[01:23] partman-base: cjwatson * r146 ubuntu/ (debian/changelog init.d/parted): init.d/parted: Use more concise syntax for building up lists. [11:27] partman-ext3: cjwatson * r750 ubuntu/ (77 files in 7 dirs): merge from Debian 56 [11:29] evand, cjwatson: Did ubiquity shrink in size over night? [11:30] partman-ext3: cjwatson * r751 ubuntu/debian/changelog: note that ext4 is mounted with relatime too [11:30] davmor2: I noticed in yesterday's daily build that its window was smaller than before; I assume that this was due to some of Evan's changes in 1.12.0 [11:31] evand, cjwatson: also use weak password now says ubiquity/text/yes and ubiquity/text/no rather than yes no [11:31] cjwatson: can it not be automated so it's size is reduced on any screen size below 900? [11:32] davmor2: weak password> that's fixed in 1.12.1 [11:32] partman-ext3: cjwatson * r752 ubuntu/debian/changelog: releasing version 56ubuntu1 [11:32] davmor2: screen size hacks generally suck maintenance-wise [11:33] they make it difficult for us to tell in any reliable way what the interface is actually going to look like for end users [11:33] cjwatson: True hadn't thought of that :) [11:33] I do think the timezone map is perhaps a little too small now, mind you [11:34] depends on your POV :) [11:34] on my 800x600 screen on the babbage it cant be to small :) [11:34] we might want to weight its size up a bit so that GTK feels free to use more space for it if it's available [11:34] ogra: obviously [11:34] * cjwatson <- focused on fixing RC bugs thouygh [11:34] though [11:39] cjwatson: I like the retrieving file indicator on d-i :) [11:39] oh, the fix for bug 290234? [11:39] Launchpad bug 290234 in apt "Intrepid: Netboot locks up at 2% installing the selected edubuntu desktop" [High,Fix released] https://launchpad.net/bugs/290234 [11:39] it's nice to that the machine hasn't locked up :) [11:40] that's the one :) [11:40] excellent [11:40] will make a difference to tasksel too [11:43] * ogra sighs about dd really tearing down IO performance [11:44] cjwatson: I'll do a netboot after for now I'm doing standard alternate install :) [11:45] yeah, it'll make a much bigger difference to netboot [11:56] cjwatson, hmm, i see mdz's bug that offers to unmount the installer medium on the babbage [11:56] wasnt that supposed to be fixed ? [11:57] and the partman UI asks me where i want to place $(RELEASE) (instead of showing the actual release name) [11:58] ogra: no, I'm working on that bug right now actually [11:58] ah, k [11:58] ogra: $(RELEASE)> can you give me the exact string so that I can grep for it? [11:58] the screen size changes are awesome btw [11:59] hmm, i'm already advanced, lets see if i can cancel [11:59] ogra: installer medium> the only thing that was fixed from that set of bugs so far is that the button behaviour is now more saner [11:59] sane [11:59] ah [12:00] user-setup: cjwatson * r167 ubuntu/debian/ (62 files in 2 dirs): Update Ubuntu-specific strings from Launchpad. [12:00] hrm, we dont have a cancel button ... i wonder how to kill it [12:01] * ogra just resets the board [12:09] ga, this time it doesnt let me unmount anything [12:09] and drops me directly into a manual partitioning window [12:10] (i didnt have the target device unmounted in advance this time, had done that last time) [12:14] cjwatson, aha, seems to be a translation issue "Where do you want to put Ubuntu 9.04?" i used german before and there it said $(RELEASE) instead of Ubuntu 9.04 [12:15] * ogra tries a german once again to confirm ... [12:15] its fine in english [12:18] cjwatson, "Wo möchten Sie $(RELEASE) platzieren?" [12:27] cjwatson: you might know this how hard would it be to get the mini.iso's renamed? as in mini-i386.iso mini-amd64.iso [12:29] currently they are all called mini.iso [12:35] davmor2: I'd prefer not since there are all sorts of things that would have to be renamed if we went down that road. Just rename them on the client side [12:36] cjwatson: I do I just wondered about it is all :) [12:37] debian-installer: cjwatson * r1078 ubuntu/ (35 files in 2 dirs): Update help text translations from Launchpad. [12:37] user-setup: cjwatson * r168 ubuntu/user-setup-ask: /target -> $ROOT [12:38] ogra: oh, idiot translators. $(RELEASE) -> ${RELEASE} [12:38] ogra: I'll get that fixed using my magic rosetta-admins powers [12:39] cjwatson, the other thing is more worrying, but i want to finish my test install now to see the bootloader stuff first before filing a bug [12:39] user-setup: cjwatson * r169 ubuntu/debian/changelog: releasing version 1.23ubuntu16 [12:39] ogra: I'm not really worried about the partitioning bit since I'm completely rewriting that code anyway [12:40] well, it didnt unmount anything and presented an empty partitioning window [12:40] but your call [12:40] I wouldn't mind looking at logs, but chances are it's covered [12:41] ok [12:52] :-'( it crashed [12:53] cjwatson, i see "No bootloader installer found" again [12:55] lol ... ubiquity has NSLU2 support now ? heh [13:00] cjwatson, oh, archdetect still returns "armel/unknown" ... [13:00] indeed that cant work [13:02] ogra, odd, base-installer should have been updated to catch that [13:02] ogra, what does /proc/cpuinfo say? [13:02] the usual babbage stuff [13:03] Hrm ... [13:03] archdetect works against libdebian-installer to determine the subarchitecture [13:04] ogra, what version of libdebian-installer is installed? [13:05] none [13:05] ah [13:05] 0.61ubuntu1 [13:05] libdebian-installer4 is the package name [13:05] I would like to talk to someone about bug 349173 i believe the sense of the choice is backwards [13:05] Launchpad bug 349173 in ubiquity "weak-password dialog should be Yes/No" [Medium,Fix released] https://launchpad.net/bugs/349173 [13:05] ogra, imx51 support landed in 0.61ubuntu2 [13:06] yes, i see that [13:06] ogra, there's your problem :-) [13:06] well, i'll wait [13:06] ogra, you could just apt-get install libdebian-installer4 and get the new version :-) [13:07] if thats in the archive already [13:07] ogra, its uploaded [13:07] mind you, thats tonights build [13:07] Fri, 27 Mar 2009 11:54:26 +0000 [13:09] not there yet [13:09] huh? [13:09] already the newest version ... [13:10] oh [13:10] * NCommander rescores on ARM [13:10] There's your problem :-) [13:10] yeah, needs building :) [13:10] that was uploaded ages ago. How come it hadn't built yet? [13:10] ogra, ok, its bumped in the build queue [13:11] jtholmes: please explain? [13:11] busy buildd i guess [13:11] cjwatson, ARM had a bit of a buildd backlog, if I noticed it hadn't built, I would have rescored it ages ago. [13:11] s/had/has/g [13:11] arent we in archive rebuild as well atm ? [13:11] archive rebuilds happen on separate buildds [13:11] that might keep them busy [13:11] ogra, libdebian-installer should have taken presentant [13:11] ah [13:11] surely? [13:11] cjwatson, mass givebacks don't [13:11] sure, but that isn't a full rebuild [13:11] cjwatson, and armel been plowing through universe with a builder down. [13:12] cjwatson, when you have 600 FAILED/DEPWAIT? [13:12] still not a full rebuild, I'm not denying that it keeps it busy :) [13:12] cjwatson, I'm just curious when we're going to get the last arm builder back, we're down one, and its showing [13:12] send your babbage to infinity ;) [13:13] ogra, my babbage is going to FSL [13:13] I think the magic smoke blew the SD slot, because I can't get it to work reliably [13:13] oh, why is that ? [13:13] ogra, I had the magic smoke escape ;.; [13:13] ouch [13:13] Yeah [13:13] cjwatson, there really isnt a question in the dialog, there are basically two statements and then the Yes No choices No lets you change the password and Yes takes you to Ready to install dialog [13:13] i thought that was only the UART [13:13] jtholmes: does it not say "Use weak password?" in the title? [13:14] ogra, the UART works oddly enough, but I've yet to successful start a d-i image, and my SD cards seem to be OK [13:14] jtholmes: the sense is not backwards AFAICS, but it could probably do with clearer labelling [13:14] jtholmes: you're correct in ubiquity the text does look wrong I've gotten use to the text in d-i which says use weak password [13:14] ogra, the board starts, but then promptly fails with filesystem/SD card errors [13:15] cjwatson, the words 'use weak password' do not exist in that dialog [13:16] let me fire up a vm and have a look [13:16] libdebian-installer is now building, so it should be in the next publisher run :-) [13:16] (assuming it doesn't FTBFS ...) [13:16] cjwatson, ok thanks [13:16] cjwatson: the text in the dialogue for use weak password reads "You entered a password that consists of less than eight characters, which is considered too weak. You should choose a stronger password. (no) (yes) [13:18] the action Yes No are correct the Text is a little skewed [13:18] ogra: I mailed ubuntu-translators@ about the translation problem; similar things affect rather more languages than I'd hoped, for a variety of strings [13:18] Yes moves you forward No lets you change the password [13:18] that behavior is correct [13:19] cjwatson, anyway, w.r.t. to your question about archive rebuilds, they hwerer traditionally done on ubuntuwire, but they haven't been done since early intrepid as far as I know [13:19] jtholmes: are you installing Ubuntu or Kubuntu? [13:19] kubuntu [13:19] NCommander: aside from the way they were traditionally done in dak ... [13:19] jtholmes: it's good to mention that up-front rather than me wasting time with the wrong frontend :( [13:20] in the future will do [13:20] cjwatson, ow. Archive rebuilds in dak would be painful unless the Ubuntu dak had some features to help it [13:21] NCommander: infinity appears to cope fine *shrug* [13:21] jtholmes: here's how it looks in Ubuntu: http://people.ubuntu.com/~cjwatson/tmp/weak-password.png [13:22] jtholmes: can you show me a screenshot? [13:23] your paste of ubuntu is what i see in kubuntu [13:23] cjwatson, I know on tradition dak, there was a hard assumption that there would ALWAYS be source; I'm not saying its impossible, but it doesn't sound like my idea of fun :-/ [13:23] jtholmes: I don't believe that - the frontend is completely different [13:23] but there isnt a question 'do you want to use weak password' [13:23] jtholmes: "Use weak password?" should be in the window title bar [13:23] it is [13:24] right then [13:24] Use weak password is in the title bar [13:24] anyway, the button text should be changed (again) [13:25] d-i handles this sort of thing a bit differently - what we render as the window title in ubiquity is actually placed at the end of the text in d-i, immediately before the buttons [13:25] it would seem to me the dialog text should say 'Do you want to choose a stronger password' instead of 'You should choosea a stronger password' [13:26] I'd prefer to change the button text [13:26] sure to what [13:26] human interface guidelines generally say that buttons should describe actions, rather than simply being "yes" and "no" [13:27] agreed [13:27] putting "yes" and "no" in here was just a stopgap measure since we're past string freeze, which means that translations are going to be difficult to manage [13:27] but if I'm going to get confused people turning up thinking the sense is wrong, I probably have little choice [13:28] yes i see the problem [13:28] "Continue" is probably fine in place of "Yes", and perhaps "Choose another password" in place of "No" [13:29] that looks good to me or "Continue with weak password' [13:29] whatever [13:29] I would prefer not that on two grounds [13:29] (1) the buttons would end up very wide and I think it would look odd [13:29] (2) it's one more thing that translators would have to do at short notice [13:30] good catch that is why you are in your current position [13:30] hmm, unfortunately "Continue" is translated as roughly "forward" in at least some languages [13:30] well, hmm, "Weiter" could be understood either way I guess [13:31] in any case i will try to test more alpha releases in the future to try and ferret out these things before string freeze [13:31] thanks [13:31] but please in future try to keep the questions to a single channel; it's very distracting to find that three windows have lit up regarding a single topic [13:32] will do i was not aware of #ubuntu-installer channel [13:32] and in general bug reports belong on LP [13:33] i did add to to 349173 but was afraid that the late state of the release it would possibly get overlooked [13:34] comments on already-fixed bugs are in general liable to be ignored [13:34] ubiquity: cjwatson * r3157 ubiquity/ (debian/changelog partman/commit.d/01unmount_busy): [13:34] ubiquity: Add a comment to partman/commit.d/01unmount_busy to explain why this [13:34] ubiquity: script is still here despite similar code in partman-base. [13:34] it's better to file a new bug [13:35] ok i am still learning the ropes [13:35] thanks [13:35] I think you've been infected by the idea that the number of bugs must be kept to a minimum [13:35] yes i am [13:35] what we actually need is for each bug to describe a single, coherent, fixable issue [13:36] i will adhere to that in the future [13:36] it tends to impede development when a bug describes more than one problem [13:36] thanks [13:38] ubiquity: cjwatson * r3158 ubiquity/ (3 files in 2 dirs): [13:38] ubiquity: Using "Yes" and "No" buttons for the weak-password question turns out to [13:38] ubiquity: be more confusing than I'd hoped. Instead, use "Continue" and "Choose [13:38] ubiquity: another password" (breaking the string freeze, but at least for only one [13:38] ubiquity: string). [13:44] ubiquity: cjwatson * r3159 ubiquity/ubiquity/frontend/gtk_ui.py: i18n todo [15:11] cjwatson: prog bar is much, much better on netboot :) [15:14] excellent [15:28] cjwatson, is there an easy way to respin a bootable ISO with a custom d-i initrd + kernel? I'm testing the new kernel config for lpia with a USB stick, but I'd like to test it with an actual CD to see if it works ... [15:29] mkisofs -r -V 'Ubuntu 9.04 i386' -o jaunty-alternate-i386-hacked.iso -cache-inodes -J -l -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 4 -boot-info-table new-i386 [15:29] NCommander: I have a file full of runes like that, which I copy and paste :) [15:30] NCommander: (for i386, it prefer to run as root) [15:30] s [15:30] I assume that will just work for lpia? [15:30] yep, should be the same [15:30] Perfect [15:30] I just want to make sure my new kernel and RAMdisk work sanely. [15:53] cjwatson, wooo, success [15:53] cjwatson, (and the new lpia kernel config works for the CD-ROM :-)) [15:54] NCommander, Thanks for chasing the kernel bits to test that :) It's been annoying me for several months now. [15:54] oh good [15:55] cjwatson, thats a great one liner, is it documented anywhere? [15:55] I think it's on help.ubuntu.com/community/InstallCDCustomization or some such [15:57] cjwatson, once the kernel upload goes through, will d-i need a bump upload (there is no ABI bump, it will just be a normal revision) [15:58] yes [15:58] there's stuff in bzr to upload anyway [16:00] cjwatson, ah, ok; I just wanted to make sure. [16:01] it's not a problem, we always make sure d-i is up to date for releases [16:01] cjwatson, I just find the logic to get stuff into installer-* is a bit of a blackbox [16:02] oh, that's easy, installer-* is only updated by way of debian-installer uploads [16:02] assuming you mean dists/jaunty/main/installer-* [16:03] cjwatson, right, I meant the Soyuz logic behind it [16:03] cjwatson, I stopped trying to figure out how that worked in dak ages ago :-) [16:04] it's easy, it's well-encapsulated in both dak and soyuz [16:05] Fair enough [16:06] * NCommander was referring to the actual code. [16:06] so was I [16:07] it's scripts/debian/byhand-di in dak [16:07] and it's lib/canonical/archivepublisher/debian_installer.py in LP [16:09] cjwatson, hrm, I always thought it was a dak function, I didn't realize it lived in a script. The only time I ever looked at BYHAND in dak was when investigating 3.0 source package support, and when I fixed the DM code [16:12] * NCommander didn't know cjwatson knew Soyuz's internals [16:14] reasonably [16:14] it used to be in Python when we put that "auto-byhand" (an oxymoron) code for d-i into Ubuntu originally [16:14] when aj took it for dak upstream, he rewrote it in shell [16:15] I assume he had a reason [16:15] I wrote the guts of the code for both dak and Soyuz, way back when [16:15] I mean, the guts of the d-i handling stuff, not the whole thing :) [16:30] ubiquity: cjwatson * r3160 ubiquity/.bzrignore: update ignores [16:33] HOORAY !!! [16:33] finally ubiquity isnt complaining anymore in apt-setup [16:56] wooo, fully successful lpia installation from CD [17:08] cjwatson, oh, I didn't see you in the authors file (sorry, stepped away) [17:08] cjwatson, I was working at removing some of the more braindead pieces of code (such as removing the hardcoded 'unstable'' lines) [17:36] ubiquity: cjwatson * r3161 ubiquity/ (debian/changelog ubiquity/misc.py): Suppress unnecessary exception if find_in_os_prober fails. [17:40] cjwatson, unfortunately still having troubles with grub2 and ubiquity 1.12.1. i'll try to get an apport bug submitted again [17:42] ok === nxvl_ is now known as nxvl [17:59] ubiquity: cjwatson * r3162 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): [17:59] ubiquity: * GTK frontend: [17:59] ubiquity: - Don't bother calling find_in_os_prober for partitions without real [17:59] ubiquity: devices (e.g. free space). [19:20] argh, I made a change to ubiquity's advanced partitioning page, and now the entire page apart from the heading is blank for no readily explicable reason [19:21] I hate when that happens [19:21] * cody-somerville isn't being facetious either. === Nicke_ is now known as Nicke [19:38] ubiquity: cjwatson * r3163 ubiquity/ (debian/changelog ubiquity/components/partman.py): Fix handling of "yes" response to partman/unmount_active. [22:40] ubiquity: superm1 * r3164 ubiquity/ (debian/changelog scripts/mythbuntu/mythbuntu_install.py): [22:40] ubiquity: Make sure mythbuntu installs have users installed into the 'video' [22:40] ubiquity: group. (LP: #351773)