/srv/irclogs.ubuntu.com/2020/06/17/#ubuntu-devel.txt

mwhudsonanyone happen to know how grub finds its configuration when booting under uefi?03:31
stgraberthere's normally a minimal grub.cfg alongside grubx64.efi in /boot/efi/EFI/ubuntu/ which has enough logic to source the main config from whatever partition has /boot/grub.cfg03:37
mwhudsonstgraber: thanks, turns out i was not testing the thing i thought i was testing, sigh04:00
tewardbryyce: rbasak: when the nginx package has been git-ubuntufied can you give me the location to git-ubuntu clone it?05:11
teward(thank you!)05:11
bryyceteward, sure.  It'll be findable at https://code.launchpad.net/ubuntu/+source/nginx, or just run 'snap install git-ubuntu; git ubuntu clone nginx'05:24
bryyceteward, I finished importing it to git ubuntu today, but need to make sure what remains is accurate, and build/autopkgtest it.05:27
tewardcool.  i have git-ubuntu installed already so i'll just git ubuntu clone it as you indicated.05:32
tewardand again if there's anything you want me to peek at let me know, there's a few changes also in the works - moving some of the SSL components to a conf.d snippet instead of keeping in nginx.conf (been requested many times)05:32
tewards/SSL components/default SSL components/05:32
tewardso that's a packaging level change but i'll let the merge happen first :)05:32
bryyceteward, thanks, yeah would be helpful to get your eyes on the result to make sure I haven't messed any important bits up :-)05:35
bryyceteward, btw I noticed debian seems to have some of the php version stuff under /run rather than /var/run, so wondered if that's led to any issues with anyone, or if it's just a theoretical issue05:37
tewardbryyce: it's more of an issue when we do backports to releases where everything isn't in /run or /var/run or such isn't symlinked to /run or a few other cases - i haven't heard of it causing any issues in Debian so long as the configuration is updated properly to know the different locations.  The NGINX PPA has been a direct-from-debian sync since eternity and has not had any issues with /run vs. /var/run so long as the PID config is done05:45
tewardright.05:45
tewardbut i'll take a look after i wake up05:46
teward01:46 local and last night I didn't sleep well so i need to get to bed heh05:46
bryyceteward, sleep well05:46
=== pieq__ is now known as pieq
tewardbryyce: ideally what we're going to do here is match the Ubuntuisms with the configs (php paths, etc.) so if you want to do me a solid and see where php-fpm in Groovy ends up dropping its socket, that'd be great, I *think* it's a /var/run path, but I forget exactly.  :P05:49
tewardi'm also thinking of moving the PHP configs to a snippets/ config snippet for easier inclusion into other configs as well, just as a heads up for future planned ideas/changes that are minor but make things look nicer and easier for end users heh05:50
bryyceteward, that'd be great, thanks05:54
bryyceyeah I can look into what php-fpm does05:55
tewardbryyce: I don't have upload / push privs it seems, heh.  not sure whether that's an issue, but I can just dump merge requests if necessary06:04
bryyceteward, Mp's are fine, that's basically our standard procedure already.  afaik push privs are just == core dev06:06
tewardheh.  well i am coredev so xD06:06
tewardalls good.  actually going to sleep now.  /off06:07
bryycenight o/06:07
knoctedo gtk2 apps have the same theme as gtk3 apps in ubuntu 20.04? or is that a known issue?08:05
seb128rbasak, hey, if you doa  SRU shift today could you review the gnome-shell stack in the focal queue? it's waiting for 10 days and including some fixes we would like to see landing10:19
seb128rbasak, also for an easier one gnutl28/focal would be nice10:20
seb128rbasak, gnutls28 uploaded to bionic as well, would be nice to also review, it's a simple patch but it should fix connection failing to yahoo/aol/verizon emails servers (following server changes so it used to work and recently stopped working in bionic/focal)10:27
rbasakseb128: sure - if I get to my SRU shift today I'll look at those.10:55
cpaelzerhi doko, if I'd file a LP bug to get a fix for https://sourceware.org/bugzilla/show_bug.cgi?id=23465 into Bionic is ther any chance this will happen or should I stop right away?11:13
ubottusourceware.org bug 23465 in gas "wrongly scale non-8-bit x86 displacements" [Normal,Resolved: fixed]11:13
cpaelzerfnordahl: ^^ FYI11:13
cpaelzerwell I blindly assumed binutils would usually be you doko, let me actually check the changelog  ...11:14
dokocpaelzer: including the referenced commit in the bug report?11:16
cpaelzerdoko: yeah I guess that is what would need to be backported11:19
cpaelzermostly changing the self tests afaics11:19
cpaelzerI think the effective functional change would just be https://sourceware.org/git/?p=binutils-gdb.git;a=blobdiff;f=gas/config/tc-i386.c;h=b8042ba434207efe9d70c9de9654a2ff110a2c9d;hp=cd69321bdb25f698f7764e5648d10253472724a3;hb=4e518864c879be2e6af4c64415e8775d9a20deaf;hpb=e521dc888158a6cdbdccef0397e663c437450a8011:20
fnordahlcpaelzer: ta for heads up11:25
cpaelzersure fnordahl - if doko says it appears to be generally doable I'd file a proper LP bug for it and reply that to the mail11:27
cpaelzerasking them to chime in on the bug for e.g. testing11:27
mantas-baltixAnyone noticed, that Snap store launcher is untranslated in Applications menu (Activities Overview)? I've registered a bug #1882929 , but no answer from developer in 7 days :(11:31
ubottubug 1882929 in snap-store "Ubuntu Software (Snap Store) desktop launcher Name isn't translated in Ubuntu 20.04" [Undecided,New] https://launchpad.net/bugs/188292911:31
cpaelzerdoko: I'll file a bug as that will be quick - could you state ther eif you consider this doable or not?11:33
cpaelzerthat will also serve as a single place to unite communication on the topic11:33
cpaelzerdoko: https://bugs.launchpad.net/ubuntu/+source/binutils/+bug/1883880 it is11:34
ubottuLaunchpad bug 1883880 in binutils (Ubuntu Bionic) "fix non-8-bit x86 displacements breaking AVX512 builds on Bionic" [Undecided,New]11:34
cpaelzerfnordahl: ^^ for you to subscribe to11:34
seb128rbasak, thanks!11:36
ahasenackseb128, doko: hi, I'm starting my day, +1 maintenance for the whole week. I updated the wiki with what I have going on11:55
seb128ahasenack, ack, thanks, I'm about to start as well, I will check what you put there and update the channel as needed11:56
dokoahasenack, seb128: manually cleaned up some NBS binaries in -proposed, now working from the bottom, on ftbfs. lib3mf11:59
ahasenackdoko: did you happen to see mail-stack-delivery (from dovecot) in those nbs?11:59
ahasenackit's not in the report yet, I think because current dovecot (a sync) is an ftbfs11:59
dokono12:00
ahasenackok12:00
ahasenackone thing I'm noticing with this +1 rotation is that my inbox is getting out of control12:09
ahasenacksince I'm not focusing on other tasks12:10
ahasenackhttps://github.com/golang/go/issues +5000 issues open12:19
ahasenackwhat chance to I have12:19
ahasenack*do12:19
ahasenackxnox: about that golang issue from yesterday, should I file a bug upstream with go? I fear a rabbit hole if I keep digging, like starting a matrix of tests with multiple versions of unpatched go (i.e., not from ubuntu, but straight upstream)12:20
ahasenackI could file that issue upstream, and upload the yaml.v2 package with that crazy patch of mine with a link to that issue12:20
ahasenackthat would unblock our migration, and help prometheus to migrate as well I'm told12:21
ahasenackdoko: do you have an opinion? Do you fancy go? :)12:21
ahasenackcrazy patch is https://bugs.launchpad.net/ubuntu/+source/golang-yaml.v2/+bug/1883770/comments/412:21
ubottuLaunchpad bug 1883770 in golang-yaml.v2 (Ubuntu) "Tests fail on s390x and go >= 1.13" [Undecided,New]12:21
dokoahasenack: you should ask mwhudson ;p   but yes, I think we need a more general solution for all the go packages12:22
ahasenackit's the first I see of this, can't say it's affecting others12:23
ahasenackmwhudson is many hours away12:23
ahasenackI'll file the upstream bug, see how they respond over this week while I'm still on +1 maint12:23
* ahasenack thinks we need an acronym for "+1 maintenance"12:23
rbasakPOM12:25
rbasakWhen there are two people on rotation, you would therefore have POM POM.12:25
ahasenacksounds like PAN PAN12:25
ahasenackinside airplane pilots joke12:25
=== alan_g is now known as alan_g_
ahasenackxnox: I filed https://github.com/golang/go/issues/3965113:39
ahasenackxnox: I also just tried upstream's s390x build and the test fails in the same way, so it doesn't look any of our patches are the culprit14:11
cpaelzerdoko: the upsream people already chimed in on bug 1883880 to help wit ha testcase15:06
ubottubug 1883880 in binutils (Ubuntu Bionic) "fix non-8-bit x86 displacements breaking AVX512 builds on Bionic" [Undecided,New] https://launchpad.net/bugs/188388015:06
cpaelzerdoko: that should make an SRU easier15:06
sforsheeLocutusOfBorg: dkms 2.8.2-1ubuntu1 is breaking building dkms builds in our kernel packages, and it looks like the change which breaks us has been reverted upstream - https://github.com/dell/dkms/commit/885f8275aa65fb11be1e17bc28a0b0ea734dc58515:17
sforsheecan we get that reverted in our package too?15:17
sforsheeapw: ^15:19
apwutter shit15:19
apwoops :) applogies15:20
LocutusOfBorgohhhhhhhhhhhhhh15:54
LocutusOfBorgthansk sforshee15:54
LocutusOfBorgI was trying hard to find that commit!15:54
LocutusOfBorgthanks a lot!15:54
LocutusOfBorgdamn, I didn't look at irc :/15:54
LocutusOfBorgI have another possibly breaking change, lets see how far we go15:56
=== ijohnson is now known as ijohnson|afk
=== ijohnson|afk is now known as ijohnson
ahasenackI have a curious go test problem. In a PPA, it fails because it cannot download a certain go module19:38
ahasenackbut locally in lxd, with network *disabled*, it doesn't try that, and it works19:38
ahasenackhttps://pastebin.ubuntu.com/p/G3FprRXztN/19:38
ahasenackany ideas?19:38
ahasenacksame package, version, proposed enabled19:38
ahasenackah, wait, I don't have proposed enabled in the lxd container19:38
LocutusOfBorgsforshee, looks really better now :D19:39
sforsheeLocutusOfBorg: great, let me retry these kernel builds and see how it goes19:40
ahasenackhm, no change with proposed19:47
ahasenackfor some reason, locally it doesn't reach out19:47
ahasenackI wonder if it's about a controlling terminal during the build, if it behaves differently in each case19:47
ahasenackaha!19:54
ahasenackit was the terminal19:54
ahasenackif I run dpkg-buildpackage -uc -us > ../log 2>&119:54
ahasenackthen it fails locally as well19:54
ahasenacksneaky!19:54
ahasenackhm, scratch that19:57
=== nobuto8 is now known as nobuto
=== xedniv_ is now known as xedniv
=== CarwynNelson5 is now known as CarwynNelson
ahasenackmwhudson: hey, saw you are online20:35
mwhudsonahasenack: hi20:35
ahasenackmwhudson: if you have an idea, I have this problem: in a ppa, the go test tries to pull in a module from the internet20:35
ahasenackbut locally (lxd), it doesn't: https://pastebin.ubuntu.com/p/G3FprRXztN/20:35
=== balloons7 is now known as balloons
ahasenackhttps://launchpad.net/ubuntu/+source/golang-github-fsouza-go-dockerclient/1.6.5-1/+build/19442767 is the package/build in question20:35
ahasenackand locally I tried without a network (used lxc exec to enter the container, and disabled eth0)20:36
mwhudsondoes the package depend on golang-golang-x-crypto-dev?20:36
mwhudsonbut hm yeah that's strange20:37
ahasenackmwhudson: no, we don't even have that package20:38
mwhudsonoh i bet golang-github-docker-docker-dev depends on golang-golang-x-crypto-dev in debian but not ubuntu20:38
ahasenackmwhudson: looks like it was added by http://paste.ubuntu.com/p/23vMxTBWCV/ once upon a time, and upstream took it20:39
ahasenackmwhudson: oh, so golang-golang-x-crypto-dev provides that ssh/terminal bit?20:39
mwhudsonyes20:39
ahasenackbut still wouldn't explain why it works locally20:39
mwhudsonthat's true that is strange20:40
ahasenackhm, so I have that installed20:40
ahasenackbut the builder log doesn't show it being installed20:40
ahasenackI'll investigate that20:40
ahasenackmwhudson: indeed, the debian build log shows golang-golang-x-crypto-dev being installed20:41
ahasenackours doesn't20:41
ahasenackinteresting20:41
ahasenackmwhudson: I have enough to go on now, many thanks20:41
mwhudsonahasenack: it'll be related to the golang-docker-*-dev stuff20:41
=== smoser1 is now known as smoser
mwhudsonin any case, if the package imports it directly, it should be in b-d, iow that debian patch should have come with a change to control as well20:43
ahasenackmwhudson: check out this fun golang bug I filed today, btw: https://github.com/golang/go/issues/3965120:44
mwhudsonahasenack: mundaym is pretty good at fixing this kind of stuff20:44
mwhudsonahasenack: did you try 1.15beta1 fwiw? it's available as a snap now20:48
ahasenackmwhudson: I noticed he is from ibm, and he mentioned he fixed a similar bug in the past20:48
ahasenackmwhudson: I didn't, I just tried the upstream binary tarball for s390x, I saw they had one20:49
mwhudsonahasenack: yeah, he did the s390x port20:49
mwhudsonahasenack: https://dl.google.com/go/go1.15beta1.linux-s390x.tar.gz is a thing too, might be worth a try if you have an environment set up20:49
ahasenackoh, I do20:49
ahasenacklet me fetch that20:49
ahasenackmwhudson: 1.15b1 failed the same way20:55
mwhudsonahasenack: ah well20:55
ahasenackworth a try20:55
mwhudsonleave it to the ibm elves then :)20:56
ahasenackhehe20:56
mwhudson(unless you want to start staring at disassembly)20:56
ahasenackmwhudson: there is a shocking diff between the Depends of golang-github-docker-docker-dev in debian and ubuntu: https://pastebin.ubuntu.com/p/9d8XmsfGqT/20:59
ahasenackwe have all that stuff vendorized?20:59
mwhudsonahasenack: ah hah um yeah20:59
mwhudsoni don't know, tbh20:59
ahasenackwe don't use dh-golang to build21:04
ahasenackand debian declares a *lot* of manual Depends packages21:05
ahasenackok, something for later21:05
ahasenacklet me fix this build21:05
ahasenackmwhudson: https://code.launchpad.net/~ahasenack/ubuntu/+source/golang-github-fsouza-go-dockerclient/+git/golang-github-fsouza-go-dockerclient/+merge/38595721:46
ahasenackconfirmed fixed in https://launchpad.net/~ahasenack/+archive/ubuntu/fsouza/+packages21:47
mwhudsonahasenack: +121:47
ahasenackmwhudson: can you do a quick +1 in the mp?21:47
xnoxahasenack:  we have outofsync containerd/docker-dev stuff between debian and ubuntu21:49
xnoxahasenack:  i think we are newer, and that's causing divergence in the build-depends.21:49
ahasenackmwhudson: awesome, thanks21:49
xnoxbecause stuff got reorged21:49
ahasenackxnox: it's quite the divergence21:49
ahasenackI filed a bug to investigate that21:49
ahasenackhttps://bugs.launchpad.net/ubuntu/+source/docker.io/+bug/188397821:50
ubottuLaunchpad bug 1883978 in docker.io (Ubuntu) "Possibly missing dependencies in golang-github-docker-docker-dev" [Undecided,New]21:50
ahasenackxnox: it was a difference in depends, not just build-depends21:50
ahasenackhaven't checked b-d21:50
mwhudsonalso the debian maintainer has a king sized pair of jfdi boots21:50
ahasenackit's weird to have golang.*-dev packages as depends21:53
ahasenackrbalint: doko: osomon: seb128: hello fellow +1maintainers for tomorrow (18th), I kept my wiki entry up-to-date under the "Progress" title22:01
* ahasenack -> EOD22:02
sarnoldahasenack: nice find22:08

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