| jbicha | cortina and meta-gnome3 are the only 2 I see that will need new uploads now then for that | 00:01 |
|---|---|---|
| jbicha | so I'll do that now | 00:01 |
| infinity | Well, ostree clearly does, or it wouldn't have s390x binaries. | 00:01 |
| infinity | Not sure I want to hunt through the rest of the bug to see if others are in such a state. | 00:01 |
| jbicha | yes but as long as we can get the newer ostree to migrate that will be fixed | 00:01 |
| infinity | jbicha: Eh? | 00:02 |
| slangasek | infinity: when people are retrying builds for things that ftbfs on random archs that have never built, and I get email for those, I absolutely reserve the right to not like the emails and not commit to fixing the builds | 00:02 |
| infinity | jbicha: The new ostree has s390x binaries. | 00:02 |
| infinity | slangasek: Okay, it's fair that we could do with some regression history, so we could maybe tag FTBFS mails in a filterable way. | 00:02 |
| jbicha | infinity: no, it's in depwait (and my diff is unnecessary) https://launchpad.net/ubuntu/+source/ostree/2017.11-1ubuntu1 | 00:02 |
| infinity | jbicha: Oh, there's a VERY new one. :) | 00:03 |
| infinity | jbicha: I was looking at rmadison, silly me. | 00:03 |
| jbicha | ok | 00:03 |
| infinity | jbicha: Alright, removing away. | 00:03 |
| infinity | jbicha: Thanks for weathering my rant with rational responses. :P | 00:03 |
| jbicha | infinity: since you're opinionated, do you prefer I have cortina build-depend on gjs (so depwait) or specify a list of architecture to build for? | 00:05 |
| slangasek | infinity: to your rant though, I want to double-check I'm not doing something wrong - my usual response to "please remove these binaries" is "please give me a new version in -proposed that isn't just going to bring those binaries back." so I'm in the clear, right? | 00:05 |
| infinity | slangasek: Right. Binary removal should only happen if the binary is NBS. | 00:06 |
| jbicha | cortina/s390x has an unsatisfiable depends on gnome-shell, which confuses britney I believe | 00:07 |
| slangasek | ok | 00:07 |
| infinity | slangasek: I mean, there should also be due diligence to make sure that version in proposed will actually migrate, but basically the right track there. | 00:07 |
| infinity | jbicha: That doesn't confuse britney, britney's correct. | 00:07 |
| infinity | jbicha: ostree cleaned up. | 00:08 |
| infinity | (And so noted in the bug) | 00:08 |
| infinity | jbicha: As for the question, I prefer the unnecessary build-dep method (though, really, it should build-dep on gnome-shell, since that's what it depends on) because, while it's a bit heavy-handed and gross, it has the advantage of "fixing itself" if the stack suddenly becomes buildable again. | 00:10 |
| infinity | jbicha: And also doesn't need adjusting for new arches. | 00:10 |
| jbicha | ok | 00:11 |
| doko | slangasek, infinity: I gave back all packages last week, because it wasn't done since April | 00:11 |
| infinity | The amount of the archive that isn't built on Random Arch X purely because the maintainer chose an Architecture line 13 years ago that no one has since reviewed for accuracy is larger than you'd think. | 00:11 |
| jbicha | there's a chance someone in Debian will figure out gettings mozjs52 to work on other arches… | 00:11 |
| infinity | doko: Thanks. | 00:11 |
| doko | and yes, it built a handful more packages sucessfully | 00:12 |
| infinity | jbicha: Well, afaik, newer mozjs is entirely dependent on rustc being non-crap, which arch maintainers are working on, so it might just kinda fix itself over the next 6-12 months on a few arches. | 00:12 |
| doko | well, working on rust is wishful thinking | 00:13 |
| infinity | doko: I don't mean downstream arch maintainers, I mean IBM has people working on it. | 00:13 |
| infinity | I certainly am dedicating exactly as much time to rustc as I did to golang. | 00:14 |
| jbicha | I think mozjs52 is pre-rust but yeah, the future will be fun | 00:14 |
| infinity | jbicha: Ahh, if it's pre-rust, it may just be a legit silly endian bug or something and, indeed, Debian might fix it. | 00:15 |
| infinity | jbicha: The future actually looks pretty good for mozilla after the switch to rust, IMO. And that's coming from someone who's sick of "oh look, a new compiler that needs porting." | 00:16 |
| jbicha | https://buildd.debian.org/status/package.php?p=mozjs52 ( you can probably ignore mips64el but the other test failures are really bad) | 00:16 |
| infinity | jbicha: Cause at least this compiler WILL be ported (over time). | 00:16 |
| infinity | jbicha: The most complex part of firefox that ALWAYS broke on non-primary arches was mozjs, so switching to a highly-portable thing that needs less babysitting by port maintainer might make firefox Just Work a lot more often than it used to. | 00:17 |
| infinity | (Again, in a wonderful future of a year or three when rust's ports are all solid) | 00:18 |
| jbicha | it's nice that GNOME is now on the latest mozjs (ESR) instead of mozjs24 like we were a year ago too | 00:19 |
| infinity | Of course, if you take my above analysis to extremes, I may have just accidentally advocated writing a javascript interpreter in Perl. | 00:21 |
| infinity | Which I don't recommend. | 00:21 |
| infinity | (Though it wouldn't be any worse than jbailey's pet project to write a libc in js) | 00:22 |
| jbicha | if you're in a removing mood, want to remove some or all of the packages at LP: #1710318? we might leave a few webkitgtk rdeps in artful but it doesn't need to be so many | 00:23 |
| ubottu | Launchpad bug 1710318 in xiphos (Ubuntu) "Please remove webkit1 rdepends removed from Debian Testing" [Undecided,New] https://launchpad.net/bugs/1710318 | 00:23 |
| infinity | jbicha: I have to run out to the hospital shortly, but I have a browser tab open to look at it. | 00:24 |
| infinity | karstensrage: Oh, ditto to you. I got sidetracked a bit here. But when I get back, I'll have a poke at your stuff. What timezone are you in? | 00:24 |
| karstensrage | Pacific | 00:25 |
| jbicha | infinity: what time zone are you in? lol | 00:25 |
| infinity | karstensrage: Ahh, kay. I'm Mountain. So if we need some real-time back-and-forth on this (and your NSS module), let's talk tomorrow. | 00:25 |
| karstensrage | hope all is well infinity and thank you for your patience with me | 00:25 |
| karstensrage | will do | 00:25 |
| infinity | jbicha: I'm in both all of them and none of them. And even moreso the last week or two with the above mentioned hospital issue. :/ | 00:25 |
| jbicha | best wishes | 00:26 |
| infinity | There are many crossed fingers. | 00:26 |
| infinity | Mom's scheduled for all the open heart surgery ever next week. | 00:26 |
| infinity | Which is a bit scary. | 00:26 |
| infinity | A lot scary. | 00:27 |
| gaughen | infinity, slangasek - would you, or someone else with a release team hat, have a look at - https://bugs.launchpad.net/ubuntu-power-systems/+bug/1706248 | 00:47 |
| ubottu | Launchpad bug 1706248 in slof (Ubuntu) "[FFE] SLOF dhcp request doesn't include client architecture code 93" [Medium,Confirmed] | 00:47 |
| infinity | gaughen: I'm conceptually okay with a new SLOF from an HWE perspective, but I think I'd like to give it a once-over/test-drive, and chat with Aurelien about doing it in Debian too. | 00:55 |
| infinity | gaughen: So putting that open browser tab on my "will look a bit later". | 00:55 |
| gaughen | fair enough | 00:55 |
| infinity | Also super curious about this "new boot menu" thing, since firmware presenting boot menus is usually a bad thing, in my experience, not a good one. :P | 00:56 |
| gaughen | cpaelzer, ^^ fyi re: ffe on slof | 00:56 |
| gaughen | infinity, from the perspective of users select options they shouldn't? | 00:57 |
| infinity | gaughen: I want to see what they mean by "boot menu". If it's "hit a key combination to enter forth prompt", fair enough, if it's some nasty attempt to present what grub should be presenting, ick. | 00:59 |
| gaughen | yeah that is ick | 00:59 |
| infinity | gaughen: Anyhow, the motivation for the bug report (the dhcp thing) feels like something we'd want to cherry pick back to (at least) xenial, so worth someone looking at that while I also evaluable the wholesale version bump. | 00:59 |
| infinity | evaluate, too. | 00:59 |
| infinity | evaluable? Really, fingers? THAT'S NOT EVEN A WORD. | 01:00 |
| slangasek | UEFI has a boot menu, it must be a good idea | 01:01 |
| infinity | slangasek: Yeah, and it's STELLAR. | 01:01 |
| infinity | Except the other thing. | 01:01 |
| infinity | Also, see petitboot as exhibit suck. | 01:02 |
| slangasek | yes, I was just seeing petitboot this afternoon | 01:02 |
| infinity | My condolences. | 01:02 |
| slangasek | for half a second, before its 10 second boot timeout triggered after I had waited a half hour for it to load | 01:02 |
| cjwatson | Evaluable is totally a word. Just not the word you wanted. | 01:32 |
| infinity | Err, yes it is. But not in the way I was thinking. | 01:33 |
| infinity | e-valuable is how my brain broke it apart. | 01:33 |
| cpaelzer | good morning | 05:35 |
| Unit193 | Heya. | 05:35 |
| === Guest19724 is now known as RAOF | ||
| jamespage | o/ | 09:12 |
| doko | stgraber: autopkg test failures triggered by lxc | 09:47 |
| doko | jamespage: cinder autopkg test failure on s390x | 09:47 |
| jamespage | doko: looking now | 09:48 |
| LocutusOfBorg | hello chrisccoulson, doesn't rust need a sync from debian? because of firefox? | 10:54 |
| LocutusOfBorg | (actually a merge probably) | 10:56 |
| chrisccoulson | unless you're also going to do cargo, please do not sync or merge rust | 11:01 |
| ricotz | LocutusOfBorg, chrisccoulson, notice/remember libgit2 | 11:40 |
| ricotz | chrisccoulson, please see PM | 11:40 |
| LocutusOfBorg | chrisccoulson, I was asking about what firefox requires, I don't plan to do it :) | 12:00 |
| LocutusOfBorg | ricotz, what is the libgit2 problem? | 12:01 |
| LocutusOfBorg | is somebody requesting a ffe? | 12:01 |
| ricotz | LocutusOfBorg, debian is not shipping it internally and relies on it as system-dep which won't work on ubuntu | 12:02 |
| ricotz | since backport down to trusty are required | 12:02 |
| ricotz | LocutusOfBorg, so don't plainly take it from debian | 12:02 |
| ricotz | basically the toolchain for firefox 56 is already done | 12:03 |
| ricotz | the version of rustc/cargo which are in debian currently will be required beginning with firefox 57 | 12:03 |
| LocutusOfBorg | ok, anyway, I don't plan to do cargo/libgit2/rust work | 12:12 |
| LocutusOfBorg | I was just asking in case you want to prepare for FF57 | 12:13 |
| LocutusOfBorg | libgit2 will be autosyncd btw when 18.04 opens | 12:16 |
| ricotz | LocutusOfBorg, I think rustc/cargo should not be synced, e.g. even likely that ff56 won't build with rustc 1.20 | 12:19 |
| rbasak | xnox: any progress on bug 1691096 please? AFAICT, it's a bug in systemd and I'd like to start considering it release-critical for server. | 12:45 |
| ubottu | bug 1691096 in systemd (Ubuntu Artful) "mysql in lxd fails to start with systemd 233, 234: failed at step KEYRING" [High,Triaged] https://launchpad.net/bugs/1691096 | 12:46 |
| rbasak | ahasenack: ^ | 12:46 |
| ahasenack | thx | 12:46 |
| === giraffe is now known as Guest431 | ||
| sil2100 | hm, we seem to be having some issues with our autopkgtests in our ubuntu-image CI | 14:22 |
| sil2100 | Looks like it has problems installing haveged? | 14:22 |
| sil2100 | https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-artful-canonical-foundations-ubuntu-image/artful/amd64/u/ubuntu-image/20170918_021026_e3fc6@/log.gz | 14:22 |
| * sil2100 tries re-running | 14:24 | |
| xnox | sil2100, universe is not enabled? | 14:26 |
| sil2100 | Did it get disabled from the infra? | 14:27 |
| sil2100 | Since I don't think this is anything we control | 14:28 |
| stgraber | doko: I know, I've been watching :) | 14:47 |
| === dupondje_ is now known as dupondje | ||
| nacc | jbicha: infinity: ok, setup a mate VM of 16.04.3. In the 'boutique', if i search for virtualbox, both virt-manager and oracle's virtualbox show up (confusingly not the one from multiverse). No repository is setup for virtualbox initially (that I can see in boutique). If I click on install oracle virtualbox, it queues it and then if i click on apply changes, the third party virtualbox repo has been | 15:54 |
| nacc | added and a non-ubuntu package has been installed. Now, it does say the source is oracle's repo (by URL), but I'm not sure a regular user would konw it's unsupported software (by Ubuntu) | 15:55 |
| nacc | infinity: not sure if that still constitutes a policy violation or not; it's opt-in, but not at all obvious to an end-user what they are opting in to, imo | 15:57 |
| rbasak | ddstreet: sorry, I didn't realise you had a new patch for me there. I don't think forcing use of gcc-6 is a suitable fix for the development release though. I suspect doko wouldn't be happy about that. | 16:14 |
| rbasak | ddstreet: I think I'd already made suggestions on this channel? | 16:14 |
| ddstreet | rbasak it's being worked by niedbalski who is actually out today, but i'll pass that on to him, so he may ping you tomorrow | 16:15 |
| === tkamppeter_ is now known as tkamppeter | ||
| pitti | wgrant: trouble with lgw again? most builders seem to be stuck in "cleaning" again (some spot checks like https://launchpad.net/builders/lgw01-amd64-041/+history says 6 hours) | 20:53 |
| karstensrage | infinity, very sorry to bother you, no worries if you have more pressing issues, any preliminary comments? Id like to get started on the new NSS packaging | 23:52 |
| karstensrage | i assume since thats new it will be an ITP and then a RFS so it goes through the normal process | 23:53 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!