[05:14] <mborzecki> morning
[06:45] <mvo> zyga, mborzecki: good morning
[06:45] <mborzecki> mvo: hey
[06:45] <mup> PR snapd#8600 closed: tests: run smoke test with different bases <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8600>
[06:45] <mborzecki> mvo: pinged you in some prs
[06:46] <mvo> mborzecki: yeah, saw it in the first, will look for the others now :)
[06:46] <mborzecki> ok
[06:46] <mborzecki> maybe we should just make those pulseaudio tests manual
[06:47] <mborzecki> mvo: this one too https://github.com/snapcore/snapd/pull/8562
[06:47] <mup> PR #8562: store: implement DownloadAssertions <Bulk assert refresh :scroll::scroll::scroll:> <Created by pedronis> <https://github.com/snapcore/snapd/pull/8562>
[06:51] <mup> PR snapd#8562 closed: store: implement DownloadAssertions <Bulk assert refresh :scroll::scroll::scroll:> <Created by pedronis> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8562>
[07:02] <pstolowski> morning
[07:02] <zyga> Good morning
[07:04] <mvo> pstolowski: hey, good morning
[07:04] <mborzecki> zyga: pstolowski: hey
[07:11] <mup> PR snapd#8557 closed: c/snap-bootstrap: check mount states via initramfsMountStates <UC20> <Created by pedronis> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8557>
[07:34] <mborzecki> zyga: did you get a link to activate your suse account?
[07:37] <mup> PR snapd#8436 closed: configcore,tests: use daemon-reexec to apply watchdog config <Squash-merge> <Created by pedronis> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8436>
[07:43] <zyga> mborzecki: checking
[07:44] <zyga> hmm
[07:44] <zyga> no?
[07:44] <zyga> I got an email earlier
[07:44] <zyga> about the new account system
[07:47] <zyga> mborzecki: I _hope_ that with the current changes the pulseaudio tests will no longer fail this way
[07:47] <zyga> now that both tests kill PA reliably via systemd
[07:50] <zyga> remarkable little failures
[07:50] <zyga> >0 but usually just one
[07:50] <zyga> and a pretty short quuee
[07:54] <zyga> hmmm
[07:54] <zyga> upgrade/basic
[07:54] <zyga> - Download snap "core" (9066) from channel "stable" (read tcp 10.240.0.214:45966->91.189.88.179:443: read: connection reset by peer)
[07:54] <zyga> eh, just network woeas
[07:54] <zyga> *woes
[07:55] <zyga> https://github.com/snapcore/snapd/pull/8571 is super-simple and needs a 2nd review
[07:55] <mup> PR #8571: overlord/snapstate: warn of refresh/postpone events <Created by zyga> <https://github.com/snapcore/snapd/pull/8571>
[07:55] <zyga> just warnings behind an experimental flag (though the flag is check in the code that is not in the diff)
[08:04] <mup> PR snapd#8571 closed: overlord/snapstate: warn of refresh/postpone events <Created by zyga> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8571>
[08:05] <mup> PR pc-amd64-gadget#46 closed: gadget.yaml: bump edition <Created by xnox> <Merged by mvo5> <https://github.com/snapcore/pc-amd64-gadget/pull/46>
[08:07] <zyga> \o/
[08:07] <zyga> thank you mvo!
[08:11] <mborzecki> https://github.com/snapcore/snapd/pull/8563 needs a 2nd review, pstolowski maybe you cuold take a look, since we both reviewed the other bulk assertions PRs? :)
[08:11] <mup> PR #8563: asserts/internal: introduce Grouping and Groupings <Bulk assert refresh :scroll::scroll::scroll:> <Created by pedronis> <https://github.com/snapcore/snapd/pull/8563>
[08:13] <mup> PR snapd#8603 opened: tests: pair of follow-ups from earlier reviews <Skip spread> <Created by zyga> <https://github.com/snapcore/snapd/pull/8603>
[08:13] <pstolowski> mborzecki: yes absolutely, i meant to, thanks
[08:14] <pstolowski> ok :)
[08:22] <zyga> huh
[08:22] <zyga> gnome 3.36.2 has this changelog entry
[08:22] <zyga> - Reenable TLS 1.0/1.1 protocols due to COVID-19.
[08:22] <zyga> !??!
[08:23] <pedronis> mvo: pstolowski: hi, #8602 will need to be cherry picked for 2.45, right?
[08:23] <mup> PR #8602: configcore: only reload journald if systemd is new enough <Created by stolowski> <https://github.com/snapcore/snapd/pull/8602>
[08:24] <mvo> pedronis: it looks like it, yes
[08:25] <pedronis> mvo: I put the milestone on it
[08:25] <zyga> pedronis: I requested your review on https://github.com/snapcore/snapd/pull/8566 -- it is the principal part of the refresh-app-awareness that is yet to be designed
[08:25] <mup> PR #8566: cmd/cmdutil: add run inhibition operations <Created by zyga> <https://github.com/snapcore/snapd/pull/8566>
[08:25] <zyga> this PR contains my take on the idea
[08:28] <pedronis> ok, I probably need to look at older PRs first... hopefully I can go back doing some reviewing later today
[08:29] <zyga> ok
[08:29] <zyga> if you want to talk about it apart from the review and have some time just ing me
[08:29] <zyga> *ping me
[08:29] <zyga> I have enough to work on that it is not a problem
[08:29] <zyga> (not urgent)
[08:30] <zyga> having an idea for next week would be brilliant
[08:31] <pstolowski> pedronis, mvo: yes
[08:34]  * zyga reboots to unbreak his touchpad :/
[08:35] <zyga> Those ThinkPads....
[08:36] <zyga> re
[08:37] <mborzecki> Those touchpads :P
[08:37] <mborzecki> zyga: pedronis: when is the desktop sync meeting? can you invite me in case we discuss the fonts issue?
[08:37] <mborzecki> s/pedronis/mvo/ ^^
[08:37] <zyga> IIRC tomorrow
[08:38] <zyga> let me check
[08:39] <zyga> note, the meeting tomorrow is the code review sync
[08:39] <zyga> invited you
[08:40] <mborzecki> zyga: thanks!
[08:41] <pedronis> zyga: fwiw, pulseaudio failed again here: https://github.com/snapcore/snapd/pull/8563/checks?check_run_id=648645627
[08:42] <mup> PR #8563: asserts/internal: introduce Grouping and Groupings <Bulk assert refresh :scroll::scroll::scroll:> <Created by pedronis> <https://github.com/snapcore/snapd/pull/8563>
[08:42] <pedronis> with master from this morning
[08:42] <zyga> looking
[08:43] <zyga> hmmm, interesting
[08:43] <zyga> I was adding some precondition code
[08:43] <zyga> but perhaps something else is going on here: look that we check for the socket first
[08:43] <zyga> and initially it is not there
[08:43] <zyga> then it shows up
[08:43] <zyga> pulseaudio --check looks if the daemon is there
[08:44] <zyga> but does not connect to it IIRC
[08:44] <zyga> thanks, let me try something and I'll send a PR shortly
[08:45] <zyga> it reads /run/user/UID/pulse/pid
[08:45] <zyga> so perhaps we are socket activated but not really running
[08:50] <zyga> pedronis: I have the logs now, feel free to restart or override
[08:51] <pedronis> mvo: did you squash merge the watchdog one? it had the label
[08:51] <mvo> pedronis: yeah, I overlooked it and realized when pressing merge :(
[08:51]  * mvo is a muppet
[08:52]  * zyga hugs mvo
[08:52] <mup> PR snapd#8604 opened: interfaces/builtin/desktop: do not mount fonts cache on distros with quirks <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/8604>
[08:52] <pedronis> mvo: we need to ping Taipei when it's in edge
[09:01] <mvo> pedronis: I added a comment to the bugreport asking for testing
[09:01] <mvo> pedronis: I can also add it to the sync doc
[09:01] <pedronis> thx
[09:02] <mvo> pedronis: added to the sync doc as well, if we need to revert I will take care of this (to make up for my blunder to not squash it)
[09:03] <pedronis> let's see how it goes
[09:18] <zyga> fascinating :)
[09:31] <zyga> some seriously weird stuff is happening
[09:36] <zyga> so I ran the pulseaudio test and tweaked the prepare section a little
[09:37] <zyga> no other tests, no other stuff
[09:37] <zyga> on *some* runs pulse starts and works
[09:37] <zyga> on some runs pulse starts, dies and cannot recover
[09:37] <zyga> starting it again fails
[09:37] <zyga> I have a debug shell, can strace and stuff
[09:37] <zyga> just wondering what the real issue is
[09:40] <mborzecki> zyga: i have a fascinating story to share, https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1876583 i can reproduce it in a terminal that has 80 colums, with that specific locale, because the unicode messages add up to fewer bytes than expected
[09:40] <mup> Bug #1876583: "panic: runtime error: slice bounds out of range" while installing any snap package <amd64> <apport-bug> <focal> <snapd (Ubuntu):In Progress by maciek-borzecki> <https://launchpad.net/bugs/1876583>
[09:40] <zyga> heheh
[09:41] <zyga> are we slicing utf-8 by bytes?
[09:42] <mup> PR snapd#8603 closed: tests: pair of follow-ups from earlier reviews <Skip spread> <Created by zyga> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8603>
[09:43] <mborzecki> zyga: i think the translation is wrong, not what we expected
[09:44] <mborzecki> mvo: zyga: where are the non-packed translations for snapd?
[09:44] <zyga> dunno
[09:44] <zyga> what do you mean by non-packed?
[09:44] <zyga> probably in rosetta
[09:44] <zyga> but just guessing
[09:46]  * zyga tests a fix for pulseaudio tests
[09:47] <zyga> and I suspect I know why it passed sometimes too
[09:47] <zyga> the long story short is that pulse quits when idle
[09:47] <zyga> it starts up on specific request (socket activated)
[09:47] <zyga> and just quits
[09:48] <zyga> the two sanity chceks up front are checking the socket and the service
[09:48] <zyga> and if fast enough, pulse is still running after starting the session
[09:48] <zyga> (it takes a moment to start and quick)
[09:48] <zyga> *quit
[09:48] <zyga> so the solution is just not to run pulseaudio --check
[09:49] <mborzecki> #. TRANSLATORS: this needs to be a single rune that is understood to mean "minutes" in e.g. 1m30s
[09:49] <mborzecki> msgid "m"
[09:49] <mborzecki> msgstr "хв"
[09:50] <mborzecki> heh
[09:50] <mborzecki> so this breaks our message/column calculation
[09:50] <jamesh> and the translator probably said "what on earth is a rune?"
[09:51] <mborzecki> nevertheless, we shouldn't panic :P
[09:53] <mvo> mborzecki: nice
[09:53] <mvo> mborzecki: also this requirement that it needs to be a single rune is silly, I mean, do we know that it always is in all the languages :) ?
[09:54] <mvo> mborzecki: (not your fault of course, just saying)
[09:54] <mborzecki> mvo: well, yeah agreed ;)
[09:54] <mvo> mborzecki: thanks for looking into this!
[09:54] <mvo> jamesh: lol, indeed
[10:09] <zyga> pedronis: https://github.com/snapcore/snapd/pull/8605 fix for what you reported
[10:09] <zyga> mborzecki: ^ fun
[10:09] <mup> PR #8605: tests: fix racy check for pulseaudio activity <Test Robustness> <Created by zyga> <https://github.com/snapcore/snapd/pull/8605>
[10:10] <mup> PR snapd#8605 opened: tests: fix racy check for pulseaudio activity <Test Robustness> <Created by zyga> <https://github.com/snapcore/snapd/pull/8605>
[10:26] <zyga> mborzecki: you need to adjust spread test in the fix for fonts https://github.com/snapcore/snapd/pull/8604/checks?check_run_id=648902472
[10:26] <mup> PR #8604: interfaces/builtin/desktop: do not mount fonts cache on distros with quirks <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/8604>
[10:26] <mborzecki> aah right
[10:31] <jamesh> zyga: I've left a comment on your PA PR
[10:31] <zyga> thanks, looking
[10:33] <zyga> jamesh: replied
[10:34] <jamesh> zyga: waiting for default.target would probably be worthwhile, yeah.
[10:35] <zyga> yeah, let's do it
[10:35] <zyga> I will keep the check for "active" as a sanity check
[10:36] <zyga> and add an extra patch that waits in prepare
[10:37] <jamesh> zyga: you should be able to rely on the return value of "systemctl is-active" rather than needing the extra MATCH call
[10:38] <zyga> jamesh: yeah, just more obvious what went wrong when it does fial
[10:38] <zyga> though I guess the log would contain that as well
[10:39] <zyga> hmm, default.target does not seem to be related to sockets
[10:39] <zyga> perhaps I should wait for sockets.target as well?
[10:41] <zyga> jamesh: something like this:
[10:41] <zyga> https://www.irccloud.com/pastebin/qZYsG5ek/
[10:43] <jamesh> zyga: probably.  That's the one I try to start in e.g. tests/main/snap-session-agent-service-control/task.yaml
[10:44] <zyga> ah, indeed
[10:45] <jamesh> zyga: actually, default.target should handle it
[10:45] <zyga> I _suspect_ so but I've started both
[10:45] <jamesh> default.target requires basic.target, which requires sockets, timers, and paths.target
[10:45] <zyga> ah
[10:45] <zyga> I can drop sockets then
[10:50] <zyga> jamesh: please look again
[10:51] <mborzecki> heh, that fix isn't as simple as i thought
[10:51] <zyga> mborzecki: isn't that my line? :D
[10:52] <zyga> nothing is
[10:53] <jamesh> zyga: looks good!
[10:55] <zyga> thanks!
[10:56] <zyga> quite a journey to test session-level user code
[11:06] <jamesh> zyga: fwiw, I think the session-tool changes have made the tests in my user-daemons PR more reliable.  I'm still seeing an occasional failure where connecting to the session agent socket fails, but that seems less common too
[11:06] <zyga> jamesh: do you have an idea why that could be the case?
[11:07] <jamesh> zyga: sadly no.  I've added some extra debug output, but I'm not sure what the cause was.
[11:07] <zyga> ok
[11:07] <zyga> well, little by little :)
[11:07] <zyga> I think we made a lot of progress
[11:08] <zyga> apart from gradual growth of reliability
[11:08] <zyga> writing tests is actually easier
[11:08] <zyga> less mucking with environment varibales, chown, quoting su -l ... test correctlyt
[11:08] <jamesh> zyga: I thought it might have been a stray session from a previous test, but you fixed things to shut down the user slice
[11:08] <zyga> yeah
[11:08] <zyga> it's a surprise it doesn't shut down though
[11:08] <zyga> so many weird things across releases of systemd
[11:12]  * zyga -> afk
[11:35] <mup> PR snapd#8606 opened: progress: fix progress bar with multibyte duration units <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/8606>
[11:35] <mborzecki> mvo: ^
[11:38] <mvo> mborzecki: nice
[11:52] <mup> PR snapcraft#3108 opened: tools: install wheel in snapcraft-dev <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/3108>
[11:55] <cmatsuoka> niemeyer: could you have a look at https://github.com/snapcore/spread/pull/104 when you have time? it's needed for the encrypted partition creation tests on gce
[11:56] <mup> PR spread#104: Adding option for systems created on google to start with secure boot enabled <Created by sergiocazzolato> <https://github.com/snapcore/spread/pull/104>
[11:59] <zyga> re
[12:00] <cmatsuoka> cachio: could you have a look at the error message mvo reported in https://github.com/snapcore/snapd/pull/8551? it seems that we have an unsupported system error there
[12:00] <mup> PR #8551: snap-bootstrap: remove create-partitions and update tests <UC20> <Created by mvo5> <https://github.com/snapcore/snapd/pull/8551>
[12:01] <cmatsuoka> (is it caused by the still missing spread PR?)
[12:07] <cachio> cmatsuoka, I know the  problem
[12:08] <zyga> kenvandine: hey, is there any plan to surface snap warnings to the user in a desktop session?
[12:09] <cachio> cmatsuoka, in nested.sh
[12:09] <cachio> in functinos get_google_image_url_for_nested_vm and get_ubuntu_image_url_for_nested_vm
[12:09] <cachio> plase add a * at the end of any system
[12:09] <cachio> cmatsuoka, I can fix it otherwise
[12:10] <cachio> as you prefer
[12:10] <cmatsuoka> cachio: the name would become ...-.img*?
[12:11] <cachio> cmatsuoka, no
[12:11] <cmatsuoka> or ...*.img?
[12:11] <cachio> ubuntu-16.04-64) -> ubuntu-16.04-64*)
[12:11] <cmatsuoka> ah, got it
[12:11] <cmatsuoka> thanks
[12:11] <cachio> for all the systems in both functinos
[12:12] <cachio> cmatsuoka, yaw
[12:14] <mup> PR snapd#8607 opened: tests: remove user.sh <Created by zyga> <https://github.com/snapcore/snapd/pull/8607>
[12:20] <mup> Bug #1876482 changed: There is no way to trace the upgrade history of a snap <Snappy:Invalid> <https://launchpad.net/bugs/1876482>
[12:46] <cmatsuoka> degville: this is the line I use for non-encrypted installs (it's a stripped down version of the other one, possibly it can be further reduced): https://pastebin.ubuntu.com/p/8tfnS642yM/
[12:47] <cmatsuoka> (smp 4 because I have a lot of cpus, the typical user will probably want smp 2)
[12:59] <degville> cmatsuoka: thank you! perfect.
[13:03] <mup> PR snapd#8608 opened: configcore: issue a warning on core16 when journal.persistent option is set <Created by stolowski> <https://github.com/snapcore/snapd/pull/8608>
[13:15] <zyga> pstolowski: https://github.com/snapcore/snapd/pull/8608/files#r420780119
[13:15] <mup> PR #8608: configcore: issue a warning on core16 when journal.persistent option is set <Created by stolowski> <https://github.com/snapcore/snapd/pull/8608>
[13:15] <pstolowski> zyga: thanks!
[13:51] <zyga> ijohnson: could you look at https://github.com/snapcore/snapd/pull/8605/commits/e3f01f69bb67376ac0f9065493b31d9e4efcd3f0 -- it's a part of thet fix but I pushed it after you approved
[13:51] <mup> PR #8605: tests: fix racy check for pulseaudio activity <Test Robustness> <Created by zyga> <https://github.com/snapcore/snapd/pull/8605>
[13:51] <ijohnson> sure
[13:51] <zyga> the *essential* change is https://github.com/snapcore/snapd/pull/8605/commits/e3f01f69bb67376ac0f9065493b31d9e4efcd3f0#diff-a8fce2de6a6829d98f798117050b6c93R134 line
[13:51] <zyga> everything else is the consequence of that
[13:56] <kenvandine> zyga: we haven't really put any thought into that
[13:56] <zyga> hey!
[13:56] <kenvandine> zyga: got an example?
[13:56] <zyga> sure
[13:57] <zyga> one main example is the warnings we show when an app was not refreshed when it was open
[13:57] <zyga> and when it got refreshed despite being open
[13:57] <kenvandine> ah
[13:57] <zyga> but as we just stared to use warnings now, there will be many more
[13:57] <kenvandine> well that is something there are plans for
[13:57] <zyga> we *could* convert those two to special events that are handled differently as well
[13:57] <zyga> just wanted to start a conversation about it
[13:57] <kenvandine> the whole "firefox needs a refresh, do you want to restart to refresh now?"
[13:58] <zyga> not quite, that's not what this is
[13:58] <zyga> what you describe is more like an interacting application that is snap aware
[13:58] <zyga> let's say gimp is snapped
[13:58] <zyga> and it has a refresh
[13:58] <zyga> and we just tried but postponed it because you had gimp open
[13:58] <zyga> without gimp being snap aware, we emit a warning
[13:59] <zyga> now that warning shows up in command line
[13:59] <zyga> when you interact with "snap" in any way
[13:59] <zyga> that's the context
[13:59] <zyga> you *can* see this today on edge
[13:59] <zyga> just run an app, switch to a different channel and refresh
[14:00] <zyga> (with refresh app awareness enabled)
[14:02] <zyga> kenvandine: so in my naive thought
[14:02] <zyga> I had two ideas
[14:02] <zyga> 1) we could show a desktop notification that just goes away
[14:02] <zyga> it's a one off it flies off
[14:02] <zyga> but you see it (if you look at the screen)
[14:02] <zyga> 2) we could show them in gnome-software
[14:02] <zyga> bonus points for showing this in the context of a particular app
[14:03]  * roadmr awards zyga 💯  bonus points 👾
[14:03] <zyga> woot :)
[14:04]  * zyga notices his ideas left people speechless ;-)
[14:27]  * zyga -> lunch
[14:35] <kenvandine> zyga: we're looking to monitor snapd for those refresh warnings and provide a notification with a button to restart the app.  This way the app doesn't need to know how to talk to snapd
[14:35] <kenvandine> we could surface other warnings the same way, perhaps with a different action attached.
[14:40] <pstolowski> pedronis: mind if i merge master into #8564?
[14:40] <mup> PR #8564: asserts: introduce Pool <Bulk assert refresh :scroll::scroll::scroll:> <Created by pedronis> <https://github.com/snapcore/snapd/pull/8564>
[14:40] <pedronis> pstolowski: no
[14:54] <mup> PR snapd#8607 closed: tests: remove user.sh <Created by zyga> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8607>
[14:54] <zyga> thank you
[14:54] <zyga> mvo: after https://github.com/snapcore/snapd/pull/8605 we will have much less red PRs
[14:54] <mup> PR #8605: tests: fix racy check for pulseaudio activity <Test Robustness> <Created by zyga> <https://github.com/snapcore/snapd/pull/8605>
[14:55] <zyga> just a little more :)
[14:55] <zyga> brb
[14:56] <mvo> zyga: nice one
[14:56] <mup> PR snapd#8609 opened:  Adds missing paths to desktop, solves lp:1876804 <Created by sklei4> <https://github.com/snapcore/snapd/pull/8609>
[14:58] <mup> PR snapd#8601 closed: Adds missing paths to apparmor, solves lp:1876804 <Created by sklei4> <Closed by sklei4> <https://github.com/snapcore/snapd/pull/8601>
[14:59] <mvo> pedronis: 8563 has two +1, should I merge or do you want to push the tweaks to it (some style suggestions in there)?
[15:00] <pedronis> mvo: please merge, I can do the tweaks in a follow up or the follow up, depending
[15:01] <cachio> ijohnson, hey, did you have this issue with pi image https://paste.ubuntu.com/p/f6C5BD9NK8/ ?
[15:02] <mup> PR snapd#8563 closed: asserts/internal: introduce Grouping and Groupings <Bulk assert refresh :scroll::scroll::scroll:> <Created by pedronis> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/8563>
[15:10] <ijohnson> cachio: how big is the SD card ?
[15:10] <cachio> ijohnson, 32GB
[15:11] <ijohnson> cachio: how big is the image you are trying to write?
[15:12] <cachio> 8GB
[15:12] <ijohnson> hmm I have only see that issue with no space left on device when the SD card is physically too small
[15:12] <ijohnson> cachio: maybe try deleting all partitions on the SD card first before flashing?
[15:12] <cachio> ijohnson, I already did it
[15:13] <ijohnson> cachio: does it happen everytime for you that it fails to write with No space left on device ?
[15:13] <cachio> ijohnson, first time trying to write a core 20 image for a pi3
[15:13] <ijohnson> cachio: I'd say just try again, I don't know why that would happen for you
[15:14] <cachio> I'll download another image
[15:25] <cachio> ijohnson, I see the problem
[15:26] <cachio> the systemd detects my sdcard as it'd have 3.3GB
[15:26] <cachio> instead of 32GB
[15:26] <cachio> not sure why
[15:26] <cachio> it is doing the same with all mys cards
[15:29] <zyga> there's one more race in that pulse test
[15:29] <zyga> pulse server writes the cookie
[15:29] <zyga> but we race trying to read it
[15:29] <zyga> I fixed that locally, will push in a while
[15:30] <ijohnson> cachio: ahh that's weird
[15:31] <pedronis> pstolowski: I tried to answer your questions in the review
[15:33] <pstolowski> pedronis: thanks, checking
[15:42] <zyga> ijohnson: I pushed to https://github.com/snapcore/snapd/pull/8605 again, to fix one more race
[15:42] <mup> PR #8605: tests: fix racy check for pulseaudio activity <Test Robustness> <Created by zyga> <https://github.com/snapcore/snapd/pull/8605>
[15:42] <zyga> ijohnson: explained in the commit message
[15:42] <ijohnson> I haven't had a chance to look at it, it will be a while before I can look at it today unfortunately
[15:43] <ijohnson> need to get uc20 things in order asap for the beta
[15:43] <zyga> ijohnson: this test keeps on giving but it also shows us that ~test matters
[15:43] <zyga> ijohnson: specifically, one test can leave home data that is relevant (as in, influences) another test
[15:43] <zyga> ijohnson: no worries, thank you
[15:43] <zyga> ijohnson: I will merge it iff it is green
[15:43] <zyga> ijohnson: and will happily iterate on feedback
[15:43] <zyga> just because red master is a disaster
[16:09]  * zyga EODs
[16:19] <pstolowski> mvo: do you think you could merge https://github.com/snapcore/snapd/pull/8520 directly (see my comment there re cla-check)
[16:19] <mup> PR #8520: data: fix shellcheck warnings in snapd.sh.in <Created by jelovac> <https://github.com/snapcore/snapd/pull/8520>
[16:28] <zyga> weird apt error https://www.irccloud.com/pastebin/y0zj5y5y/
[16:29]  * zyga bangs his head against pulseaudio wall
[16:31] <mvo> pstolowski: looking (sorry, was in meetings)
[16:31] <pstolowski> np, ty
[16:47] <mup> PR snapd#8610 opened: many: use /run/mnt/data over /run/mnt/ubuntu-data for uc20 <UC20> <Created by anonymouse64> <https://github.com/snapcore/snapd/pull/8610>
[16:53]  * cachio lunch
[18:05] <zyga> re
[18:34] <zyga> test leakage is the bane of our testing stack
[18:49]  * ogra hands zyga a bucket
[18:49] <zyga> and a mop
[18:50] <ogra> oh, i had hoped i was in time for not needing that ... didnt notice your floor is full of leaked tests already :)
[18:54] <zyga> one of those days where it's both leaks and tears
[18:57] <ogra> oh my
[19:03]  * ijohnson hugs zyga
[19:03] <zyga> thanks :)
[19:03] <zyga> I started my side server to have more spread slots while others have to re-start tests more often
[19:03] <zyga> in the end we all learn, as a team
[19:04] <ijohnson> I see you pushed some debug messages to your pulseaudio branch, should I review those ?
[19:09] <zyga> yeah
[19:09] <zyga> please have a look
[19:09] <ijohnson> ok, cool
[19:09] <zyga> my current working theory is that ~test/.config is root-owned
[19:09] <zyga> and that this makes pulse start but not write the cookie
[19:09] <zyga> which makes subsequent unconfined paplay fail
[19:09] <ijohnson> well that's silly
[19:10] <zyga> I found ~test/.cache/go-build is huge and left over by other tests
[19:10] <zyga> in general, look at the commit messages please
[19:10] <zyga> I need to try with -seed the next time it fails
[19:10] <zyga> I entirely forgot about that
[19:11] <zyga> it's not perfect because of job-stealing but it's better than random
[19:11] <zyga> ijohnson: I stand by all prior commits that improve the test
[19:12] <zyga> now some policykit test failed on something random
[19:12] <zyga> er, package kit
[19:15] <ijohnson> what do you mean by job-stealing ? for spread ?
[19:16] <zyga> spread has a inherently indeterministic scheduler
[19:16] <zyga> out of a pool of jobs, threads that are idle steal tasks to run
[19:17] <zyga> so if server side timings are not identical, you don't get the exact same ordering
[19:17] <zyga> it's not bad but it's not a guarantee to reproduce a failure
[19:22] <ijohnson> interesting I didn't know that
[19:26] <zyga> ijohnson: obviously, after adding the debug section, the tests pass
[19:26] <zyga> heh
[19:30] <ijohnson> hooray schrodenbug problem solved by looking at
[19:32] <zyga> hehe
[19:32] <zyga> 13 tests are in progress now
[19:32] <zyga> so some chances to reproduce it
[19:39] <tianon> anyone know whether "getent group foo" inside a snap environment will check /var/lib/extrausers/group appropriately? O:)  (my google-fu is failing me, and I'm not well-versed enough to easily test /o\)
[19:40] <zyga> tianon: on a core system yeah, it should
[19:40] <tianon> zyga: awesome, thank you! :D  (that makes my life easier)
[19:40] <zyga> tianon: because /etc/nsswitch.conf has extrausers entries
[19:41] <tianon> right, makes sense :)
[19:43] <ijohnson> hey tianon o/
[19:44] <tianon> ijohnson: hello :D  I bet you can imagine why I'm asking about snap-related things O:)
[19:44] <ijohnson> I can indeed!
[19:44] <ijohnson> let me know if I can be of any help
[19:44] <tianon> for sure, thank you! <3
[19:44] <ijohnson> :-)
[19:45] <tianon> hopefully this bump is uneventful, but you know, interesting things happen when you run a containerization platform inside a container, so who knows where we're headed xD
[19:46] <ijohnson> tianon: yes, hopefully uneventful
[19:46] <ijohnson> tianon: also btw I was meaning to ask you about a different project of yours, gosu, would you accept a patch to gosu to make it work with extrausers so we can use it on ubuntu core devices from inside snaps ?
[19:48] <tianon> ijohnson: really curious about the use case there -- you've got root inside the snap to do some initial setup and want to use gosu to step down after startup like is commonly done in Docker land?  or something more esoteric?
[19:50] <ijohnson> tianon: basically yes, we have now some support for dropping to users, specifically to the snap_daemon user, but the catch is that you have to use setgroups(...,NULL), which libc does not, and so to use standard linux tools you need to patch them or do a LD_PRELOAD hack to get it to work
[19:50] <ijohnson> tianon: gosu however doesn't suffer from that problem, but it doesn't understand extrausers
[19:51] <ijohnson> see https://snapcraft.io/docs/system-usernames for example
[19:52] <zyga> ijohnson: btw, about https://github.com/snapcore/snapd/pull/8605
[19:52] <mup> PR #8605: tests: fix raciness in pulseaudio test <Test Robustness> <Created by zyga> <https://github.com/snapcore/snapd/pull/8605>
[19:52] <zyga> you indicated you a review request there, is that something you still plan to carry on
[19:52] <zyga> I'd like to merge it when green, just to expand the reach of the fixes so far
[19:53] <ijohnson> zyga: yes hopefully this afternoon, but feel free to merge if it goes green before I get to it
[19:53] <zyga> and to get more data from other runs as people develop and we  get more scheduling runs
[19:53] <zyga> ok
[19:53] <zyga> thanks!
[19:53] <zyga> I'm happy to send follow-ups of any kind
[19:55] <tianon> ijohnson: interesting :D  have you tried setpriv from util-linux (2.32.1-0.2+)?  it's pretty flexible, and is essentially gosu but more "standard" O:)
[19:56] <ijohnson> tianon: iirc setpriv suffers from the same or a different problem and doesn't just work ootb
[19:57] <ijohnson> tianon: anyways I have a patch to gosu I could propose, and Ialso started an issue with upstream libcontainer but it would be more work to make it work upstream libcontainer
[19:58] <tianon> ijohnson: certainly open to further conversation -- maybe open the PR and we can discuss more there? :)
[20:01] <tianon> (then the conversation / thoughts can be recorded more persistently :D)
[20:07] <ijohnson> tianon: sure when I get some spare cycles I'll throw something up in GitHub
[20:08] <tianon> :+1:
[20:11] <zyga> ijohnson: I KNEW IT
[20:11] <zyga> 2020-05-06T19:49:53.9291698Z May 06 19:49:53 may061928-848515 pulseaudio[52998]: Failed to create secure directory (/home/test/.config/pulse): Permission denied
[20:11] <zyga> 2020-05-06T19:49:54.2159546Z + echo 'Files present in the test user'\''s home directory, that are owned by root'
[20:11] <zyga> 2020-05-06T19:49:54.2160021Z Files present in the test user's home directory, that are owned by root
[20:11] <zyga> 2020-05-06T19:49:54.2160438Z + test -d /home/test
[20:11] <zyga> 2020-05-06T19:49:54.2160837Z + find /home/test -user root
[20:11] <zyga> 2020-05-06T19:49:54.2160999Z /home/test/.local
[20:11] <zyga> 2020-05-06T19:49:54.2161142Z /home/test/.local/share
[20:11] <zyga> 2020-05-06T19:49:54.2161313Z /home/test/.local/share/applications
[20:11] <zyga> 2020-05-06T19:49:54.2161455Z /home/test/.config
[20:15] <ijohnson> haha wow that's broken
[20:15] <zyga> I'll find the guilty test tomorrow
[20:15] <zyga> small fix to unbreak master
[20:16] <ijohnson> when every day we have to land multiple small fixes to unbreak master are they really small anymore
[20:16] <ijohnson> and is master really "master" at that point
[20:16] <ijohnson> clearly something else is in control at this point
[20:20] <zyga> ijohnson: pushed
[20:20] <zyga> ijohnson: and EOD
[20:20] <zyga> and tomorrow I just, dunno, sleep longer
[20:20] <zyga> see you :)
[20:21]  * zyga waves
[20:22] <ijohnson> see you
[21:23] <zyga> cmatsuoka, ijohnson: please merge master / rebase rather than retrigger
[21:23] <zyga> it's merged!!!
[21:23] <cmatsuoka> zyga: \o/
[21:24] <ijohnson> hooray
[21:24] <mup> PR snapd#8605 closed: tests: fix raciness in pulseaudio test <Test Robustness> <Created by zyga> <Merged by zyga> <https://github.com/snapcore/snapd/pull/8605>
[21:24] <ijohnson> zyga: also what do you mean merge master rather than retrigger ?
[21:25] <zyga> I often retrigger tests
[21:25] <zyga> just to roll the dice again
[21:25] <zyga> if you merge master *this* particular failure can hopefully go away for good
[21:27] <cmatsuoka> merged and pushed a couple of branches, let's see what happens... :)
[21:27] <zyga> thanks!
[21:27] <zyga> I pushed to one of mine
[21:27] <zyga> one of my? or one of mine?
[21:27] <cmatsuoka> mine I guess
[21:27] <cjwatson> one of mine, yes
[21:28] <cmatsuoka> dinner time! will be back to see what happened to the PRs...
[22:50] <cmatsuoka> zyga: yay, it worked
[22:51]  * cmatsuoka EODs