[00:49] PR snapd#10602 closed: release: 2.51.4 [01:24] PR snapd#10603 opened: packaging: merge 2.51.4 changelog back to master [07:00] PR snapd#10603 closed: packaging: merge 2.51.4 changelog back to master [08:45] PR snapd#10601 closed: .github/workflows/test.yaml: use snapcraft 4.x to build the snapd snap <⚠ Critical> [10:06] mvo: hi! I added the workaround for 16.04, and it seems to work fine. However I now noticed that 14.04 doesn't work at all. It looks like mount units are not properly created there (still investigating). [10:16] looks like a bug in "systemctl show": if I run "systemctl show .mount" then all is fine, but if I pass "*.mount" as last parameter, it finds only one mount unit, with a bogus ID (Id=\x2a.mount) [10:25] I'm also getting an error on ArchLinux: [10:25] 2021-08-10T08:35:36.2606742Z + test-snapd-mount-control.cmd grep /var/snap/test-snapd-mount-control/common/target /proc/self/mountinfo [10:25] 2021-08-10T08:35:36.2608249Z cannot change profile for the next exec call: No such file or directory [10:25] 2021-08-10T08:35:36.2609165Z snap-update-ns failed with code 1 [10:25] (from https://github.com/snapcore/snapd/pull/10473/checks?check_run_id=3288775107) [10:25] PR #10473: interfaces/builtin: add mount-control interface [11:26] mardy: sorry, was in various meetings. if 14.04 does not work it's fine if we just error clearly if someone tries to use the feature on 14.04 - we keep 14.04 just for live-patch around [11:35] yesterday i asked about application confinment with the help of snap, a series of modules such as lsm=landlock,lockdown,yama,apparmor,bpf and a hardened kernel [11:35] but then i left. so now that i am here, could somebody shade some light? [11:36] (oh the good old days of grsecurity...) === marcustomlinson_ is now known as marcustomlinson [12:07] mardy: actually the Archwiki mentions Apparmor for snaps, and in the Apparmor article it suggests the above lsm kernel paramaters [12:35] georgios: I might be wrong, but as far as I know only one security module can be active at a time. So, you are free to enable as many as you like when you build the kernel, but at run-time only one will be used [12:37] georgios: oh, it looks like I was wrong: https://www.kernel.org/doc/html/latest/admin-guide/LSM/index.html [12:38] "yama" and the others you mentioned are "minor" LSM modules, so multiple of them can coexist [12:38] georgios: so yes, using the line given in the ArchLinux wiki might be a good idea [13:31] PR snapd#10600 closed: configcore: fix a bunch of incorrect error returns [14:16] mvo: I asked about this after we did 2.51.3 but apparently you didn't answer since we didn't follow through, but can I cherry-pick https://github.com/snapcore/snapd/pull/10315 to release/2.51 ? I have to manually fix the changelog and date formats for releases without that on the release branch, or what I just did yesterday was apply the changes locally being very careful not to commit those changes when I go to create the changelog commit [14:16] PR #10315: release-tool/changelog.py: misc fixes from real world usage [14:16] PR snapd#10604 opened: wrappers: measure time to enable services in StartServices() === sarnold_ is now known as sarnold [14:44] mvo: also if you could force land #10542 that would be appreciated [14:44] Bug #10542: Hoary: synaptic asks for root's passwd [14:44] PR #10542: tests/nested/manual: enable serial assertions on testkeys nested VM's [14:46] mvo: also that bug that mup pointed out with the same number as the PR I mentioned for Ubuntu 5.04 was triaged by you 17 years ago 😀 [15:04] ijohnson[m]: sure, will do [15:04] ijohnson[m]: hahaha - scary actualy [15:06] PR snapd#10542 closed: tests/nested/manual: enable serial assertions on testkeys nested VM's [15:06] PR snapd#10599 closed: configcore: fix early config timezone handling [15:07] thanks mvo ! [15:08] ijohnson[m]: yw [15:09] PR snapcraft#3568 closed: cli: enable SNAPCRAFT_TARGET_ARCH envvar matching --target-arch [16:52] PR snapd#10605 opened: many: do not re-check snaps on disk during uc20 install <⛔ Blocked>