[01:00] RAOF: you using mk-sbuild for btrfs snapshot-based chroots much? [01:00] Yes. [01:00] It's my default setup now. [01:00] I was trying to build new ones today and ran into issues with running the finish script [01:00] I've just started using that setup ;) [01:00] Hm. I think that last time I ran mk-sbuild I had a problem with the finish script. [01:00] looks like the source chroot can't be referred to as $chroot_name-source anymore [01:01] Yeah, that'd be right. [01:01] ok [01:01] in that case, I'll send a merge request for ubuntu-dev-tools to fix this [01:01] I think. [01:02] What? chroot-source no longer works? What are we supposed to use then? [01:03] TheMuso: source:chroot [01:03] * TheMuso sighs. Schroot is changing too much for my liking, thanks. [01:04] TheMuso: looks like it's not actually a very new change, just something that wasn't really made to be used in mk-sbuild [01:06] it only appears to me as being an issue when you try to use the lvm or btrfs based chroots [01:14] Right. [01:21] grr the way script-config works has changed too... [01:21] Its now config-profile or some such, and it has to be a directory, and not a single file. [01:22] I was using a script-common-build to augment the mount script to add an additional fstab line for a build LV. [01:22] In the cases where I don't use tmpfs for building. [01:26] yeah, I also was using a script from the security team's sbuild howto, and that was exploding in my face [01:27] yup [01:27] What did you have to re-implement to get the same functionality back, or did you just remove it for now? [01:28] I just dropped it, but I suspect it probably works kind of like a simple list of scripts to run [01:29] anyone using UEFI on precise here? [01:29] TheMuso: the script I was using was just to get ddebs, so it wasn't critical [01:30] TheMuso: Hm. I think I've always used the config-profile stuff [01:30] Why would you want ddebs in the chroot? [01:30] Oh, unless it's for reproducing bugs. [01:30] RAOF: getting them out of the chroot [01:30] hrm ok, I used a file that Kees wrote to use an LV for package building, i.e /var/cache/buildtmp, and a dir was created inside that for every schroot instance to host files that was bind mounted to /build. [01:31] So I could build multiple packages at once with the files on disk. [01:31] Now I use tmpfs for all smallish stuff, but still, its nice to have around. [01:31] cyphermox_: mk-sbuild by default installs pkgbinarymangler, which strips out the ddebs and dumps them in ~/Builds [01:31] RAOF: perhaps you know: do you have a trick to turn on and off universe when building packages? [01:31] * TheMuso drops for now. [01:31] cyphermox_: I do not, no. [01:31] RAOF: ddebs: not sure what the exact idea was, again, jsut stolen from security without thinking much. Perhaps it was a script to steal my passwords ;) [01:32] cyphermox_: Actually, I think it'd be pretty easy to do a main-only builder config... [01:32] yes, it would [01:33] I haven't *got* one, but I can see how it'd go ? [01:33] but switching between both so that I don't need to have twice as many schroots is what I'd like to do :) [01:33] Oh, that's what I mean. [01:33] or without having to go muck around in the source chroot manually every time [01:33] sbuild runs apt-get update by default, so all you need to do is replace sources.list pre-build. [01:33] RAOF: yes, I'm just refining some sort of idea of how to do it [01:34] http://paste.ubuntu.com/1026080/ in /etc/schroot/setup.d/60add-local-repository is almost all the way there. [01:35] indeed [01:36] Oh, but you need to remove the fatal with the new schroot. [01:36] where does this file come from? [01:36] It's what I cooked up to handle build-against-built-packages. [01:37] Really I should fold this into mk-sbuild, because it's broadly useful. [01:38] yup, really useful, thanks! [01:41] cyphermox_, RAOF: fyi, the security team uses a tool called 'umt' (in a bzr branch) that among a lot of other things, will handle building with whatever components you want (with sane defaults that mimic the buildds) https://wiki.ubuntu.com/SecurityTeam/BuildEnvironment [01:42] people are aware of that page, but often they stop at the umt step... [01:42] jdstrand: thanks. I just found that umt was getting in the complicated side [01:42] jdstrand: yeah, that's what I did [01:42] oh? [01:42] umt build [01:43] that's it :) [01:43] * cyphermox_ goes back to reading [01:43] setting it up, I mean [01:43] I see. maybe-- I know we tried to make that easier. feel free to ask if you have questions or the directions are unclear [01:43] but far more than that I wish I didn't have to put files in /etc; but it seems like most of these steps are no longer necessary [01:44] jdstrand: I'll review and do a honest try at umt and let you know ;) [01:44] cool! :) [01:45] the whole team has standardized on it so we can do reproducible builds. but beyond that, there are neat things like 'compare-log' and 'compare-bin' [01:45] and 'check' [01:45] all that really helps make sure the packaging is sane and to verify you changed what you meant to change [01:46] anyhoo, you are already sold on trying it, so I should stop selling it :) [01:50] sigh :( [01:50] ubuntu server does not like me [01:51] jdstrand: I was already referring to your page for setting up the chroots, just not using umt [01:52] jdstrand: though now I've scripted my system setup to a point where I almost don't have to do anything [01:53] cyphermox_: scripted the setting up a system to do builds or scripting the builds? [01:53] setting up my development system(s) [01:54] next will be setting up my "local buildd", when I have spare time [01:55] yeah-- we could definitely script setting up umt, but it just hasn't been a priority [01:55] I have a build VM that I use when I am building untrusted sources. it could be useful to script all this so others could do the same... [01:56] jdstrand: well, what I used to have was a setup with reprepro, pbuilder and rebuildd to get to a point very much like a local PPA [01:57] that's kinda cool. I didn't bother with any of the ftp upload business [01:57] but I'm using sbuild now instead of pbuilder, so since that has been torn down in a reinstall some time ago, I haven't taken the time to set it up again [01:58] * jdstrand nods [01:58] it should be trivial to fix up rebuildd to not require pbuilder [01:58] we like sbuild because it is what the buildds use (though I think they still use a really hacked up version based on an ancient script [01:58] ) [01:59] but then we use umt to do all the buildd hacks like making sure main only builds with main, restricted only witn main and restricted, -security only with release and -security, etc [02:31] Can anyone remember what GPU Robert has on his primary devel system? It's intel, right? [02:32] Robert as in Sarvatt? [02:43] Robert as in Robert Ancell. [02:43] I'm just seeking to be lazy in the construction of my system-compositor PPA. [02:44] RAOF: I *think* last I read that he had radeon, but that may be incorrect. [02:44] Hm. That actually rings a bell. [02:47] If so, boo. It's the least well tested one :/ [03:51] Good morning [03:52] Morning pitti. [04:05] morning pitti [04:14] Whoops, wrong window. === jalcine is now known as Jacky === zyga-afk is now known as zyga [06:00] good morning === zyga is now known as zyga-gone [06:28] hey, somebody able to look at #1008682 ? (that's a dup with a better description) [06:28] it's possible that it happens with xfwm (xfce) only, but it would be nice to confirm that first [07:39] morning mvo [07:53] hey glatzor, good morning! [07:54] glatzor: you mentioned apt-key and chroot support, we can add that to the apt apt-key [07:55] mvo, a -o parameter that would be add to the apt-config call? [07:55] mvo, an -o parameter that would be added to the apt-config call? [07:57] mvo, but this would not result in parsing the apt configuration of the chroot [07:58] glatzor: hm, hm, could we simply do a "chroot /path/to/the/chroot apt-key foo" ? [07:59] glatzor: for in code: pid=os.fork(); if pid == 0: os.chroot(); subprocess.call() ? [07:59] mvo, this would require that the chroot contains a vaild apt installation? [07:59] glatzor: meh, indeed [07:59] glatzor: you can set APT_CONFIG in the environemnt, that should get around the config parsing problem [07:59] mvo, but it would only load the main config and not the snippets [08:00] mvo, if the main config isn't available it will even fail completely [08:00] mvo, furthermore the APT_CONFIG config is only loaded additionally to the configuration of the main system [08:00] glatzor: right, we would have to make sure there is a main config that is valid and that contains Dir::Etc::parts to point to either a empty dir or to the chroot snippets [08:01] mvo, dir::etc::parts would have to specify an absolute path then [08:01] glatzor: right, and if APT_CONFIG contains Dir::Etc::main then the default conf would not be read [08:01] glatzor: indeed, its a bit cumbersome, but possible, not sure if its worth the hassle :/ [08:02] glatzor: but I do like the idea of not dulicating all the gpg driving code in two places [08:02] mvo, rights perhaps a chroot is the cleanest approach [08:02] mvo, we also require this for dpkg [08:02] in python-apt [08:03] mvo, and in a test suite i could drop a fake script [08:05] mvo, would this be ok for you? [08:08] hey [08:08] glatzor: yes, that sounds ok, so we would generate the config on the fly basilcy in the _TMPDIR ? [08:16] salut seb128, ça va ? [08:17] lut didrocks, ouais, et toi ? [08:17] seb128: nickel ;) === tkamppeter_ is now known as tkamppeter [08:30] mvo, hm. not exactly. I would just call apt-key with the chroot and it would have to get the config from the chroot completely [08:33] subprocess.Popen(["chroot", "apt-key", "add", "BLIBLALBU"], preexec_fn=lambda: os.chroot("/tmp/mychrot")) [08:33] mvo, oh wait. then we cannot read the key file anymore :) [08:34] mvo, I think that you used gnupg calls directly since apt-key doesn't any keys on stdin [08:36] mvo, what about tmpconfig.write(apt_pkg.config.dump()) os.environ["APT_CONFIG"] = tmpconfig_path ? [08:37] mvo, that way we could make sure that apt-key uses the same config as the apt.cache.Cache() [08:38] glatzor: sounds sensible [08:38] good morning everyone [08:40] mvo, the wait statement is ok? [08:41] mvo, I would also like to add a parse_list_keys_output() to allow an async list_keys() [08:41] is poppler-data on desktop's plate ? can you look at bug 1009052 ? [08:41] Launchpad bug 1009052 in poppler-data "CJK Installation fails with error: poppler-data : Breaks: cmap-adobe-gb1 (<= 0+20090930-2)" [High,Confirmed] https://launchpad.net/bugs/1009052 [08:43] mvo: Should the path of the apt-key binary be configurable? Dir::Bin::Apt-key? [08:43] glatzor: I think so [08:43] chrisccoulson, hey, did you manage to avoid the queen for the whole w.e? ;-) [08:43] glatzor: "the wait statement is ok"> which one was that? sorry, multitasking is hard :) [08:44] seb128, yeah, pretty much :) [08:45] mvo, the functions in the auth module feature an additional wait statement which is false by default. if true the method will return the subprocess.Popen instance of apt-key. I would like to poll the subprocess from the gobject main loop in aptdaemon to not block on the call [08:45] hey chrisccoulson! [08:45] mvo, but on the other side I could fork in the worker [08:45] mvo, perhaps this is cleaner. since we don't have got any wait statements somewhere else in python-apt [08:46] * glatzor tends to write 1000 things in one APIs :) [08:58] glatzor: yeah, having it in the worker would be better I think as it makes the API easier === zyga-gone is now known as zyga [10:40] http://www.uberlin.co.uk/what-i-know-about-germans/ <- if there ever is a UDS in germany, this is the survival guide ... [10:58] Sweetshark, that seems rather advertisement than a guide :p [11:00] well the myth about punctuality for example can be dispelled by me acting as a counterexample ;) === MacSlow is now known as MacSlow|lunch === MacSlow is now known as MacSlow|lunch [12:48] good morning === MacSlow|lunch is now known as MacSlow [13:01] bcurtiswx, hey [13:01] hi seb128 [13:02] chrisccoulson, https://launchpad.net/ubuntu/precise/+source/firefox/+bugs [13:02] chrisccoulson, is that list of test issues uptodate and do you really want to track them for precise still? === Ursinha` is now known as Ursinha [14:44] jibel: sorry, running out of time for the zh image fix, I'm afraid; something for next week [14:44] jibel: feel free to assign the bug to me [15:00] anyone know why the max resolution Ubuntu sees on a desktop is 1600x1020 4:3 but the monitor has 2560x1600 16:10 [15:00] im using the NVIDIA driver [17:08] hi [17:08] i have installed ubuntu 12.04 on a few intel desktops allready and i still have the same problem, my system is a intel i7 3930k and asus p9x79 deluxe motherboard, the other intel desktop i installed it on was a i7 920 with the x58 chipset mobo. They both had the same problem: [17:08] the problem is that i cant click/select some buttons/windows [17:08] it's like the gui is freezing but its not if i check in a terminal [17:08] everything works perfectly if i go non-gui [17:08] The same problem persists if i install the nvidia drivers so i dont think its gpu related [17:08] I've searched google for hours now and i installed several other linux distro's, they all had the same problem [17:08] any chance i can fix this? [17:11] bryceh: I am trying to test bug 825624, but I am confused by (1) how I select latin; how I see it working. Can you help? [17:11] Launchpad bug 825624 in xkeyboard-config "patch: added dead_hook and dead_horn to latin keyboard layout" [Low,Fix committed] https://launchpad.net/bugs/825624 [17:16] hggdh, yeah one sec [17:22] hggdh, I've updated the bug description with a more detailed set of steps. [17:35] jkbe: this isn't really a help channel. have you filed a bug report about it? [17:39] i did a few months ago but it's dead now, there are many bugposts about it but not a fix [17:41] ok [17:41] is there any channel where i can go with this problem? [17:56] jkbe: #ubuntu is the main help channel. i think there might be an ubuntu-xorg channel too perhaps, which would be a better place [18:10] Sarvatt, you there? [18:38] mvo, hello. So one of my work items is to split out the release-upgrade stuff in update-manager to a separate source. Did you have a preferred name for that? Left to my own devices, I might name it "ubuntu-upgrader" but welcome whatever thoughts you have [18:41] mterry: yeah, something like ubuntu-release-upgrader sounds good to me [18:42] mvo, ok [18:51] bryceh: ả does not look like a horned 'a' to me. Is this what I should see? [18:58] Greetings... So if I wanted to talk with somebody about Network Manager, who would be the right person? :) [18:59] BigWhale, cyphermox_ [18:59] hey hey [18:59] awesome! [18:59] mterry: thanks, I was just about to throw him under the bus [18:59] hggdh, I see some sort of a with something on top. Not sure what horned means? [19:00] thanks [19:01] hggdh: what I see is 'a' with a kind of cedila on top [19:02] jcastro: thanks for the support :) [19:02] mterry, cyphermox_ : this is bug 825624 -- adding Vietnamese accents [19:02] Launchpad bug 825624 in xkeyboard-config "patch: added dead_hook and dead_horn to latin keyboard layout" [Low,Fix committed] https://launchpad.net/bugs/825624 [19:02] sweet fun :D [19:03] cyphermox_, anyway, it appears that there is some race condition with networking and VPN. you connect to a wifi, then VPN starts to connect, but before VPN connection is established, some applications can already connect to the network, bypassing the VPN [19:03] BigWhale: that's expected [19:04] BigWhale: there is no connected-to-vpn signal, just a connected signal [19:04] yes, I know. but I was thinking if there is anything that can be done about it [19:04] right now, not without a lot of work [19:05] BigWhale: fortunately, that work is "kind of" planned already; there needs to be some rework to support having connections depend on other connections and prioritization [19:06] hello desktopers [19:07] chrisccoulson, hi :), how come that there is already a firefox 14.0 beta 6? [19:07] hggdh: maybe try with o or u; you should see it much more easily [19:08] cyphermox_, ok, I am glad that it is already noted and people know about it. Awesome. [19:08] ỏ ủ :-) [19:08] it is a vietnamese character, but not the horn; that's the above hook [19:12] cyphermox_: thanks. I will fail the SRU, then [19:14] hggdh: does it do the same with the 3rdlevel and 4-level keys? [19:15] cyphermox_, there's someone local here volunteering to fix this if nobody gets it done in two months or so. :) [19:15] BigWhale: you can tell them to hang around in #nm and speak to dcbw about it :) [19:16] cyphermox_, noted and done. thanks! [19:19] cyphermox_: fourth level shows a blank (or a non-displayable char) [19:19] hggdh: ah, then yes, totally fails verification :) === cyphermox_ is now known as cyphermox === Zdra is now known as xclaesse [21:34] hggdh, ả does look horned to me actually, so yes === bryceh_ is now known as bryceh [21:46] hggdh, actually maybe it's not quite right; as per http://www.fileformat.info/info/unicode/char/01a1/latin_small_letter_o_with_horn.png [21:46] but I didn't find a+horned yet [21:47] bryceh: yeah, I expected it to be like a o cut horizontally at the middle [21:47] hggdh, aha ả is hooked [21:47] ả [21:47] aye [21:47] OK [21:47] reverting to -done [21:48] ảa [21:48] ả a [21:48] http://www.fileformat.info/info/unicode/char/1ea3/latin_small_letter_a_with_hook_above.png [21:49] hggdh, try doing it with shift held too. [21:49] ā [21:50] aa [21:50] Ā [21:50] bryceh: but nothing under the letter [21:51] I wonder if it would be better to have someone that reads/speaks Vietnamese to check on the bug... [23:03] Hggdh, afaik there is no horned a, only u and o. But I'm very much a novice at vietnamese [23:04] Also, neither the horn or hook should be under, there might only be a dot under