=== salem_ is now known as _salem === karstensrage_ is now known as karstensrage [06:19] Good morning [06:44] o/ pitti, how are you? :) [06:46] tsimonq2: quite well, thanks! back home now. how about yourself? [06:46] great :) [07:36] is there a way of producing the changelog merge-o-matic produces locally? [07:41] ah maybe ~rbasak/ubuntu-git-tools/git-reconstruct-changelog [07:43] no, more git-reconstruct-changelog === sits_ is now known as sits [08:04] mwhudson: also git-merge-changelogs? That's probably the trickier bit. [08:26] doko, hey - my recent uploads for ryu address your original MIR review points for bug 1500950 [08:26] bug 1500950 in ryu (Ubuntu) "[MIR] ryu" [High,New] https://launchpad.net/bugs/1500950 [08:26] it would be nice to get that worked through if possible to unblock neutron from proposed migration [08:27] :) === King_InuYasha is now known as Son_Goku [09:25] doko: is the python3-defaults in https://launchpad.net/ubuntu/trusty/+queue?queue_state=1 still actually relevant? bug 1348954 does not have a trusty task for it nor a description of the changes, and the changelog doesn't say either [09:25] bug 1348954 in python3-defaults (Ubuntu) "update Python3 for trusty" [Undecided,Confirmed] https://launchpad.net/bugs/1348954 === greyback_ is now known as greyback === hikiko is now known as hikiko|ln [12:55] on a side note - i really hate libvirt & qemu source code === _salem is now known as salem_ [14:47] speaking of it in 14.04->16.04 apparmor prevents libvirt from starting, missing capability net_bind_service [14:54] sarnold, ^ [15:21] Odd_Bloke: hey i was real busy with $dayjob yesterday and didnt have a chance to reply. i'm really excited to see the merge accepted. let me know what the next steps are if I need to do anything for the xenial SRU. thanks!!! [15:21] semiosis: So if you'd like to, I was hoping I could convince you to prepare the SRU patch. :) [15:22] semiosis: (I have a sprint week of meetings next week, so will be unlikely to get to it any time soon) [15:23] * semiosis arm doesnt take much twisting [15:23] i'll do it! is there a doc I can read about the process? any tips? [15:24] semiosis: https://wiki.ubuntu.com/StableReleaseUpdates is the documentation. [15:24] You'll need to get someone to target https://bugs.launchpad.net/cloud-images/+bug/1565985 to xenial (I don't have the powers, unfortunately). [15:24] Launchpad bug 1565985 in cloud-images "vagrant vb ubuntu/xenial64 cannot mount synced folders" [Undecided,In progress] [15:29] Odd_Bloke: thanks I'll read the docs. how much longer will you be around today in case I have questions? [15:31] semiosis: Around 90 minutes, though I'll probably see pings in IRC after that. :) [15:48] I'm intending to make update-initramfs reference https://help.ubuntu.com/community/Lubuntu/Documentation/RemoveOldKernels if the disk is full -- should I link to it directly (moved out of Lubuntu) or reference some other URL redirector? [15:48] I'm intending to make update-initramfs reference https://help.ubuntu.com/community/Lubuntu/Documentation/RemoveOldKernels if the disk is full -- should I link to it directly (moved out of Lubuntu) or reference some other URL redirector? === debfx_ is now known as debfx [17:30] jtaylor: that seems surprising, I've used libvirt remotely on 16.04 LTS without having to do anything.. got a bug number? [18:01] sarnold: do you have that cap in your apparmor file? [18:06] jtaylor: no, there's no net_bind_service. Now I'm insanely curious.. [18:09] sarnold: and its in enforce? [18:11] jtaylor: yeah, and I can't spot any DENIED entries in my logs [18:12] sarnold: I'll try my desktop [18:15] hm I got a different apparmor profile [18:17] oh no it was just reorded by aa-logprof [18:17] and it works which is weird [18:17] maybe an upgrade from 14.04 leaves some configuration file which triggers some other behavior? [18:23] it could be you've got a different libvirt config, I'm surely not a power user.. [18:34] me neither, i'll file a bug with the logs [18:34] thanks [18:43] sarnold: bug 1605727, ubuntu-bug was run on my desktop so its not the same as the affected machine [18:43] bug 1605727 in libvirt (Ubuntu) "libvirt-bin start prevented by apparmor" [Undecided,New] https://launchpad.net/bugs/1605727 [18:45] jtaylor: do you have "network netlink," in your profile? [18:46] sarnold: the apparmor profile? [18:47] yes its in there [18:49] ah, good. hmm. [18:55] Akuli: how did you decided on 50? [18:56] Akuli: are you 100% sure there are no commands that are 50 characters? [18:56] nacc: I think we need to clarify "command" here [18:56] just a random number [18:56] Akuli: that's probably not a suitable fix then :) [18:56] teward: it's whatever is passed to "command-not-found" afaict [18:56] nacc: "If I enter an invalid command that is a few thousand characters long" <-- what's the likelihood of this? [18:56] but that part is easy to change, preferably it should be in a setting file [18:57] well if someone else like me feels like hey i have boring lets try that :) [18:57] nacc: right, but what's the likelihood of something thousands of characters long really being paseed through to the command-not-found? [18:57] well [18:57] teward: cats do it all the time [18:57] today i froze my system twice with that so [18:57] there you go [18:57] dobey: cats. :P [18:57] teward: it's any arbitrary string you type at the terminal, i think (by default) [18:57] Akuli: not my point [18:57] teward: i'm *guessing* it's a page overflow [18:57] Akuli: why were you trying to type an arbitrary string that's so long into the terminal? [18:58] or some other buffer overflow [18:58] nacc: right, but what's the practical likelihood that someone's going to have to type in 2000+ characters [18:58] i have no idea :) [18:58] i don't think its a buffer overflow [18:58] teward: practically, no, but it shouldn't crash the system if you do, right? [18:58] i mean. its python [18:58] then why is it crashing your system? [18:58] because there's no check for the length [18:58] you mean it OOMs your system? [18:58] its' not crashing the system [18:58] * teward grabs a lorem ipsum string 3000 long [18:59] has spaces [18:59] it's making the system unusable [18:59] won't do [18:59] dobey, indeed unusable [18:59] dobey: ah, ok [18:59] 5 minutes of fighting with running out of ram [18:59] four seconds of looking four the executable /usr/src/linux-headers-4.4.0-21/tools/testing/selftests/rcutorture/bin/kvm-recheck-rcu.sh [18:59] well really, it should have a length equivalent to that of the filename length restriction [19:00] sarnold: :) [19:00] E: Cannot Reproduce @ 1000, 2000. [19:00] i have no idea what that is, people who know that better than i do can change that [19:00] dobey: ack, that seems most appropriate for a true fix [19:00] teward: you have too much RAM [19:00] teward: cat /dev/urandom | command-not-found [19:01] i think their command searching function should be O(1) anyway [19:01] *snort* /usr/lib/x86_64-linux-gnu/ubuntu-system-settings/private/Ubuntu/SystemSettings/SecurityPrivacy/UbuntuSecurityPrivacyHelper [19:01] not O(n) like it seems to be [19:01] dobey: is that 255 characters still? or is that the best common default (iirc, it's fs-specific) [19:01] dobey: i'm doing this in a 512 RAM VM? [19:01] at 6500+ and not able to reproduce [19:02] with that low resources it should die off [19:02] it needs to be command [19:02] not arguments [19:03] well [19:03] nacc: 255 sounds good to me [19:03] it took 35 seconds for my 8GB system to run, and only took 400MB [19:04] Akuli: i think that's well understood :) [19:04] dobey: 255 sounds good to me too, because what sane command would be more than 255 characters long [19:04] excluding malware executable strings [19:04] teward: nothing sane is even that long [19:04] um [19:04] what sane program name would be more than 255 characters long [19:04] Akuli: none. [19:05] right [19:05] dobey: then could we not use even less than 255?> [19:05] ones in german might have a "name" that long, but the exectuable command shouldn't be :) [19:05] TBH [19:05] http://paste.ubuntu.com/20490401/ [19:05] I really think this is a 'corner case' [19:05] teward: very certainly a corner case [19:05] it's a usability issue, arguably [19:05] teward: well i think the limit should match what is allowed by dpkg [19:06] nacc, you have a syntax error in your code [19:06] Akuli: bah [19:06] unbalanced parenthesis [19:06] dobey: what's the upper limit there then? [19:06] http://paste.ubuntu.com/20490547/ [19:06] Akuli: --^ [19:06] bah almost out of power >.< [19:06] (on my laptop) [19:06] teward: i'm not entirely sure. it would be whatever the tar format limit is, for the tar format version being used [19:06] i dont' think it's accurate to print 'command not found' as it wasn't searched for [19:06] which i /think/ is 255 [19:07] i mean, 255 is a sane limit; there might be more optimal choices [19:07] nacc, it was [19:08] bash searches the command before command-not-found runs [19:08] someone should hunt around java applications that install things into /opt. if anyone's going to have ultra-long command names that seems like the group that would do it :) [19:08] without command-not-found it would print 'command not found' no matter what [19:08] well it wasn't found and it's not going to be searched for in the package archives, either [19:09] sarnold: but they have to all get wrapped in a shell script with a shorter name, becuase it's the arguments to java that are long, there [19:10] hmm [19:10] i tried that in a virtual machine, and it started attempting to kill command-not-found [19:11] but the vm is frozen anyway [19:17] reading the code more, the fix is actually pretty awful [19:17] it ignores options.ignore_installed [19:18] that's a bit rude [19:18] indeed [19:18] no, i mean your statement is [19:19] i meant, it doesnt do anything with options.no_failure_message [19:19] sorry i was confusing two things === freyes__ is now known as freyes [19:23] i meant, you called nacc's work "actually pretty awful" when the time was taken to help create a patch for your problem, which was only reported today, and is really an extreme corner case, while you refuse to "download the source for a 3 line change" to make a patch [19:23] ie, it's rude :) [19:24] i made the original implementation [19:24] nacc just changed one number [19:24] oh wait.. [19:25] he changed much more than i thought he did [19:25] that looks great to me [19:34] how is php 7.1 plans ? [19:35] need an rebuild everiting ? [19:38] another transition, and bug everyone againg for ftbfs ? [19:39] ups, I mean again [20:33] emanuel7: there aren't any php7.1 plans right now? [20:34] emanuel7: i don't think it's packaged in debian or ubuntu yet [20:37] thanks for rapid feedbak, @nacc [20:39] emanuel7: np, i think php7.1 is still in beta anyways [20:39] emanuel7: let me see if debian has anything cooking [20:39] emanuel7: with the understanding that if it exists anywhere, it would be 16.10+, probably 17.04 [20:39] you are the packager ? also in debinan ? [20:41] emanuel7: i did the php7 transition (with help) in ubuntu [20:41] emanuel7: and work with the debian maintainers a bit [20:41] emanuel7: https://anonscm.debian.org/cgit/pkg-php/php.git/log/?h=master-7.1 looks like they have some stuff in git, but not yet packaged/published [20:43] it will help to pan early... [20:43] plan [20:43] emanuel7: what version of ubuntu are you referring to? [20:44] of course the latest one [20:44] emanuel7: 16.04 will never have php7.1 [20:44] emanuel7: unless you mean 16.10? [20:44] emanuel7: that's why i asked. [20:46] I mean the developing one 16.10, and of course debian unstable [20:48] or testing [20:48] look at golang [20:49] emanuel7: debian unstable has no php7.1 either [20:49] they get testing [20:49] at beta now [20:50] emanuel7: i'm not sure what you're saying [20:50] emanuel7: there is no packaged php7.1 in debian or ubuntu [20:52] https://lists.ubuntu.com/archives/yakkety-changes/2016-July/004343.html [20:53] ok golang is not the latest version [20:53] now is rc3 [20:53] emanuel7: i thought you were asking about php? [20:54] the same about php [20:54] same what? [20:54] why not testing the latest version ? [20:54] !latest | emanuel7 [20:54] emanuel7: Packages in Ubuntu may not be the latest. Ubuntu aims for stability, so "latest" may not be a good idea. Post-release updates are only considered if they are fixes for security vulnerabilities, high impact bug fixes, or unintrusive bug fixes with substantial benefit. See also !backports, !sru, and !ppa. [20:56] there are fixes for bugs, not changes in language [20:56] and that is all I will say about [20:57] emanuel7: you are saying random statements, I'm having trouble following. [21:03] @nacc usually the newer version is better, fixes bugs, etc. [21:03] Error: "nacc" is not a valid command. [21:07] emanuel7: php7.1 is not bc with php7.0 necessarily, they add features, etc. It's not just bugfixes, if you simply glance at the NEWS file upstream. [21:07] emanuel7: and, again, php7.1 is not yet published in debian, so we don't have it in ubuntu [21:08] that is why I made an parallel to an other language like golang. [21:08] ok [21:09] and just like php5, php7.0 will presumably continue to get bugfixes [21:10] and you're free to build newer versions in a PPA to test if you need to [21:10] dobey: +1, good point [21:11] or get involved with the packagers of php in debian if you want more testing of it in unstable. [21:11] but you need to go to the debian channel for that, not here :) [21:17] what I want to say is deep respect to people doing the transitions, tons of work going unnoticed [21:19] emanuel7: ^5, it's a pain :) [21:19] emanuel7: i wills ay, i don't think php7.1 will be the same [21:20] emanuel7: what will probably ahppen is php7.1 will be an alternative in the metapackages (e.g. php) and it will "just work" [21:20] that's an eventuality, though === Foxtrot is now known as foxtrot === salem_ is now known as _salem [23:20] Odd_Bloke: the SRU -- https://bugs.launchpad.net/ubuntu/+source/livecd-rootfs/+bug/1605795 [23:20] Launchpad bug 1605795 in livecd-rootfs (Ubuntu) "[SRU] livecd-rootfs ubuntu-cpc vagrant image builder" [Undecided,New] [23:22] flexiondotorg, infinity: Know what's up with the whole -core thing? Not heard anything about it for a while is everything still pending on the reviewers? [23:29] Unit193: Yeah, I need to revisit it. Especially since flexiondotorg bought me beer. [23:29] Hah! What type? [23:30] Unit193: Whatever random hefeweizen the hotel bar had on tap. :P [23:30] infinity: FWIW, core is pretty popular with our users, got an email about a point release too. :P [23:31] s/core/base/ ;) [23:31] And noted. [23:31] Hopefully I can find some down time when I get home. [23:31] Eh... [23:32] Right now it's pretty well known as 'Core', though it seems it'll get renamed soon™ Thanks, hope you do get time.