[01:41] jdstrand, kyrofa, popey Wimpress https://github.com/snapcore/snapcraft/pull/2234 [05:03] morning [05:55] Good morning [05:55] Summer feels over now === chihchun_afk is now known as chihchun [06:19] good morning [06:20] again [06:20] this time from a computer [06:20] man, Monday will be brutal :) [06:22] zyga: good morning [06:22] zyga: monday is back to school? [06:23] yes [06:27] mvo: reviewed systemd env generator PR [06:27] zyga: nice [06:32] mvo: zyga: hey [06:33] yeah, i'll likely be starting late on monday (or have a break sometime around morning) with the school and all [06:36] mborzecki: same here [06:36] two kids to sort out :) [06:36] mborzecki: good morning [06:36] mborzecki: no worries [06:44] and back to the 'breakfast, work, kids from school, help with home homework, , dinner, sleep, goto 1' routine [06:44] zyga: curious, why a man-page for the generator? this thing is pretty well hidden [06:47] All generators have them [06:47] zyga: aha, ok [06:48] It is good courtesy to users since this is an opaque compiled program [06:49] systemd-environment-d-generator has it too, although it's not very useful [06:53] I’m moving outdoors [06:53] Network at home sucks [06:53] And I need to do a trial school run to see how bus traffic looks like === pstolowski|afk is now known as pstolowski [07:10] morning [07:24] pstolowski: heyah [07:31] re [07:40] mborzecki: the arch failures on 2.35 look real [07:41] mvo: that's unfortunate, do you have link to the logs? [07:41] mborzecki: I mean, they are the same as yesterfday, any idea ? and if this does not affect master we probably just need to find the right commit that fixed it in master [07:41] mborzecki: https://api.travis-ci.org/v3/job/422560948/log.txt [07:43] mvo: ssh keygen? [07:43] mborzecki: ohhh [07:43] mborzecki: yeah, let me find htis one and cherry-pick it [07:43] mborzecki: thats definitely a good candidate [07:46] mborzecki: cherry-picked and pushed, lets see if that helps [07:46] mborzecki: thank you! [07:48] mvo: hopefully it'll be green now :) [07:49] breaking compat one by one :) === chihchun is now known as chihchun_afk [08:02] moin moin [08:03] any idea why #5744 still shows up in github as 'in progress' (and no coverage report) when the tests passed? [08:03] * zyga works on trespassing unit tests [08:04] working from suse :) [08:04] I think we have enough people with ubuntu [08:04] (in the team, go ubuntu :) [08:06] haha [08:10] hmm [08:11] when snapd dies, "snap info" lies [08:11] it says "no snap found" when it should really say that it couldn't contact the store [08:11] https://www.irccloud.com/pastebin/MErmQt88/ [08:11] popey: it didn't say there weren't any, it said it didn't find any [08:11] popey: like when I tell my boys to go get something [08:11] "Man looking" [08:12] popey: "I LOOKED NOWHERE AND I CAN'T FIND IT" [08:12] :D [08:12] Should it not also warn that it can't get to snapd? [08:13] the "I looked nowhere" bit [08:13] maybe snap should learn to pick things up and look under things, especially if it just put that thing down [08:14] It should at least pick up the boxes and give them a bit of a shake. [08:14] that's like level 3 [08:14] no chance [08:15] snapd needs to lean the Jan Hankel Flank Pat system for finding things [08:15] I see what's going on, I'll push a pr to address it in a bit [08:15] https://www.youtube.com/watch?v=DY-Zdgo0OXo [08:15] Thank you! [08:16] popey: the first time I saw that vid I didn't notice how old it was [08:16] maybe it was a long time ago [08:16] hm [08:18] Yeah, Mitchell & Webb Look aired 10 years ago! [08:19] mborzecki: aarch is greeeeeeeeeeen [08:19] yay, rejoice :P [08:19] mborzecki: merged into 2.35 thanks again [08:19] popey: if you did 'snap info foo bar' in that situation, what would you expect to see? [08:24] We'll. Shut down snapd and run snap list. You get a completely different experience [08:24] I would prefer snap list had a nicer error too [08:24] When snapd is down [08:25] (in this case snapd was down due to lack of disk space) === chihchun_afk is now known as chihchun [08:30] popey: the problem is 'snap info' does a lot more stuff than 'snap list' [08:30] popey: so each query to 'snap info' can produce up to 2 errors during normal operation [08:31] popey: and up to 3 errors when snapd is down [08:32] popey: note info _can_ produce useful output with no snapd [08:32] popey: e.g. 'snap info /var/lib/snapd/snaps/*' [08:34] (I never knew that) [08:34] (I bet nobody else does either) [08:34] popey: 'snap info /snap/core/current/' also tries to be informative [08:35] popey: or 'snap info ./prime' [08:43] I need to make the errors from client a little more structured before I can hope to make this work [08:43] but we've been wanting this for a while now [08:43] so, yay [08:44] \o/ [08:44] yeap. Just, not a quick fix [08:44] no rush [08:44] is it done yet? [08:48] popey: https://www.youtube.com/watch?v=RYVlJKSEsyM [08:49] awwww <3 that video [08:49] Those were happy times [08:52] popey: back when we were part of ESA you mean? [08:52] zyga: niemeyer: https://github.com/snapcore/snapd/pull/5747 [08:55] +1 [08:57] zyga: is there a way to make apparmor refuse to stat/etc/lsb-base-logging.sh ? [08:57] zyga: context is https://bugs.launchpad.net/snapd/+bug/1779416 [09:03] main.go:192: cannot change mount namespace of snap "mattermost-desktop" according to change mount (/snap/gtk-common-themes/319/share/sounds/communitheme /snap/mattermost-desktop/78/usr/share/sounds/communitheme none bind,ro 0 0): cannot create writable mimic over "/snap/mattermost-desktop/78/usr/share/": permission denied [09:03] ^ anyone seen that before? What does it mean? [09:03] zyga: ^ [09:03] I am launching mattermost-desktop from edge, and have the communitheme install [09:07] popey: hey [09:07] popey: known issue, please show me the snapcraft yaml, I will show you how to work around it [09:07] it's a trivial trick [09:22] zyga: https://github.com/snapcrafters/mattermost-desktop/blob/master/snap/snapcraft.yaml [09:23] looking [09:23] right [09:24] if the snap ships $SNAP/usr/share/{icons,themes,sounds} as empty directories [09:24] this bug goes away [09:24] hah! [09:25] ooh, new mycroft [09:25] * diddledan gets compiling [09:25] popey: the issue is that the rules for poking holes are super precise [09:25] and they should be more generic in cases like this where there are many more directories to create [09:26] popey: I'm working on fixing that (2.36) [09:26] ok, thanks. [09:28] oh, looks like it's not rolled-out yet (mycroft 18.08) [09:29] zyga, snapd 2.35 rolled out to fc28 [09:29] still waiting for karma for fc27 [09:29] Son_Goku: Thank You! [09:29] Son_Goku: I can give it a spin today [09:29] I'm trying to get out of a very long unit test change [09:29] for a .1 release [09:30] Son_Goku: btw, I want to participate in the golang sig [09:30] I have my own proposal to contribute [09:30] I was working on it after hours recently [09:30] cool [09:30] something that fits upstream and downstream at the same time [09:30] is easy to use for regular go developres [09:30] and does all that distros need [09:30] (magic, right?) [09:30] I will have it ready in a few days [09:31] I think there's a real chance to unify fedora and suse helpers [09:31] heh [09:33] zyga: no more shell wrappers for each go command? [09:33] mborzecki: exactly [09:34] anyway, back to work, I'll demo it next week [09:35] heh going through opensuse wrappers was interesting, but to be fair OE also has some hacks spefifically for go packaging [09:36] Fedora's wrappers are pretty thin [09:36] the macro directly calls the go command [09:43] Son_Goku: well, suse has an elaborate shell script(s) :/ [09:43] it used to be ruby [09:43] so be happy it's not that [09:43] hehehe [09:43] Son_Goku: but ruby is like haiku [09:44] Son_Goku: or better, the code reads like plain english [09:44] suse's golang-packaging being rewritten from ruby to shell was the result of an ugly dispute with the original creator of suse's golang packaging standard [09:46] Son_Goku: otoh the languages bundling their dependencies do not really fit that well with the packaging model used by some distros, i can't imaging splitting dependencies of a trivial node app [09:46] you don't have to imagine [09:46] it happened for a while in Fedora [09:46] and it still does in Debian [09:47] yeah, the number of ITP mails which are "please package this node module" is too damn high [09:58] mborzecki: Thank you! And sorry for the late reply [09:58] And good morning :) [10:03] niemeyer: you got to say 'good morning' by two minutes :-) [10:03] niemeyer: good afternoon sir [10:03] Aftnoon! [10:04] :-) [10:05] niemeyer: sure, thanks for the review [10:05] zyga: pstolowski: if you have a window for an easy review, #5744 is kinda blocking me (I could merge it and then let git sort it out, and I'll do that if it's not reviewed soonish, but i'd rather avoid it) [10:06] looking [10:06] woo [10:06] thanks [10:06] zyga: tests are green despite github not knowing it [10:06] kk. and what happend to mup btw, is it gone again? [10:06] mvo: my dragon apparently has android on emmc [10:06] pstolowski: ^ zyga's on it, thanks [10:06] niemeyer: pushed the changes to https://github.com/snapcore/snapd/pull/5713 too (hopefully got all of it now) [10:07] also, how can I merge a PR if travis didn't let github know it was green? [10:07] :-( [10:07] Chipaca: ask mvo [10:08] Chipaca: is bytes.Bytes expensive? [10:08] er [10:08] buffer.Bytes [10:08] zyga: only a little bit more than doing the maf ourselves [10:08] zyga: memorywise it's about the same [10:08] (especially as it's already used elsewhere) [10:09] Chipaca: I can [10:09] why did you move w.partial = nil; to be a w.partial.Reset() a few lines below (after } ) [10:09] Chipaca: which one do you need [10:09] 5713 [10:09] er [10:09] not [10:09] 5744 [10:09] mvo: 5744 (but it doesn't have zyga's second +1 yet) [10:10] Chipaca: https://github.com/snapcore/snapd/pull/5744/files#diff-7c96d8cb550f40c45f4ded10cc30f5eeL54 about this line [10:10] zyga: yep ye [10:10] it is clear that it is just a p.Reset but why did you move it? [10:10] is it because of https://github.com/snapcore/snapd/pull/5744/files#diff-7c96d8cb550f40c45f4ded10cc30f5eeL58 [10:10] which is not 1-to-1 (write is append, = was reset + write) [10:10] zyga: because I'd be doing it in the other paths as well [10:10] ok [10:10] zyga: that is: w.partial = p -> w.partial.Reset(); w.partial.Write(p) [10:11] +1 [10:12] mvo: ok now yes please [10:12] done [10:12] mvo: 3/4 flashed [10:13] mvo: https://github.com/snapcore/snapd/pull/5744 is green and has two +1's [10:13] * Chipaca ~> break [10:14] Chipaca: done thank you [10:14] mvo: whee, thanks [10:14] wait i was supposed to be on a break [10:14] * Chipaca hides [10:14] hm, is there any way to see which snap pulled in a core18 ? [10:15] we should include that in the error (there's a // TODO) [10:15] (in the "snap remove core18" error i mean) [10:15] popey: snap info --verbose should show the base of a snap [10:15] yep, it does [10:16] maybe snap list should include it as well [10:16] or snap info core18 could list the rdepends [10:16] mvo: oooh [10:16] mvo: I'll add that [10:16] that would be nice [10:17] could also look at slots/plugs [10:18] diddledan: you moved to core18 on gimp? :) [10:18] yes. I've got the snap building on launchpad for now [10:18] zyga: where was the mapping thing that changed system -> core and back? [10:19] just this second told it to do a new build for libx11-related security patches overnight [10:19] mborzecki: in overlord/ifacestate [10:19] man, SD cards are slow [10:20] zyga: i'm thinking if we could have something similar instead of snap.Info.ExpandSnapVariables(), like a mapped that does this in 'inside' snap mount ns or 'outside' (cc niemeyer) [10:20] s/mapped/mapper/ [10:20] hmmm [10:20] * zyga thinks [10:21] zyga: then for eg. snap.MountNSView().ExpandVariables(info) [10:22] mborzecki: The problem tastes a bit simple for that [10:22] mborzecki: What changes between the two cases, and how often we need either? [10:23] diddledan: that'll be why core18 number of installs rocketed up over the last few days then :) [10:23] niemeyer: yeah, the only use cases now are 'inside' mount ns, so we could go yagni about that just have one function for this (as it is right now) [10:23] mborzecki: ! [10:23] rocketed? [10:23] one snap. yeesh :-p [10:24] one - popular - snap :) [10:24] aha [10:24] I see what you mean [10:24] pics or it didn't happen! [10:24] * niemeyer wants pictures too [10:25] https://snapcraft.io/core18 [10:25] look at the distro graph [10:26] I don't see a timeline :-( [10:27] On a tangent, but the map just reminded me of https://xkcd.com/1138/ [10:28] mvo: it's booting [10:28] MattJ: hah! :D [10:28] niemeyer: hey! will you find a moment today for a quick HO re udev & stop channel? [10:28] mvo: same as last time [10:28] mvo: let me look at logs [10:29] pstolowski: Yeah, definitely [10:29] mvo: same == "snap command not found" [10:29] I'll enable journal and see what we get [10:32] xnox: did you get my message about https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1778936 ? [10:32] xnox: i.e. can/should I sru this or will you do a sru with more changes anyway where this could be included? [10:32] mvo: I reflashed and created /var/log/journal [10:32] zyga: meh, that sucks [10:32] zyga: thanks [10:32] it failed [10:32] let me recover the journal after a second, I want to make sure this is written to the card [10:32] zyga: ta [10:33] zyga: the error you got a couple of days ago should be fixed so I wonder what is going on [10:33] mvo: I'll bring it for the sprint [10:34] sie 31 12:32:18 localhost run-snapd-from-snap[738]: stateengine.go:102: state ensure error: devicemgr: cannot use gadget snap because its base "" is different from model base "core18" [10:34] mvo: :D [10:34] let me know if you want another test run [10:35] mvo: I streamlined my setup so I can test it very quickly [10:35] there are more errors, I'll give you the full og [10:35] *log [10:37] mvo: boot log https://paste.ubuntu.com/p/pkhHBrzFNm/ [10:37] back to coding [10:37] ping me for another run [10:38] zyga: thanks, fixing [10:59] ouch, a panic from prepare-image [10:59] * zyga is hungry [10:59] Chipaca: bt? [11:00] Chipaca: yeah, I saw this a couple of time in the lsat day [11:01] Chipaca: I pushed a fix for the panic [11:01] Chipaca: its an open pr, that hopefully gives us clues what is going on [11:01] Chipaca: 5743 if you look for a simple review ,) [11:01] mborzecki: https://pastebin.ubuntu.com/p/2ZTtNDngWp/ [11:02] mvo: nice [11:02] Chipaca: the question is really why we get into this situation, its impossible :) (well, clearly not but) [11:04] mvo: because it's only added to the map if typ == snap.TypeKernel || local.Name(snapName) == baseName { [11:05] mvo: and then it looks in that map for 'snapweb' [11:05] Chipaca: is it snapweb its crashing for? [11:05] Chipaca: I thought that was invisilble without the other PR [11:05] (but maybe I hvae not looked hard enough) [11:05] mvo: it prints "fetching %s" before looking at the map? [11:06] bah, i don't realluy know this code [11:06] so i could have it backwards === pstolowski is now known as pstolowski|lunch [11:07] hmm hmm [11:07] mvo: so, with your change, the test will still fail at this point, right? [11:08] Chipaca: yes [11:08] Chipaca: it will fail but report on what snap it failed [11:09] Chipaca: I can add some more code to diagnose this I think [11:10] Chipaca: heh, you suggested this too :) let me add it [11:10] mvo: +1'ed, but yeah [11:11] Chipaca: ta [11:23] re [11:34] * Chipaca ~> lunch === chihchun is now known as chihchun_afk [11:42] niemeyer: https://github.com/snapcore/snapd/pull/5749 [12:02] zyga: hey, I'm not really here. I didn't get to the /mnt PR yesterday but will try today [12:03] jdstrand: hey [12:03] jdstrand: mvo and I decided to merge it (ahead of an important release) but if you find anything we will do a point release [12:03] sorry about that, I hope that's okay [12:03] jdstrand: so take your holidays for real and let's do this next week [12:11] mvo, there is an sru outstanding to do, yes. [12:11] mvo, however, it appears that snapd is failing autopkgtests now that said patch is in, in cosmic-proposed. [12:12] https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-cosmic/cosmic/amd64/s/snapd/20180830_184959_a6800@/log.gz [12:12] 2018-08-30 18:49:42 Failed tasks: 1 [12:12] - autopkgtest:ubuntu-18.10-amd64:tests/main/dirs-not-shared-with-host:alternatives [12:12] mvo, as if i should back that change out.... if it is related. [12:16] mborzecki: less than 20 errors left, brace yourself for that long code review :D [12:21] jdstrand: who else can approve snaps? I have some gadgets with "base: core18" that neeed approval [12:21] xnox: its not related I will push a fix for this [12:22] xnox: its a core snap change [12:22] mvo, oh, ok! [12:22] mvo, that's good =) cause it means i need not to back out things in systemd package then! [12:22] xnox: yeah, I'm very happy if the fix is included too :) === pstolowski|lunch is now known as pstolowski [12:26] wheee [12:27] all those failed-to-release: https://www.youtube.com/watch?v=0BHQT3Omqtw [12:27] err [12:27] wrong link [12:27] file:///home/dllewellyn/Pictures/Screenshot_20180831_132719.png [12:27] bah [12:27] THIS TIME! https://usercontent.irccloud-cdn.com/file/hKp3lTzL/Screenshot_20180831_132719.png [12:29] diddledan: i had some store upload rejections as well, mine were some new fields from snapcraft iirc [12:30] mborzecki: it's green :) [12:30] mborzecki: I'll chop it [12:30] https://www.youtube.com/watch?v=TkZFuKHXa7w [12:34] ^^ without comment [12:39] mvo, hey [12:39] https://paste.ubuntu.com/p/Gr835VXzcF/ [12:39] did yo see that error on master? [12:41] cachio: yes [12:41] I pushed a PR to make it more obvious what is going on [12:41] mvo, great, thanks [12:42] cachio: 5743 - its ready [12:42] cachio: that hopefully gives us a clue [12:42] cachio: what is going on there [12:42] cachio: I saw it a couple of times today and yesterday [12:43] diddledan: let me try to find out what is going on with the rejections - do you also get it for nknown keys in snap/manifest.yaml: snapcraft-os-release-id [12:45] no idea why it's failed to release [12:45] everything appears normal [12:46] for example, micropolis failed to release, but: https://usercontent.irccloud-cdn.com/file/eZSJeaNF/Screenshot_20180831_134635.png [12:47] mvo, yes it started yesterday suddenly, some branches without any change started failing [12:48] gimp just got rejected with "unknown keys in snap/manifest.yaml: snapcraft-os-release-id,snapcraft-os-release-version-id,snapcraft-version lint-snap-v2_snap_manifest" [12:54] brb, see you at the call [12:58] diddledan: yeah, I had the same [13:00] apparently I singlehandedly skyrocketed the number of core18 installations (according to popey ) [13:00] I call shenanigans [13:01] as if my gimp is really _That_ popular!?! [13:01] "skyrocketed" :D [13:02] oh, right, so in the last 3 days 77 thousand people have updated to core18ified gimp... [13:03] htf have I managed to become influential to 100k people?! [13:04] Turns out people like gimp. who knew? :) [13:04] nuts [13:15] left with no comment: https://usercontent.irccloud-cdn.com/file/gOjXXEh5/Screenshot_20180831_141513.png [13:16] actually, question regarding that, is nil gonna be treated as string 'nil' or as no-value? [13:17] I wanted it to be no-value [13:17] i.e. don't you dare think of giving me a base damn you! [13:19] * diddledan hacks all the things [13:22] mvo, zyga: https://github.com/snapcore/snapd/pull/5750 [13:22] mvo, also, your script for generating changelogs is broken [13:23] I had to add the "- New upstream release " line for the last few changelog entries [13:23] Son_Goku: shall we sort the changelogs like I did in suse? [13:23] I need to update my script [13:23] sort them? [13:23] the dates are broken [13:24] at least in suse they were, due to various historic mistakes [13:24] yeah [13:24] the dates and authors need to be fixed [13:24] https://gitlab.com/zygoon/rpmtools [13:24] I thought about fixing it the last time I went in and whacked the changelogs === zarcade_droid is now known as ^arcade_droid [13:24] err fixed the suse packaging [13:24] just need to handle embedded changelogs [13:24] changes format changed last week [13:25] zyga: spot the difference: https://build.opensuse.org/package/view_file/system:packagemanager:dnf/dnf/dnf.changes?expand=1 [13:25] Son_Goku: thanks, in a meeting right now, but thanks for the PR - and sorry for the broken script I have a look and fix it [13:26] Son_Goku: in a meeting, but I will look after [13:26] mvo, zyga, clearly you're both in the same meeting :P [13:26] yep [13:26] standup [13:29] zyga, also: https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7-beta/html-single/7.6_release_notes/#BZ1461652 & https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7-beta/html-single/7.6_release_notes/#BZ1564775 [13:39] cachio: uploaded [13:39] niemeyer, tx [13:40] Son_Goku: how do the two rhel notices you linked to affect us? [13:41] zyga, well, the first one means that the code I'm writing for integration in fedora infra will work for producing centos7 base snaps eventually [13:41] as for the second one, I'm not sure, it sounds like something we might want? [13:42] at least with the mmap() support, that means that the policy will compile again on el7 [13:42] zyga, it seemed like the socket stuff is something along the lines of what the socket mediation stuff in apparmor does [13:43] Son_Goku: apparmor does packet inspection (so to speak) [13:43] more like a firewall than a traditional "you can or cannot use that socket" control [13:43] ah [13:43] but I understand what you meant now [13:49] diddledan: I got feedback about the store rejections, they are working on it and it should be back to normal soon. out of curiosity, is your gimp snap core18 based? [13:50] yes, it is [13:50] diddledan: thanks for confirming. it looks like only the 18.04 build env is currently in unhappy land [13:50] hence 77k new core18 downloads this week :-p [13:50] * mvo hugs diddledan [13:51] diddledan: thats cool [13:51] ooh cuddles! [13:52] I wanna see the graph of core18 to see what impact that's had overall [13:53]