[00:15] Bug #1839230 opened: duplicate login message on raspberry pi core18 image [05:24] morning [05:34] PR snapd#7186 closed: gadget: ensure filesystem labels are unique === pstolowski|afk is now known as pstolowski [07:07] mornings [07:32] pstolowski: hey [07:32] wow, tmux died [07:32] as in segfauled and all [07:36] good morning [07:36] * zyga is curing his sleep deficit by falling asleep at random times :/ [07:36] how are you doing guys? [07:44] sil2100: hello [07:44] sil2100: do you think we can ship bash-completion on core18? [07:44] sil2100: this is reported as https://bugs.launchpad.net/snappy/+bug/1825254 [07:44] zyga: hey [07:44] Bug #1825254: auto-complete doesn't work on ubuntu core 18 [07:45] zyga: tried to investigate the http2 PROTOCOL_ERROR we see randomly, but there's quite a number of places in net/http that generate an error matching the log [07:45] zyga: left a topic though https://forum.snapcraft.io/t/snap-download-failures-with-protocol-error/12677 [07:46] zyga: fwiw, nothign significant in curl output, also a minial http request in tried in go did not raise anytying suspicious [07:47] mborzecki: thank you [07:47] mborzecki: I wonder if it could be a go thing? [07:47] but we didn't change anything on our end, did we? [07:47] no now go compiler rev? [07:48] zyga: hm maybe, i can try with older version of go, 16.04 uses 1.10? [07:48] mborzecki: I don't remember, just wonder if it's a bug in go in general and we're hitting it along with others now [07:48] mborzecki: the alternative is that it is a bug in our web side [07:48] mborzecki: but again, I don't have any data to sway either side [07:49] sil2100: on the same vein, is https://mail.google.com/mail/u/1/ something we can fix? [07:49] sil2100: seems like a pair of tweaks to core18 [07:50] er :) [07:50] https://bugs.launchpad.net/snappy/+bug/1839230 :) [07:50] Bug #1839230: duplicate login message on raspberry pi core18 image [07:50] silly copy paste [07:52] zyga: no change with 1.10, wish there was a way to dump the http request/response headers in go in a reliable way [07:52] mborzecki: would using pcap help? [07:52] or are those all ssl? [07:52] zyga: uses tls [07:56] zyga: about that reddit thread: https://www.reddit.com/r/linux/comments/cmse4r/interview_why_canonical_views_the_snap_ecosystem/ the comemnts are pretty depressing [07:56] yeah, internet is as it was [07:56] I think it's important to have our own voice [07:57] I hope that degville martin and alan can come up with the daily snap podcast or something like that [07:57] I think a good fraction of those comments are just misinformed [07:58] also funny how some particular users seem to be repeat the same thing over and over again under a number of submissions [07:59] i mean, i understand how one can not like something, but tbh it seems as if some poeple have plenty of free time ;) [07:59] kenvandine: hey, do you know about this issue where the mouse cursor changes appearance inside some snap applications? do you know how this part of the stack works? [08:00] kenvandine: there are some bug reports about it [08:00] one I found just now is: https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1579268 [08:00] Bug #1579268: Mouse cursor is different inside graphical windows of snaps (snaps not using system theme) [08:00] zyga: iirc electron was a suspect there last time [08:00] mborzecki: yeah but electron inside snaps [08:01] mborzecki: we could call the podcast "responding to internet trolls" ;) [08:11] I wish there was one click toggle for dark mode / light mode [08:12] terminal profile, browser profile, terminal window color, etc [08:12] it's annoying to have to do this little by little by hand [08:13] PR snapd#7218 opened: tests: measure behavior of the device cgroup [08:13] I'm pushing a PR with the tests that show the bug from Friday [08:13] https://github.com/snapcore/snapd/pull/7218 [08:13] PR #7218: tests: measure behavior of the device cgroup [08:13] working on the fix, I gave it some thought and while the 1st part of the fix is indeed very easy [08:13] the second part is hard [08:13] the first part is when you "snap connect" the cgroup is confining [08:14] the hard part is when you snap disconnect the last interface affecting udev [08:14] you really need to switch the cgroup to unconfined [08:14] but that's not easy in our model [08:14] it's also related to udev [08:14] so we really need a the equivalent of snap-update-ns for cgroups [08:14] and that's not trivial [08:21] zyga: ah, yes, should be a easy fix! Let me note that down [08:29] sil2100: thank you! [09:00] pulled https://github.com/snapcore/snapd/pull/7166 and the unit tests fail consistently, will investigate more at the airport [09:00] PR #7166: client: add doTimeout to http.Client{Timeout} [09:01] * zyga works on fixing https://bugs.launchpad.net/snapd/+bug/1838937 [09:01] Bug #1838937: device cgroup enforcement is partially ineffective for snap services [09:01] mborzecki: thank you! [09:03] pstolowski: https://github.com/snapcore/snapd/pull/7209#issuecomment-519009878 [09:03] PR #7209: firstboot: queue service commands before mark-seeded [09:04] mborzecki: hmm, interesting, ty [09:07] ok, wrapping it up and heading to the airport in a bit and the off to Flock [09:10] * zyga designs snap-update-cgroup [09:10] it's actually good [09:10] because that's the only part that needs to talk to udev [09:10] * zyga thinks about getting breakfast first [09:24] ondra: ping [09:28] the easy half is fixed [09:28] now thinking about the "hard" half [09:29] I think I cannot avoid snap-update-ns for cgroups :/ [10:09] I wrote https://forum.snapcraft.io/t/the-need-for-snap-update-device-cgroup/12679 [10:09] it's a bit hard to follow, I will rewrite that for clarity [10:09] but I think it is unavoidable [10:11] pstolowski, ondra: ^ [10:11] ideas welcome [10:13] Hi! I am having a problem with launching graphical snaps in ubuntu 19.04. Can you help me? https://paste.ubuntu.com/p/2RnQWq3b8Q/ [10:15] Aavar: hey, can you run "snap version" please? [10:15] zyga: ack [10:15] zyga, https://paste.ubuntu.com/p/2tkZnVB3pg/ [10:16] Aavar: are you running wayland? [10:16] note that snap-store should not be launched as root [10:16] zyga, No, x11 [10:16] Aavar: can you run "snap connections inkscape" [10:18] zyga, https://paste.ubuntu.com/p/2DRKFM6qcY/ [10:18] (note that I am running Unity7, byt I had the exact same issue when running xfce) [10:19] do you get any apparmor denials? those can be checked by running "dmesg | grep DENIED" [10:19] (it will take me 10 minutes to download inscape on my connection) [10:20] zyga, yes i do. https://paste.ubuntu.com/p/mzDHfTfg5K/ [10:20] zyga, inkscape was just an exaple. I get similar errors on all apps with a gui. [10:26] * zyga tries to open the link but struggles [10:27] ok, I see it now [10:27] confinement rejects unix sockets [10:27] hmm [10:27] * zyga checks [10:28] zyga, is it better if I paste to pastebin instead? [10:28] Aavar: can you run "snap connections --all inkscape" [10:29] no, it's just my network, I'm sorry about that [10:29] ok :) [10:29] zyga, "error: cannot use --all with snap name" [10:31] zyga, https://paste.ubuntu.com/p/3fQgngzpvd/ [10:31] d'oh :/ [10:32] zyga: cursor themes are just icon themes. So just more icon themes to add gtk-common-themes [10:32] Aavar: hmmm, [10:32] kenvandine: hey [10:33] kenvandine: can you try to look at the behavior Aavar is reporting, graphical apps don't work, it seems they cannot talk to x11 [10:33] kenvandine: one example is inkscape [10:33] Aavar: 5 minutes to inkscape [10:34] zyga, :) [10:34] 5 minutes till I pull the snap [10:34] I have very slow network this week [10:35] zyga, I understand :) [10:36] Connections look fine [10:36] Not sure what's up there [10:37] unity7 gives access to x11 [10:37] I think [10:37] I can't look more now, time to get my son off to school [10:37] But when I get to work I can [10:38] thank you! [10:38] Aavar: installed [10:38] started fine, apart from fractional scaling [10:39] kenvandine, thank you. I could just reinstall ubuntu, but it would be great to figure out what I have done to cause this :P [10:39] Aavar: I have the same connections as you did [10:39] Aavar: when did this start to happen? [10:40] zyga, I few days ago, but I didn't think much of it (just installed the app I needed via apt) [10:40] did any graphical snaps work for you? [10:40] as in before some moment when they all broke? [10:41] zyga, yes, I used a few (for example inkscape) before they broke. [10:41] can you recall any change at the time it broke? [10:42] zyga, No, thats the thing. I cant remember anything that would cause this. [10:42] can you look at /var/lib/snapd/apparmor/ and then at the file related to inkscape [10:42] can you pastebin that file (snap.inkscape.inkscape) [10:45] zyga, cat /var/lib/snapd/apparmor/profiles/snap.inkscape.inkscape [10:45] zyga, https://paste.ubuntu.com/p/4Zndvv5ZzX/ [10:45] yeah [10:47] (ps: Is it a good idea for the purpouse of testing to stop the apparmor service?) [10:47] no [10:47] it's not a real service [10:47] stopping it won't do us any good [10:47] ok [10:47] give me a moment [10:50] Aavar: can you pastebin /etc/apparmor.d/abstractions/X [10:51] zyga, yes. https://paste.ubuntu.com/p/mmMyxsz7sQ/ [10:52] in a terminal, can you echo $DISPLAY and ls /tmp/.X11-unix [10:55] Aavar: ^ [10:57] zyga, echo $DISPLAY gives: :0 [10:57] ok [10:57] and ls /tmp/.X11-unix gives: X0 [10:57] PR snapd#7219 opened: devicestate/firstboot: check for missing bases early [10:58] hmmm [10:58] so far everything looks good [10:58] can you run [10:58] sudo apparmor_parser -r /var/lib/snapd/apparmor/profiles/snap.inkscape.inkscape [10:58] and then "snap run inkscape" [11:00] brb [11:02] zyga, still the same errors: https://paste.ubuntu.com/p/F8dV2x3WHR/ [11:12] hmmm [11:12] * zyga is lost [11:12] sorry, I have no idea what might be wrong [11:12] I it clear it cannot talk to X but I don't see why [11:13] * pstolowski lunch [11:16] zyga, do you think it could help if I removed everyhing regarding X and reinstall? [11:16] no, I don't think that would help [11:16] I think we need another pair of eyes to understand what is going on [11:16] let's check with kenvandine once he is back [11:17] zyga, yes :) [11:25] pstolowski hi [11:25] pstolowski sorry I'm still in Canada this week [11:29] ondra: hey :) [11:30] zyga hey [12:00] * cachio afk [12:03] ondra: ah, i had no idea, sure [12:13] sergiusens (cc kyrofa): I looked into why kyrofa was getting snap USN emails for snapcraft. there are entries in the store dump for revisions for which kyrofa was the uploader. those revisions aren't published to any channel in the dump I get though [12:16] sergiusens (cc kyrofa): I'd rather not second guess why the store dumps those. would have to talk to roadmr to see (eg, 1549, 1550, 1551, 1552, 1723, 1724, 1725, 1726, 1727) [12:18] sergiusens (cc kyrofa): I could change the algorithm for detecting who to send to, but would wan't to hear from roadmr why those revisions are in the dump [12:18] want* [12:20] jdstrand: thanks for taking a look! Will wait on what the situation is. === ricab is now known as ricab|lunch [12:47] PR snapd#7220 opened: cmd/libsnap: don't leak memory in sc_die_on_error [12:48] + snap install --dangerous /home/gopath/src/github.com/snapcore/snapd/tests/lib/snaps/test-snapd-timedate-control-consumer/test-snapd-timedate-control-consumer_1.0_all.snap [12:48]

[12:48] hmmmmm [12:48] master is a sad puppy [12:50] * zyga breaks === ricab|lunch is now known as ricab [13:46] xnox: hey [13:56] zyga Aavar: sorry, forgot i have back to back meetings all morning [14:04] roadmr: hey, can you look at https://paste.ubuntu.com/p/8m8dzfFRTz/ and get back to me whenever it is convenient? [14:04] jdstrand: sure, let's see [14:09] Aavar: please pastebin the output from this: [14:09] ls -ltr /run/user/`id -u`/ [14:10] roadmr: oh, you know what, I looked at this week's dump (where those revisions are in there with no channel), but the dump in question happened on July 31 [14:10] roadmr: (if that helps) [14:10] jdstrand: hmm so maybe they were published at that point in time? [14:10] roadmr: note that the USN email that went out on July 31 did not include those channel-less revisions, so I suspect they were channel-less at that time as well [14:11] hm. ok, jdstrand, we're checking the logic in the dump script [14:11] roadmr: thanks. I wouldn't consider this urgent work, so, however you want to prioritize it is fine [14:12] roadmr: I thought I remembered there were some brand stores or something that might've pinned old snapcraft revisions for some internal issue that has long since been resolved... [14:13] I don't have any more details on that [14:28] back from a walk but I need to eat something first [14:28] because not in a rush I will return in >1hour [14:28] jdstrand: the fix works, I will give first RFC patches today [14:28] jdstrand: it's more complex than before but also easier than I thought :) [14:28] jdstrand: (than in our initial discussions) [14:29] jdstrand: I think overall it's not super complex though [14:29] that's cool [14:29] jdstrand: more streamlined I would say :) [14:29] now for that lunch, I'm starving :) [14:29] I like the more deterministic aspect [14:29] zyga: note, I'm stepping away [14:29] for a little bit [14:29] jdstrand: that's okay [14:29] I'm very sleepy all day long, cannot readjust [14:31] zyga: yeah, hope you get back on track soon [14:55] zyga, hey, I see 2 unit tests faiing https://paste.ubuntu.com/p/Ky4pw9G38D/ [14:55] seem to be related both fails [14:55] zyga, do you know if it a new thing? [14:56] cachio: it's a thing maciej was investigating lately [14:56] he sent a PR to fix it [14:57] zyga, nice [14:57] I'll take a look [14:57] thanks [14:57] cachio: but I don't know if that's enough, I think we have some more bugs in this area [15:00] zyga, ok, thanks for the info [15:12] cachio: I his PR has landed, it was 7215 [15:13] zyga, in that case there is something wrong because the next Pr merged failed wiht the error I showed [15:13] the #7186 [15:14] PR #7186: gadget: ensure filesystem labels are unique [15:14] yeah, I think it's just buggy still [15:34] * cachio lunch === pstolowski is now known as pstolowski|afk [16:57] PR snapd#7213 closed: cmd/snap-confine: implement snap-device-helper internally [17:04] * zyga takes a break [19:09] * cachio afk [20:29] PR snapd#7221 opened: tests: split the sbuild test in 2 depending on the type of build [20:36] kenvandine [20:37] ha i ctrl+f and that doesn't work in polari apparently [20:39] PR snapcraft#2653 closed: scriplets: run override-pull on update_pull [23:30] PR snapcraft#2654 closed: click: update to 7.0 [23:42] PR snapcraft#2656 opened: appstream: xslt support for ul nested in p [23:57] PR snapcraft#2657 opened: Release changelog for 3.7.2