[00:45] ubiquity: cjwatson * r4025 ubiquity/ (debian/changelog ubiquity/components/ubi-language.py): [00:45] ubiquity: Don't lower-case the locale attribute on frontends. This was added a [00:45] ubiquity: long time ago; since then, it's become unnecessary since i18n.get_string [00:45] ubiquity: lower-cases the locale itself, and it prevented us from finding [00:45] ubiquity: slideshow files for language variants (LP: #476269). [00:51] ubiquity: cjwatson * r4026 ubiquity/ (d-i/manifest debian/changelog): [00:51] ubiquity: Automatic update of included source packages: partman-auto 89ubuntu4, [00:51] ubiquity: partman-ext3 58ubuntu2. [00:57] cjwatson: Hi, I'm here about bug 476269 :) [00:57] Launchpad bug 476269 in ubiquity "Chinese, Brazilian Portuguese, and English variant translations aren't shown during Karmic and Lucid installation" [Undecided,Fix committed] https://launchpad.net/bugs/476269 [00:57] dylan-m: hey [00:58] I can fix that issue you mentioned on the slideshow's end (and I would much prefer to). Just didn't realize that was the way we were meant to do the fallback :) [00:59] IMO it should match slideshow_get_available_locale as closely as possible [00:59] that method should be correct - it was just getting screwed by locale being pt_br not pt_BR [01:00] Oooh, I see [01:00] gee, I completely missed that! Nice catch [01:00] Yep, that isn't a problem. I had it as a full implementation at first and for some reason decided that falling straight back to the default made more sense [01:01] (All guessing, though; is there some kind of reference guide for localization best practices?) [01:02] I'm the wrong person to ask, I've picked it up over ten years and no longer remember :) [01:02] slideshow_get_available_locale isn't perfect of course (way too much parsing of locale bits, and it doesn't do a complete job) - it's just good enough for the job at hand [01:02] Oh well, no problem [01:03] 'info gettext' is useful reading [01:03] does http://paste.ubuntu.com/407324/ look right? I haven't written any JS for a long time [01:04] Yep, looks good [01:04] I guess it'll also need to compare using .lower, too [01:05] I think I'd like to assume that the locale is in the canonical case [01:06] rather than messing about with case-canonicalisation which will get fiddly [01:06] Oh, yes, I guess it's on the disk in the right way anyway [01:06] whoever is requesting the locale just needs to say the right thing [01:08] (on which topic, that's a reason why it makes sense for this to happen within the slideshow: it crashes and burns if it is expecting a locale that doesn't exist, so the fallback routine always needs to be in there anyway) [01:09] mm, I admit it's nicer not to have this in ubiquity [01:09] OK, testing with my .lower removal plus that slideshow change plus your branch [01:09] if that works, I'll upload all three; OK with that? [01:10] Okay! [01:10] Thanks, Colin [01:14] you caught me just in time, I was this far from uploading ubiquity [01:15] right, that seems to work fine [01:17] ubiquity-slideshow-ubuntu: cjwatson * r258 ubiquity-slideshow-ubuntu/ (debian/changelog slideshows/link-core/slideshow.js): [01:17] ubiquity-slideshow-ubuntu: Fall back from language variants (e.g. de_DE) to languages (e.g. de) [01:17] ubiquity-slideshow-ubuntu: (LP: #476269). [01:21] ubiquity-slideshow-ubuntu: cjwatson * r259 ubiquity-slideshow-ubuntu/debian/changelog: releasing version 20 [01:26] dylan-m: oh, err [01:26] dylan-m: what about the RTL stuff? [01:27] ltr = i18n.get_string('default-ltr', slideshow_locale, 'ubiquity/imported') [01:27] if ltr == 'default:RTL': [01:27] that doesn't look straightforward to port [01:27] slides += '?rtl' [01:27] hate to say, that one isn't [01:27] I was aiming to fix that with 10.10 (since I'm hoping to redo how localization works) [01:27] I think we'll have to leave the ubiquity locale selection there for now, in that case [01:27] Yep, that one does [01:28] oh well [01:28] this'll fix the bug, at least [01:28] unless we can store if a locale is an RTL locale in the fantastically ugly directory.js :) [01:28] if you figure it out, I'll be happy to merge your ubiquity branch; I'm not going to touch that though :) [01:29] With the current approach it unfortunately makes en right to left if it falls back that far, but hopefully it just won't get that far [01:30] * dylan-m just reminded myself there is some extra stuff to handle there, but there's no hurry on it [01:32] ubiquity: cjwatson * r4027 ubiquity/debian/changelog: releasing version 2.2.12 [01:32] thanks for the prod and the advice! [01:33] Thank you for solving that bug! [04:37] ev: bug 528678 ... is that even reproducable? [04:37] Launchpad bug 528678 in ubiquity "UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 13: ordinal not in range(128)" [High,Confirmed] https://launchpad.net/bugs/528678 [08:53] just want to extra point this out, because it's been bugging me for a while and I hope it can be fixed in the next ubuntu version: https://bugs.launchpad.net/ubuntu/+source/debian-installer/+bug/160366/comments/1 [08:53] Ubuntu bug 160366 in debian-installer "Add xen netboot support" [Wishlist,Triaged] [08:53] (that is, point out that the solution would take about 5 minutes) [09:36] shtylman: that's definitely a bug. Wrapping an 8-bit string in unicode() without an encoding just doesn't work. [09:36] shtylman: I'm not sure why I suggested it was QString's fault. This is before it's passed to the QString constructor. [10:06] cjwatson, superm1: http://paste.ubuntu.com/407480/ - does this roughly look like what you were after? I still need to check to make sure it doesn't rip out the langpack, but it otherwise works for my test of installing gnome-do and vim before running oem-config-prepare. [10:07] :/ [10:11] Kevin`: it might take about five minutes, but it's something that needs to be done in the kernel packaging - we don't control those files [10:12] it's already packaged in your stock kernel, just not in the installer [10:12] Kevin`: I know what I'm talking about. :-) [10:12] the kernel delivers udebs for use in the installer [10:13] well, I don't know much about the politics, heh. who handles it [10:13] it's not politics, it's just where the data files live [10:13] debian/d-i/ in the linux source package [10:13] sorry, debian.master/d-i/ these days [10:14] I've reassigned the bug, thanks [10:15] #ubuntu-kernel ? [10:15] if it were changeable in the installer, I'd have done it already :) [10:15] yes, tell them I sent you [10:17] I'd like it to be fixed, certainly [10:17] xen-blkfront should go in scsi-modules, and xen-netfront should go in nic-modules [10:18] (feel free to copy and paste that) [10:20] ev: looks plausible enough [10:20] hooray [10:21] just need to put it through its paces, then beg the release team for sympathy [10:21] it's been a recorded/tracked work item for a while, you shouldn't have to beg to hard [10:22] *too [10:36] cool [10:38] oh hm, one part of this situation actually may be in your area, although it's from upstream debian: the installer won't let you select to use a disk unpartitioned unless there's already a filesystem on it. normally people just ask 'wtf would you want to do that', but it makes management a ton easier, and disk alignment a non-issue for weird storage [10:38] or I just don't know how to use the partitioner [10:38] :) [10:47] shtylman: I think there may be a problem with your event loop changes. I don't seem to be progressing through the install if I click next and leave the mouse pointer still. As soon as I move the mouse, the installer continues on. This is most apparent on the timezone page, where the time stops updating until I move the mouse. [10:59] Kevin`: mm. I do think that's probably a bug but it's not trivial to change - the UI gets very tricky. Alignment issues should be fixed in 10.04 though [11:03] what did you do to fix the alignment problem, query the disk for it's sector size (for eg the new consumer drives), or just pick a more sane default then the ms-dos standard that's actually a multiple of anything besides 512? [11:18] Kevin`: modern parted queries the disk; it defaults to 1MiB if it doesn't get specific information [11:19] the legacy DOS standard is cylinder alignment, i.e. 63*512 [11:19] sounds good :) [11:26] shtylman: if debian/changelog isn't UNRELEASED at the top, then you need to create a *new* entry, not add to the existing one [11:28] * cjwatson fixes [11:28] ubiquity: cjwatson * r4031 ubiquity/debian/changelog: fix changelog [11:53] dch -U does this magically for you if this is a native package (ubiquity is) [11:56] for what it's worth, bug 553184 is the freeze exception request for oem-config/remove_extras [11:56] Launchpad bug 553184 in ubiquity "Freeze exception request for removing extra packages in oem-config" [Undecided,New] https://launchpad.net/bugs/553184 [12:24] partman-crypto: cjwatson * r696 ubuntu/ (blockdev-keygen debian/changelog): Allow preseeding partman-crypto/weak_passphrase too (LP: #546405). [12:26] partman-crypto: cjwatson * r697 ubuntu/debian/changelog: releasing version 40ubuntu3 [13:27] usb-creator: rgreening * r291 trunk/debian/changelog: Updated with LP: Bug for the QVariant issue [13:30] ev: we really need to release usb-creator again to fix bug 553243, which makes usb-creator unusable (fix is in lp). [13:30] Launchpad bug 553243 in usb-creator "usb-creator-kde mostly broken" [Critical,Fix committed] https://launchpad.net/bugs/553243 [13:30] rgreening: just working on some other bug fixes first. I'll upload a new version today. [13:31] * rgreening thinks I should get per package upload rights to usb-creator :P [13:31] hhe. cool [13:31] apply [13:31] okies, as long as we get something out. ty ev. [13:31] sure thing [13:31] ya, I probably will. [13:33] let me know when you do and I'll put a good word in [13:33] cool [14:11] usb-creator: evand * r292 usb-creator/ (bin/usb-creator-helper debian/changelog): Encode ISO path with UTF-8 when mounting (LP: #460298). [14:44] ubiquity: evand * r4032 ubiquity/ (3 files in 2 dirs): [14:44] ubiquity: In oem-config, support removing packages that were not part of the [14:44] ubiquity: base install and are not needed in the final system by preseeding [14:44] ubiquity: oem-config/remove_extras to true (LP: #315644, LP: #553184). [15:26] cjwatson: my bad... I thought that when I ran dch it would do the right thing :/ [15:26] ev: no problem... we just have the user play "wack-a-mole" while it installs to ensure the mouse is always in motion [15:29] cjwatson: a few weeks ago, you mentionned a script in the installer which takes arguments like "disk" or "cd" and returns the devices detected on the system. where was that script again? [15:29] list-devices [15:29] if you want the source, it's in the debian-installer-utils source package [15:40] cjwatson: after having a quick look at the script, it seems that it relies on ID_TYPE output by udevadm. however, this environment variable is not set when using virtio devices in kvm :( [15:59] usb-creator: rgreening * r293 trunk/ (6 files in 5 dirs): Update version info to 0.2.21 in preparation for release. [16:10] usb-creator: rgreening * r294 trunk/debian/ (changelog control): Add explicit version requirement to depend on latest usb-creator-gtk for usb-creator. [16:32] rgreening: ^ why? [16:33] ev: a bug on upgrade from old usb-creator meta would be satisfied with an old usb-creator-gtk and not the latest, which it should [16:34] it certainly doesn't hurt [16:34] as we expect it to be that version anyway [16:34] huh? usb-creator is a virtual package. It's job is to simply install usb-creator-gtk, which will be updated on its own [16:35] hmm.. didn't seem to work in at least on case [16:35] dunno why [16:35] is there a bug number for this? [16:35] 1 sec. [16:36] ev: it was in comment to this bug 503808 [16:36] Launchpad bug 503808 in usb-creator "usb-creator-gtk does not start (GtkFrontend instance has no attribute 'cancelbutton')" [Undecided,Fix released] https://launchpad.net/bugs/503808 [16:37] okay, I can see the utility in that [16:38] sorry for the tough questions, I just want to be sure we're just fixing serious bugs at this point :) [16:38] :) [16:38] no prob [16:38] I fixed the version too [16:39] :) [16:40] ev: how close are you to pumping out a release? we really need usb-creator-kde working for the beta [16:41] rgreening: I'll do it now [16:41] and ubiquity while I'm at it [16:41] cool. [16:41] ty ev. [16:43] usb-creator: evand * r295 usb-creator/po/ (40 files): update-po [16:44] ev, looks about right to me. i'm curious what happens if you pull ubiquity and oem-config out from under their feet at that time though? hasn't everything not been loaded into memory by then? [16:44] superm1: in the event that we someday decide to open a file shipped with ubiquity after that point [16:44] was just being cautious [16:45] it is a kind of unfortunate experience that it done in the shell script for oem-config-firstboot right now, because after that pretty oem-config wizard gets done it drops you to a black screen with the machine spinning for a minute or two [16:45] and you might not want them removed, if you set the preseed accordingly [16:45] hm [16:46] perhaps we could meet halfway by involving plymouth? [16:46] so if it doesn't "break" it would be better if that preseed's logic was moved into the same place as all this remove extras logic [16:46] yeah that would probably be sufficient too [16:49] Looking at the code, I don't have a strong feeling either way. cjwatson? [16:51] I'm not sure I follow the question [16:54] the question is moving the code to remove oem-config from oem-config-firstboot into remove_extras (pulling the feet from under oem-config/ubiquity, hopefully it all being in memory at that point) or adding some plymouth magic to not show a black screen after oem-config is done before gdm comes up while it's removed there [16:55] if the former is selected, oem-config/late_command would need to be moved up too [16:55] oh, right [16:55] I'm dubious that it will actually work reliably; I'm OK with it if it does [16:56] I don't like relying on things being in cache more than we have to, especially for memory-hungry things like ubiquity [16:56] on a possibly related note, I was working on splitting out a ubiquity-common package, which might reduce the impact of having oem-config remain installed [16:57] and would be useful to avoid a bunch of dependencies - but I haven't finished that yet [16:57] it's still loading all sorts of debconf templates after that point of remove_extras, so i'm starting to think it wouldn't work properly either [16:58] when installing ubuntu-server on kvm with virtio, /bin/list-devices disk returns nothing and yet it installs just fine on /dev/vda. how does the installer find the disk if not using list-devices? [17:00] it normally does use list-devices, and IMO it's a bug that it doesn't work there - I'll investigate when I have time [17:00] but in some cases it uses parted_devices instead [17:02] cjwatson: parted_devices works on kvm with virtio, fyi [17:05] could use another set of eyes on this (internet connectivity check for update link) - http://paste.ubuntu.com/407636/ [17:07] cjwatson: where does parted_devices come from? it doesn't seem to be in debian-insaller nor debian-installer-utils [17:08] cjwatson: it seems to be part of ubiquity but I'm installing ubuntu-server. does ubuntu-server now use ubiquity? [17:31] usb-creator: rgreening * r296 trunk/ (debian/changelog usbcreator/frontends/kde/frontend.py): [17:31] usb-creator: * Get rid of the needs-format warning for now. It was always shown [17:31] usb-creator: for parent block devices and the confusing resulting from that was [17:31] usb-creator: quite bad. This was previously done for gtk but not kde version. [17:45] cjwatson, superm1: could one of you sort out a ubiquity upload? I was hoping to squeeze in that internet connectivity check, but I've got to run to catch a train down south. [17:45] have a great weekend [17:45] rgreening: I *think* I've sorted the usb-creator upload [17:45] see the irclogs.ubuntu.com page for #ubuntu-release for the details [17:46] If you need help getting it though, I can poke Riddell [17:46] :) [17:48] ev: ty [17:51] cr3: parted_devices> partman-base source package [17:52] cr3: no, ubuntu-server doesn't use ubiquity; ubiquity incorporates partman-base [17:53] cjwatson: Wubi will not work when ported to linux OS with no ntfs support right or is there a turn around ? [17:54] ubiquity: superm1 * r4033 ubiquity/debian/changelog: releasing version 2.2.13 [17:55] kusum: I don't see how you could work around the lack of NTFS, given that the point of Wubi is to include support for installing inside Windows filesystems which are normally NTFS! [17:55] kusum: ntfs-3g shouldn't be that hard to incorporate [17:56] if we integrate ntfs-3g getting wubi working shouldnt be a hard task right ? [17:59] about as hard as I explained :-) [17:59] it's not trivial software [17:59] yeah [19:18] usb-creator: rgreening * r297 trunk/ (debian/changelog usbcreator/frontends/kde/frontend.py): [19:18] usb-creator: * Get rid of the needs-format warning for now. It was always shown [19:18] usb-creator: for parent block devices and the confusing resulting from that was [19:18] usb-creator: quite bad. This was previously done for gtk but not kde version. [19:18] usb-creator: * Update kde frontend format option to match gtk frontend/backend [19:18] usb-creator: (LP: #553460) [19:18] usb-creator: * Add warning dialog to format button for usb-creator-kde, to match gtk. [19:35] usb-creator: rgreening * r298 trunk/ (5 files in 4 dirs): Update version strings for next release: 0.2.22 (let's try and be consistent) [20:00] usb-creator: rgreening * r299 trunk/ (debian/changelog usbcreator/frontends/kde/frontend.py): Add BusyCursor to kde frontend while format operation is underway. [21:06] ev: what do you think we should do about the event stuff? how do you handle the tight loop on the gtk side? [21:07] anybody knows when can i find xivulon online ?? [21:07] If somebody knows please answer [23:45] grub-installer: cjwatson * r851 ubuntu/ (debian/changelog grub-installer): (log message trimmed) [23:45] grub-installer: Historically, Ubuntu's default behaviour in the absence of command-line [23:45] grub-installer: arguments to the installer has been to pass 'splash' to the installed [23:45] grub-installer: system so that it gets a splash screen (and similarly 'quiet'). There's [23:45] grub-installer: some argument that this behaviour should require passing 'splash' to the [23:45] grub-installer: installer as well, but it's too late in the Lucid cycle for this change. [23:45] grub-installer: Nevertheless, we need a way to configure images such that they don't get [23:47] grub-installer: cjwatson * r852 ubuntu/debian/changelog: releasing version 1.49ubuntu10