[00:26] PR snapcraft#3086 opened: repo: restore marked-install strategy for apt-cache [00:26] Issue core20#34 closed: please provide dbus-launch [00:26] PR core20#43 closed: extra-packages: add dbus-user-session for user-session dbus [03:26] PR snapd#8555 opened: bootloader/uboot: use secondary ubootenv file boot.sel for uc20 [05:39] morning [06:16] PR snapd#8531 closed: secboot,cmd/snap-bootstrap: add model to pcr protection profile [06:18] damn wasps [06:18] already flying everywhere [06:37] errand, back in a bit [07:00] mborzecki: wasps? [07:00] szerszenie? [07:01] I wonder if I should merge master to resolve those build failures from yesterday [07:02] mborzecki: perhaps we should patch the search test to skip on release weeks? :P [07:03] centos-8 is still busted, I had no chance to look at htis yet [07:07] morning [07:09] mborzecki: mvo: hi, I'm staring at #8552 and thinking what to do there [07:09] PR #8552: cmd/snap-bootstrap: measure epoch and model before unlocking encrypted data <⛔ Blocked> [07:09] pedronis: thank you [07:17] hey mvo, pedronis, pstolowski [07:17] good morning zyga [07:21] mvo: should I review the UC20 branches that samuele mentioned yesterday or work on the GUI for refresh? [07:22] zyga: what uc20 PR did he ask you about? if he did not ask you specifically I would say GUI is more important [07:22] I think we are ok, most things that could land have landed [07:22] no not me [07:22] during the standup [07:22] ok, I think the gui is within reach today [07:22] what's left atm need a bit more than reviews [07:23] assuming we can land things today [07:23] ok [07:23] mvo: mborzecki: can we have a quick sync in 10 mins? [07:24] I need a review for the GUI front-end https://github.com/snapcore/snapd/pull/7700 [07:24] PR #7700: many: wait while inhibition file is present [07:24] I will work on the glue logic now [07:30] pedronis: I don't think mborzecki is around just then [07:30] mborzecki: but I can be there if you want [07:32] mvo: ok, let's do one quick and we might need one with him later though. I want to bounce some ideas before I go off on a tangent [07:32] pedronis: but I can be there if you want, just need 1min [07:32] pedronis: ok, let me just push this one PR and I'm there (1min) [07:32] mvo: he said he's out on an errand [07:33] PR snapd#8556 opened: tests: ensure $cache_dir is actually available [07:34] mvo: I'm in the standup HO [08:39] re [08:40] i'll grab a coffe and back to work [08:40] installing windows is slow, snaps are faster ;-) (this is a joke with extra meanings) [08:42] pedronis: sync in 5-10? [08:43] mborzecki: yes [08:52] pedronis: mvo: i'm in the standup HO [08:52] omw [10:06] Good morning folks [10:08] hey ian :) [10:10] Hey zyga [10:12] mvo: btw, the refactoring is goinf well, but we have a ton of those MockOsutilIsMounted [10:13] pedronis: could we make it part of the suite setup? [10:13] not really [10:13] pedronis: but I'm in a meeting right now so might be a bit slow to reply [10:13] just saying it takes a while to refactor [10:22] pedronis: +1 [10:52] PR snapd#8556 closed: tests: ensure $cache_dir is actually available [11:23] PR snapd#8557 opened: c/snap-bootstrap: have a small struct for checking mount states [11:24] mvo: mborzecki: done ^ [11:39] afk, small child invasion [11:41] PR snapd#8550 closed: ubootenv, uboot: support new uc20 style text bootenv <⛔ Blocked> [11:52] mborzecki: now my vendor.json does change [11:52] weird [11:52] zyga: govendor --version? [11:52] 1.0.9 [11:53] the changes I see are on secboot and sys/unix [11:53] master was updated with a PR from claudio, and vendor.json changed [11:53] zyga: checksum? [11:53] es [11:53] yes [11:53] hm maybe cmatsuoka had a differetn govendor version? [11:53] idk, way over my head :P [11:58] pedronis: thank you [11:58] is master broken, is should I hold with any new PRs/ [11:59] xerrors.Is() is kinda silly with os.PathError or i'm doing something wrong here [12:00] cmatsuoka: what's you `govendor --version` ? [12:04] mborzecki: what are you doing with path error? [12:04] maybe you pass some syscall result? [12:11] zyga: maybe i'm doing/reading it wrong, but xerrors.Is(err, &PathError{}) compares whether *(err(*os.PathError)) == *(&PathError{}) [12:11] zyga: while i really want to know wetherh os.PathError is wrapped somewhere in the error stack [12:13] hmmm [12:13] just looking at it quickly [12:13] either err is really PathError [12:13] or you need to provide Unwrap [12:13] mborzecki: xerrors cannot really do it's job if the go stdlib isn't collaborating [12:13] to look deeper into err [12:14] mborzecki: it gives you some of the future, but if go itself is too old some things won't work [12:14] I think [12:15] mborzecki: you could manually repackage an error you get from somewhere [12:15] into something that implements the Wrapper interface [12:15] though Chris has given Unwrap to some stuff [12:15] then you could have "nice" properties from that point on [12:15] or maybe only something else I need to check [12:15] but it's some extra work in a specific ase [12:15] *case [12:15] there are different orthogonal things you can provided [12:15] for Is vs As [12:16] pedronis: yeah, As seems to work for now [12:23] mborzecki: hmm, 1.0.8. Too old? [12:23] cmatsuoka: looks like it's generating a different checksum that 1.0.9 does [12:23] * cmatsuoka updates [12:24] mborzecki: are you using the edge snap? [12:24] cmatsuoka: the edge snap? [12:24] of govendor? [12:24] I was using stable, switched to edge now [12:25] cmatsuoka: idk, i go get'ed it [12:25] oh there's a snap for that? [12:25] mborzecki: I don't mind using any version but we should agree on a standard one [12:26] well it's archived [12:26] we should use the latest [12:26] available [12:26] and switch to modules when possible [12:28] ok, I switched to 1.0.9 from the archive [12:39] PR snapcraft#3087 opened: meta: remove snapd workaround for classic for core20 onwards [13:09] PR snapd#8558 opened: tests: make the nested library usable independently of spread [13:32] mborzecki: so the issue with the /run/mnt/data PR is that it hangs after booting run mode in the initrd? [13:32] mborzecki: should I just give it a try and see what falls out?> [13:38] ijohnson: sure, go ahead and play with it [13:38] ijohnson: i've merged master there this morning, so it shoudl be fairly up to date [13:39] okay thanks mborzecki I'll let you know how it goes [13:40] ijohnson: cool, thanks! [13:45] PR snapd#8424 closed: cmd/snap-bootstrap/initramfs-mounts: cross-check partitions when mounting [13:52] hmm looks like there's a problem during reboot with the measurement PR on uc20 [13:55] but pc-kernel & snapd snaps repacked with snap-bootstrap still work, wth? [14:04] mborzecki: are they using the right version? [14:05] mborzecki: did it ask you to enter a recovery key ? [14:06] ijohnson: pstolowski: the nodes hang on reboot for some reason, i'm tying to download the image and check it locally [14:07] we don't have tests that encrypt [14:07] there's might be a tpm though and things explode [14:14] pedronis: yeah we have /dev/tpm0 in gcp [14:16] mborzecki: maybe write a small bit of code that does the two measurements and see how it fails there [14:54] hm somewhat unsuprisingly, the image i downloaded works locally [14:55] mborzecki: given the changes in your PR, I expect you need to try the same stuff on gce [14:56] mborzecki: definitely something wrong in the image I created from current 8552, will debug/bisect after lunch to see what's happening [14:56] cmatsuoka: where does it hang? [14:57] mborzecki: somewhere inside the-tool [14:57] meh :/ [14:57] cmatsuoka: is there a write up of what i need to try the emulated tpm with qemu locally? [14:58] mborzecki: the difference with the previous code is that you don't ignore all connect error no? [14:59] mborzecki: maybe you can revert to that but log the error somewhere that stays [14:59] mborzecki: I think we don't have it documented, but this is the script I use: https://pastebin.ubuntu.com/p/Nd9642FkgR/ [15:00] will be back after lunch [15:00] cmatsuoka: thanks, i'll try that [15:00] mborzecki: you won't need the -serial stdio, I added it to debug things recently but it's not needed anymore [15:02] fwtw i still have no luck running basic test on master. running for over 25 minutes (a few ssh attempts so far). no sure if this is the same problem? [15:03] (i mean core20/basic) [15:09] fwiw the split out bit that measures epoch and model does not return any errors when i run it manually [15:09] and it's definitely touching /dev/tpm0 [15:15] mborzecki: on gce ? [15:15] pedronis: yes, i'm calling this basically https://paste.ubuntu.com/p/DWrPSgDqR4/ [15:16] and it makes the stamps? [15:16] pedronis: yup, i'm looking at strace too, quite some traffic on the fd that /dev/tpm0 was opened with [15:17] ok, so maybe the tpm stuff is not the issue and is some other change? [15:18] uhh and can't installs swtpm-mvo snap, because the store times out [15:22] ha ok, i see a bug now [15:24] PR snapcraft#3088 opened: repo: add interface to get packages from base [15:27] I made some comments on the PR, let me know what the bug is [15:28] mborzecki: do you think you might be able to review #8555 before you EOD? [15:28] PR #8555: bootloader/uboot: use secondary ubootenv file boot.sel for uc20 [15:29] pedronis: i'm using incorrect path to recovery system model, ubuntu-seed/