[00:48] hello [00:49] I got ubuntu 17.10 on my machine and got installed skype as snap package [00:49] also latest nvidia card with propriety drivers [00:50] i got no error no the terminal when i start the skype but it just simply doesn't start [00:50] no hanged process on the ps aux. it's like kills instantly at the start without any error [00:51] how can i get the error log at least? [01:18] Hello, How Do I get A Spotify Icon (from the Spotify snap) in my menu ? [01:23] CodeMouse92__: sup ? [01:23] ManxPies: Well HELLO there! I know you, methinks ;) [01:25] CodeMouse92__: Can you not see my PM's or something ? [01:25] ManxPies: Hmm? I just PM'd you. [01:25] OH! I have non-registered nicks blocked from PMing. Spam [01:26] CodeMouse92__: So you cannot see this ? https://paste.ubuntu.com/p/6sYv2fKfVm/ [01:27] ManxPies: Yeah, no, don't see that. Here, let's move this to an ot, so we don't muck up the room! ##c++-share [01:27] So what #channel - should we talk in (apart from u-o) ?? [01:27] okay [01:30] Also, is there a Brave Browser snap ? [03:01] PR snapcraft#2004 closed: errros: add a specific error when running commands from plugins [05:48] zyga-ubuntu: http://kernel.ubuntu.com/jj/ubuntu-artful.git#lp1755563 [06:11] morning [07:10] jjohansen: thank you! [07:10] mborzecki: good morning [07:10] PR snapd#4859 opened: tests: autopkgtest may have non edge core too [07:11] zyga-ubuntu: hey [07:11] zyga-ubuntu: hey, good morning - and good morning to mborzecki [07:11] mvo: morning [07:11] I should not be working today but I just jumped in to look at some things briefly [07:11] are we ready for 2.32 yet? :) [07:12] not fully [07:12] woah [07:12] https://travis-ci.org/snapcore/snapd/builds/354083620?utm_source=github_status&utm_medium=notification [07:13] the test ran for 6 hours and 58 minutes [07:13] the log says it succeeded [07:14] but the job didn't end [07:14] interesting, travis didn't kill it at 60 minutes mark [07:16] looks like a bug [07:56] zyga-ubuntu: from the screenshot that panic from last night is with 2.23.1 (the original dragonboard image?) [07:56] pedronis: I believe so === pstolowski|afk is now known as pstolowski [07:57] mornings [07:57] https://forum.snapcraft.io/t/panic-on-dragonboard/4490 [07:57] pedronis: ^ [07:57] do we make new ones? [07:57] can we even fix it? [07:57] PR snapd#4856 closed: release: 2.32~pre2 changelogs [07:57] I don't know who manages those images [07:59] PR snapd#4859 closed: tests: autopkgtest may have non edge core too [08:01] zyga-ubuntu: I think it's a bug we fixed since a long while [08:02] https://github.com/snapcore/snapd/blob/2.23.1/store/auth.go#L106 := vs = [08:03] I think we just need to refresh the image with new snapd build [08:04] I don't who makes or what's the refresh policy of those [08:05] it is foundations managed now, we need to talk to steve if we need a refresh [08:05] the policy is that it gets only refreshed on regular point releases of the distros [08:05] (so not very often) [08:07] mvo: zyga-ubuntu: mmh, interesting, http://cdimage.ubuntu.com/ubuntu-core/16/stable/current/ everything seems from January except dragonboard that is older [08:07] heh (or not) [08:07] woah [08:07] March 2017 [08:08] yeah [08:08] slangasek: ^ is the dragonboard image expected to be much older than other images? [08:09] * pedronis breakfast [08:16] moinsen [08:25] PR snapd#4847 closed: osutil,testutil: add symlinkat(2) and readlinkat(2) [08:29] PR snapd#4808 closed: cmd/snap: use shlex when parsing `snap run --strace` arguments [08:30] PR snapd#4738 closed: snap: unify snap name validation w/python; enforce length limit [08:30] mvo: I looked into what is involved to backport #4829 , I got conflicts related to #4722 , as we discussed yesterday I will try to look how involved/annoying is backporting that [08:30] PR #4829: store: Sections and WriteCatalogs need to strictly send device auth only if the device has a custom store [08:30] PR #4722: store: cleanup test naming, dropping remoteRepo and UbuntuStore(Repository)? references [08:31] * zyga-ubuntu rebased 4851 and keeps it in a debug loop [08:32] I saw /etc error again so there must be _another_ reason for it [08:37] PR snapd#4860 opened: spread,tests: move suite-level prepare/restore to central script [08:40] zyga-ubuntu: hi. I was having a look at the "safe bind mount" problem, and came up with some initial code based on what we discussed last week [08:40] zyga-ubuntu: https://github.com/jhenstridge/snapd/blob/safe-bind-mount/cmd/snap-update-ns/secure_bindmount.go [08:41] it isn't identical: I found mount(..., MS_MOVE) doesn't work in the presence of shared mount namespaces [08:43] looking [08:45] the remount to read only, that will not handle recursively bind mounted things, will it? [08:46] you can also use filepath.Clean to simplfy some of the logic [08:46] lastly, the MS_MOVE issue [08:46] zyga-ubuntu: yeah. Just noticed that. I can also move the remount to the stashDir too, which gets rid of the FIXME [08:47] I' [08:47] ll check out filepath.Clean [08:48] so when is MS_MOVE forbidden? [08:48] when crossing namespace boundaries? [08:48] or when crossing shared peer group kind? or when? [08:49] zyga-ubuntu: I was just getting EINVAL from the system call [08:49] pedronis: thank you [08:49] I got a more expressive error from /sbin/mount, which seems to be due to it doing some checking before hand: https://git.kernel.org/pub/scm/utils/util-linux/util-linux.git/commit/?id=fcc0413a12efde3f413064a622ea0a0595ede49f [08:50] I see [08:50] but [08:50] can we make the internal stash private? [08:50] bind from original place [08:50] to stash [08:50] make stash private [08:50] MS_MOVE is supposed to be atomic, and I guess they decided that making the mount propagation atomic was too hard [08:50] and move to destination [08:51] ah [08:51] the destination is shared too [08:53] jamesh: I'll pick this up next week when [j]dstrand is back [08:53] it looks like the parent of the stash directory would need to be !MS_SHARED [08:54] if I'm reading the util-linux patch right [08:54] yes [08:54] or the stash dir can be a private bind mount [08:54] we do something similar to /run/snapd/ns [08:54] it is privately shared [08:54] or we would not be able to bind mount mount namespaces there [08:54] I wrote a patch for linux that explains which of the EINVAL code paths is taken [08:54] but I did that for pivot_root [08:55] I think I wrote one for mount too but I cannot find it [08:55] the one for pivot is in the tree of snapd in snap-confine/misc [08:55] do you think the openNoFollow() logic will help cover some of jdstrand's concerns? [08:55] yeah [08:55] we have similar code like that in many places [08:56] it means we're working with paths that didn't contain symlinks at some point in time, and then only reference those directories going forward as "." after fchdir()ing into them [08:56] yeah, that is a nice improvement [08:58] * Chipaca isn't here [08:58] haha, hey John [09:03] Chipaca: so you have a day off too? just like zyga-ubuntu :) [09:03] mvo: it's not too terrible backporting #4712, it's slightly smoother if we backported #4715 as well , your call [09:03] PR #4712: release: version 2.31.1 [09:03] PR #4715: store: reorg auth refresh [09:04] pedronis: +1 for 4715 [09:04] pedronis: I get the feeling 2.32 will be a frank2.33 ;) but thats ok, its special circumstances so close to release [09:05] mvo: yea, anyway 4715 it's also a fix, we run out of stack (admittedly because of strange store bug) once [09:06] mvo: anyway as I said my maing argument for backporting these, is that if we get asked to do a close 2.32.x to bring in something to 18.04 it's better if the delta is small (with reason for safety) [09:09] pedronis: agreed [09:09] pedronis: I think this is fine [09:15] meh, everything is terrible - autopkgtest bionic/i386 fails in strange ways (/bin/bash: line 104: 12090 Killed snap connect ... [09:16] out-of-memory [09:17] * zyga-ubuntu hugs mvo [09:17] mborzecki: can you look at https://github.com/snapcore/snapd/pull/4860 - it should be a helper NO-OP change [09:17] PR #4860: spread,tests: move suite-level prepare/restore to central script [09:17] PR snapd#4861 opened: store: reorg auth refresh (2.32) [09:17] https://github.com/snapcore/snapd/pull/4851 is green and needs reviews now [09:17] PR #4851: cmd/snap-update-ns: fix creation of layout symlinks [09:17] zyga-ubuntu: my old friend lowmem_reserve again [09:18] what is lowmem_reserve? [09:22] PR snapd#4862 opened: store: cleanup test naming, dropping remoteRepo and UbuntuStore(Repository)? references (2.32) [09:23] mvo: I created backports in 4861 and 4862 [09:24] pedronis: thanks, I have a look [09:26] zyga-ubuntu: its the same problem we had before, let me try to find the papertrail. essentially the issue is that on i386 each time we load a new profile we eat up kernel memory and on i386 that is lowmem which is apparently not plentiful [09:26] https://forum.snapcraft.io/t/oom-for-interfaces-many-on-bionic-i386/4101/4 [09:29] Chipaca: could you look at #4829 and #4849 [09:29] PR #4829: store: Sections and WriteCatalogs need to strictly send device auth only if the device has a custom store [09:29] PR #4849: many: propagate contexts enough to be able to mark store operations done from the Ensure loop [09:33] Bug #1756285 opened: Gamepads only working in devmode [09:36] mborzecki: hah! i did it :-) [09:36] Chipaca: hmm? [09:37] mborzecki: … wait it's buggy [09:37] mborzecki: the string version of wrap1 [09:37] Chipaca: haha challenge accepted :) [09:37] Chipaca: i have something too, but it's wrapping a bit too far, idk why yet [09:37] mborzecki: i've done something very similar in my playing around with widths, so i know it's doable, it's just very very fiddly [09:38] mborzecki: I promise not to include fullwidth characters in my testing of your thing [09:39] Chipaca: i have this atm https://paste.ubuntu.com/p/HB9HzJzrHb/ but it wraps too far, probably missed something [09:40] mborzecki: can you pastebin the code, not the diff? [09:41] * Chipaca 's seen too many variations over the last 24h to cope with the diff atm [09:41] Chipaca: https://paste.ubuntu.com/p/FtxH8gPgsk/ [09:43] mborzecki: idx = strings.LastIndexFunc(text, unicode.IsSpace) [09:43] mborzecki: that's finding you the last space in the whole thing [09:43] mborzecki: not the last space that is <= width [09:43] damn, right [09:43] mborzecki: and... good luck because finding that one needs runes, with that approach [09:43] :-) [09:44] (because width is the number of runes(ish), not bytes) [09:44] Chipaca: yeah, i see what you meant [09:45] PR snapd#4798 closed: snap: don't create empty Change with "Hold" state on disconnect [09:45] i'll give it 15 minutes, if i cannot come up with something i'll leave it :P [09:45] mborzecki: https://pastebin.ubuntu.com/p/SMK5NKTMd6/ [09:45] mborzecki: that's where I'm at [09:45] mborzecki: i've still got a bug in there but it's close [09:45] pstolowski: if we need 4798 for 2.32, can you please create a PR that targets 2.32 (if not, just ignore me :) [09:46] Chipaca: truly worthy of /r/dailyprogrammer/ [09:46] heh [09:47] tis a nice problem though, it'd be great to have a TextFormatter that is an io.Writer [09:48] I'll push chipaca/forms sometime when i'm happy enough with it [09:48] mvo: it's na old bug. would be nice to have in 2.32. I'll prepare a PR [09:48] Chipaca: you +1 but not approved apparently [09:49] pedronis: yes [09:49] pedronis: I thought I understood the other pr until I saw the 'call snap find before testing stuff' thing :-) [09:49] * Chipaca needs to give it a better read [09:50] * Chipaca is having fun with code on his day off [09:51] Chipaca: we had a problem related to switching to use the fakstore before we have a macaroon before, this just make it more likely, snap find blah is a bit of a hack, but the real solution is annoying too (teach how to fake registration to the store just not to use it at all) [09:53] PR snapd#4849 closed: many: propagate contexts enough to be able to mark store operations done from the Ensure loop [09:53] Chipaca: if you ever saw strange 418 errors around getting a nonce from the fakestore that was the reason [09:56] Chipaca: am I talking gibberish ? :) [09:57] Chipaca: https://paste.ubuntu.com/p/ZRqDHcbJ9s/ [09:58] pedronis: no, that makes sense [09:58] pedronis: and although i haven't seen teapots, many of us have [09:58] pedronis: so thank you [09:59] Chipaca: :) this branch would increase the teapot spotting probability [10:00] before the "hack" [10:00] pedronis: thank you again [10:00] :) [10:01] mborzecki: youch [10:06] * Chipaca goes for a walk [10:33] PR snapd#4863 opened: Fix abbreviated disconnect 232 [10:33] mvo: ^ [10:34] pstolowski: thank oyu [10:34] you [10:53] mvo, already testint beta [10:53] testing [10:54] mvo, it should be ready today [10:54] cachio: excellent, thank you [10:54] yaw [10:58] zyga-ubuntu: hey, do you remember that issue that was causing the desktop interface font's not to be mounted on refreshs? Did you find anything related to that? As I'd like to promote a snap using it to stable, but I'm a bit worried that it could break things [11:31] mvo, 1 unit test is failing for bionic, how do you do to print stuff inside the unit tests or debug it? [11:33] seb128, I tried what we discussed in Budapest and moved the preseeded snap from filesystem.squashfs to the CD filesystem directly but there is not so much gain. Only 2s on a 77s boot. unsquashfs and snapd are still the main CPU and IO consumers. [11:33] cachio: which test is that? [11:33] seb128, it's on VM, I'll try on HW to compare [11:34] mvo, https://paste.ubuntu.com/p/7RRpWVCf3V/ [11:34] mvo, last test to fix I think [11:36] cachio: interessting. I can try this on my bionic system in a bit, otherwise I think a Commentf("Unexpected build date %s", snap.BuldDate()) " next tothe test [11:46] mvo: how are you today [11:48] Trevinho: hey, I remember the issue. I didn't manage to reproduce it. [11:48] although I found something in a related area [11:49] zyga-ubuntu: mh, I see.. I wasn't able to reproduce it either [11:49] zyga-ubuntu: switching from a snap without desktop-interface to one with it could suffice sometimes [11:49] I didn't investigate the other thing yet, I just found it yesterday [11:49] zyga-ubuntu: like switching from `telegram-desktop` stable to edge [11:50] but... Well, as I said it's not something happening all the times [11:52] interesting, I will setup a loop [11:52] and try to reproduce that [11:53] thank you for this advice, please hold over weekend ok? [11:53] I will keep looking and if I find something we can fix it or stop the update [11:54] mvo: +1 on merging 4851? [11:54] I think it's not everything, I saw one failure but I wasn't able to reproduce locally yet [11:57] mborzecki: would you be interested in looking at 4765 [11:57] it's not short though [11:57] so a bit of warning [12:01] hey, am using "snap try ./prime" and getting "Ensure prerequitites for "snap-name" are available *snap not found)" - any hint what that actually means [12:04] greyback: one of the default providers for your content interfaces is wrong [12:05] zyga-ubuntu: aha ok, that is a possibility [12:25] mvo, Unexpected build date 1969-12-31 23:59:59 [12:25] hehehe [12:25] a bit old [12:33] mvo: backports #4861 and #4862 are green now but need to be looked at/approved [12:33] PR #4861: store: reorg auth refresh (2.32) [12:33] PR #4862: store: cleanup test naming, dropping remoteRepo and UbuntuStore(Repository)? references (2.32) [13:13] * zyga-ubuntu runs all of main in a loop since morning and ... no failures [13:24] PR snapd#4864 opened: daemon: support 'system' as nickname of the core snap [13:37] mborzecki: https://pastebin.ubuntu.com/p/yPNh9YB2xQ/ [13:40] mborzecki: … and that's why I made it take a []rune :-) [13:50] Chipaca: are you updating the pr with this code? [13:51] mborzecki: should I? [13:51] mborzecki: is it better? [13:51] cachio: this looks like we did not get the new squashfs-tools in bionic-proposed in this test [13:51] it was fun to write :-) [13:51] aah ok ;) thought that you were planning to push it or sth [13:52] not of my own volition [13:52] i'm open to be convinced though :-) [13:52] mvo, ok, [13:53] mvo, is it ok if we add proposed until bionic release date? [13:53] * Chipaca wanders off to to not-work things [14:02] * zyga-ubuntu runs tests since 8 and no failures [14:04] PR snapd#4860 closed: spread,tests: move suite-level prepare/restore to central script [14:05] I swear this worked last week, but this week it fails (using layouts). Anyone mind having a quick look https://pastebin.ubuntu.com/p/vJHMh3DGxN/ [14:07] PR snapd#4851 closed: cmd/snap-update-ns: fix creation of layout symlinks [14:09] greyback: not sure if this is related [14:09] greyback: https://forum.snapcraft.io/t/layouts-re-mapping-snap-directories/1471/63?u=zyga [14:09] if you follow edge you must set a feature flag to enable layouts [14:09] zyga-ubuntu: if I create the directory in advance, it appears to work [14:10] yep, done that [14:10] other than that, not sure, I'm not really working today so I cannot help now [14:10] if you can drop me a note on IRC about how to reproduce that I will try [14:16] zyga-ubuntu: oh, sorry for bothering you! [14:23] No worries:-) [14:26] pstolowski: I did a pass over the interface hooks PR [14:27] pedronis: ty! [15:06] mvo, which is the package "squashfs-tools" to install and which version ? [15:11] cachio: hm, it looks like there is a new sync in bionic, let me check what is going on there [15:12] mvo, I ask because squashfs-tools deb package was not installed in bionic [15:12] cachio: what is the version installed in the vm? [15:12] mvo, there was not [15:12] cachio: that is strange, do you have /usr/bin/mksquashfs? [15:12] I installed 1:4.3-6 [15:12] cachio: in the image? [15:12] sorry [15:12] cachio: that looks correct [15:12] cachio: the version [15:13] it was installed 1:4.3-4ubuntu2 [15:13] I updated with 1:4.3-6 [15:13] is it ok [15:13] is it the right version? [15:15] yes, this is the right verson [15:16] mvo, ok, running again [15:16] tx [15:16] good luck [15:16] mvo beta validation is going very well btw [15:17] cachio: great to hear [15:40] zyga-ubuntu: hmm I'm not sure why dragonboard 410c never saw an update in the latest round, would need to check my logs. In any case, there is a new round of image updates needed for 16.04.4 [15:42] slangasek: that sounds good, when is that, roughly? when 18.04 gets released? [15:42] zyga-ubuntu: no, it should be done already (in line with the 16.04.4 point release that has happened), and I'll be looking at it next week [15:43] ah, perfect, thank you for looking into that === om26er_ is now known as om26er === icey_ is now known as icey === rumble is now known as grumble === apw_ is now known as apw === ahayzen_ is now known as ahayzen === chihchun_afk is now known as chihchun === sarnold_ is now known as sarnold === pstolowski is now known as pstolowski|afk === zyga-ubuntu is now known as zyga [18:25] I switched my debug refresh loop to use -reuse, in case that makes the magic bugs show up [19:10] hi all. I'm testing Xubuntu 18.04. Just installed skype from "software" (the xubuntu app to install things). Skype is a snap package. It installed. But it won't start. [19:10] The error in the log says: libGL error: unable to load driver: i965_dri.so [19:10] libGL error: driver pointer missing [19:11] Any advice on what could cause this, and what I can do about it? [19:21] PR snapd#4865 opened: many: propagate contexts enough to be able to mark store operations done from the Ensure loop (2.32) [19:36] PR snapcraft#2005 opened: pluginhandler: special case go patchelf failures for classic confinement [20:36] ah, found one more issue affecting symlinks [20:36] actually, some code that does nothing but hurt now [22:34] PR snapcraft#2006 opened: many: use packaging logic to get patchelf