[07:14] ubiquity: superm1 * r3664 ubiquity/ (debian/changelog scripts/mythbuntu/mythbuntu_install.py): [07:14] ubiquity: Drop old hack for reconfiguring mythbuntu-default-settings since [07:14] ubiquity: this is fixed in user-setup now. [13:14] debian-installer: cjwatson * r1202 karmic-proposed/ (build/config/common debian/changelog): Use udebs from karmic-security, karmic-proposed, and karmic-updates. [13:15] debian-installer: cjwatson * r1203 karmic-proposed/ (10 files in 3 dirs): Move to 2.6.31-17 kernels. [13:17] debian-installer: cjwatson * r1204 karmic-proposed/ (build/config/armel/imx51.cfg debian/changelog): Move iMX51 images to 2.6.31-107 kernels. [13:17] debian-installer: cjwatson * r1205 karmic-proposed/ (build/config/armel/dove.cfg debian/changelog): Move Dove images to 2.6.31-210 kernels. [13:18] debian-installer: cjwatson * r1206 karmic-proposed/debian/changelog: fix version [13:19] debian-installer: cjwatson * r1207 karmic-proposed/debian/changelog: releasing version 20081029ubuntu70.1 [14:41] ubiquity: cjwatson * r3665 ubiquity/ (12 files in 6 dirs): [14:41] ubiquity: Only restart debconf-communicator when changing language, not every time [14:41] ubiquity: we switch page. This speeds up page transitions quite a bit. [14:51] hrm, perhaps target-config would've been a better place for the apparmor cache generation stuff. [14:51] debconf-communicator> oh, nice! [15:16] hi [15:16] anyone known how to bypass gpg check in preseed ? [15:17] i lost the d-i line :/ [15:19] saispo: https://help.ubuntu.com/9.10/installation-guide/i386/preseed-contents.html search for "gpg" [15:20] good ! thanks cjwatson [15:29] ev: I asked James about the GeoIP ticket today; we've tentatively set a due date of 1st Feb, and he's going to hunt around for somebody to implement it. Will that be time enough for you? [15:29] cjwatson: yes [15:29] great, thanks [15:29] and thanks for chasing it [15:37] hmm, there's something odd about translation handling right now. plugin pages aren't getting translated [15:39] ubiquity: evand * r3666 ubiquity/ (debian/changelog scripts/install.py): [15:39] ubiquity: * Don't try to generate an apparmor cache in oem-config. [15:39] ubiquity: * Slightly better exception printing in scripts/install.py [15:42] I'd put money on it being that the widgets for each plugin page are not added to frontend.all_widgets. I worked around that for removing labels from the focus ring, but didn't think to do the same for translations. [15:43] that was my first guess but it seems not to be true [15:43] aha [15:43] I wonder if this is because each plugin instantiates its own gtk.Builder [15:43] which means that of course self.builder.get_objects() in the frontend doesn't return plugin objects [15:46] * cjwatson tries http://paste.ubuntu.com/358553/ [15:48] not quite [15:52] hmm, that's no good. we do genuinely need a different builder for each plugin due to how signal connection works. [16:09] ev: actually you're right, I was confusing frontend.all_widgets with mod.all_widgets [16:09] ahh [16:13] ubiquity: cjwatson * r3667 ubiquity/ (6 files in 3 dirs): [16:13] ubiquity: * GTK frontend: [16:13] ubiquity: - Check plugin builders as well when adding widgets (setting their [16:13] ubiquity: names, making them callable by the toplevel, etc.). This fixes [16:13] ubiquity: translations of the language, timezone, and keyboard pages. [16:15] ubiquity: cjwatson * r3668 ubiquity/ubiquity/frontend/ (gtk_ui.py kde_ui.py): spacing [16:55] ubiquity: cjwatson * r3669 ubiquity/ (3 files in 2 dirs): Fix immediate retranslation when changing language. [16:57] ubiquity: cjwatson * r3670 ubiquity/ubiquity/frontend/kde_ui.py: spelling [16:57] ubiquity: cjwatson * r3671 ubiquity/ubiquity/frontend/kde_ui.py: show_browser died a long time ago, and I think this was a GTK thing anyway [18:06] ev: does http://paste.ubuntu.com/358609/ feel right to you? this (a) changes regain_privileges/drop_privileges to keep a semaphore, so that if you call regain/drop within a regain/drop pair funny things don't happen (b) makes drop_all_privileges permanent so that you get an assertion if you try to call regain/drop thereafter (c) adds a context manager so that you can do 'with raised_privileges(): BLOCK' [18:07] * ev looks [18:07] (c) sometimes results in greater nesting, but if we use it consistently then we avoid forgetting about putting drop_privileges in a finally: block and that sort of thing [18:18] looks fantastic [18:18] thanks for doing that [18:18] (and hooray for python 2.6) === dpm is now known as dpm-afk [19:32] ev: ok, cool. 2.5, BTW :-) [20:01] err yeah [20:42] Hi guys, I'm trying to preseed an installation on hardware raid and can't seem to find the d-i line to skip the "Do you want to enable RAID" question during the install. Any ideas? [20:59] ubiquity: cjwatson * r3672 ubiquity/ubiquity/frontend/kde_ui.py: use regain_privileges rather than writing out the same code [21:05] ubiquity: cjwatson * r3673 ubiquity/ (debian/changelog ubiquity/misc.py): [21:05] ubiquity: Add a semaphore to misc.regain_privileges and misc.drop_privileges, so [21:05] ubiquity: that a drop/regain pair always returns to the previous state. [21:05] ubiquity: misc.drop_all_privileges is now more clearly one-way, enforced by an [21:05] ubiquity: assertion. [21:08] cjwatson, is there a reason that /usr/share/apt-setup/release-files/archive.ubuntu.com contains tons of files from older releases on current media? [21:08] ubiquity: cjwatson * r3674 ubiquity/ubiquity/ (debconfcommunicator.py i18n.py): factor out unnecessary subprocess_setup functions that just call misc.regain_privileges [21:16] superm1: we could probably ditch those [21:33] ev: oh yeah, you had to use the with_statement pragma in 2.5, hmm [21:34] apt-setup: cjwatson * r180 ubuntu/ (61 files in 23 dirs): Remove Release files for old releases. [21:36] ah, whoops. Are you already working on that or shall I add the from future imports? [21:43] we could just require 2.6 [21:43] I don't particularly mind that [21:43] I'm fine with that as well [21:44] I found a neater way to do some of this too [21:44] @raise_privileges [21:44] def func(...): [21:44] ... [21:45] does mean you have to carefully distinguish between raised_privileges (context manager) and raise_privileges (decorator); I did find a way to make the same spelling work for both, but the cure was worse than the disease [21:48] heh, very cool though [21:48] I'm a big fan of decorators [21:48] I use one for threads_enter/_leave in usb-creator [21:50] (cure worse than the disease: it required a class with def __init__(func=None). I'm OK with default arguments but I don't like the effect of passing a defaulted argument being to make the class be a completely different category of thing) [21:50] yeah, yikes [21:57] ubiquity: evand * r3675 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): [21:57] ubiquity: GNOME Bug #56070 (can't click button twice) has long since been [21:57] ubiquity: closed. Remove hack. [21:57] Gnome bug 56070 in gtk "Can't click button after setting it sensitive." [Major,Resolved: fixed] http://bugzilla.gnome.org/show_bug.cgi?id=56070 [22:59] cjwatson: is there a particular reason why UTF characters, specifically an ellipsis, are not used in debconf description fields? Digging through my ~/bzr doesn't find any, and my google-fu is failing to find a reason. I can understand why we didn't do it historically, but what doesn't have proper unicode support these days? [22:59] s/UTF/unicode/