=== cpg is now known as cpg|away [03:02] Hi, I have a bug. It just needs to rebuild the package to solve. [03:02] How can I do that for Ubuntu quantal? [03:04] https://bugs.launchpad.net/ubuntu/+source/ibus-chewing/+bug/1028746 [03:04] Launchpad bug 1028746 in ibus-chewing (Ubuntu Quantal) "quantal alpha3: ibus-engine-chewing crashed with SIGABRT in __assert_fail_base()" [Medium,Confirmed] [03:55] hallyn: yes, the ubuntu branch [04:07] Good morning === cpg|away is now known as cpg === dendrobates is now known as dendro-afk === hdon- is now known as hdon [06:49] good morning [07:00] tjaalton, happy birthday! :) [07:01] dholbach: thx :) [07:01] happy birthday tjaalton ;) [07:02] didrocks: :) === sivatharman_ is now known as psivaa === smb` is now known as smb === tkamppeter_ is now known as tkamppeter === henrix_ is now known as henrix [08:23] jtaylor: hey, did you investigate what pwlib used to be used for? === henrix is now known as henrix_ === henrix_ is now known as henrix [08:38] FourDollars: if you want a no-change rebuild, please subscribe ubuntu-sponsors to the bug. [08:39] xnox: Thank you. :) [08:39] FourDollars: is it just your package that needs rebuilding or a wider set of packages? everything that depends on particular versions of ibus? [08:39] xnox: Not sure. [08:39] xnox: I am only sure for ibus-chewing. [08:40] FourDollars: ideally we want package uninstallable if there was an api/abi bump. Or a proper transition in place for all rdeps of ibus.... [08:40] FourDollars: that's ok. [08:41] xnox: IIRC, there is no such mechanism in Ubuntu archives. Right? [08:41] FourDollars: sure there is. [08:42] FourDollars: tick good on this page http://people.canonical.com/~ubuntu-archive/transitions/boost1.49.html [08:42] FourDollars: or see this in progress transition http://people.canonical.com/~ubuntu-archive/transitions/gnutls28.html [08:42] xnox: Thanks for your information. [08:43] we also have uninstallation reports and other bits and bobs ;-) [08:43] Or just do it if it's a small number [08:43] xnox: I don't see any ibus relative under http://people.canonical.com/~ubuntu-archive/transitions/ . [08:44] FourDollars: either it's small, not done/setup, or ibus had incompatible changes without bumping shlib depends/soname. [08:45] OK. === tkamppeter_ is now known as tkamppeter [10:22] hi! I am trying ubuntu 12.10. Where is "logout", i wish to choose some other DE [10:23] gsedej_work: support is in #ubuntu [10:23] ubuntu 12.10 beta? [10:23] for 12.10 i think its rather #ubuntu+1 [10:24] oh, sorry, i didn't see it in channel list [10:24] well, the logout button didn't change in 12.10 :) [10:24] sorry for bothering === MacSlow is now known as MacSlow|lunch === cpg is now known as cpg|away === JamesJRH_ is now known as JamesJRH [11:48] ev, the #100 crash in 12.10 is in worldofgoo [11:48] nice [11:49] The first sign of importance for giving access to independent upstreams [11:49] yeah, we can't possibly retrace those at the moment [11:50] but great that we're actually getting them [11:51] The number of people still installing beta-1 is a little scary [11:51] The number of people who install milestones more than a few days after we release them scares me. :/ [11:52] But, I guess if you have the ISO lying around. [11:52] and https://wiki.ubuntu.com/QuantalQuetzal/TechnicalOverview/Beta1 gives no clue that beta 2 is out, that I can see [11:52] mpt: Why would it? [11:52] That's more understandable; it's only relatively recently that daily builds have been worth bothering with [11:52] mpt: Do the oneiric release notes point to precise? === MacSlow|lunch is now known as MacSlow [11:53] mpt: Ah, now that at least is readily fixable, although the question presupposes a charming belief that users read the tech overview ... [11:53] infinity, why wouldn't it? [11:53] I think it's reasonable for milestone notes to point forward [11:53] infinity, for people who have heard that beta 1 has been released, and haven't heard that beta 2 has also [11:53] mpt: I'm not saying that obsolescence notes on release notes might not be a cool idea, just that we've never done it before that I know of, hence my wondering at your wondering. :) [11:54] * cjwatson adds a bold-face note to the top [11:55] cjwatson: And the alphas! [11:55] cjwatson, good -- is there also a checklist for "things to do when a beta/final is released" to which that step can be added? [11:55] Probably ought to but have other things to do [11:55] mpt: BetaProcess. But I'm not bothering since I hear milestones may not be happening next cycle [11:56] * infinity isn't quite sure why we have one set of notes per milestone at all, rather than just a cumulative document for Q that ends up being the final notes. [11:56] Two eminently plausible ways to solve the problem [11:56] Wouldn't hurt to archive some more of the beta-1 images, either [11:56] I'll do that later today [11:58] cyphermox, yubiserver ftbfs === dendro-afk is now known as dendrobates [12:27] cjwatson, a lot of the amd64 only build failures (found now 7) seem to be caused by changed dpkg-dev behaviour. all have in common that either the install target is called in the build step, or that there are missing dependencies for the build-arch target. these build failures are not seen in unstable [12:31] well, unstable or not, they are surely bugs [12:32] I don't see any obviously-relevant changes in dpkg on either side, although I've only checked changelogs [12:32] But at this point I'd rather fix those packages than change dpkg anyway [12:35] looking through the amd64 only ftbfs now [12:37] I had something like that with addons, e.g. dh_sphinxdoc -a fails [12:38] That doesn't sound related. [12:38] so I had to make dh_sphinxdoc call only in binary-indep [12:38] Unless DH_OPTIONS is wrong as a result of the above. [12:39] oh and that one was reproducible in sid, so different problem. [12:43] kees: Next meeting is 2012-19-15 %-) === mcclurmc_away is now known as mcclurmc [12:52] cyphermox, do you think i'm way off assuming bug 1058517 and bug 1058987 [12:52] Launchpad bug 1058517 in dbus (Ubuntu) "upgrade of dbus causes unclean unmount of /" [High,Confirmed] https://launchpad.net/bugs/1058517 [12:52] Launchpad bug 1058987 in network-manager (Ubuntu) "In Quantal, the root filesystem is not cleanly unmounted at shutdown or reboot" [High,In progress] https://launchpad.net/bugs/1058987 [12:52] are related [13:07] smoser: not necessarily [13:08] it looked weird too, since I could have sworn that piece of code (in NM to stop dnsmasq) used to work, and certainly didn't change significantly enough to cause this [13:08] and in the cloud imageswhere i see this, i have no network manager [13:08] but do have dbus [13:09] (and this was in precise) [13:09] ok [13:09] I'm still going to look at dnsmasq/NM closely to be sure [13:09] (since the reporter days they can avoid this by removing dnsmasq-base) [13:10] cyphermox, right. that didn't make much sense to me either. [13:10] i'm not even sure how i became aware of your bug. [13:11] hehe [13:12] fwiw, I don't understand because dnsmasq is reportedly no longer running, otherwise things would hang earlier and get caught by report_unkillable in sendsigs [13:13] or at the very least I should always be able to see it on all my systems [13:14] cyphermox, well, any thoughts you have, i'd be interested in. [13:14] because this is pretty serious a regression in precise [13:14] smoser: "Also, doing a fully up-to-date netinstall works fine until I install network-manager-gnome (which pulls in dnsmasq-base). That's when this bug starts to show up." [13:14] @pilot in === udevbot changed the topic of #ubuntu-devel to: Quantal Quetzal development | Beta 2 released! | Archive: Frozen | Dev' of Ubuntu (not support or app devel) | build failures -> http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and dicussion of hardy -> precise | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: Laney [13:14] also, this is a bug in quantal, not precise [13:14] maybe it's really two different things [13:14] cyphermox, right. [13:27] Hm, the #1 crash for Ubuntu users overall right now is from software that is unmaintained [13:27] Launchpad bug 821233 in indicator-weather (Ubuntu) "indicator-weather crashed with AttributeError in export_location_details(): Location instance has no attribute 'location_code'" [Medium,Confirmed] [13:33] pitti, https://launchpadlibrarian.net/118057586/buildlog_ubuntu-quantal-amd64.vala-0.10_0.10.4-1ubuntu1_FAILEDTOBUILD.txt.gz [13:34] this is the same error as for ubuntu-drivers-common ... [13:34] not for the same service, though; but yes, timed out while waiting for a service to get activated [13:35] but I can't see how this connects to a swap space issue [13:36] swapping makes everything grind to a halt, so everythingthat does I/O (that includes d-bus service activation) just takes ages [13:36] or at least slow enough to trigger some race conditiosn [13:36] but this is amd64, not arm [13:36] pitti: do you know it is swapping? I doubt it [13:36] so apparently this is a different problem [13:37] the arm build is only 12min, so it doesn't like to be a timeout [13:37] Damn, rebootstrapping eclipse-mylyn didn't fix eclipse-egit [13:37] Wonder what's going on there [13:37] doko: well, it's "slowness"; I just parrot infinity when he said that since the upgrade our arm builders have to do a lot more swapping [13:38] pitti: but here I can't see any reason for that [13:38] doko: right; pitti | so apparently this is a different problem [13:40] cjwatson, I think I did address all binary-arch issues, maybe uw-imap is another one [13:55] infinity: would you like to take care of https://bugs.launchpad.net/ubuntu/+source/eglibc/+bug/1000498 ? [13:55] Launchpad bug 1000498 in eglibc (Ubuntu Quantal) "fmod() incorrectly returns NaN for (some?) denormalized inputs" [Medium,In progress] [13:55] Laney: Yeah, it's on my TODO today, since I need to re-do my SRU upload ANYWAY. [13:55] (Thanks, security team) [13:55] (I'm not bitter) [13:56] Pfft, security, what do we need THAT for? [13:59] jamespage: I could use help with the eclipse-egit build failure; it's not making a lot of sense to me, as those plug-ins do seem to exist. But I'm not sure where Eclipse might be looking. [14:01] cjwatson, looking now [14:03] cjwatson, ergh - that looks like osgi terror - lemme see if I can dig out the DM [14:08] Where can one apply for steam linux beta? [14:09] jamespage: Hmm, some dependencies were dropped between Debian and Ubuntu [14:10] jodh: ping [14:12] jamespage: Aha, installing those missing dependencies did it. So eclipse-mylyn has misbuilt [14:13] * cjwatson hopes this wasn't the result of his bootstrapping attempt [14:13] cjwatson, not sure I understand - I would suspect some sort of odd bootstrap sequence to get this stuff working in Debian from scratch so that might make sense [14:13] cjwatson: What did you do to break it? :) [14:14] Installed eclipse-egit from Debian [14:14] The -mylyn build log has stuff like: [14:14] jh_installeclipse: Cannot determine the package providing /usr/share/java/httpclient.jar. [14:14] Which looks possibly relevant [14:14] Missing build-deps? [14:15] libhttpclient-java is there [14:15] JDK 6 vs. 7? [14:15] That seems likely. === sivatharman_ is now known as psivaa [14:17] But it just uses dpkg -S for that check [14:17] barry: howdi [14:18] jodh: hi! sorry about crossing signals on that bug. i tried to find you on irc but i think it was past your eod [14:18] cjwatson: Which would imply the files don't exist... [14:18] And yet. [14:18] barry: np! [14:18] jodh: did you see the patch and test i uploaded? i wonder if it makes sense to pull those into your mp for lp:python-apt? [14:20] * infinity scratches his head... [14:20] Trying to construct a matching build locally now. [14:21] barry: I think mvo has already taken both your + my code and put it into the ubuntu branch. [14:22] please can someone reject https://code.launchpad.net/~schumski-deactivatedaccount-deactivatedaccount/ubuntu/quantal/kmess/1045090/+merge/122427 per the comment there [14:22] jodh: great! i still have a few comments about the mp, so i'll just follow up there and let mvo dtrt. [14:22] Laney: marked as work-in-progress [14:23] stgraber: if that gets it off the list, cool [14:23] (I could have done that too) [14:23] * dholbach hugs Laney [14:23] Laney: no you can't do that yourself, you need to be in ~ubuntu-branches (so basically be a TB member) [14:23] I get the option [14:23] Work in progress, Needs review, Merged [14:23] * xnox too [14:24] oh, is that new? [14:24] that is not "Rejected" though. [14:24] anyway, yes, work in progress will get it off the report [14:24] and sounds nicer =) [14:25] barry: I merged a bunch of stuff this morning, I did not see a MP from you though? did I miss that? (just from from jodh) [14:25] cjwatson: I bet it's just sketchy parsing of "dpkg -l" in perl that the new dpkg may have broken. [14:25] It's possible. I should be able to tell once I get far enough through the build (unfortunately it's a bit slow). [14:26] mvo: no, i didn't get around to doing the mp. was going to do that today, but i think i won't need to. you grabbed the package patch i applied for LP: #1030278 right? [14:26] Launchpad bug 1030278 in the python apt bindings "Quantal failed to install: ubiquity crashed in apt/progress/text.py in pulse: OverflowError: Python int too large to convert to C long" [Undecided,New] https://launchpad.net/bugs/1030278 [14:28] jodh: would you care to continue reviewing https://bugs.launchpad.net/ubuntu/+source/zabbix/+bug/1047837 ? [14:28] Launchpad bug 1047837 in zabbix (Ubuntu) "Upstart support for Zabbix services" [Wishlist,Triaged] [14:29] jodh, mvo i added a few comments to https://code.launchpad.net/~jamesodhunt/python-apt/test-for-size_to_str/+merge/127435 [14:30] cjwatson: Or not. A simple copy-and-paste reduced test case from javahelper spits out package and version just fine. [14:30] It's not locale-dependent or something, is it? [14:31] (I do wish eclipse* didn't have 100MB-plus build-deps from universe.) [14:31] LANG=C seems to behave fine too. [14:31] C.UTF-8? [14:31] cjwatson: Installing just libhttpclient-java is small. [14:32] Yeah, but since that apparently doesn't reproduce the problem ... [14:32] Laney: comment added. [14:32] Well, yes. But I'm stumped, cause it clearly has the right string there. [14:32] ty [14:32] Which it then passes to the sub that I copied verbatim. [14:32] And it works. [14:32] >:( [14:32] barry: thanks, I'm in a call now wil comment in a bit [14:33] barry: thanks. [14:38] cjwatson: Any chance those strings could be null-terminated or something wonky? [14:38] cjwatson: That would asplode the dpkg -S [14:39] Yeah, not sure yet [14:42] Riddell: ping [14:42] hi mfisch [14:43] Riddell: hey, I'll test my deb in a bit here and let you know in the defect [14:44] mfisch: point me at it and I can test too [14:44] Riddell: okay, need to finish up a phone call first [14:48] ugh, yes python-apt, what I really mean by "extract this deb" is "create a lot of empty directories for each symlink you encounter." [14:51] oh, that fix to bug 390247 looks tempting [14:51] Launchpad bug 390247 in ubiquity (Ubuntu) "Change I/O scheduler to NOOP when installing in SSD Drive or virtual machines" [Wishlist,Confirmed] https://launchpad.net/bugs/390247 [14:51] * ogra_ wonders if there are any drawbacks using udev rules like that [14:53] ogra_: hmm... is the scheduler per drive? what if I have one rotational and one SSD ? [14:53] jcastro_: do you know when the HP cloud trial finished/finishes? [14:53] yes, it is [14:53] ogra_: that's ok then. Upload on day one or r-series? [14:54] xnox, well i wonder if it could help with bug 864051 as well somehow [14:54] Launchpad bug 864051 in partman-target (Ubuntu) "TRIM support not enabled for SSDs" [Medium,Confirmed] https://launchpad.net/bugs/864051 [14:55] ogra_: I like the bug 390247 solution better. If we can enable trim/ change scheduler on the fly the better. [14:55] Launchpad bug 390247 in ubiquity (Ubuntu) "Change I/O scheduler to NOOP when installing in SSD Drive or virtual machines" [Wishlist,Confirmed] https://launchpad.net/bugs/390247 [14:55] Riddell: it will be built in a bit, I haven't tried it yet: https://launchpad.net/~mfisch/+archive/testing/+packages [14:56] Riddell: it's looking like it will be a few hours before I can install it [14:56] ogra_: the caveat here, if we get it wrong we may wear one of the other out. [14:56] s/of/or/ [14:56] ogra@anubis:~/Devel/packages/nvidia-graphics-drivers-tegra-16.0$ cat /sys/block/sdd/queue/scheduler [14:56] [noop] deadline cfq [14:56] ogra@anubis:~/Devel/packages/nvidia-graphics-drivers-tegra-16.0$ cat /sys/block/sda/queue/scheduler [14:56] noop [deadline] cfq [14:56] wrt your first question [14:57] Laney, they have not given me an exact date. [14:57] Laney, basically, keep an eye on it, lmk if they start charging you. [14:57] jcastro_: OK, I'm kind of scared to use it any more for this reason [14:57] xnox, hmm, i thought if the HW doesnt support TRIM it simply wont be used [14:57] you're an employee, we have a cloud policy you can expense. [14:58] heh [14:58] ogra_: but schedulers are available for all drives? [14:58] xnox, the question is, can we tell it somehow through sysfs to use TRIM or is that only possible as mount option [14:58] And that would enable us to know if they did start charging us so I can tell community members to either stop or keep going. [14:58] ogra@anubis:~/Devel/packages/nvidia-graphics-drivers-tegra-16.0$ ls /sys/block/sd*/queue/scheduler [14:58] /sys/block/sda/queue/scheduler /sys/block/sdb/queue/scheduler /sys/block/sdc/queue/scheduler /sys/block/sdd/queue/scheduler [14:58] grmpf ... IRC [14:58] jcastro_: let me know, too, haha. they also ended their 50% discount on cloud usage. [14:59] jcastro_: Laney: don't they invoice monthly with no updates and that means, if they do invoice it's too late... [14:59] xnox, at least for all sdX ones [14:59] xnox, yes, exactly [14:59] well, you'll then have to claim it back [14:59] as long as random usage is covered by this policy then it's cool by me [14:59] ah, also for sr0 loop and ram devices that i have here [14:59] like, I'll use it now to do this patch piloting since my PC sucks [15:00] Laney, don't take my word for it, please read the internal cloud policy document (I don't have it handy) [15:00] Laney: didn't you brag about your new i7 on facebook?! [15:00] yeah but I'm working [15:00] can't be building PCs now can I :P [15:00] I am kind of hoping they just love Ubuntu and leave the accounts open and free for the benefit of the project. :) [15:02] xnox, hah ! [15:02] xnox, http://lwn.net/Articles/515790/ [15:04] jcastro_: yeah just read it, seems like it would clearly be fine [15:04] so seems trim/discard should be tweakable through sysfs too with that patch [15:04] I'll give it a go then, thanks [15:04] ogra_: lol =) interesting. [15:05] Laney, yeah, unfortunately I have no way of finding out when/if they shut it off, I figure it's better for us to find out than having a community member stuck with a bill [15:05] xnox, did you ever do the rebuild thing on HP cloud? === salem_ is now known as _salem [15:08] jcastro_: i did do recompress https://blueprints.launchpad.net/ubuntu/+spec/foundations-r-dpkg-xz [15:08] for that blueprint [15:08] of amd64, all, dbgsym packages. [15:08] using hadoop cluster on top of HPCloud. [15:09] isn't there a problem with using xz in the base system, or something? [15:09] * Laney can't remember exactly [15:09] something to do with debootstrap maybe [15:10] There certainly was at one point, and may still be. [15:10] * lfaraone|sh is reducing his usage to ~3 XS instances, as a precaution. [15:13] Laney: cjwatson: but it is awesome for -dbgsym packages =) [15:13] infinity: Argh. eclipse-mylyn got the right dependencies when built locally. [15:13] yeah, I'm sure the ddeb archive will thank you [15:13] xnox: You could fix -dbgsym easily independently, in pkg-create-dbgsym. [15:13] cjwatson: yes. [15:14] cjwatson: ... [15:14] (And no errors from jh_installeclipse.) [15:15] cjwatson: Well, that fits with my by-hand unit-test of test.pl, but that's pretty unhelpful. [15:15] I might try another build in a PPA to see if it's cosmic rays. [15:15] cjwatson: That's some serious cosmic rays... [15:15] cjwatson: I guess filesystem corruption could have thrown a null in the perl script... Somehow? [15:15] cjwatson: and then we can discuss if we fill flip the xz switch elsewhere e.g. (amd64, i386, all, armel, armhf, powerpc) x (debs, ddebs, udebs) x (lamba function of your choice) [15:16] xnox: Personally I'm not interested in doing this independently of Debian. [15:16] Too much hassle for too little gain. [15:16] ok. [15:16] Debian seems to be showing signs of wanting to do it in the near future anyway. [15:17] So if we do it ourselves, that's potentially a non-trivial amount of work that might well be superseded in not very long anyway ... [15:17] Indeed, it was hotly discussed at Debconf. [15:17] pitti, any opinion about bug 390247 ? i think we shoudl ship it, but i'm not sure where to add it, udev or udisks ... [15:17] Launchpad bug 390247 in ubiquity (Ubuntu) "Change I/O scheduler to NOOP when installing in SSD Drive or virtual machines" [Wishlist,Confirmed] https://launchpad.net/bugs/390247 [15:17] So why not contribute our effort to making it work right in Debian, if that's what we want to do [15:17] Though, mostly just from the POV of twiddling debhelper's defaults. [15:17] (And their defaults may still end up not being what we want, given that people think "slow arches" don't deserve good compression, so we'll see) [15:18] I tried to fight that mentality as hard as I could. [15:18] or look at the stats and identify packages that are not gaining anything from compression, and e.g. initiate fixing those pre-compressed packages. [15:19] xnox: Yeah, a lot of compressed stuff shouldn't be. [15:20] xnox: And some stuff should be gz instead of xz, and, and.. But the per-package tweaks are orthogonal to picking a sane default. [15:20] infinity: especially after we already optimise our pngs for example. [15:20] I realise there are some sub-block files where it makes no difference; but "fixing" many others might well entail making the installed system bigger. [15:20] Which isn't a good side-effect. [15:21] cjwatson: We're talking debs specifically, hard to make the installed system bigger by changing compression of debs... [15:21] (At least, I think we are) [15:21] cjwatson: how would the installed size change? [15:21] * xnox agrees with infinity [15:22] "fixing those pre-compressed packages" - any pre-compression is going to be individual files inside those .debs, surely. [15:22] cjwatson: When I talk about "some stuff shouldn't be compressed", I literally mean some data.tar.gz components should be data.tar (openclipart is a great example). [15:22] Oh, I see [15:23] Apparently nobody's doing that right now since LP doesn't support it [15:23] Yeah, we noticed that in the BOF. ;) [15:23] So you might want to fix that first [15:23] dak supports it because they tore out the component filename check a while ago. [15:24] soyuz should likely just do the same. [15:24] Probably. [15:24] (Or maybe you pulled it out when you decided it was crusty and unnecessary...?) [15:24] I dunno, I don't have a checkout in front of me. [15:25] should I upload at tar compressed package to a ppa? [15:25] or does it need to be archive? [15:25] xnox: PPA would explode the same way. But I'm not sure I care deeply to test this morning, and finding the relevant bit of code is minimal effort. :P [15:26] * xnox goes to make coffee instead =) [15:26] coffee ! [15:28] slangasek: If you need any more debugging info on that mountall bug, shout soon because this system is going to be flattened tonight [15:39] Laney: flattened? [15:39] slangasek: replaced [15:40] Laney: so one thing I noticed in the output was /media/debian/var/lib/schroot/union/overlay mounted before /var/lib/schroot/union/overlay, which looks like it's probably wrong, but I'm not sure [15:41] infinity: Search for verifyFormat (at least) [15:41] Laney: would be interested to know if you get different results when commenting out either or both of that mount point, and the non-standard /cgroup one [15:41] OK, can do that [15:41] Probably ought to run dpkg-deb -I or something [15:41] no clue where the second overlay came from [15:42] cjwatson: Yeah, I feel like we've had this conversation before. [15:42] It's not impossible [15:42] cjwatson: Cause I remember verifying that dpkg-deb -I exits non-zero on broken archives. [15:43] cjwatson: Of course, that approach also requires making sure soyuz's dpkg always support the latest and greatest deb formats, but I suspect that's a sane thing to require anyway. [15:46] infinity: ICBW but I think I landed something else a while back which means that it has to nowadays. [15:49] Laney: I presume that schroot did some session recovery and mounted it within a chroot for you? But it looks like it's doing this too early [15:50] /media/debian is an installation of sid on another partition [15:51] I can't remember how I set this up :P [15:52] who's in charge of errors.ubuntu.com? [15:53] brendand_: ev [15:53] * ev hides [15:53] ev, who can have access? there's an upstream maintainer who wants to access it [15:54] ev, at least they say they are [15:54] that's me actually :) [15:54] mitya57, i'm sure you are :) [15:54] I'm Ubuntu maintainer, not only upstream [15:54] brendand_: at the moment it's only people in ~canonical or ~ubuntu-bugcontrol [15:55] so getting into the latter group would be my suggestion. :) [15:55] we do want to open it up to a wider audience, but there are serious privacy concerns [15:55] as anything at all can be in the crashes [15:56] I definitely plan to do that, but I want to become a Ubuntu member first, which won't be possible until November :( [15:56] ogra_: I don't have an opinion about this, I'm afraid; I know nothing about schedulers :( [15:56] ogra_: I do know that linux sucks while doing heavy I/O, especially on my rather fast SSD [15:56] pitti, no, but you know where device rules usually live := [15:56] so maybe this alleviates it a bit [15:56] :) [15:56] mitya57, no reason to do it that way around. i was in bug-control first and then only after ubuntu-members [15:56] mitya57, in fact bug-control membership can be valuable collateral for your application [15:56] pitti, the rule is correct ... its just about where to put it [15:57] ogra_: well, if you mean that, a rule like this would never go into udev itself; it's a workaround for bad kernel defaults (if it is a bad default) [15:57] at least not upstream [15:57] its not [15:57] its something the kernel cant detect (yet) [15:57] so if the kernel can't detect it, how can userspace? [15:57] but anyway, if you want to put rules into udev, you know where the bzr tree lives :) [15:57] theoretically the kernel shoudl just notice the device is an SSD and set the right default scheduler [15:57] brendand_: ok, so I'll apply when I have some time, can you still show me that crash? [15:58] well, the kernel exposes the feature to userspace ... but doesnt do any proper automated handling of it yet [15:58] Hi, would someone review this bug and (hopefully) mark it public? Apport has identified it as the master bug for an issue, I'd like to see the master bug. https://bugs.launchpad.net/bugs/883615 [15:58] Error: malone bug 883615 not found [15:58] pitti, so rather udev than udisks ? [15:58] (that was my actual question ;) ) [15:59] mitya57, anyway i have the screenshot here (of the python stack trace). i'm pretty sure there's nothing confidential/private shown so unless ev wants to veto it i can post the link here... [15:59] rather "fix the kernel", I'd say, but it's a nasty workaround rule no matter where it lives :) [15:59] smagoun: done [16:00] stgraber: thank you! [16:00] brendand_: presumably you mean paste the text or put up a pastebin url of the contents - you shouldn't be able to get to the stack trace without going through open ID [16:00] but I don't mind if you share the contents :) [16:00] pitti, whats nasty about it ? i prevents your SSD from wearing out and increases performance :) [16:01] ogra_: I mean having userspace rules to fix broken defaults (that's why it wouldn't go upstream) [16:01] andi agree it should be fixed in kernel eventually [16:01] ogra_: i'd stick it into hdparam package or something like that. [16:01] well, its not defaults, but rather missing handling of them [16:02] ogra_: so for packaging, I guess it's fine to put it into udev, as that has an ubuntu specific packaging anyway, so doesn't introduce a delta; but I don't mind much which package it ends up in [16:02] xnox, hmm, i think that would add a depends on udev, no ? [16:02] pitti, thanks :) [16:02] no, it's just shipping a file [16:02] if udev isn't installed, the rule wouldn't work anyway [16:02] ev, heh - pastebin would be even better :) assuming mitya57 is only interested in the stack trace? [16:02] but then again, you can't uninstall udev [16:03] * pitti waves good night [16:03] need to run now [16:03] mitya57, http://paste.ubuntu.com/1256228/ [16:03] thanks brendand_! [16:03] ciao pitti and thanks again [16:05] mitya57, and do get bug-control sooner rather than later :) make your life easier [16:05] brendand_: sure [16:08] jbicha, hi. I'd like to have your advice on https://bugs.launchpad.net/unity-lens-photos/+bug/1049090. It's a OEM team request to change a lens shortcut (Super+P to Super+C). [16:08] Launchpad bug 1049090 in OEM Priority Project quantal "The shortcut Super + P of photo lens is conflict to Video out hotkey" [High,Confirmed] [16:11] jbicha, is this shortcut documented? [16:14] pitti: any objections to depending on python3-debian? I had to rework that merge [16:14] python-debian is in main [16:14] I'll need to work out a MIR for python3-debian [16:15] No, you don't [16:15] ORLY [16:15] MIRs are by source package [16:15] oh, right [16:15] cheers [16:15] python3-debian can be promoted as soon as something depends on it [16:15] woohoo [16:16] (Or just in advance, if that's more convenient - but not too much in advance so that we don't spend too long with junk in component-mismatches) [16:17] davidcalle: yes http://bazaar.launchpad.net/~ubuntu-core-doc/ubuntu-docs/quantal/view/head:/ubuntu-help/C/unity-dash-photos.page [16:20] slangasek, i'm guessing that this is fallout of the mountall change [16:20] https://bugs.launchpad.net/ubuntu/+source/coreutils/+bug/1060296 [16:20] Launchpad bug 1060296 in coreutils (Ubuntu) "'df /' reports Filesystem '-'" [Undecided,New] [16:21] but Super+P as a video out shortcut isn't in ubuntu-docs [16:24] jbicha, I see. It's apparently a standard for some OEM. [16:24] smoser: please get me the boot-time output of mountall --verbose [16:25] slangasek, isnt it clear this is a fallout? [16:25] davidcalle: we might be able to sed replace it, do different languages have different keyboard shortcuts? [16:25] on the other hand, the Ubuntu Manual is almost string-frozen too [16:25] i can do that though. [16:25] smoser: knowing that it's fallout doesn't tell me how to fix it [16:25] jbicha, no, same shortcut. [16:26]