/srv/irclogs.ubuntu.com/2020/01/23/#ubuntu-devel.txt

CarlFKto a new thing that uses yaml ... are there images using this yet?00:00
tjaaltonCarlFK: it's called subiquity, and the current focal images should use it aiui08:19
CarlFKtjaalton: thanks08:19
tjaaltonserver image that is08:21
xnoxmwhudson:  arrrrrrrrgh08:30
JackFrostInternational pirate day?08:30
mwhudsonxnox: p. much08:31
mwhudsonxnox: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/bionic/ubuntu-server-live/+build/200376 has new casper, not sure image has been published yet though08:32
mwhudson(it hasn't)08:34
CarlFKtjaalton: any idea why this is now an iso9660?  (which is RO, so I can't mount and tweek it's boot parameters) http://archive.ubuntu.com/ubuntu/dists/focal/main/installer-amd64/current/images/hd-media/08:56
tjaaltonno08:57
* CarlFK crys 08:58
alkisgHi all. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=792687 recommends NOT regenerating the .pot file on each build, yet python-distutils-extra's build_i18n.py does that unconditionally. Any way around it, or any other advice?09:36
ubottuDebian bug 792687 in src:gettext "gettext: please support timestamps from environment" [Wishlist,Fixed]09:36
=== cpaelzer__ is now known as cpaelzer
xnoxjibel:  bdmurray: is this critical for .4 release? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1860559 and do we need ask kernel team to fix this; or should we promote finalrd to main and make casper depend on it. As far as I can tell it's installer failing to reboot, rather than installed systems.10:10
ubottuLaunchpad bug 1860559 in linux (Ubuntu Bionic) "[18.04.4] System fails to reboot after installation" [Critical,Confirmed]10:10
xnoxsil2100:  ^10:11
sil2100xnox: I would like the kernel team to take a look and see what caused the regression, at least for now. Didn't dive into the bug so I'm not saying no to the finalrd approach, but before making a decision it would be wise to know what the fudge happened10:17
rbalintupdate_excuses is quite outdated :-(11:00
rbalintsil2100, , could you please trigger it? ^11:01
cjwatsonupdate_excuses is updated automatically as fast as it can go.  If it's outdated then it's not just a matter of triggering it.11:02
cjwatsonAlso it's only 14 minutes old.11:03
cjwatsonUpdated seconds after you asked, I think (the timestamp that's 14 minutes old is probably when the process started; the mtime on the filesystem is about three minutes old)11:04
rbalintcjwatson, thanks!11:13
sil2100rbalint: which update_excuses do you mean? Since if it's about the focal one, it's as cjwatson mentioned11:14
rbalintsil2100, cjwatson when i asked it was from 2020.01.22 21:06:24 +000011:14
rbalintsil2100, cjwatson i think ~11 hours is a bit long between updates, maybe something regressed and became very slow11:16
sil2100uh, sounds very unlikely, you sure you didn't have some cached version on your browser? I don't think britney ever took so long11:16
sil2100I can actually check the logs, wait11:16
sil2100rbalint: looking at the logs, it looks like britney was rather running as expected11:17
rbalintsil2100, strange, i tried refreshing11:18
sil2100rbalint: https://people.canonical.com/~ubuntu-archive/proposed-migration/log/focal/2020-01-23/11:18
rbalintsil2100, i'll check with wget next time11:18
rbalintsil2100, cjwatson sorry for the noise then11:19
sil2100rbalint: no worries! It still might have been something fishy indeed, but at least from the britney POV it seemed to be working as intended11:19
bdmurraysil2100: who will chase down the kernel team regarding that 18.04.4 bug?11:27
sil2100bdmurray: if xnox didn't yet, I will poke them in a moment11:29
sil2100bdmurray: jibel poked them already, I'll follow up as well11:31
rbasakwaveform: we should discuss this in public really, so let's chat here11:33
waveformrbasak, sure11:33
rbasakThe planned flash-kernel change fixes a race that already sort of exists I think11:34
rbasakSo perhaps we can land the u-boot-rpi changes now, certainly in Focal11:34
waveformrbasak, a race? Where's that?11:34
rbasakThat without the Depends it's possible flash-kernel will be deconfigured11:35
waveformah yes - of course, and now it'll ensure it runs anyway11:35
rbasakDoes the Breaks then need to be bumped to refer to the version where we fix this in flash-kernel - for Bionic or Focal or both?11:36
waveformhmm - the Breaks was mostly to do with the issue that the proposed u-boot will not accept the uboot.env from older f-k's (only applicable to the bionic pi2 image, but still) - but if we're relying on the new f-k trigger behaviour then it probably would be sensible to bump it up11:37
rbasakIt would ensure that an older flash-kernel won't be running at the time the postinst runs11:38
rbasakAnd flash-kernel will be upgraded if was installed at some point during a full upgrade11:39
waveformyup - sounds good - I'll tweak that commit; would I be right in thinking focal will be 2019.07+dfsg-1ubuntu4 and bionic will be 2019.07+dfsg-1ubuntu3.18.04.1 ?11:41
rbasakHang on11:41
rbasakI'm not certain the Breaks needs to be bumped!11:41
rbasakAnd do we know the version of flash-kernel that will be uploaded?11:42
rbasakThis mess is complicated :(11:44
rbasakddstreet: ubuntu-dev-tools imported11:45
rbasakIt'll be manual until my whitelist change lands11:45
waveformrbasak, doh! Was looking at the u-boot versions - just a mo11:48
waveformrbasak, well - f-k will be jumping from 3.90<stuff> to 3.98<stuff> on bionic so presumably it would be sufficient to just have Breaks: << 3.9811:50
=== Wryhder is now known as Lucas_Gray
ahasenacktribaal: hi, about your kitty tip, nice terminal, but I'm finding myself having to install kitty-terminfo in all remote servers I log into :/12:28
tribaalahasenack: for me it's only a problem if I want to tmux inside of (Ubuntu) remotes12:28
tribaaljust SSH seems to work out of the box12:28
ahasenacktribaal: saw that with tmux, yes, but just now saw that backspace didn't work without the terminfo package12:29
tribaaloh?12:29
* tribaal tries12:29
ahasenackmaybe something in focal, though. It was a focal lxd I started up12:29
tribaalworks fine here on a 16.04 remote12:29
tribaalhaving the terminfo installed by default in focal would be nice12:30
tribaalbut I guess that'd be a thing to send to upstream terminfo if there's such a thing :)12:30
* tribaal tries on a focal lxd as well12:32
tribaalahasenack: silly question: how do you launch a focal LXD? I see to remember there was a special alias for in-devel distros, but I can't remember what12:34
ahasenacktribaal: lxc launch ubuntu-daily:focal12:34
tribaalahhh ubuntu-daily: that's the trick :)12:35
tribaalahasenack:thanks12:35
tribaalahasenack: indeed, backspace doesn't work via SSH on eoan/focal apparently12:42
tribaalworks fine on bionic however12:42
ahasenackinteresting12:42
* tribaal tries disco12:42
tribaaldisco fails as well12:44
tribaalyep, just tried on our platform as well - it's not LXD-related12:47
tribaalworks on bionic, but not on disco+12:47
ahasenacklxd or not?12:49
tribaalahasenack: correct. For disco+ the backspace key doesn't work on LXD and on "normal" virtualisation12:50
tribaalfor bionci the backspace key seems to behave properly in both cases12:50
tribaal(LXD and non-LXD)12:50
ahasenacktribaal: I also seem to have lost my prompt colors12:53
ahasenacknoaha12:55
ahasenackcase "$TERM" in12:55
ahasenack    xterm-color|*-256color) color_prompt=yes;;12:55
ahasenackesac12:55
ahasenackbut TERM is xterm-kitty12:55
tribaalah ha12:56
tribaalahasenack: does it behave like this on bionic as well? I wonder why the backspace works there and not later12:57
ahasenackI'm on eoan12:57
ahasenackTERMs are complicated12:59
=== ricab is now known as ricab|lunch
tribaalahasenack: I'm on eoan as well, but connecting to bionic13:06
rbalintapw, could you please merge this hint for systemd? https://code.launchpad.net/~rbalint/britney/hints-ubuntu/+merge/37799814:28
=== ricab|lunch is now known as ricab
apwrbalint, could you check line 1415:03
rbalintapw, thanks, fixed!15:04
rbasakwaveform: on u-boot, give me a few minutes to think this through16:08
waveformrbasak, no prob16:08
rbasakWhich branches exactly are ready?16:09
rbasaksru-pi4 preferred for both?16:09
rbasakwaveform: do you have an LP bug ready? We'll need that for the SRU at a minimum16:09
rbasakBut probably best to write everything up there16:10
rbasakTogether with the required SRU paperwork16:10
waveformrbasak, yup - oh, I need to update the branch links though - just a mo16:10
waveformrbasak, LP: #184632916:11
ubottuLaunchpad bug 1846329 in u-boot (Ubuntu Bionic) "[SRU] 2019.07 to support Pi4 boot" [Undecided,In progress] https://launchpad.net/bugs/184632916:11
waveformrbasak, and just for reference the f-k one is LP: #184758716:11
ubottuLaunchpad bug 1847587 in flash-kernel (Ubuntu Bionic) "[SRU] Add entries for Pi 4" [Undecided,Fix committed] https://launchpad.net/bugs/184758716:11
waveform(should have the correct branch links in there too now)16:12
rbasakwaveform: you still need the trigger thing, don't you?16:22
rbasakBecause if flash-kernel is upgraded first and entirely, then the u-boot-rpi upgrade won't cause flash-kernel to run again otherwise?16:23
waveformrbasak, that doesn't matter - u-boot copies itself into the boot partition (it shouldn't but as discussed with infinity on -release last night that's a bigger fix than we want to do right now); so if flash-kernel goes first then it'll already have updated the boot scripts, then u-boot updates (and updates itself)16:24
rbasakOK, thanks.16:26
rbasakSorry, I had my head round this on Tuesday, but keeping my head around the complexities is quite difficult!16:26
waveformrbasak, no prob - I know the feeling :)16:27
rbasakdannf: o/16:31
rbasakdannf: do you have any interest currently in u-boot or flash-kernel?16:31
dannfrbasak: hi - no, i don't16:32
rbasaknp, thanks.16:32
rbasakWe're about to do a major SRU, so I thought I'd check.16:32
* ahasenack has a pi4 with focal16:39
ackkhi, I have a tricky deb question. package A is using dbconfig-common to manage a postgres database. now I want to replace A with B (which has Breaks/Replaces/Provides: A). when A gets removed, it asks if the database should be purged. I need that not to happen, is there a way to preseed that answer from scripts in  B?16:41
rbasakwaveform: so I'm uploading f8f21a5 to Focal, right?16:53
waveformrbasak, yup - that looks right16:53
rbasakYou don't mention the shebang drop in that chagnelog entry. Intentional?16:53
rbasakwaveform: ^16:55
rbasakIn Bionic, you want 2019.07+dfsg-1ubuntu3~18.04.1 instead of 2019.07+dfsg-1ubuntu3.18.04.1 since we need it to be lower than the version in Focal. I can tweak that.16:56
waveformrbasak, hmm - what's the rules surrounding purely cosmetic changes?16:56
rbasakwaveform: fuzzy. I don't mind either way.16:57
rbasakWhat about Disco and Eoan?16:57
waveformin that case, let's say intentional (rather than "Dave forgot at whatever the unholy hour was last night" :))16:57
rbasak:-)16:57
rbasakDisco is EOL now16:58
rbasakBut usually we need to SRU Eoan too16:58
rbasakSorry I hadn't looked at that before.16:58
rbasakvorlon: opinion on Eoan? ^16:58
rbasakI'm happy to upload your Focal branch16:58
rbasakBionic is ready to go with that tweak.16:58
rbasakJust the Eoan question. That's more of an SRU hat thing.16:59
waveformDisco's already EOL and Eoan already has pi4 support so from that perspective there's no absolute requirement - but the f-k changes are probably sufficiently important to SRU; the u-boot changes may be relevant for anyone upgrading from bionic16:59
rbasakOh, the versioning is good16:59
rbasakBecause you're using ubuntu316:59
waveformwell ... I'm using whatever dch -i gave me frankly ... really hadn't thought that hard about it17:00
rbasakLet's say we land your branch as-is right now (with the minor version string tweak to ~ for Bionic)17:00
rbasakThen what will happen if a user upgrades to Eoan?17:00
rbasakEoan has 2019.07+dfsg-1ubuntu3, which doesn't include your planned changes in Focal 2019.07+dfsg-1ubuntu417:01
rbasakActually let's defer that.17:02
rbasakI'll leave a note in the bug.17:02
waveformjust checking something ... okay, there's a potential issue in the case that they're upgrading from the weirdo pi2 image with the differing root file-system label, and the fix for the 3A+ booting is also missing17:02
rbasakLet's at least put the thing in bionic-proposed assuming the SRU hat is happy. I don't think an Eoan fix will require a change to the Bionic upload.17:02
rbasakYou might need to backport the fixes in Focal back to Eoan in an SRU.17:03
rbalintsil2100, coult you please merge this for the libnfs transition? https://code.launchpad.net/~rbalint/britney/hints-ubuntu/+merge/37800317:03
waveformokay - I'll prep and Eoan branch17:03
rbasakThat should be checked before the Bionic SRU is released I think.17:03
waveform*an17:03
rbalintdoko, could you please do that? LP: #186069817:11
ubottuLaunchpad bug 1860698 in skiboot (Ubuntu) " Please remove i386 binaries for 6.5-1" [Undecided,New] https://launchpad.net/bugs/186069817:11
rbasakwaveform: uploaded17:11
rbasakwaveform: please get yourself PPU for u-boot :-P17:11
waveformrbasak, thanks - will try!17:12
rbasakwaveform: FTBFS :-/17:17
rbasakLooks like some Python thing17:17
rbasakCan I leave that with you?17:18
waveformrbasak, bizarre ... built in the PPA - yup, leave it with me17:19
rbasakwaveform: maybe a component mismatch?17:19
rbasakIf your PPA has universe visible. u-boot being in main does not for archive builds.17:19
rbasakAlthough17:20
waveformbah "Depends: libpython-dev but it is not going to be installed"17:20
rbasakActually, didn't that chagne a while back?17:20
rbasakAnyway, I'll leave it with you :)17:20
ahasenackwas in main in disco, not so much for eoan17:20
dokorbasak: I'll leave the i386 removals with vorlon17:21
ahasenackdisco and before, that is17:21
waveformno idea (if that changed a while back) - but yeah, I'll dig into it17:21
dokowaveform: libpython2-dev17:21
dokofor a quick fix17:21
ahasenackalso in universe for eoan+17:21
dokowaveform: which package is this?17:23
waveformdoko, u-boot-rpi (from u-boot)17:23
dokoI'll have a look to built it17:24
sil2100rbalint: ACK17:33
dokowaveform: https://paste.ubuntu.com/p/rhDdKKFfny/   not checked. please make sure that the binary package has the correct dependencies and uses python2 as the shebang17:34
waveformdoko, thanks - will do17:36
waveformrbasak, I've incorporated doko's python2 patch (plus some extra fixes for any/all bare "python" shebangs I could find) into the focal-clean-up branch, rebased and pushed that - just building the pkg in a PPA again18:31
waveformrbasak, okay - ignore me - still got a build failure - having a look18:40
ahasenackautopkgtest is driving me nuts, why isn't it adding deb-src lines anymore?19:04
ahasenackfailed with stderr "E: You must put some 'source' URIs in your sources.list19:04
ahasenackautopkgtest -o dep8 -U -s --apt-pocket=proposed=src:php7.3 ./php-doctrine-dbal-2.10.1/ -- lxd ubuntu-daily:focal19:04
ahasenackthat is my command line19:04
ahasenackbefore I had -B, because this package doesn't need rebuilding19:04
ahasenacksame thing19:04
waveformrbasak, if you're still around: got a u-boot version that builds in -proposed now - I've pushed the rebased branch which has one additional commit at the top to handle the python 2 switchover; all subsequent commits are unaltered20:04
coreycbtjaalton: Hi, I see you're on sru rota tomorrow. if you have a chance would you be able to take a look at the the openstack uploads in unapproved queues for bug 1858933 and bug 1723030?20:48
ubottubug 1858933 in octavia (Ubuntu Eoan) "[SRU] train stable releases" [High,Triaged] https://launchpad.net/bugs/185893320:48
ubottubug 1723030 in python-oslo.policy (Ubuntu Bionic) "Under certain conditions check_rules is very sluggish" [Medium,Triaged] https://launchpad.net/bugs/172303020:48
tjaaltoncoreycb: ok20:48
coreycbtjaalton: thanks20:48
* mwhudson rubs his eyes20:56
mwhudsonautopkgtest --apt-pocket=proposed -U . fails in the build step20:57
mwhudsonsbuild -d focal passes20:57
mwhudsonhow does that happen?>20:57
mwhudson(for python-jsonschema fwiw)20:57
mwhudsonif this pure python test depends on the running kernel version i am going to be very angry20:57
mwhudsonah i think the failing tests are skipped in the sbuild case for some reason20:58
mwhudsonah because python3-json-pointer is not installed in sbuild21:36
mwhudsonwhy the heck is it in cloud-images21:36
mwhudsoncloud-init -> python-jsonpatch -> python-json-pointer21:37
vorlondoko, rbasak: ugh why is that skiboot binary there?  looking to see why my scripts haven't already removed it22:42
vorlonuh weird it's an arch: all binary22:42
vorlonso that's why22:42
rbasakwaveform: can you give me a commit with an archive-suitable changelog please?22:42
rbasakWith 2019.07+dfsg-1ubuntu4 in focal-proposed, it needs to be 2019.07+dfsg-1ubuntu5 presumably22:43
=== ben_r_ is now known as ben_r
vorlonLaney: do you have any idea about the autopilot-gtk/arm64 autopkgtest failure with new glib2.0? autopilot-gtk itself is uninteresting and will probably be removed as unmaintained (python2), but I'm not sure whether the failure represents a glib2.0 regression23:53

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!