[02:51] PR snapd#11331 opened: gadget: mv modelCharateristics to gadgettest.ModelCharacteristics === amurray_ is now known as amurray [06:31] morning === ackk is now known as ack [08:06] morning [08:21] pstolowski: hey [09:03] PR snapd#10166 closed: systemd: implement restart as a single systemctl restart call [09:08] PR snapd#9846 closed: [RFC] systemd, many: no reload when enabling or disabling services <:birthday:> [11:03] PR snapd#11200 closed: usersession/userd: query xdg-mime to check for fallback handlers of a given scheme [11:03] PR snapd#11264 closed: overlord: remodel getting stuck on prerequisites <â›” Blocked> [11:03] PR snapd#11273 closed: o/devicestate: verify that the new model is self contained before remodeling [11:18] PR snapd#11332 opened: gadget/install/partition.go: include DiskIndex in synthesized OnDiskStructure [11:18] PR snapd#11333 opened: gadget/install: rm unused support for writing non-filesystem structures [11:47] mborzecki, pstolowski: any of you is able to tell me why adding a s.settle(c) on this line causes the test to fail (no updates are found)? https://github.com/snapcore/snapd/blob/master/overlord/snapstate/snapstate_update_test.go#L3892 [11:48] mardy: what is the error? [11:48] mardy: are you sure it's line 3892? [11:53] yes, there, I get an error on the 4th scenario: https://paste.ubuntu.com/p/FYZZktSNkX/ [11:54] pstolowski: basically the update does not generate any tasks [11:54] without the s.settle(c) line, it works; I just don't understand what settle() does :-) [12:09] lots of lxd fun this morning it seems [12:09] PR snapd#11065 closed: cmd/snap: remove use of zenity, use notifications for snap run inhibition [12:14] PR snapd#11334 opened: gadget/install: add unit tests for makeFilesystem, allow mocking mkfs.Make() [12:14] PR snapd#11335 opened: gadget/install: add unit tests for install.Run() [12:42] mardy: if you add settle() there you are actually executing the tasks. as is now they are created but not run afaict [12:47] ijohnson: hi, have you seen somehing like this maybe? https://pastebin.ubuntu.com/p/STBZMQwZmh/ [12:48] bboozzoo: yeah we don't have a solution other than to retry [12:48] https://github.com/snapcore/snapd/blob/09b4d9128ef9c9584fffebc17ee84900fad5c800/tests/lib/nested.sh#L87-L97 [12:49] but Sergio has more details [12:52] pstolowski: I finally found the solution, the state was leaking from one iteration of the tests to the next one: https://github.com/snapcore/snapd/pull/11321/commits/9f50b31c04d0e869a0e640fae66b75282c1d809e [12:52] PR #11321: overlord: fix issue with snapshot restoring on CentOS [12:55] mardy: nice find. maybe you can replace s.settle(c) there with explicit chg.SetStatus(Done) there? [12:57] pstolowski: I'll try, thanks [13:03] mardy: or just make this a function helper like we sometimes do, eg. `func (s *suite) TestFoo{1,2}()` call `func (s *suite) testFoo(c *C, some param)` [13:04] or have new state on every iteration (do not use s.state), unless it needs complicated setup in SetUp() [13:18] PR snapcraft#3631 opened: parts: add early base parsing (CRAFT-766) [13:54] PR snapd#11336 opened: cmd/snap: add --debug to snap run [14:49] PR snapd#11337 opened: many: fix leftover empty snap dirs [18:26] PR core20#130 opened: static: remove perl dependency [18:42] PR core20#131 opened: No ssh key in install mode