, but possibly the best option [16:23] as in everything console-setup needs to run, for example [16:23] I guess some dh hackery built on top of the existing build, I haven't thought of the implementation much [16:24] hmm [16:24] I suspect the latter will involve less time spent swearing at build systems [16:24] unreliable> as in if the archive is in an inconsistent state, though I did have an idea to leverage an ISO [16:24] indeed [16:25] it would be great if we had a reliable userspace union mount [16:25] then we could mount the ISO, copy on write with the union mount, then copy over the bits for the test and run [16:25] which would be much, much quicker [16:27] doesnt unionfs-fuse work for that ? [16:28] oh yeah, hrm [16:28] will have to toy [16:28] "quicker" may not be the right adverb for unionfs-fuse [16:28] heh [16:29] indeed [16:30] at least not the "much, much" part :) [18:15] ubiquity: evand * r4391 trunk/ (3 files in 3 dirs): [18:15] ubiquity: Correctly use the plugin_translate hook in the KDE frontend [18:15] ubiquity: (LP: #645092). [18:18] ev, I tried to create a 10.04.1 netbook usb disk using usb-creator (on Maverick) and it does not boot on my mini9 [18:18] ev, bug against usb-creator? known one= [18:18] ? [18:22] ara: is there a 'ui' keyword in your /syslinux/syslinux.cfg file? [18:25] ev, no [18:30] I want to say "change the part that reads 'gfxboot' to 'ui gfxboot'", but I seem to recall it being slightly more complicated than that [18:30] otherwise superm1 would've fixed that with his commit [18:30] I suspect incompatible syslinux binaries [18:30] but my brain is somewhat melted from the day [18:31] ev, OK, I'll call it a day as well [18:31] ev, will file the bug tomorrow [18:32] good night! [18:32] g'night! [18:32] it's incompatibilities for syslinux binaries afaik [18:32] ah ha! [18:32] so yeah, not fixable for Maverick [18:32] yeah don't think so [18:36] Sounds like release notes material. [18:37] indeed [18:37] usb-creator Maverick, so cutting edge that it doesn't look back (because it breaks if you try) [18:38] well it might be a worthwhile solution to provide an error in the 10.10 usb creator that it's known not to work with earlier media or something when selected [18:40] added to the release notes [18:40] and indeed [19:50] hey guys, got a d-i question [19:51] cjwatson, you are probably the best guy for that :) [19:51] i am trying to preseed apt-setup from an early_command using debconf-set, and it does not work [19:52] from the shell, if I do, say, "debconf-get apt-setup/local0/key", it returns nothing [19:53] but I did set it using "debconf-set apt-setup/local0/key ..." in the preseed/early_command [19:53] am i missing something? [19:53] run the entire installation with DEBCONF_DEBUG=developer and send me the log [19:53] cjwatson, doing that! [19:53] cjwatson, thanks, must be late in your place [19:54] I bet you forgot to register the question. there are smarter tools but I'll need to check your log [19:54] also send me your preseed file [19:55] cjwatson, sure thing [19:55] cjwatson, but ... what is this "registering the question" thingy ... [19:55] should I declare the value outside of my early_command too? [19:55] that just might be my mistake [19:56] I'll explain when I've looked at the log [20:02] cjwatson, you need only /var/log/syslog? [20:02] (and the preseed) [20:04] correct [20:05] cjwatson, in your inbox [20:06] cjwatson, just fyi, what I am trying to do is to have a location-aware preseed [20:06] based on the location, the mirror is at well-known, such as 10.x.10.10 [20:06] where x is variable [20:07] so I look at the default route, figure out x, and then preseed apt-set/local0/... accordingly [20:07] if there is a better way to do that, I am all ears! [21:27] cjwatson, poke, have you had time to look at what we discussed earlier? (I know it is very late for you now, so it can wait until tomorrow if you did not) [21:28] it's evening, no [21:28] mostly been putting kids to bed [21:38] cjwatson, ah, the routine! I have been away from home for three weeks, looking forward to do some of that.