[06:40] morning [07:28] school run, back in 30 [07:48] re [07:48] mvo: hey [07:49] good morning mborzecki [07:50] PR snapd#9920 closed: many: add Delegate=true to generated systemd units for special interfaces (master) <⚠ Critical> [07:55] PR snapd#9925 opened: release: snapd 2.49 [07:57] mvo: can you upload the release tarballs to github releasese page? [07:58] mborzecki: sure, just done [07:58] thanks! [08:01] morning [08:18] pstolowski: good morning [08:28] pushed out 2.49 to arch [08:29] mvo: btw. not really needed at this point, but i noticed the source tarballs for 2.48.1 are missing from github ;) [08:29] it's probably for posterity only [08:51] #9893 needs a 2nd review [08:51] PR #9893: store: support validation sets with fetch-assertions action [09:00] yep, i was just going to ask for 2nd review :} [09:02] mvo: do you need me in the desktop sync? [09:02] mborzecki: no, it's fine [09:03] ok === mcphail2 is now known as mcphail [09:55] PR snapd#9926 opened: interfaces/seccomp: allow copy_file_range by default [09:55] mvo: ^^ quick fix [10:28] mborzecki: how do you feel about doing 2nd review of 9893? [10:33] pstolowski: have it open, but haven't looked in detail yet, will try to go through it today [10:33] ty [13:09] ijohnson: lol @ your file range copy comment [13:10] :-) [13:11] PR snapd#9927 opened: packaging/fedora: sync with downstream packaging in Fedora <⛔ Blocked> [13:12] ijohnson: hahaha, yeah, felt kind of familiar but i saw we only had it in the docker iface, so assumed that whatever we hd for master did not land [13:13] ijohnson: can we land it? :) node14 seems to use that syscall via libuv [13:13] i mean #9702 [13:13] PR #9702: interfaces/seccomp/template.go: allow copy_file_range [13:16] PR snapd#9926 closed: interfaces/seccomp: allow copy_file_range by default [13:17] mborzecki: still needs security review :-/ [13:19] amurray: gentle ping ^^ [13:51] PR snapd#9928 opened: boot: cmd/snap-bootstrap: handle a candidate recovery system <⛔ Blocked> [14:06] PR snapd#9925 closed: release: snapd 2.49 [14:39] ijohnson: I think I found some go code that does what you where trying to do [14:39] *were [14:39] ah nice [14:39] well as you mentioned what I ended up doing is actually not what we need to do so maybe it is all moot [14:39] but I would be curious to see in any case [14:40] ijohnson: it's a bit old so it nees changes but I can put it somewhere [14:42] pedronis: in any case this is what I settled on to create a self-signed account-key, but I'll just ignore that for now https://gist.github.com/anonymouse64/22fd11b48676e1641eb51bd6d7fe4031 [14:43] ijohnson: https://gist.github.com/pedronis/d4ba5849e40814f1fa5ee22d8b525471 [14:44] pedronis: ah nice I missed some of those fuctions in assertstest I guess [14:45] pedronis: is there one to output the private gpg encoded private key that can be imported by fakestore to sign the repair assertion ? [14:48] ijohnson: that's the dumpRSAKey thing, see it's consumed later by ReadPrivKey [14:48] to check it works [14:49] ah right got it [14:51] ijohnson: basically once fixed, the output would need to be embedded in systestkeys [14:53] ijohnson: tbh, I don't remember why I didn't do that at the time I wrote this [14:53] maybe I used the result only locally [14:53] anyways yes this seems to do what I need for the repair assertion account-key [14:54] ijohnson: hey, requested a teensy change for copy_file_range [14:54] hey jdstrand, thanks! I'll have a look in a bit, I have quite a few meetings this morning [14:55] ijohnson: np at all. have a nice rest of the day :) [14:55] amurray: you may be interested in ^ [14:56] ijohnson: I think it needs a couple of fixes because some assertstest helpers have changed [14:56] hey pedronis :) [14:56] jdstrand: hi [14:58] ijohnson: if we haven't grown something like it since, it might be worth adding dumpRSAKey to assertstest [14:59] pedronis: ack [16:11] ijohnson: are unblocked on this? [16:11] sorry still in meetings [16:14] sorry, my fault [16:44] hi, snapcraft question: in a snap, I have libblas.so.3 (pulled in as a dependency of a stage-packages) ending up in $SNAP/usr/lib/x86_64-linux-gnu/blas, while on a regular install it's under /usr/lib/x86_64-linux-gnu. is that snapcraft doing? [17:02] hi, is it intentional that the 'current' symlink of a snap doesn't seem to be changed to the latest version until it's actually run? [17:04] i ask because i was hoping to jiggle a few things around in a snap's $XDG_CONFIG_HOME in a post-refresh hook, but i don't think it's working :-/ [17:14] any thoughts/suggestions diddledan? [17:15] bandali: no. current symlinks should be updated when post-refresh hook runs [17:15] the user-level ones aren't, pstolowski [17:16] diddledan: you're right! [17:16] only the non-user-specific ones are. i.e. $SNAP, $SNAP_DATA and $SNAP_COMMON are updated [17:16] err, i'm kind of confused :-/ do you have an example for me to check, for how to update stuff in XDG_CONFIG_HOME in a post-refresh hook? [17:16] (if it's at all possible ?) [17:16] it's not possible [17:17] oh, dang.. my use case was adding a symlink in XDG_CONFIG_HOME/autostart/ to a desktop file somewhere in $SNAP/ [17:18] but the symlink would need to be updated/fixed in each upgrade because $SNAP changes [17:18] symlink it through /snap/$SNAPNAME/current instead of $SNAP [17:20] ohh okay cool that might work! thanks for the idea, i'll give that a shot [19:05] PR snapcraft#3434 closed: build providers: clean environment if project directory is changed [19:10] PR snapcraft#3320 closed: python v2 plugin: consistent linking for interpreter [19:10] PR snapcraft#3363 closed: repo: apt sources management refactor [19:10] PR snapcraft#3370 closed: repo: address issue with fix_symlink() when pointed at directory [19:16] bandali: diddledan: mind filing a bug about only switching $SNAP_USER_* current symlinks at runtime? [19:16] I know why we don't do it because it's $HARD but it would be nice to track that we don't do it in case it comes up again [19:20] just in case soneone wonders - snapd (and more) in hirsute broken because of https://bugs.launchpad.net/ubuntu/+source/fakeroot/+bug/1915250 - only happens if you run hirsute-proposed [19:20] Bug #1915250: buildd file owner/group for shared libraries [20:40] is there a way to see what underlying event triggered builds in https://snapcraft.io/xxx/builds ? (whether it was a GitHub webhook, manual trigger, etc) [21:57] tianon: builds from snapcraft.io/builds will end up in launchpad at https://launchpad.net/~build.snapcraft.io/+snaps, so you could try to find the snap you are looking for in there, otherwise the #launchpad channel might have more tips [21:57] I think this was a feature that the old build.snapcraft.io used to have that was lost in the rewrite. [21:57] IIRC Launchpad was never told the triggering reason for the build - it was only ever stored on the BSI side. [21:58] right, the builds themselves show up there, but the trigger doesn't [21:58] And I think the rewrite decided that they wanted to go from not much persistent state to no persistent state. [21:58] I can go into the build logs on snapcraft.io and see the launchpad link for a specific build even (it's the first line) [21:59] To restore the feature I guess we'd need to extend LP's data model to have some kind of a triggering-reason column, and have snapcraft.io/build pass it so that it could be stored and rendered [21:59] but the issue I'm having is that my last commit was 5 days ago and there have been two all-architectures rebulid triggers since that I'm scratching my head about :) [22:00] maybe there's a wishlist tracker somewhere this could be filed in? (or already is?) O:) [22:02] https://github.com/canonical-web-and-design/snapcraft.io/issues [22:02] I think [22:03] (Maybe they do have a database on the snapcraft.io side that could be used without a lot of complicated back-and-forth with LP. I'm not sure - I was involved quite a bit with the old BSI but not much with the rewrite.) [22:05] searching those now, and I'll file something if I don't find anything; thanks :) [22:24] (ended up with https://github.com/canonical-web-and-design/snapcraft.io/issues/3408 for anyone curious or interested)