/srv/irclogs.ubuntu.com/2019/10/30/#snappy.txt

mupPR snapd#7691 closed: builtin/browser_support.go: allow monitoring process memory utilizati… <Created by Erick555> <Merged by anonymouse64> <https://github.com/snapcore/snapd/pull/7691>00:21
mupPR snapd#7693 opened: Enabled GPG plug to access extra socket <Created by Kedstar99> <https://github.com/snapcore/snapd/pull/7693>01:20
mupPR snapd#7693 closed: Enabled GPG plug to access extra socket <Created by Kedstar99> <Closed by Kedstar99> <https://github.com/snapcore/snapd/pull/7693>01:35
mupPR snapd#7693 opened: Enabled GPG plug to access extra socket <Created by Kedstar99> <https://github.com/snapcore/snapd/pull/7693>01:38
mupPR snapd#7689 closed: client,daemon: pass sha3-384 in /v2/download to the client <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7689>05:47
mupPR snapd#7692 closed: tests: opensuse tumbleweed has similar issue than arch linux with snap --strace <Created by sergiocazzolato> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7692>05:50
mborzeckimorning06:15
mvohey mborzecki06:19
mborzecki-3 in the morning06:20
zygaHey06:20
zygaI’ll start sometime after 806:21
zygaSorry about all the bug spam from last night06:21
mborzeckizyga:  hey06:28
mupPR snapd#7684 closed: tests: disable mount-ns test in release/2.42 <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7684>07:05
mupPR snapd#7661 closed: packaging: tweak handling of usr.lib.snapd.snap-confine <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7661>07:06
=== pstolowski|afk is now known as pstolowski
pstolowskimorning07:51
zygahey pawel08:09
zygahey mvo08:12
zygarebooting08:31
mborzeckipstolowski: hey08:39
pedronis#7679 needs a 2nd review08:40
mupPR #7679: many: changes to testing in preparation of Core 20 seed consuming code <Created by pedronis> <https://github.com/snapcore/snapd/pull/7679>08:40
mupPR snapd#7685 closed: snapstate,devicestate: make OldModel() available in DeviceContext <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7685>09:17
ondrasergiusens can you translate for me this error "Snap 'grade' was set to 'stable' but must be 'devel'." and "Set 'grade' to 'devel' or use a stable base for this snap."09:49
ondrasergiusens candidate channel and something what worked yesterday09:50
mupPR snapd#7679 closed: many: changes to testing in preparation of Core 20 seed consuming code <Created by pedronis> <Merged by pedronis> <https://github.com/snapcore/snapd/pull/7679>09:53
sergiusensondra: can you share your snapcraft.yaml?10:11
sergiusensand build-environment10:11
ondrasergiusens https://git.launchpad.net/~ondrak/ondras-snaps/+git/openhab-snap/tree/snap/snapcraft.yaml?h=milestone-ondra10:11
ondrasergiusens build in LP with candidate channels for core/core18/snapcraft10:12
sergiusensondra: oh, building with candidate channel of core18 could do it, let me look into it10:13
ondrasergiusens that seems like way to constraining10:13
sergiusensondra: it was done to support core20, non-stable base means you can create a non-stable snap10:14
sergiusensondra: and yes, it is a way of constraining, it is exactly what grade is10:15
sergiusensbut I will need to think if your use case is one we want to allow for a stable grade10:15
ondrasergiusens so I will be only able to build stable grade with stable base snap installed on build machine?10:15
* sergiusens notices it is well before working hours and parts10:16
ondrasergiusens because we build and push into non stable channels and then promote, this will essentially kill use-cases when we want test build agains future stable core snap10:17
mupPR snapd#7694 opened: many: load/consume Core 20 seeds (aka recovery systems) <Created by pedronis> <https://github.com/snapcore/snapd/pull/7694>10:52
mupPR snapd#7695 opened: o/devicestate: the basics of Core 20 firstboot support with test <Created by pedronis> <https://github.com/snapcore/snapd/pull/7695>11:01
* Chipaca takes a break11:03
mupPR snapd#7696 opened: cmd/snap,image: initial support for Core 20 in prepare-image with test <Created by pedronis> <https://github.com/snapcore/snapd/pull/7696>11:06
pedronisChipaca: mvo: ^ those are the current PRs,  7694 is the base and the larger one11:07
mvopedronis: thank you11:07
Chipacapedronis: ack11:09
mupPR snapd#7697 opened: interfaces/apparmor: handle pre-seeding mode <Prebaking> <Created by stolowski> <https://github.com/snapcore/snapd/pull/7697>11:11
mupPR snapd#7650 closed: o/ifacestate: unify code into autoConnectChecker.addAutoConnections <Created by pedronis> <Merged by pedronis> <https://github.com/snapcore/snapd/pull/7650>11:13
* pstolowski lunch11:52
=== ricab_ is now known as ricab
=== ricab_ is now known as ricab
mupPR snapd#7698 opened: seed/internal: doc comment fix and drop handled TODOs <Simple 😃> <Created by pedronis> <https://github.com/snapcore/snapd/pull/7698>12:07
mupPR snapd#7674 closed: interfaces: de-duplicate emitted update-ns profiles (2.42) <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7674>12:17
mborzeckipedronis: i've updated #766512:21
mupPR #7665:  devicestate: add support for gadget->gadget remodel  <Remodel 🚋> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/7665>12:21
pedronismborzecki: I put it into my queue again12:30
mupPR snapd#7699 opened: release: 2.42.1 <Created by mvo5> <https://github.com/snapcore/snapd/pull/7699>12:43
=== ricab is now known as ricab|lunch
mborzeckicachio: hi, can you check whether fedora 31 images are already available in gce?13:29
cachiomborzecki, sure13:30
ijohnsonmorning folks, how is travis doing this morning13:32
cachiomborzecki, sorry, I forgot13:32
ijohnsonsome really weird failures it seems overnight, I had a unit test fail on me for snapstate.SnapState not having a Channel field but it definitely has that field?13:32
cachiomborzecki, we build fedora images13:33
cachiobecause are not published for gce13:33
mborzeckicachio: oh, using the cloud.fedoraproject.org ones?13:33
cachiomborzecki, we use the cloud ones yes13:33
mborzeckicachio: then 31 are up at https://alt.fedoraproject.org/cloud/13:34
cachiomborzecki, I could create one by using http://fedora.c3sl.ufpr.br/linux/releases/31/Cloud/x86_64/images/13:35
mborzeckicachio: sgtm, thanks!13:41
Chipacazyga: https://github.com/whitequark/unfork/blob/master/README.md13:57
zygaChipaca: I read that13:57
zyganeat stuff13:57
Chipacazyga: i thought you might've13:58
Chipacabut, yeah13:58
Chipacabonkers13:58
zygamvo: I simplified this stuff, no snap-confine changes required14:27
zygamvo: it's just a small go change now14:27
mvozyga: oh?14:31
mvozyga: how did you do that?14:31
zygamvo: it works super nice already14:31
zygamvo: working on snapd side now14:31
zygamvo: we can actually wire it in cmd_run.go14:31
mvozyga: what is the mechanism that you use there?14:31
mvozyga: just looking for a flag file?14:31
zygamvo: the same as before, yes14:31
mvozyga: nice14:32
zygamvo: if there's a feature flag enabled14:32
zygamvo: and the inhibit file is present14:32
zygamvo: loop on the presence14:32
* mvo nods14:32
zygamvo: // TODO: ask snapd for details14:32
zygamvo: it even _runs_ the app later14:32
zygait is easier and better than the idea that snap-confine execs that14:32
mvozyga: nice14:33
Chipacain unrelated news, don't do an image search for 'grub size', if you're thinking about bootloaders.14:35
jdstrandwoo, looks at those small snap-update-ns profiles in 2.42.114:39
zygajdstrand: indeed :)14:39
jdstrandzyga: ^ (hi)14:39
zygafixing bugs is the bests part of this job14:39
jdstrandzyga: I did notice, and this isn't a big deal, that there is a blank line between every rule14:39
zygaI noticed14:40
zygaI'll fix it up, just a bit annoying at the time14:40
* jdstrand nods14:40
zygasnippets are joined with "\n" automatically14:40
jdstrandyeah14:40
jdstrandzyga: not a big deal. loving the reduced ruleset :)14:40
zygarequires either some test adjustment or some careful coding to prevent spurious \n14:40
* jdstrand hugs zyga 14:40
zygayeah, I will second what you said in a review14:40
zygait's surprisingly readable14:41
zygabecause it builds a context14:41
zygaand adds incremental changes14:41
zygaI didn't plan for that, it just ended up working better than I expected14:41
jdstrandzyga: I know, right? that OrderedSet worked wonderfully14:41
jdstrandme too :)14:41
zygajdstrand: I'm doing a small prototype but I will be pushing that cgroup PR forward soon14:41
zygajdstrand: with fixes that make lxd work14:41
jdstrandzyga: sweet :)14:42
jdstrandI want to update my test vms to 2.42.1 now that the dupe rules are gone, but I can what for stable14:43
jdstrandI'll just have to 'settle' for fast refreshes on my laptop for the next week or so :)14:43
=== ricab|lunch is now known as ricab
zygajdstrand: I noticed that it's also useful for development14:47
zygajdstrand: because system key regeneration is super fast now14:47
zygajdstrand: it was mainly slow because of this, apparently14:48
* zyga breaks for some food14:48
ijohnsonmvo: pedronis: assuming y'alls calendars are up to date, there's a 30m window in about an hour where we could meet to discuss if you want (I don't think the discussion would take 30m though)14:58
ijohnsonalso different topic, but should I get another review on #7581, or am I good to merge that? I had to tweak the tests slightly to disable core16 tests of the Info D-Bus method, not sure if that warrants more reviews15:01
mupPR #7581: tests: add netplan test on ubuntu core <Created by anonymouse64> <https://github.com/snapcore/snapd/pull/7581>15:01
mupPR snapd#7700 opened: many: wait while inhibition file is present <Created by zyga> <https://github.com/snapcore/snapd/pull/7700>15:07
pedronisijohnson: it works for me, don't know about mvo15:17
mvoijohnson: yes, works for me15:18
cachioChipaca, hey https://paste.ubuntu.com/p/qPWP4nF48F/15:18
cachioon pi3, snap list does not show latest/15:18
cachioand it breaks the listing test15:18
ijohnsonCool I'll just send out a calendar invite just to be sure15:19
cachioChipaca, it should show latest/edge?15:19
mupPR snapd#7698 closed: seed/internal: doc comment fix and drop handled TODOs <Simple 😃> <Created by pedronis> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7698>15:19
mupPR snapd#7581 closed: tests: add netplan test on ubuntu core <Created by anonymouse64> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7581>15:20
Chipacacachio: it depends, but probably yes15:21
Chipacammm15:21
* Chipaca looks at what's on edge15:21
cachioChipaca, I just refreshed and then did snap list15:23
Chipacahmmm15:23
Chipacathere might be a bug somewhere15:24
Chipacaon edge15:24
ChipacaI'm seeing a tracking of "latest"15:24
Chipacanot "stable", not "latest/stable", just "latest"15:24
Chipacathat's not good15:24
Chipacacachio: what channel are you looking at?15:24
cachioI have created an image with channel stable15:25
cachioand then refreshed core to edge15:25
cachioand I see what I pasted15:25
Chipacacachio: let me think a bit15:29
Chipacacachio: also, let me start with something sane because my local snapd has seen some shit (and might be doing its best)15:29
cachioChipaca, hehee15:29
cachiosure15:29
cachiotake your time15:29
Chipacacachio: also, also, most of my channels work is not on master yet15:30
Chipacacachio: and there's no code to add latest/ on master15:31
Chipacaafaik?15:31
* Chipaca checks again15:31
cachioit si merged to master I think15:31
Chipacaah, yes, there is, for the tracking channel, yes15:31
Chipacain state15:32
Chipacacachio: what does 'snap info' say for that snap?15:32
Chipacacachio: in 'tracking:'15:32
cachioChipaca, https://paste.ubuntu.com/p/TWHsVz6j6g/15:32
cachiotracking : edge15:32
zygare15:33
Chipacacachio: ok give me a bit15:34
cachioChipaca, sure15:34
* cachio lunch15:41
Chipacacachio: ok, so i'm seeing this "latest" bug on stable; after your lunch we can talk15:41
Chipacaone dge i mean15:41
cachiosure15:42
cachiotell me now15:42
cachiolunch can wait15:42
Chipacacachio: so. tell me what you did, on the pi315:42
cachioChipaca, so, is it a bug?15:42
Chipacathere is a bug, but it's not the one you told me :) so you tell me15:42
cachiohehe15:43
cachioChipaca, I created a stable image15:43
Chipacawith you so far15:43
cachiothen I booted that image in the pi315:43
Chipacaok15:43
cachiothen I refreshed the core snap from edge15:43
Chipacaok15:43
cachiorebooted15:43
cachioand did snap list15:44
cachiothat's it15:44
Chipacaok15:44
Chipacathere is no code that would've magically converted the 2.42 "stable" to "latest/stable" in 2.4315:44
Chipacaif you install something new, say: snap install --beta http15:44
Chipacathen you'd see latest/beta15:44
Chipacathe code to fix state is still up for review15:45
Chipacacachio: are you thinking of cutting edge to candidate or something?15:45
Chipacabecause i wouldn't ship with part of this channels work15:46
Chipacai'd recommend backing out the change, or merging the rest of it15:46
cachioChipaca, add15:46
cachiook15:46
Chipacacachio: the bug _I_ saw was that the code for 'snap list' turns "latest/stable" into "latest", which could be confusing, but it's only for display so it's not a biggie15:47
Chipacaand it would not be a problem with the rest of channels15:47
cachioin that case the problem is the test, but should automatically be fixed soon15:47
Chipacaok15:47
Chipacawell, should be, but reviews are hard to come by :)15:48
cachioChipaca, agree :)15:49
cachiothanks for the help15:49
* cachio lunch for real now15:52
* Chipaca steps away for a while16:33
* Chipaca lied16:36
Chipacaok, now yes16:48
Chipacabbl, ttfn16:48
jdstrandzyga: in case you didn't see and possibly relevent for your pr: https://bugs.launchpad.net/bugs/185066717:03
mupBug #1850667: cgroup v2 is not fully supported yet, proceeding with partial confinement <docker.io (Ubuntu):New> <lxd (Ubuntu):New> <snapd (Ubuntu):New> <https://launchpad.net/bugs/1850667>17:03
mupPR snapcraft#2779 closed: ci: switch to travis workspaces <Created by sergiusens> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2779>17:06
mupPR snapcraft#2780 opened: remote-build: rename `--arch` to `--build-arch` <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2780>17:12
zygajdstrand: thank you17:15
zygajdstrand: relevant to this is https://github.com/snapcore/snapd/pull/761417:23
mupPR #7614: cmd/snap-confine: implement snap-device-helper internally <Created by zyga> <https://github.com/snapcore/snapd/pull/7614>17:23
zygajdstrand: not urgent but if you look at the PR text you can see where this is going17:24
zygajdstrand: I have code that is close to handling v2 as well17:24
zygajdstrand: but not fully working, it's been touched a while ago the last time17:24
zygajdstrand: but this moves us towards real v2 support end to end, without partial17:25
jdstrandzyga: ack, thanks. it is on my list but not at the very top. hopefully late this/early next week17:26
zygajdstrand: but first, that agent needs to land17:26
zygajdstrand: no rush, I'm off on Friday17:26
zygajdstrand: really no rush, it's post vancouver topic17:26
jdstrandok, thanks17:26
zygajdstrand: but after that we could get full v2 support in about three weeks17:26
jdstrandok even better. cause the stuff in front is all pre-vancouver :)17:26
jdstrandzyga: that's really cool :)17:27
zygajdstrand: yeah17:27
=== pedronis_ is now known as pedronis
mvopedronis: I pushed an update to 764917:41
mvopedronis: if that looks good I will push the same approach to 768217:41
mvopedronis: should be small and simple to review17:41
pedronismvo: do you want me to look tonight? or tomorrow?17:42
mvopedronis: tomorrow is fine of course17:42
pedronisok17:42
=== pstolowski is now known as pstolowski|afk
* cmatsuoka steps into another strange race-like behavior when partitioning disks18:57
=== ondra_ is now known as ondra
* Chipaca AFKs for another while19:03
* cachio afk19:03
Saviqjdstrand: hey, I'm trying to make snapcraft compatible with strict multipass, which means piping file contents via stdio rather than throwing file paths around19:12
Saviqbasically I need the equivalent of: cat /tmp/tmp6q6ur_co/snapd.snap | multipass transfer - snapcraft-multipass:/var/tmp/snapd.snap19:12
Saviqbut when I do this in python, I'm getting a denial:19:12
Saviqhttps://pastebin.ubuntu.com/p/6nnw6GNjWG/19:13
Saviqit seems like Python is making some kind of a short circuit between the confined process and the file being read?19:27
jdstrandSaviq: hey, that's funny you should mention that as I was doing an investigation on https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/184975319:34
mupBug #1849753: AppArmor profile prohibits classic snap from inheriting file descriptors <amd64> <apport-bug> <focal> <wayland-session> <AppArmor:Confirmed> <snapd (Ubuntu):Confirmed> <https://launchpad.net/bugs/1849753>19:34
jdstrandSaviq: it is the same issue you are seeing ('file_inherit' is blocked with snap-confine)19:35
jdstrandSaviq: you could probably do: multipass transfer - snapcraft-multipass:/var/tmp/snapd.snap < /tmp/tmp6q6ur_co/snapd.snap19:36
jdstrandSaviq: well, maybe not since you said python is doing this to you19:37
jdstrandSaviq: I can't think of a way to do that besides making python read from /dev/stdin19:38
jdstrand(and I'm not sure otoh how to make it do that)19:38
jdstrandSaviq: dan from our team might have a suggestion19:39
Saviqjdstrand: right, in this case the calling process (snapcraft) is classic and the callee is strict, but I suppose the same thing applies19:40
Saviqbut yeah I may be able to do the shell dance19:41
jdstrandSaviq: that doesn't matter, it is that you are calling something in /snap/bin which calls snap run, which calls snap-confine and it doesn't have access at the profile transition when the path is revalidated19:43
jdstrandSaviq: you could also trying calling something in /snap/whatever/current/... instead, outside of confinement, but that is likely not ideal19:44
jdstrands/outside/which ends up being outside/19:44
Saviqyeah, a "cat file | multipass…" works fine19:44
Saviq"multipass… < file" has the same denial19:45
Saviqwhich I suppose makes sense19:45
jdstrandSaviq: ah, yes on both counts19:51
mupPR snapcraft#2781 opened: yaml_utils: move _load_yaml from project for re-use (and add tests) <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2781>19:52
Saviqnow… what's the reverse of "cat"… like "tee", but without printing to stdout :D19:53
Saviqah, `dd` to the rescue :P19:54
Saviqbut all that really won't be cross-platform…19:56
jdstrandthere won't be an immediate fix for that bug. but there are some options to investigate before we have proper fd delegation in apparmor at least19:57
jdstrandthat came out wrong19:58
jdstrandproper fd delegation would allow us to fix that. before we have that, there might be some things we can do, but that needs investigating19:58
Saviqit looks like I'll really have to read the file into memory to feed it in again…20:00
Saviqwhat's weird(ish) is that python claims that from 3.4 onwards the fds are non-inheritable, which I thought for a bit could avoid this…20:01
mupPR snapcraft#2782 opened: snapcraft: introduce click-based YAML configuration file support <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2782>20:04
mupPR snapd#7701 opened: overlord: add kernel rollback accross reboots manager test and fixes <Created by mvo5> <https://github.com/snapcore/snapd/pull/7701>20:26
techalchemyjdstrand, how's this20:46
jdstrandtechalchemy: thanks! :)20:46
mupPR snapd#7702 opened: tests: adding fedora 31 to google-unstable backend <Created by sergiocazzolato> <https://github.com/snapcore/snapd/pull/7702>20:47
mupPR snapcraft#2783 opened: cli: run review tools before pushing to the store if available <Created by sergiusens> <https://github.com/snapcore/snapcraft/pull/2783>21:22
mupPR snapcraft#2780 closed: remote-build: option renames for arch, user, and accept-public-upload <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2780>22:40
mupPR snapcraft#2781 closed: yaml_utils: move _load_yaml from project for re-use (and add tests) <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2781>22:40
mupPR snapcraft#2784 opened: [multipass] use stdio to get data in/out of Multipass <Created by Saviq> <https://github.com/snapcore/snapcraft/pull/2784>23:16
mupPR snapd#7703 opened: cmd/snap: make 'snap list' shorten latest/$RISK to $RISK <Channels 🏷️> <Simple 😃> <Created by chipaca> <https://github.com/snapcore/snapd/pull/7703>23:43
mupPR snapcraft#2785 opened: remote-build: add initial command unit tests <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2785>23:46

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