[00:05] bdmurray: I hope my comments in bug 1039484 and the rationale for why I don't think this should be done in python-apt make sense [00:05] Launchpad bug 1039484 in ubuntu-release-upgrader (Ubuntu Quantal) "Accents in sources.list breaks do-release-upgrade" [High,In progress] https://launchpad.net/bugs/1039484 [00:05] bdmurray: You mention "two places" in the MP - what's the other? [00:21] robert_ancell: I responded to your question on gnome-icon-theme-symbolic [00:25] mfisch, my point is neither the gnome-icon theme git log (http://git.gnome.org/browse/gnome-icon-theme/log/) or tarball show any of these changes for gnome-icon-theme [00:31] infinity, jockey and nvidia-common fixed and re-uploaded [00:33] infinity, I renumbered nvidia-common to 0.2.44.1 which I suspect is more correct since it's an SRU of a native package; the previous upload was 0.2.46... I'm assuming it's not too late to redo the number but if so I can redo it as .46. [00:33] hi bryceh [00:33] * rickspencer3 waves back === fisted_ is now known as fisted [00:43] robert_ancell: sorry, I had it backwards, updating the bug [00:44] robert_ancell: looks like this is only a version change [01:11] bryceh: That should be fine, lemme give them a once-over right now. [01:12] cjwatson: yes, it makes sense that merge proposal was a bit stale. I believe it was also crashing in logging.debug() although it isn't happening with my test now [01:14] It'll depend on the other arguments. In general, in Python 2, attempting to do unicode_value % utf8_byte_string_value will fail. [01:15] (Which logging.debug may do under the covers.) [01:15] bryceh: The fix for #976779 ... Shouldn't the package ship the /var/lib/nvidia-common directory? [01:15] bryceh: Or does hybrid-detect write the directory as well as the file? [01:16] lemme doublecheck [01:17] bryceh: Yeah, looks like it just does an fopen() on the filename, so that's going to fail without the directory being present. [01:17] bryceh: Probably needs fixing in quantal too, if this was a straight backport. [01:19] in setup.py is: ("/var/lib/nvidia-common/", glob.glob("share/last_gfx_boot")), [01:19] infinity, let me check the deb but I think it establishes the dir at package install time [01:19] bryceh: Ahh, I didn't notice that setup.py was creating the path. That'd do it. [01:20] -rw-r--r-- root/root 5 2012-09-20 07:04 ./var/lib/nvidia-common/last_gfx_boot [01:20] yep, it's in the .deb [01:21] Shiny. [01:21] Accepting both, then. [01:21] thanks :-) [01:22] bryceh: Thanks for fixing all the nit-picking. ;) [03:39] robert_ancell: someone registered the wrong upstream project for sonic, I'm going to fix it [03:41] mfisch, yeah, there's a suprising number of those - I don't know if that was people or a script that linked them [03:43] it caused me some confusion, but it's fixed now [03:55] robert_ancell: what's the fix for this? dpkg-source: error: Version number suggests Ubuntu changes, but Maintainer: does not have Ubuntu address [03:55] robert_ancell: there's a control field I think I can use, can't remember what it is [03:56] mfisch, it's just the maintainer field [03:56] we just set them to Ubuntu Desktop if they're maintained by the desktop team [03:58] the last upload was NMU as well [03:58] so it somehow worked [03:58] I did talk to the maintainer BTW [03:59] is there an easy way to unpack the source of a given package to see how it's built? [04:00] without having to setup a whole development environment? [04:00] jetsaredim: apt-get source [04:01] where does it get deposited? [04:01] in your current directory [04:01] and you don't need to use sudo [04:01] mfisch: That warning is implying that you should run "update-maintainer(1)" [04:04] infinity: so that just un-made Bill Cox the maintainer [04:04] mfisch: Bill Cox being the Debian maintainer? [04:05] I assume so [04:05] he's the upstream maintainer [04:05] Yes, well. That's correct. [04:05] Ubuntu-modified packages are maintained by ubuntu-devel. [04:05] I'm still curious how the last upload worked, but I'll make this change [04:06] thanks infinity [04:06] And the original maintainer is set as Original-Maintainer in the control info. [04:06] The last upload may have worked if you didn't have DEBEMAIL set when you built it. [04:06] must have been (it wasn't me who did it) [04:06] Which makes the hard error turn into a warning, for reasons the devscripts maintainer has never convinced me make sense. [04:07] that script set "XSBC-Original-Maintainer", I assume that's right [04:07] Either way, it's Ubuntu policy to change maintainer when we change packages, so the Debian maintainers don't get hassled about things we change/break in Ubuntu. [04:07] mfisch: It does it right, yes. [04:07] its been a while since I wrote/reviewed Makefiles but where would the configure script be called? [04:07] thanks, I was worried that changing it was a rude thing to do maybe ;) [04:07] jetsaredim: usually in ./configure? [04:07] mfisch: No, not changing it tends to be the rude thing. This policy was at the request of Debian. :) [04:08] (Or, rather, many many Debian developers) [04:08] mfisch: no i mean where is that called from [04:08] like where are the options defined for calling configure [04:08] jetsaredim: If you're talking about packaging, in debian/rules [04:09] ah, well it's done durin... [04:09] what he said [04:09] jetsaredim: If you're building by hand, nothing calls configure, you do. [04:11] infinity: i'm trying to figure out what is causing an issue I'm seeing with a given package [04:11] tcdecode which is part of transcode is complaining about not being able to handle mpeg2 [04:11] (in quantal) [04:12] was just trying to see if the configure option had been set to include the --enable-libmpeg2 [04:12] which is seems to have - so now i'm a bit baffled [04:12] jetsaredim: sounds like you'll want to setup a build environment and see how it was built [04:14] mfisch: i'm a little over my depth at this point [04:14] jetsaredim: how did you figure out if that flag was enabled? [04:14] its listed in the rules file [04:14] doko_: more like 0:8 :( [04:14] confflags += --enable-libmpeg2 [04:15] Good morning [04:15] mfisch: ldd seems to think that none of the transcode utils are linked to libmpeg2 [04:16] pitti: thanks for the talk this morning [04:16] jetsaredim: I'm not familiar with that package unfortunately and it's my bedtime, did you file a bug and attach what you figured out? [04:16] mfisch: only thing I can think of is that the configure needs both --enable-libmpeg2 AND --enable-libmpeg2convert [04:17] jetsaredim: that should be easy for you to test [04:17] mfisch: i'm in the process of re-opening an old bug for the same issue [04:17] jetsaredim: do you know how to build it and install build-deps? [04:18] jetsaredim: actually I can build it for you for quantal real quick [04:18] i was hoping to not turn this system into a build box [04:18] if you don't mind that would be awesome [04:18] jetsaredim: whats the package again? [04:18] transcode [04:19] i386 or amd64? [04:19] i think just adding a line to debian/rules to add the --enable-libmpeg2convert might do it [04:19] i'm amd64 [04:19] transcode 3:1.1.7-2build1 is what I'm seeing [04:20] yep - that's the current 12.10 version [04:20] okay, pulling build-deps [04:21] i suppose if i'm going to play around with this stuff i should probably learn more about packaging [04:21] been a while since i messed with any debs [04:21] doko_, infinity: OK, after 8 or 9 tries I give up, and make tests non-fatal on armel/hf [04:29] jetsaredim: you want a binary or package? [04:30] let me upload the package [04:30] mfisch: package works [04:39] jetsaredim: glad that option worked for you [04:39] yea - now i can make my perfectly legal backups of my dvds [04:39] :D [05:15] doko_: Would you please retry ktorrent on armhf in the rebuild. I think it will work now. [06:24] infinity: meh, now glib2.0 failed on armhf during compilation; shall I disable that part as well? :-) [06:24] * pitti just sticks fingers in ears, sings "lalala" and hits "retry" [07:04] good morning [07:04] hey dholbach, guten Morgen [07:04] hey pitti [07:11] What do you guys think of the Amazon Ubuntu Search feature? Should it be included in 12.10? [07:11] I think it should be in the dash search, but not in the launcher as an icon [07:11] sabdfl: good morning, how are you? [07:12] sabdfl: amazon icon in the launcher certainly stirred some "discussion"? [07:12] pitti, that is reasonable [07:12] * pitti mubmles "billboard" [07:12] My amazon icon disappeared as soon as I clicked on it. [07:12] Which was odd. [07:12] yup its call unity lens shopping [07:12] I wonder if the U1music icon will do the same. [07:12] I'm actually still using 10.04 , works well enough for me [07:13] RAOF: amazon icon does the same here, u1 ms WFM [07:13] I am interested in all the new changes [07:13] Hm. Nope. Apparently the U1music icon doesn't do anything, though. [07:13] RAOF: for me it opens firefox with the u1 music page [07:13] (shop) [07:14] Doesn't seem to want to do that here. [07:14] I tried it in a guest session [07:14] penreturns, is there a way to turn it off? [07:15] yup u can remove the package [07:15] my user doesn't have any of the default launchers except nautilus [07:15] sudo apt-get remove unity-lens-shopping [07:15] penreturns, earthling_: Or twiddle the setting in Privacy [07:16] (That'll exist once didrocks' package lands after B2) [07:17] I've never seen so much press about a development issue [07:17] other than Unity, heh [07:17] Or... what was the other one? [07:18] Oh, Banshee-amazon-referral. [07:18] There seems to be roughly one per release :) [07:25] Shuttleworth's discussion on the Amazon Dash issue http://www.markshuttleworth.com/archives/1182 === smb` is now known as smb === mcclurmc_away is now known as mcclurmc [07:54] pitti, sure, not unexpectedly [07:55] we need to do a good job of giving people privacy options, and security [07:55] but i think 12.10 will have all that [07:55] right, I specifically asked about the launcher icon [07:55] then the tin foil hat crowd can go somewhere else, where they will be disappointed by their new overlords i'm sure :) [07:55] pitti, oh, is there an issue with the icon? [07:56] sabdfl: I think the dash search is fine (modulo privacy options blabla, but we have that), I just don't particularly like the icon in the launcher which looks like an "Amazon" billboard [07:57] the u1 music store is still closely related to Ubuntu, but I have a feeling that adding the amazon icon there breaks a dam [07:58] but anyway, I guess that was discussed over and over, so perhaps I STFU to not open up old wounds again === henrix_ is now known as henrix [08:07] pitti: I believe we ship http://start.ubuntu.com containing a Google logo [08:08] sladen: that doesn't clutter up the default desktop, though; not only do we already have too many starters by default, they now also contain non-ubuntu ads [08:09] Is the amazon thing in the default (super only) dash? [08:09] sladen: I hope we at least get some good dimes from amazon for that :) [08:10] there's a webapp icon in the launcher too [08:13] jussi: the Amazon lens, is "just" the Amazon lens---however all lens get combined sourced to the default action [08:14] jussi: eg. a bit like Google searching everything by default; or you can select just images, just usenet, just patents [08:14] aye. [08:15] jussi: and it seems to be this aspect that people maybe concerned about. However, pitti's point is seemingly(?) about the icon being there in-your-face [08:15] pitti: ^^can you clarify? [08:15] sladen: yes, that and that we already abuse the launcher too much and put too many default icons there [08:15] you might calm the flames a bit if you introduce a bunch of others, not just amazon, and do what browsers do - a first run service selection for your dash [08:16] ie. when I install chrome, it asks me which search engine I want [08:17] pitti, oh, fighting this battle again? ;-) [08:17] seb128: I didn't really see the previous ones; I just raised my eyebrow when I saw the new icon yesterday [08:17] I clicked it once to see what would happen and then it disappeared on its own and never came back :P [08:17] seb128: the "overfull launcher", yes, that comes up all over again and nobody seems to care [08:18] pitti: design argued that the concertina effect is there for just this purpose, so it's not a problem [08:18] the one positive suggestion, is that below the installer MP3 tick box we have another default-ticked "enable handy online stuff [you probably want this too]" [08:18] (the one positive suggestion that I've heard) [08:19] https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1054167 that's this [08:19] Launchpad bug 1054167 in unity (Ubuntu R-series) "Unity launcher for Quantal Beta 2 has too many items, scrolls off the screen on 1366x768" [Undecided,Confirmed] [08:19] sladen: yeah, I like that suggestion [08:29] doko_: https://launchpad.net/~laney/+archive/webkit2/+packages === mcclurmc_away is now known as mcclurmc === tkamppeter_ is now known as tkamppeter [08:40] sladen: while it might work out to be a nice approach, it's punting a UIFE off onto foundations and it's a fair number of moving parts to get settled down in time for final freeze ... [08:40] so, well, if that turns out to be the answer I just hope it occurs to somebody that we need actual notice of this kind of thing and some time [08:41] this is exactly why this should have landed before FF, to allow time for consequential changes elsewhere *grump* [08:42] mpt: what are your thoughts on including a projection line on the graph? Something akin to this http://baykeeper.org/data_viz/official-sea-level-rise-projection [08:45] ev, look at the Quantal graph. Imagine a projection line on that. It would be slowly accelerating upwards, hitting about 0.15 on release day. [08:46] call me pessimistic, but that's what I expect will happen [08:46] heh [08:46] ev, we know that's not going to happen because we humans know stuff about the release process that a projection line does not. [08:46] :) fair enough === Zdra is now known as xclaesse [09:03] ev: Do you have any idea why https://errors.ubuntu.com/bucket/?id=%2Fusr%2Fbin%2Fsoftware-properties-gtk%3AKeyError%3A%3Cmodule%3E%3A__init__%3Ashow_drivers is showing up in 0.92.6? That line of code was changed in 0.92.4 [09:03] (and should have fixed that bug) [09:04] I could understand that versions might be out of sync if it were a daemon process or something, but I wouldn't have expected software-properties to be long-running [09:04] cjwatson: it's still a pretty low count, so it could be the "they upgraded while the program was still running" problem [09:04] we filter those out for crashes sent to Launchpad, but we're still working through the solution for daisy: [09:05] mm, right [09:05] https://bugs.launchpad.net/ubuntu/+source/apport/+bug/1039220 [09:05] possible I suppose [09:05] Launchpad bug 1039220 in apport (Ubuntu Quantal) "don't report crashes for programs that don't match the file on disk (like for kernel crashes)" [Medium,In progress] [09:05] and https://code.launchpad.net/~brian-murray/apport/no-unreportable/+merge/126082 [09:06] ev, do you know when the next rollout will be? [09:07] ev: that bug appears to be waiting for you to follow up :-) [09:07] mpt: of what feature? :) [09:07] ev, the graph mainly [09:07] (or maybe I'm misreading) [09:07] cjwatson: I followed up in the merge proposal, but I'll note that in the bug [09:07] mpt: I'll file and RT and ask for a rollout today [09:07] an RT* [09:07] ok [09:09] ev, btw did you realize the legend has disappeared? [09:09] mpt: I made it do so, per your mockup [09:09] I need to implement moving it to the bottom [09:20] mpt: since we will soon be able to show graph and table data for just packages in -proposed, I'm questioning the need for being able to filter to arbitrary versions of specific packages [09:20] well, questioning it a bit more [09:21] and of course we already split out release by release data [09:21] I guess it helps because we haven't quite solved updates [09:22] err never mind, you'd see what's important in the normal view there [09:26] Off switch being made for the shopping lens http://www.omgubuntu.co.uk/2012/09/is-an-off-switch-for-the-shopping-lens-in-the-works [09:40] jdstrand, infinity: how are we supposed to get lpia builds for hardy updates these days? our only buildd seems disabled [09:48] can I temporarily recruit an i386 or amd64 builder to build https://launchpad.net/ubuntu/+source/postgresql-8.3/8.3.21-0ubuntu8.04/+build/3857919 ? [09:48] Yeah, I think that should work - feel free to try [09:51] ok, trying on aatxe [10:01] cjwatson: nice, that seems to work; it's past the tests in the debhelper stage [10:02] oddly reassigning a PPA builder didn't seem to work [10:02] (I asked on #launchpad-ops) [10:02] * pitti reassigns aatxe back to i386 duty [10:03] did you have to set it to manual and back or anything like that, or did you just change the processor type? [10:04] cjwatson: I did set it to manual first, but more out of a habit [10:04] to set it back I just changed the arch (it was idle then) [10:04] cjwatson: you can tell a PPA builder to build ubuntu packages? [10:05] No, to ppa/lpia [10:05] I'll try manual/auto [10:07] cjwatson: ah; we do have a working lpia PPA builder, though [10:07] we do? gold is disabled [10:07] cjwatson: so, it worked without manual in one direction [10:07] gumiho is the one I'm trying to reassign [10:08] aah, I see; that's the one I was seeing [10:08] it says lpia but is persistently idle [10:08] I guess I don't care that much; I'll reassign it back in a bit if it doesn't do anything [10:08] aatxe picked up the waiting job within seconds [10:08] yeah, I saw [10:08] so, no idea what's wrong here :( [10:09] * cjwatson is looking forward to b2 being out so that he can stress-test the new queue-accept-should-never-time-out thing [10:10] is there a target time when we unfreeze today? [10:10] not that I know of [10:10] i. e. the point of no return^H^H^H^Hspin [10:12] that reminds me, christmas wish #23: a queue tool option to accept a "quantal" upload into "quantal-proposed" instead, so that stuff can already build and get autopkgtest-ed [10:12] once we have britney operational we'll just redirect everything from devel to devel-proposed [10:12] but that'll soon fix itself anyway when all uploads get built in -proposed [10:13] right :) [10:13] https://jenkins.qa.ubuntu.com/view/Quantal/view/AutoPkg%20Test/ tests -proposed now [10:13] * cjwatson wonders how hard changing the suite in the queue would actually be [10:14] It *might* just be a matter of making PackageUpload.pocket (and PU.distroseries?) writable [10:14] tests would be an absolute git to write though [10:15] probably not worth the effort; it just came to my mind again === cpg is now known as cpg|away [10:36] doko_: do you plan to do any build log analysis / bug filing from the rebuild test? === ion_ is now known as ion === MacSlow is now known as MacSlow|lunch [12:10] Hello, please help me, I want to update a package to new version for quantal, can I do this or that is incompatible with feature freeze? [12:25] MohamedAlaa98: there needs to be a good reason for it, and it should be thoroughly tested, etc. [12:27] thumbleweed: please explain it for me, what do you mean by "a good reason"? [12:28] tumbleweed: * [12:28] MohamedAlaa98: it needs to be worth the risk [12:29] Hi.. do you think it's still possibile add youtube-dl 2012-09-27 in quantal? [12:30] l3on: any particular reason? [12:30] tumbleweed: okay thanks :D === _salem is now known as salem_ [12:30] tumbleweed, first reason I found: there's a bug downloading playlist .. something in youtube has changed in last months [12:31] l3on: sounds good. file an ffe [12:31] tumbleweed, great!. [12:32] pitti, cjwatson: so afaik, you need to do what you did. I don't have the perms to do it, so I don't know the exact steps [12:34] jdstrand: yeah, it's all right, sorted now [12:34] didrocks: hi! do you know if there is a bug already for 'ctrl+lt+t' not opening a terminal. if not, what package should that be filed against? [12:34] didrocks: google is not helping me... :\ [12:35] jdstrand: yeah, it's because of duplicated keys on compiz and g-s-d [12:35] jdstrand: there is a fix coming with next compiz [12:35] jdstrand: but you can workaround it for now [12:35] open gnome-control-center -> keyboards -> keybindings [12:35] section Launcher [12:35] IIRC, "Launch a terminal" is the one which is bound to the actual compiz code, ensure it's set to ctrl + alt + t [12:35] "Launch a Terminal" should be unset [12:36] jdstrand: once the next compiz will be here, be aware that "Launch a terminal" will disappear and you will have to set "Launch a Terminal" then to ctrl + alt + t (as you would have unset it) [12:36] didrocks: thanks! after the compiz update, will I have to set that back? [12:36] oh, heh [12:36] cool [12:37] didrocks: thanks again, this was mildly annoying :) [12:37] jdstrand: sorry for the trouble :) [12:37] didrocks: not a big deal, I just wanted to make sure it was filed/fixed [12:38] hi cjwatson [12:38] jdstrand: it's a typical case where 2 > 1 isn't a real advantage :) [12:39] :) [12:40] cjwatson: when grub-pc is installed + grub-efi-amd64-bin then grub is installed only for efi but with config for mbr [12:40] cjohnston: thats wrong, tested with debian exp -7 [12:42] cjohnston: you definitely have to modify grub-install [12:45] err cjwatson always ;), the efivars check is not optimal alone, when you use a different bootloader like gummiboot or direct uefi boot but you want to use grub for selection of other legacy systems like win [12:48] tumbleweed, these are changes in youtube-dl.. what you think about it? is it better apply only patch rather than bump a new version of package ? [12:51] tumbleweed, sorry - these are the changes -> http://paste.ubuntu.com/1230335/ ^ === doko_ is now known as doko [12:53] Laney, the analysis for main and package sets was done. asked amoog for filing the universe bug reports. however the universe rebuild will take time until after the weekend [12:54] didrocks, seb128: ping about the gnome/gtk header include fixes ... [12:54] doko, what about those? [12:54] see backlog [12:54] didrocks, seb128: is there an easy work around for #warning "Including directly is deprecated." ? [12:54] don't include it directly? [12:55] there are a lot of these in universe [12:55] seb128, please fix [12:55] what? universe buggy apps? [12:55] sorry I've enough work on the default installation issue and that's higher priority than universe random packages [12:55] darktable, gstreamermm, subtitleeditor are those in package sets [12:56] well, whoever is owning one of those set can look at them I guess [12:56] darktable is already fixed (waiting in queue) [12:57] I'm trying to through the rebuild results and fix them when I see them [12:57] geser, thanks [12:57] honestly such changes should told once new gtk/gnome changes get uploaded [12:57] be told [12:58] it's hard to know what has an impact or not, they did similar change to lot of libs over time, I don't want to spam lists with every change GNOME does to its libs [12:59] yeah, so better leave it to other to detect and fix [13:00] it's not like you were emailing the list about every single toolchain change either... [13:00] see the stl abi issue that didrocks has to track down earlier in the cycle [13:05] cjwatson: next bug: get rid of -w for efibootmgr, it just overwrote my hd uuid, that means my win install does not boot (mbr), my hd is mbr with efi as primary partition. -w kills win [13:08] seb128, bullshit. the known toolchain changes were announced, with hints how to fix. the abi issue was not known [13:08] doko, the include issue was not known either [13:08] you call yourself a tech "lead", so please be proactive [13:08] doko, I didn't call myself that [13:09] seb128, I doubt it, it did pop up for precise as well, and afairc it was reverted [13:09] doko, that was in glib [13:09] no librsvg [13:09] which has another scale and was handled properly [13:09] right, see subtitleeditor === salem_ is now known as _salem [13:09] and gstreamermm [13:09] right, you come with like 5 rdepends [13:09] can't we just fix them and move on? [13:10] it doesn't seem worth the discussion [13:11] no, I can't fix hundreds [13:12] is it really hundreds? [13:12] we should probably revert the librsvg change if that's the case [13:13] start with glib if you really do care [13:13] $ grep-available -s Package -F Build-Depends librsvg2-dev /var/lib/apt/lists/*quantal*Sources | sort | wc -l [13:13] 67 [13:14] I don't see how it can be "hundreds" [13:14] when there is less than 100 sources using librsvg [13:14] what about glib? [13:15] there are hundreds of build failures. if you address those which you introduce with new upstream versions yourself, ... it would help a lot [13:15] again see subtitleeditor and gestreamermm [13:15] I've no time for universe sorry [13:16] I can file bugs asking those to be removed from quantal if that helps [13:17] being productive today? [13:17] I was until you started that IRC discussion === MacSlow|lunch is now known as MacSlow [13:17] I'm sorry but it's just not realistic to expect me or whoever upload glib and gtk to be responsive for ever years old unmaintained universe source [13:18] I'm happy to get those dropped from the archive if nobody has time,motiviation to fix them though [13:18] that's the best I can offer you today === _salem is now known as salem_ [13:19] that said, going back to what I was working on === Guest32361 is now known as slank === slank is now known as Guest2828 === salem_ is now known as _salem [13:39] slangasek, wonder if you have a thoughts [13:40] if an interface is brought up in the kernel (ipconfig) [13:40] then when network-device-added comes through, it is 'ifup' will fail, and not call the upstart hooks for ifup [13:41] so static-networking wont fire. even though it already is set. [13:42] open-iscsi has some "helper" stuff to get this fired for the device that its iscsi root is on. but i'm wondering if there should be a more general solution. [13:46] seb128, doko: a fixed gstreamermm and subtitleeditor are waiting in the queue [13:46] geser, thanks a lot [13:47] smoser: that sounds like bug 1010439 [13:47] Launchpad bug 1010439 in ifupdown (Ubuntu) "ifup fails prematurely if the same static IPv4 address was configured per kernel command line" [Undecided,Incomplete] https://launchpad.net/bugs/1010439 [13:47] smoser: any such interface really ought to be marked as "inet manual" though, that way ifupdown should do the right thing [13:47] stgraber, yes, yes it does. [13:48] dpm, hey [13:48] dpm, do you know why https://translations.launchpad.net/ubuntu/quantal/+source/unity-lens-shopping/ is empty ? [13:48] stgraber, the issue with 'inet manual' is that it wont apply to static-networking [13:48] when imo, it should [13:49] (ie, if that nic is up from the kernel, and its the only one in /etc/network/interfaces as 'auto', then static-networking should fire) [13:49] seb128, hm, it seems there hasn't been any .pot file imported, I was wondering about this the other day, but at that point there wasn't a source package uploaded yet. Maybe the uploaded package didn't create a .pot file? https://translations.launchpad.net/ubuntu/quantal/+source/unity-lens-shopping/+imports [13:50] dpm, the build log https://launchpadlibrarian.net/116724481/buildlog_ubuntu-quantal-i386.unity-lens-shopping_6.0.0-0ubuntu1_BUILDING.txt.gz has [13:50] "Running xgettext --add-comments --directory=. --default-domain=unity-lens-shopping --flag=g_strdup_printf:1:c-format --flag=g_string_printf:2:c-format --flag=g_string_append_printf:2:c-format --flag=g_error_new:3:c-format --flag=g_set_error:4:c-format --flag=g_markup_printf_escaped:1:c-format --flag=g_log:3:c-format --flag=g_print:1:c-format --flag=g_printerr:1:c-format --flag=g_printf:1:c-format --flag=g_fprintf:2:c-format --flag=g_sprintf:2: [13:50] c-format --flag=g_snprintf:3:c-format --flag=g_scanner_error:2:c-format --flag=g_scanner_warn:2:c-format --output=unity-lens-shopping.pot --files-from=./POTFILES.in.temp --keyword=_ --keyword=N_ --keyword=C_:1c,2 --keyword=NC_:1c,2 --keyword=Q_ --keyword=g_dgettext:2 --keyword=g_dngettext:2,3 --keyword=g_dpgettext:2 --keyword=g_dpgettext2=2c,3 --from-code=UTF-8 [13:50] Removing generated header (.h) files...done. [13:50] Wrote unity-lens-shopping.pot" [13:50] ups [13:50] sorry, didn't realize the xgettext line was a long one [13:50] seb128, I see it was in universe and then went to main. Could that have affected it? [13:51] dpm, it could I guess [13:51] * dpm builds package locally to see what translations look like [13:51] smoser: hmm, my understanding is that if you configure network from the initramfs and have the interface marked as "inet manual", then when network-device-added is received, the ifup will succeed and statis-network-up will be emitted (assuming it's the only interface) [13:51] didrocks, ^ do you plan an upload soon for unity-lens-shopping? [13:51] seb128: there is one start of next week [13:52] I can make a dummy upload if needed [13:52] to have it rebuilt [13:52] didrocks, see other channel [13:52] smoser: anyway, an alternative would be for the initramfs script to do: "mkdir -p /run/network/ && touch /run/network/ifup.$INTERFACE" [13:53] didrocks, ignore that ... do you want me to do the dummy upload? [13:53] smoser: oh, and you'd also likely want "echo $INTERFACE=$INTERFACE > /run/network/ifstate" so that ifupdown doesn't try to bring it up again [13:53] seb128: fine with me :) [13:53] didrocks, ok, doing it [13:54] but that sounds like a hack to me, having the interface configured as inet manual should do the trick... [13:54] seb128: the packaging branch in lp:ubuntu/unity-lens-shopping :) [13:54] didrocks, thanks [13:54] thanks to you : [13:54] :) [13:54] seb128, hm, I can't seem to be able to build the local package, I'll leave it to you guys and the new upload [13:54] stgraber, i'll see if manual works. [13:55] you think it will get bounced? [13:55] dpm, sure [13:56] stgraber, i think you're right. manual should work. [13:56] smoser: with manual, the interface shouldn't be touched but the if-up scripts should be run, which is what I think we want [13:56] dpm, thanks [13:56] IIRC that's what I implemented in casper for netboot installs and what we have in LTSP, so if that doesn't work, we need to figure out why and fix it :) [13:56] stgraber, shoot. the one other thing is that i want resolvconfhooks to fire. [13:56] they will, right? [13:56] yep [13:57] they are if-up hooks, so they should still be called [13:58] stgraber, http://bazaar.launchpad.net/~cloud-initramfs-tools/cloud-initramfs-tools/trunk/view/head:/dyn-netconf/scripts/init-bottom/cloud-initramfs-dyn-netconf [13:58]