| === JanC is now known as Guest56745 | ||
| === JanC_ is now known as JanC | ||
| cpaelzer | wgrant: I was unable to get the dbus-test-runner hang to occur in a local qemu based riscv build of dee | 06:06 |
|---|---|---|
| cpaelzer | wgrant: I have thrown the bit of info that I found into 1891158 | 06:06 |
| cpaelzer | and unblocked (for now) src:dee by skipping the tests on riscv64 - bumping up the timeout didn't help | 06:07 |
| cpaelzer | so far just FYI in case you want to subscribe to the bug and/or put it on your "sometime I should get to this" List of things that we all have :-) | 06:08 |
| seb128 | vorlon, firebird3.0 unsatisfiable Build-Depends(-Arch) on i386: recode | 07:35 |
| seb128 | do you know what's going on there? recode seems to be build on i386? | 07:35 |
| === Wryhder is now known as Lucas_Gray | ||
| xnox | mwhudson: 1 day and 9h later hopefully rustc on riscv64 is now publishing https://launchpad.net/ubuntu/groovy/riscv64/rustc | 08:45 |
| wgrant | xnox: Nicely done | 08:45 |
| xnox | wgrant: well launchpad publication history looks very odd, if i'm doing this again for focal, i think i will do "no change rebuild" such that things all arches are built in the same bileto ppa (which happens to have bootstrap compiler for riscv64 available as a build-dep) | 08:52 |
| xnox | hopefully, it will publish fine. | 08:52 |
| wgrant | xnox: What's weird about it? | 08:54 |
| xnox | wgrant: at https://launchpad.net/ubuntu/+source/rustc/1.43.0+dfsg1+llvm-1~exp1ubuntu2 buildlog is not yet visible, and riscv64 points at the in-archive build, instead of the copy from bileto. | 08:57 |
| xnox | (and the build in bileto) | 08:57 |
| xnox | wgrant: i think i should upload no-change rebuild after this publishes. | 08:57 |
| wgrant | xnox: Both builds are there | 08:58 |
| xnox | https://launchpad.net/ubuntu/groovy/riscv64/rustc/1.43.0+dfsg1+llvm-1~exp1ubuntu2 => i can click under Downloadable files on to https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/4195/+build/19815578 | 08:58 |
| wgrant | Which is a bit weird, but it happens. | 08:59 |
| xnox | i found a UI way to find it, so i am good. | 08:59 |
| xnox | seb128: rbalint: retriggered systemd tests with plymouth isc-dhcp triggers to hopefully have them pass and everything migrate. | 09:04 |
| seb128 | xnox, thanks | 09:09 |
| === Wryhder is now known as Lucas_Gray | ||
| xnox | mwhudson: "rustc unsatisfiable Build-Depends-Indep on amd64: wasi-libc (<= 0.0~git20200114.1fad338++)" | 10:13 |
| mwhudson | xnox: is that in focal? | 10:14 |
| xnox | mwhudson: groovy | 10:15 |
| mwhudson | xnox: oh | 10:15 |
| mwhudson | it's fairly easy to hack out the wasm stuff | 10:15 |
| xnox | mwhudson: i guess we need to keep up bootstrapping rustc up? cause debian is at 46 now, and we are now at 43 | 10:15 |
| mwhudson | xnox: i guess so | 10:16 |
| xnox | and one can only jump by one? | 10:16 |
| mwhudson | so apparently jumping by more than one "usually" works | 10:16 |
| mwhudson | but officially ye | 10:16 |
| xnox | mwhudson: i guess i'll throw 44 build into my bileto PPA, to then throw 45 build into it, to finish with 46? | 10:16 |
| mwhudson | s | 10:16 |
| xnox | let me check the contol file restrictions | 10:16 |
| mwhudson | rustc updates are usually fairly simple, it's cargo ones that are a headache | 10:16 |
| mwhudson | oh but if we're just merging from debian, that's usually fine | 10:17 |
| xnox | somehow our cargo is ahead of debian | 10:17 |
| xnox | but behind upstream | 10:17 |
| xnox | we are at 44 cargo, upstream is at 46 and debian is at 43 | 10:17 |
| mwhudson | how does debian have rustc 1.46 but not the corresponding cargo? | 10:17 |
| xnox | oh 46 is in experimental | 10:18 |
| xnox | debian is at 45 rustc in sid | 10:18 |
| xnox | with cargo at 43 in sid | 10:18 |
| xnox | dunno | 10:18 |
| mwhudson | yeah was going to say, is 1.46 even out yet? | 10:18 |
| mwhudson | i only update rustc when osomon asks me to :) | 10:18 |
| xnox | hahhahahahha | 10:19 |
| xnox | well it can't migrate now, so either new rust, or rebuild with "nowasm" | 10:19 |
| xnox | but i think we want new rust, no? | 10:19 |
| mwhudson | let me find the wasm disabling thing | 10:19 |
| mwhudson | yeah rolling forward probably makes sense | 10:20 |
| mwhudson | but might be nice to make it migratable before we get there, dunno | 10:21 |
| mwhudson | xnox: https://git.launchpad.net/~canonical-foundations/ubuntu/+source/rustc/commit/?h=focal-1.43&id=1439259a505ca4053c2a81d726e821213f0c34e9 | 10:21 |
| xnox | mwhudson: why do we even build it if nothing uses it? =) | 10:25 |
| mwhudson | xnox: /me points at debian | 10:25 |
| mwhudson | xnox: although i'm not sure quite what you're asking? | 10:25 |
| xnox | oh | 10:27 |
| xnox | but rustc is, well urgh. | 10:27 |
| xnox | your merge is large. | 10:27 |
| xnox | i don't want to merge it =) | 10:27 |
| xnox | i'd like rather rollback wasi-lib or cherrypick things for new one. | 10:28 |
| xnox | i think we should drop building wasi-lib for now, I don't see the point of it. | 10:29 |
| xnox | mwhudson: hm, based on https://wiki.mozilla.org/Rust_Update_Policy_for_Firefox we will need rust 45 soon. | 10:33 |
| mwhudson | xnox: better get cracking then i guess | 10:41 |
| xnox | so i'll create bileto ppas that build a sync of 44 & 45. | 10:44 |
| xnox | which we then will be able to use for the build of rustc which is "merged" properly. | 10:44 |
| rbasak | What version would be suitable for chromium-browser in Groovy then? 1:0? | 11:22 |
| rbasak | 1:0ubuntu1 maybe? | 11:22 |
| rbasak | Looks like it's currently a non-native package | 11:24 |
| seb128 | rbalint, sorry I missed the backlog, what's the issue? | 11:31 |
| seb128 | rbalint, it's a native package, just including an ubuntu revision in the version number? | 11:34 |
| seb128 | ups | 11:41 |
| rbasak | :) | 11:42 |
| rbasak | I mean https://lists.ubuntu.com/archives/ubuntu-devel/2020-August/041127.html | 11:42 |
| seb128 | rbalint, sorry, that was for rbasak | 11:42 |
| seb128 | I just saw that | 11:42 |
| seb128 | Olivier is out this week, let's wait monday for him to be back? | 11:42 |
| rbasak | I thought it might be wise to decide now to save round tripping if I'm unsure when Olivier uploads | 11:42 |
| rbasak | Ah I didn't realise. Sure, thanks. | 11:42 |
| seb128 | maybe accept that round of SRU meanwhile | 11:42 |
| rbasak | Can do if you think it's important to land it now? | 11:43 |
| seb128 | to avoid having a concrete upgrade issue after the next security round | 11:43 |
| seb128 | well if a security update in bionic become higher than focal than upgraders from bionic to focal aren't going to get the snap | 11:43 |
| rbasak | OK sure | 11:43 |
| * rbasak looks | 11:43 | |
| seb128 | that's already the case it seems? | 11:44 |
| seb128 | bionic has 84.0.4147.105-0ubuntu0.18.04.1 | 11:44 |
| seb128 | and focal proposed 84.0.4147.89-0ubuntu0.20.04.1 | 11:44 |
| rbasak | Yep so I'll accept this one in the Focal queue, right? That'll bring it up to 84.0.4147.105-0ubuntu0.20.04.1 | 11:45 |
| seb128 | yes, that fix the immediate problem | 11:45 |
| seb128 | then we can sort out the epoch use when Olivier is back | 11:46 |
| seb128 | rbasak, thanks! | 11:46 |
| rbasak | Done - though that's only in proposed | 11:46 |
| seb128 | that's a first step | 11:47 |
| seb128 | we can maybe discuss lowering the testing to less than a week, that's only a version bump | 11:47 |
| rbasak | I agree that should be fine. | 11:47 |
| rbasak | Probably best for someone to check the binary transitional package is as expected after it's built though | 11:48 |
| rbasak | Just to try to detect any human error | 11:48 |
| seb128 | right, I will do that | 11:49 |
| rbasak | Oh and if possible, today would be better than tomorrow to release - because Friday | 11:49 |
| rbasak | Thanks! | 11:49 |
| seb128 | np, thank you for the review! | 11:49 |
| rbasak | vorlon: so you want force-reset-test over SRUing a fixed test? This is a case where the broken test is affecting the SRU of a different package. | 11:59 |
| rbasak | vorlon: documentation in https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions please? I think what's there is out of date then? | 12:00 |
| rbasak | vorlon: I'd like a table mapping "situation" to "action expected". | 12:01 |
| rbasak | Where "action expected" might be "fix the autopkgtest in an SRU", "add a hint (specifying the type of hint)" and so on. | 12:01 |
| rbasak | Right now I'm not clear on what you intend for those things. | 12:01 |
| ahasenack | good morning | 12:05 |
| seb128 | ddstreet, so yeah, I was wrong yesterday, software-properties isn't migrating, it's blocked now because it makes livecd-rootfs uninstallable on i386, I expect it's due to the new depends on python3-launchpadlib | 12:09 |
| ddstreet | seb128 i'm not sure why that is though, python3-launchpadlib is 'all' arch | 12:10 |
| seb128 | that's probably not installable on i386 currently? | 12:10 |
| Laney | correct | 12:11 |
| ddstreet | why isn't it installable on i386 if it's 'all' arch? | 12:12 |
| seb128 | ddstreet, yeah, unsure, maybe that's not it, I guess you need to set up an i386 env to try... | 12:12 |
| ddstreet | ack, i'll look into it, it's confusing though | 12:13 |
| seb128 | indeed, I wish it was easier to locally try those i386 problems :/ | 12:13 |
| Laney | arch:all packages can depend on arch-specific ones | 12:14 |
| ddstreet | Laney right i get that, and in fact software-properties-common does (and python3-softwareproperties depends on it, and livecd-rootfs depends on that...so not sure how any of this 'worked' already) | 12:14 |
| ddstreet | ah, ok i see python3-software-properties doesn't dep on any of the other pkgs | 12:15 |
| Laney | ddstreet: It's probably going to turn out to be that some packages need 'bringing back' on i386 to make it installable again, first step would be to investigate in a chroot to find out what they are | 12:25 |
| ddstreet | at least python3-cryptography and python3-simplejson | 12:26 |
| ddstreet | unfortunately the 'livepatch' additions to python3-software-properties didn't actually update the pkg deps, as the new addition of softwareproperties.LivepatchService actually imports softwareproperties.gtk without declaring the dep on software-properties-gtk | 12:28 |
| ddstreet | i dont want to get into pulling the -gtk deps into i386... | 12:28 |
| ddstreet | rcj if i'm reading livecd-rootfs right, the only thing it needs python3-software-properties for is to print out a ppa signing_key_fingerprint? | 12:30 |
| ddstreet | if that's all, i can totally redo that for the livecd-rootfs pkg so we can drop the python3-software-properties dep | 12:30 |
| ddstreet | (or we could stop building livecd-rootfs for i386...) | 12:31 |
| cjwatson | Stopping building livecd-rootfs for i386 is not very possible, as it's required as part of having LP builds for i386 at all | 12:36 |
| cjwatson | (At least in the present configuration. Other configurations are conceptually possible but would be a good deal of work) | 12:37 |
| cjwatson | I don't think the software-properties bits are needed to make buildd chroots work though | 12:38 |
| xnox | mwhudson: rustc ubuntu archive build record kicked in now. I think you will get weird emails about it failing to upload, but we shall see. | 12:54 |
| xnox | and 44 build in bileto from debian ftbfs on i386, because llvm9 is gone. So yeah, to bootstrap 44 we will need ubuntu style merge of rustc. | 12:55 |
| xnox | not sure i want to do that. | 12:55 |
| xnox | horay | 14:06 |
| xnox | rbasak: systemd + isc-dhcp are migrating now | 14:06 |
| rbasak | Was that for me? | 14:07 |
| Odd_Bloke | xnox: vorlon: I'd appreciate your thoughts on https://github.com/canonical/cloud-init/pull/514/files#r469464608 | 14:30 |
| === M_hc[m] is now known as _hc | ||
| rafaeldtinoco | FTR: on-going open-iscsi merge (https://code.launchpad.net/~rafaeldtinoco/ubuntu/+source/open-iscsi/+git/open-iscsi/+merge/389234) | 14:50 |
| rafaeldtinoco | vorlon: ^ keeping your delta (the ones I did not merge in debian yet), will cleanup iscsi-root support (and merge things in debian as accepted) as a next step | 14:50 |
| rafaeldtinoco | im also keeping history as split git-ubuntu merges to keep history to server team when 1.1.1-1 from debian gets imported into usd-importer | 14:51 |
| rbalint | xnox, \o/ | 14:52 |
| slingamn | i submitted a draft review for a change to software-properties, but i seem to have mistakenly pinged the entire "Ubuntu Core Development Team" | 17:49 |
| slingamn | any suggestions for an individual reviewer for a software-properties change? | 17:49 |
| vorlon | ddstreet: I think we're nearly there on python3-launchpadlib; looks like it should only be 3 python packages that need added to the set (python-cryptography, simplejson, python-cffi) | 18:54 |
| ddstreet | vorlon awesome, and you want to use the alternate MR, to change livecd-rootfs to use python3-launchpadlib directly? | 18:55 |
| vorlon | ddstreet: I haven't seen the MP, can you link me to it? | 18:55 |
| ddstreet | https://code.launchpad.net/~ddstreet/livecd-rootfs/+git/livecd-rootfs/+merge/389260 | 18:55 |
| vorlon | thanks, I'll have a look | 18:56 |
| vorlon | xnox: why did you copy recode from groovy to groovy-proposed? | 18:57 |
| vorlon | xnox: ... because you were trying to make firebird3.0 build on i386, I guess... should've used --force-same-destination and copied it back to groovy instead | 18:59 |
| vorlon | seb128: recode restored in groovy release, so firebird3.0 should be migratable shortly | 19:00 |
| seb128 | vorlon, thanks | 19:04 |
| vorlon | rbasak: from what I see, https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions already documents this, aside from the s/force-badtest/force-reset-test/ optimization? | 19:38 |
| vorlon | ddstreet: and python-cryptography is built now on i386 so I believe as soon as we get a publishing run + britney run, software-properties should be good to migrate; if not, feel free to yell | 19:39 |
| realtime-neil | what's a good and proper way for an _unprivileged_ user to change his/her own locale, console keymap, x11-keymap-{layout,model,variant} --- and have that choice persist across login/reboots ? | 19:41 |
| vorlon | realtime-neil: an unprivileged user doesn't get to change the console keymap, the console is privileged | 19:43 |
| realtime-neil | vorlon: understood. what about the other items? | 19:44 |
| vorlon | realtime-neil: if you're using the Ubuntu Desktop and the necessary locale packages are installed, you should be able to configure everything through Settings -> Region and Language | 19:45 |
| realtime-neil | What if the "Region & Language" is missing from gnome-control-center? | 19:46 |
| realtime-neil | vorlon: related question: How does one restore the "Region & Language" tab to the gnome-control-center? | 19:48 |
| vorlon | realtime-neil: I have no idea how you would have managed to do that; and this is not a user support channel, maybe ask in #ubuntu? | 19:49 |
| realtime-neil | will do | 19:49 |
| smoser | anyone else seen this: https://paste.ubuntu.com/p/Hdxt9hMgwG/ | 21:24 |
| smoser | INFO: pkgstripfiles: waiting for lock (cloud-initramfs-rooturl) ... | 21:25 |
| smoser | keeps repeating | 21:25 |
| seb128 | software-properties did migrate, thanks vorlon! | 21:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!