[01:44] The "server" and "alternate" CDs are subtly different, e.g. the tasksel list provided is different. [01:44] I'm doing installs using the bare netinst kernel and initrd, which seem to behave like the alternate CD. [01:45] Can I make them behave like the "server" CD, i.e. things like offering the server tasksel list? If so, how? [04:34] I'm having trouble appending preseed.cfg to a gunzipped initrd. [04:34] $ cpio -o --append -F initrd cpio: premature end of file [04:38] Looks like I need -H newc... [10:06] Morning guys are you around I have a query about reinstall grub in rescue mode? [10:07] evand, cjwatson: ^ Forgot to add your names D'oh [10:08] what is your query? [10:09] is /dev/hdxx still relevant on an ubuntu system being as all drives are now /dev/sdxx [10:10] or is grub specific to a fault? [10:12] cjwatson: ^ [10:17] firstly, (hdN) in grub does not mean /dev/hdN [10:18] secondly, some systems do still use /dev/hdN [10:21] cjwatson: I know that (hd0) isn't specific but the text also says you can type in /dev/hda as an example. I was just checking that this information was still relevant and not out of date as it may of been missed :) [10:22] Now I know I can carry on :) Thanks [10:22] yeah, it's still relevant; only IDE controllers which have been converted to the PATA subsystem use /dev/sd* [10:22] it may not be the *best* example nowadays, but it isn't wrong [10:23] cjwatson: Cool, I only really noticed it because I'm writing up the testcase and wondered about it :) [10:53] cjwatson: Just FTR we've looked at the NSLU2 boot issue last week and ogra is continuing research on this; there are misc hints we're trying out [10:54] Such as endianess of the kernel, apex fis partitioning, apex script [10:55] Hmm right CONFIG_CPU_BIG_ENDIAN is set on Ubuntu and not on Debian [11:29] lool: ok, thanks [12:05] cjwatson: is it best to enable pc cards if you have them at the find cd stage in expert mode? [12:08] it's best to accept the defaults in expert mode unless you have specific requirements [12:09] cjwatson: Thanks [12:39] ubiquity: cjwatson * r3000 ubiquity/debian/ (57 files in 2 dirs): Update translations from Launchpad. [12:43] ubiquity: cjwatson * r3001 ubiquity/debian/po/ (58 files): more complete translation update [12:52] user-setup: cjwatson * r151 ubuntu/debian/ (64 files in 2 dirs): Update translation templates. [12:59] user-setup: cjwatson * r152 ubuntu/debian/changelog: releasing version 1.23ubuntu10 [14:50] cjwatson: ok to add the multipath patch for grub? (bug 312447) [14:50] Launchpad bug 312447 in grub "support for multipath" [Medium,Triaged] https://launchpad.net/bugs/312447 [15:03] tjaalton: that seems OK [15:06] cjwatson: thanks [15:11] ubiquity: evand * r3002 ubiquity/ (10 files in 5 dirs): [15:11] ubiquity: * Replace the previous iteration of the timezone map (zoommap) with a new [15:11] ubiquity: version (timezone_map) that should be easier to use, in accordance with the [15:11] ubiquity: JauntyUbiquityUsability specification. [15:11] ubiquity: * Use wrap_label, a replacement for GtkLabel from libview that does [15:11] ubiquity: height-for-width text wrapping to work around bgo 101968. [15:12] That still needs quite a bit of polish and bug fixing, but I wanted to make sure it landed ASAP [15:34] ubiquity: evand * r3003 ubiquity/bin/ubiquity: Create the debug log if it does not already exist. [15:38] ubiquity: evand * r3004 ubiquity/gui/glade/stepLocation.glade: The first drop down box on the timezone selection page now describes a region, not a time zone. [15:43] evand: r3003> oops, good point [15:43] evand: could you do that to oem-config too? [15:43] will do [15:43] thanks [15:45] evand: so that'll be in the installer tomorrow correct so I can see if I can break it :) [15:48] oem-config: evand * r606 oem-config/oem-config: Create the log file if it does not already exist. [15:49] ubiquity: evand * r3005 ubiquity/debian/ubiquity.templates: Change the first drop down box to reflect the region in the templates file as well. [15:50] davmor2: only if we release a new ubiquity today. I wouldn't worry too much about hammering it just yet. There are a few glaring issues that I need to fix first (the timezone points do not line up properly and we need to remove the country borders, for example) [15:50] Still if you test it and find some other issue, I'm keen to hear about it [15:51] evand: I'm smoke testing each afternoon so if it's in it'll get tested :) [15:51] wonderful [15:52] :) [16:04] mpt: http://people.ubuntu.com/~evand/tmp/ubiquity-timezone-9.04-dev.png - if you have a free moment and see any glaring errors in that design, do let me know. Alternatively I'm a few desks away and can give you a quick demo, but I suspect you're too busy for that. Do note that we're going to remove the country borders and use better colors in the next cut of the image. I also still have to render a clock in there somewhere and possibly brin [16:04] g in the right edge a bit to make the spacing even. [16:04] ^ cjwatson likewise :) [16:05] evand, it's very ... colorful [16:06] indeed, the colors on the terrain will be removed [16:06] well, merged into a single color [16:06] but noted :) [16:06] good good [16:07] The intro blurb says "the indicated current time" but I don't see any [16:08] indeed, I have to add a clock still [16:08] my current thought it to put a rendered analog clock with a digital clock embedded in that either in the bottom left of the map or between south america and africa [16:08] but suggestions welcome there [16:09] worst case scenario, I could always add the text clock back in [16:09] (it was removed to allow more space for the map) [16:09] hmm [16:10] Ideally the instructions for what to do if the time is still incorrect should be pretty close to wherever the clock ends up [16:10] ok [16:12] The whole map is off-center [16:12] and (therefore?) so are the menus underneath [16:13] indeed, I believe that was originally done to be in line with the HIG (indenting underneath the explanatory text) [16:14] cjwatson: suggested indenting the right side as well [16:14] my personal preference would be to remove the indentation altogether [16:14] I agree [16:14] with which? [16:14] removing it altogether [16:14] ok [16:15] I think indentation is intended more for showing which controls belong to which heading when there are multiple headings [16:15] Here there's only one [16:15] ok, noted. Wonderful, that will provide even more space for the map to grow into. [16:16] (Do the other installer steps have the same indentation?) [16:17] Most do (http://people.ubuntu.com/~evand/screenshots/ubiquity/) [16:19] 2 part question: what is the right way to change: /boot/grub/menu.lst:85 # defoptions=quiet splash [16:19] ugh, glade-3 is thoroughly annoying sometimes (deletes custom widgets, renames things foowidget1) [16:20] so that when a new kernel gets added, I don't get a .conf collision - and is there some way to drive this from the preseed file so that I don't have to even edit the file [16:22] ubiquity: evand * r3006 ubiquity/gui/glade/stepLocation.glade: Remove the GtkAlignment (and thus the left padding of the map) from the timezone selection page. [16:22] evand, the only other one that jumps out as wrong is the keyboard layout step [16:23] oh, now I see that the "Where are you" map isn't centered either [16:23] mpt: you would suggest removing the indentation on the keyboard selection page as well? [16:24] yep [16:24] and the "Where are you" page [16:24] the map isn't centered? It's a projection of a sphere :) [16:24] It's not horizontally centered within the window [16:24] ahh, indeed [16:24] ok, I've made the where are you page change as r3006 [16:25] I'll take care of the keyboard layout page now [16:25] CIA-3 said 3006 was about timezone, not where are you [16:25] but thanks :-) [16:25] hm [16:26] I wonder if a nice way to show the selected time would be as text, overlaid at the bottom of the selected timezone [16:26] where are you is the timezone :) [16:26] hrm [16:26] oh, sorry [16:26] that could easily be done. I can at least mock it up as one possible option. [16:27] Looking at the old screenshots I thought it was a completely separate stage for some reason [16:27] Maybe that's why the DX team have started calling me Muppet [16:27] hahaha [16:28] ... Because if the selected time is near the bottom of the map, and the "If the selected time is wrong..." text is underneath the map, that will achieve the goal of having those two things near each other. [16:29] ah, indeed [16:29] The whole set of instructions could stand being under the map, even [16:30] For "Selected region" and "Selected city", what does the word "selected" achieve? [16:31] not sure, I think I was just trying to be consistent with the old design [16:31] but if "City:" and "Region:" works, I'm fine with changing that [16:32] I think it would work [16:32] Heh, looking at this map I can see why Xinjiang has its own unofficial timezone [16:34] ok, I'll make the change there as well [16:34] The main remaining thing that puzzles me is why two of the areas look selected [16:34] two of the timezones, I mean [16:34] one is selected and the mouse is over the other [16:34] the highlight moves with the mouse [16:35] currently, anyway [16:35] ah, hmm [16:35] So is there any mousedown feedback? [16:36] the x appears over the nearest timezone point to the cursor [16:36] other than that, no [16:37] there are two ways to get to the same place. You can select a timezone point by using the map, or using the drop down menus. [16:37] Using one or the other reflects the selection in both. [16:38] I think it's confusing for the selection appearance and the mouseover appearance to be 99% the same [16:39] I'm not a fan of mouseover effects in general, unless they clarify things in a way that really would be too cluttersome if applied even when you weren't mousing over [16:40] so your preference would be for the timezone highlight to only appear when the user clicks on the map? [16:40] So if you'd decided that showing all the timezone boundaries all the time was too cluttersome, I could understand showing the border of the hovered-over timezone on mouseover [16:40] but you're showing all the timezone boundaries all the time anyway [16:40] ubiquity: evand * r3007 ubiquity/gui/glade/stepKeyboardConf.glade: Remove the alignments next to the keyboard selection widget and the 'type here' GtkEntry box, per mpt's suggestion. [16:41] yes, I think so [16:42] highlight just the selection [16:42] ubiquity: evand * r3008 ubiquity/debian/ubiquity.templates: Remove 'Selected' from the labels on the timezone page. [16:42] my only concern there is that it's then not obvious that you can interact with the map [16:43] as it does nothing at all when you move over it. You have to click on it to realize that it's interactive. [16:44] Could you use the pointy-hand cursor? [16:44] (the one Web browsers use for links) [16:44] evand: can you not change the cursor to be a pointy hand or something as you hover over the map [16:44] and a cursor the rest [16:45] Hrm, I could indeed [16:45] snap :-) [16:45] mpt: did he beat you to it? [16:45] Not from where I'm sitting, maybe I'm lagging [16:45] anyway [16:45] I was typing it as mpt posted :) [16:45] What are the dark dots west of Greenland? [16:46] 3 [16:46] I was thinking they were cities that use the next-door timezone, but according to Google Maps there's nothing there [16:48] evand: for colours could you not just use a green/blue/lightish grey and pick a slightly light colour for the adjoining time zone and then back to the darker version and so on [16:48] hrm according to the map we're using they appear to be part of -5 [16:48] * evand digs [16:48] s/light/lighter [16:49] hrm, it indeed appears to be a bug with the image. I'll make a note to mention that to Ken. [16:49] davmor2: the colors are going to change, but I'm leaving the selection mostly up to Ken as he's designing the image. [16:49] * mpt guesses supporting Xinjiang time probably would upset more people than it would delight [16:50] ah okay [16:51] mpt: xinjiang time? Isn't that just +8? [16:52] ubiquity: evand * r3009 ubiquity/debian/changelog: Changelog entry for the alignment removals. [16:52] evand, officially yes, unofficially it's +6 [16:52] heh [16:53] * evand shakes his head at the mess that is timezones [16:53] evand: And you said it would be easier ;) [16:53] What timezone you use depends on what form of transport you're using, among other things [16:54] wow [17:15] ubiquity: evand * r3010 ubiquity/ubiquity/timezone_map.py: [17:15] ubiquity: Display a hand pointing cursor when over the timezone map. [17:15] ubiquity: Remove the timezone mouseover highlights. [17:50] ubiquity: evand * r3011 ubiquity/ubiquity/timezone_map.py: Add the current time to the bottom left of the selected timezone on the timezone map page. [17:51] whoops, that should say bottom right [18:04] CarlFK: easiest way is just to put the extra parameters you want at the end of the kernel parameters when booting the installer itself, after the "--" separator parameter [18:06] usb-creator: evand * r75 trunk/ (debian/changelog usbcreator/backend.py): [18:06] usb-creator: Better handle filesystem-on-disk in the bootloader installing code. [18:06] usb-creator: Thanks Loïc Minier (LP: #325375) [18:23] cjwatson: awesome. thanks. [20:09] cjwatson: pxe booted installer with: append initrd=ubuntu/jaunty/initrd.gz root=/dev/rd/0 rw locale=en_US console-setup/layoutcode=us netcfg/wireless_wep= netcfg/get_hostname= DEBCONF_DEBUG=5 url=http://shaz/ubuntu/jaunty/preseed.cfg tasksel:tasksel/first="ubuntu-standard, ubuntu-desktop" -- vga=6 [20:09] menu.lst still has: kernel /boot/vmlinuz-2.6.28-7-generic root=UUID=a5e2a23e-a2ea-4aa8-8d20-72e43612e99c ro quiet splash [20:10] does the installer save the kernel params it used to /var/something? [20:10] vga= is explicitly excluded [20:10] heh [20:10] are you *sure* you want it? [20:11] it breaks suspend/resume [20:11] I can live without vga - more important just to get rid of the splash so that if the boot hangs I can see the last thing it was doing [20:12] sounds as though the easiest approach would be to sed defoptions in preseed/late_command and run update-grub [20:12] debian-installer/framebuffer=false will also disable splash, but may have other consequences you don't want (or you may not care) [20:13] oh, I think I've said before, but you can replace your tasksel preseeding there with tasks=ubuntu-desktop [20:13] so after menu.lst gets changed (sed of vi) I run update-grub, and that will get rid of the "menu.lst has changed, what do you want to do? " when apt-get installs a new kernel? [20:13] oh yeah... forgot to touch that up] [20:13] you'll only get that if you change the comments and the autogenerated options separately [20:15] ah - I should just change the # defoptions=quiet splash and let update-grub update the kernel line? [20:16] yes [20:17] the file format sucks, sorry about that [20:17] (not that it's my fault, but collectively ...) [20:17] i can cope. [20:17] thanks for the tips - 'i'll give it a shot [20:18] partman-partitioning: cjwatson * r694 ubuntu/ (debian/changelog free_space/new/do_option): [20:18] partman-partitioning: Don't default to logical if an extended partition already exists, only [20:18] partman-partitioning: if a real primary partition exists. [20:18] and the project I am working on has an issue tracker where I am pasting what you tell me - if we are lucky I will remember where to look when it comes up in 6 months ;) [20:21] ubiquity: cjwatson * r3012 ubiquity/ (3 files in 2 dirs): [20:21] ubiquity: Don't default to logical if an extended partition already exists, only [20:21] ubiquity: if a real primary partition exists. [20:47] partman-auto: cjwatson * r279 ubuntu/ (72 files in 25 dirs): [20:47] partman-auto: Add support for $default_filesystem as the partition type in a recipe, [20:47] partman-auto: and $default_filesystem{ } as an internal specifier. These are [20:47] partman-auto: equivalent to FS and filesystem{ FS } respectively, where FS is the [20:47] partman-auto: value of partman/default_filesystem. Requires partman-base 128ubuntu6. [20:50] partman-crypto: cjwatson * r675 ubuntu/ (debian/changelog debian/control init.d/crypto): Honour partman/default_filesystem. Requires partman-base 128ubuntu6. [20:52] partman-partitioning: cjwatson * r695 ubuntu/ (debian/changelog debian/control free_space/new/do_option): [20:52] partman-partitioning: Honour partman/default_filesystem when creating new partitions. Requires [20:52] partman-partitioning: partman-base 128ubuntu6. [21:00] debian-installer: cjwatson * r1037 ubuntu/ (debian/changelog doc/devel/partman-auto-recipe.txt): [21:00] debian-installer: Document new $default_filesystem substitution and $default_filesystem{ } [21:00] debian-installer: internal specifier in partman recipes. [21:33] partman-auto: cjwatson * r280 ubuntu/debian/changelog: releasing version 83ubuntu3 [21:33] cjwatson: how do I update-grub in the installer? update-grub= -sh: update-grub: not found [21:34] partman-crypto: cjwatson * r676 ubuntu/debian/changelog: releasing version 36ubuntu4 [21:34] partman-partitioning: cjwatson * r696 ubuntu/debian/changelog: releasing version 64ubuntu3 [21:36] CarlFK: chroot /target update-grub [21:36] actually maybe 'in-target update-grub' would be better; that also mounts a bunch of filesystems and hooks up debconf plumbing [21:41] worked at the prompt - added to late_command, from the top... [21:57] * CarlFK does happy grub dance [21:58] cjwatson: thanks bunches - this little grub issue has been bugging me for a long time [22:00] glad to help