=== chihchun_afk is now known as chihchun [03:51] PR snapcraft#2132 closed: errors: generic exception for common.run[_output] === chihchun is now known as chihchun_afk [04:03] hi [04:03] How might I be able to deploy snaps without going to the snap store? [04:04] Like I have built snaps, but I'd like to distribute it only to specific users whom should be able to download it [04:57] morning [05:15] much fun when the gnome session dies out of the blue [05:21] Hey [05:29] zyga: hey [05:29] zyga: so today we merge everything that's green right? [05:32] omg, today must be some gnome shell love day [05:33] hey mborzecki [05:34] mborzecki: just saw your helper to corrupt an image, nice! [05:34] mvo: hey, yeah took some time, fat is super annoying to parse [05:34] all that legacy cruft :/ [05:36] mborzecki: I'm totally with you here :) [05:38] mborzecki: I just verified that https://github.com/dosfstools/dosfstools/pull/83 works with the corrupted image created from your script. this is really nice [05:38] PR dosfstools/dosfstools#83: [RFS] check.c: do lfn_remove on auto_rename() [05:38] yay [05:39] mvo: any word from maintainers of dosfstools? [05:44] mborzecki: unfortunatly not, I will try a debian bugreport next, maybe he reads those [05:48] mborzecki: Yes, that is the plan [05:48] zyga: good morning! [05:48] zyga, mborzecki what did I miss fri/mon ? where are we standing? [05:49] Man, it is so cold today [05:49] Mvo, nothing major. Just lots of catch up with PRs [05:49] zyga: cool [05:49] I have a question about [05:49] Your fixes for 8 [05:50] There is a new service there [05:50] zyga: sure [05:50] But there are no changes to packaging [05:50] zyga: you mean core-fixup? [05:50] Yes [05:50] zyga: its an existing service [05:50] zyga: I just extended it [05:50] Ahh [05:50] Ok [05:50] I must have mistress [05:50] Misread [05:50] zyga: conveniently we had an issue before [05:50] zyga: lol@mistress [05:50] I did a compare against 6 [05:50] Because 7 didn’t have a tarball [05:51] Yeah, phone is fun ;-) [05:51] zyga: yeah .7 was very short lived, sorry for that [05:51] Taking the dog out [05:51] zyga: I should have waited a bit longer [05:51] zyga: heh, enjoy! [05:51] No worries, just wanted to double check [05:51] zyga: thank you for this [05:52] zyga: I will go over the PRs now and see what comments need addressing, iirc I left some small ones so that I don't have to wait for tests. now is a good time to address them [06:01] I need to work from downtown today [06:01] Need to go to a government office for some paperwork [06:02] uhh [06:02] zyga: good luck [06:02] reminds me i need to replace the my car registration [06:03] btw. i'll have a longer break ~1030, my daughter has an athletic tournament at her school, going there to cheer for her [06:03] Woot, nice :-) [06:04] I will be home soon, grab my gear and commute [06:13] PR snapd#5156 opened: core-fixup: address review feedback from #5147 === chihchun_afk is now known as chihchun === pstolowski|afk is now known as pstolowski [07:00] morning [07:02] pstolowski: hey === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [07:24] #5156 travis job timed out, restarted it [07:24] PR #5156: core-fixup: address review feedback from #5147 [07:26] * zyga is on a bus, commuting feels weird now [07:26] Weird and inefficient [07:27] zyga: no tram lines going near your destination? [07:28] No, not even close [07:28] Closest tram is the subway [07:32] re [07:32] I specifically took the slower bus to be able to sit and work on the way :) [07:33] at this time the rush hour is over and bus is rather empty [07:37] oh drat [07:37] no 2fa :( [07:37] well [07:42] mvo: hi [07:43] mvo: seems autopkgtest tests sometimes fail over a flaky snapstate test [07:43] pedronis: good morning! do you have a link with the particular failure? [07:44] mvo: https://launchpadlibrarian.net/370203382/buildlog_ubuntu-xenial-armhf.snapd_2.32.8+git720.9dc7b89~ubuntu16.04.1_BUILDING.txt.gz [07:44] it's one of the prereq tests [07:44] pedronis: aha, looks like the autopkgtest vms are very slow [07:45] pedronis: I have a look and will improve [07:50] any PRs that need attention? [07:51] mborzecki: #4844 is now green [07:51] PR #4844: overlord/snapstate: allow core defaults configuration via 'system' key [07:52] pedronis: merge or do you want to have one last look? [07:52] I approved it [07:52] very well, merging then :) [07:52] PR snapd#4844 closed: overlord/snapstate: allow core defaults configuration via 'system' key [07:53] pedronis: would you like to take a look at https://github.com/snapcore/snapd/pull/5091 too? you were the last to modify the auto refresh code and know ins and outs of this [07:53] PR #5091: many: hold refresh when on metered connections [07:54] will do in a little bit [07:56] pedronis: thanks! [08:03] does anyone know where the gnome-system-monitor snap source is? I want to investigate to see if anything special is done for localisations [08:03] zyga: https://paste.ubuntu.com/p/wSN7fZCn6Q/ nice! [08:12] mborzecki: yeah, now just convince people to enable apparmor [08:12] hahaha [08:15] zyga: fwiw there's both apparmor and selinux enabled kernel packages in aur, plus a bunch of userland packages in respective flavors, still bugs me why would one want to rebuild that much of the base system and not just use some other distro [08:18] PR snapd#5157 opened: many: improve `snap wait` command [08:21] afk for 1-2h [08:22] * zyga arrived close to the office [08:22] Bug #1771299 opened: snap is a total failure, drop it [08:22] need to check where to deliver the papers [08:22] but I can enjoy a cup of coffee and work some more from a restaurant now [08:29] this is a delightful morning message: https://github.com/snapcrafters/gimp/issues/27 [08:32] diddledan: https://giphy.com/embed/nPTXD4lJmDSTe [08:32] haha [08:32] gotta love crab ideas [08:33] yeah [08:33] crab is good cooked [08:33] ;-) [08:33] diddledan: also, https://giphy.com/embed/26mfgW5tgmbV7wkUg [08:33] I love that I've seen this guy on various forums [08:33] spewing toxic waste from his mouth [08:33] zyga: s/his/their/ [08:34] PR snapd#5158 opened: boot,partition: improve tests/docs around SetNextBoot() [08:34] also the LP issue about crabs [08:34] Bug #1771299 changed: snap is a total failure, drop it [08:34] sorry for all those PRs, its leftovers from the madness last week, hopefully super trivial to review [08:35] zyga: if you go to the github's user's homepage, it's illuminating [08:35] oh? [08:35] * zyga looks [08:35] also, they're german. So it's mvo's fault. [08:35] Chipaca: what's illuminating? [08:35] hm, yeah, some crabs are german [08:35] the crab picture :D [08:36] mvo: crabs are an international exterretorial nation ;) [08:37] zyga: to my prejudiced eyes, that's a website created by an agry 13-year-old [08:37] angry, also [08:37] https://translate.google.com/translate?sl=de&tl=en&js=y&prev=_t&hl=pl&ie=UTF-8&u=http%3A%2F%2Fschwender-beyer.de%2F&edit-text=&act=url [08:37] I love the icons all crossed-out [08:37] well [08:37] * zyga gets back to work [08:37] [08:39] zyga: 🦀̸ [08:40] Chipaca: what website is this? [08:40] where is no-space combining strike-through glyph [08:40] Chipaca: I just saw the bugreport? [08:40] mvo: starting from this *other* bug report by the same crabby person: https://github.com/snapcrafters/gimp/issues/27 [08:40] zyga: U+0338 [08:41] zyga: all combinings are zero-width, fwiw [08:41] Chipaca: this bug report is zero-value, all crab [08:41] zyga: https://www.youtube.com/watch?v=PhM4VQFkw04 [08:42] * zyga looked at twitter and ... [08:42] https://twitter.com/nixcraft/status/996283804527378432 [08:42] zyga: fwiw i'd've set it to 'opinion' more than 'won't fix' [08:43] Chipaca: thanks, though I doubt we will fix this guy [08:43] zyga: "won't fix" means you think they have a point :-) [08:43] I think they lost the point a long time ago :) [08:43] mvo: ^ that's the vfat issue [08:43] zyga: "opinion" might be best, otoh its fine, one open bug is enough [08:44] mvo: rationale* [08:44] :-) [08:44] zyga: this is briliant [08:44] Chipaca: *cough* thank you [08:45] Chipaca: I should have written that in GH, then I could edit it ;) [08:45] mvo: ikr [08:45] github is doubleplusgood [08:45] zyga: thanks for your comments in the PRs \o/ [08:45] mvo: I'm going through it [08:45] I was thinking if it would be better to use present tesnse [08:45] rather than snapd will do this and that [08:46] I can propose one on top once this lands [08:46] cachio: hey, good morning! how is the QA for .8 looking? I heard there might be issues ? [08:46] zyga: sure [08:47] * diddledan spies popey is awake [08:49] mvo: https://github.com/snapcore/snapd/pull/5158/files#r188209242 [08:49] PR #5158: boot,partition: improve tests/docs around SetNextBoot() [08:50] https://github.com/anordal/shellharden/blob/master/how_to_do_things_safely_in_bash.md [08:50] ^ ^ ^ interesting read [09:09] hey zyga, just read you did some cuda tinkering with snaps, is any of your work public somewhere? [09:09] kjackal: hey [09:09] kjackal: I didn't get anything operational [09:09] kjackal: unfortunately I no longer have nvidia hardware [09:10] thats fine, I am looking for some guidance [09:10] (or one that works with current cuda) [09:10] How would cuda integration work? [09:10] kjackal: cuda is just some userspace stuff really [09:10] kjackal: the opengl interface today should give you access to nvidia character devices [09:10] How did you manage the cuda-drivers installation? [09:11] kjackal: make a cuda hello world [09:11] kjackal: that was long ago, I don't recall that anymore [09:11] kjackal: my advice is as follows: [09:11] kjackal: 1) build the simplest cuda program you can [09:11] kjackal: 2) ldd it, print that out [09:11] kjackal: 3) snap it manually, use the opengl interface [09:11] kjackal: 4) run it and see what happens [09:12] kjackal: put that on a forum and CC mborzeck1 [09:12] he has the recent hardware and should be able to help you out [09:12] mvo: when you have a moment, can we talk about #1739097? [09:12] Bug #1739097: "This leaves tracking edge." [09:12] we correctly put all the things inside the sandbox now [09:12] so the only possibly missing thin is a userspace library bit [09:12] we should copy all of those but nobody tested that [09:13] zyga: I am not sure I can follow that path, I need to make the nvidia-docker working :( [09:13] kjackal: nvidia-docker? [09:13] how is that related [09:14] E: Could not get lock /var/lib/apt/lists/lock - open (11: Resource temporarily unavailable) [09:14] E: Unable to lock directory /var/lib/apt/lists/ [09:14] wasn't this fixed yesterday ^ ? [09:14] mvo: please fetch origin and rebase your latest PRs [09:14] kjackal: I don't know anything about nvidia-docker [09:14] zyga: nvidia-docker needs the cuda drivers installed on the host machine [09:14] kjackal: I can help you with snaps, not docker [09:15] sure, forget anything I said about docker [09:15] the question is how do I deploy the cuda-drivers deb package ? [09:15] I don't know [09:16] do you mean install? [09:16] I assumed you just install that and that's it [09:16] apt-get install cuda-drivers [09:18] zyga, I hope the missing link is the cuda-drivers, the rest of the stack seems to be working, but I will know if i get cuda working [09:18] zyga: I get this error: /snap/microk8s/x1/usr/bin/nvidia-container-cli --load-kmods configure --ldconfig=@/sbin/ldconfig.real --device=all [09:18] --utility [09:18] --pid=3426 /var/snap/microk8s/x1/lib/docker/overlay2/8e97d76a3e4a0b2d05f11d4744b805ab91bad8caa62c39a942d2f110f9fc7d68/merged]\\\\nnvidia-container-cli: [09:18] initialization error: load library failed: libnvidia-fatbinaryloader.so.390.30 [09:19] I don't know what this is or what you are doing [09:19] are you running inside docker inside a snap? [09:19] can you rewind to just plain system with nvidia and cuda working and move from there [09:20] lets forget about docker, I have a binary inside my snap that needs to load libnvidia-fatbinaryloader.so.390.30 that comes with cuda-drivers [09:20] do you get any denials? [09:20] where is the library relative to the root of the snap [09:21] I am in classic confinement [09:21] oh [09:21] that's even more messy then [09:21] :) it is [09:22] Chipaca: sure, didn't we fix the message a while ago? [09:22] mvo: that's why we're having this conversation :-) [09:22] zyga: I can rebase/repush, I thought my master was up-to-date, let me double check [09:22] mvo: it was fixed to be unambiguous, yes [09:22] mvo: unambiguously wrong :-D [09:22] mvo: then the fix for apt unattended upgrades is wrong [09:22] I was thinking that during the install hook I could apt-install cuda-drivers and then update the LD_LIBRARY_PATH of my wrapper. But this does not sound right. zyga? [09:23] mvo: sorry i missed this, to be fair you did tag me and i missed it [09:23] kjackal: that's all crazy IMO [09:23] kjackal: if your snap's hook uses apt it is not a good snap IMO [09:23] Chipaca: oh, so this means "Snap %q is *now* tracking %v"? [09:23] Chipaca: or in what way is it wrong? [09:23] kjackal: can you please go back [09:23] mvo: the message is printed when the snap is tracking channel X, but the revision is from channel Y [09:23] kjackal: run a hello-world cuda app outside all magic [09:23] kjackal: share that [09:23] kjackal: and let's iterate [09:24] zyga: what fix was that? did it happen in one of the recent PRs? [09:24] kjackal: can you get that part to work? [09:24] mvo: I don't know, cachio fixed it [09:24] zyga: ok [09:24] what would that giveme zyga? What would be the next steps? [09:24] Chipaca: what will happen in this case? the snap keeps tracking channel X ? [09:24] kjackal: that would give you a binary [09:24] mvo: the snap is tracking channel X, yes [09:24] kjackal: you can inspect that binary, strace it [09:24] see what it loads [09:24] kjackal: see what it needs to operate [09:25] kjackal: then you can put that binary (built binary) into a snap you made [09:25] kjackal: run it there [09:25] see what braks [09:25] mvo: if a revision is released to channel X it'll grab that, it's only getting things from Y because X is closed [09:25] *breaks [09:25] see why, iterate [09:25] eventually you will be able to run that built binary inside your snap [09:25] kjackal: I suspect you will do more things after that but before this is achieved I don't know what we are doing [09:25] Chipaca: ok, I see. sorry for this [09:26] mvo: it's not a very clear corner case [09:26] mvo: in the bug report i suggest "this revision is from channel %s; we'll continue to check channel %s for updates" [09:26] mvo: but i don't know if that is clear [09:26] also the royal we is probably not what we want [09:27] PR snapd#5159 opened: testutil: record system call errors / return values [09:27] Chipaca: yeah, the message looks like the right direction though [09:27] https://github.com/snapcore/snapd/pull/5159 is simple, will let me build the tests that jdstrand requested [09:27] PR #5159: testutil: record system call errors / return values [09:27] anyone wants to look quickly? [09:27] popey: you're good with words, sometimes, can you think of a better way of saying this? [09:27] * Chipaca only used "sometimes" to not trigger the "aw shucks" response [09:28] * zyga goes to the goverment office to do paperwork, ttyl [09:28] kjackal: I'll be connected but I won't reply for a while, let's keep chatting, I find this topic interesting (if only I had correct HW) [09:30] zyga: kjackal: for extra fun, you can run cuda stuff on PRIME hardware when the intel card has the con [09:31] :) [09:31] zyga: and that might royally mess up the detection stuff [09:31] dunno [09:31] Chipaca: hello. what words in particular, and where? [09:31] popey: when you 'snap install --edge foo' but edge is closed right now [09:32] popey: so you get it from, say, beta instead [09:32] popey: right now we print something that's incorrect, so looking at fixing that [09:32] popey: so far best attempt is "this revision is from channel %s; we'll continue to check channel %s for updates" [09:32] popey: but we don't use the 'we' elsewhere :-) [09:32] "$channel for $snap is closed, forwarding to $channel+1" [09:33] popey: that does not convey the 'as soon as something is pushed to edge you'll get it instead of this beta crud' [09:33] you will? [09:33] I thought you get bumped to the next channel down and stay there [09:33] that's what tracking means [09:33] nope [09:33] like, if you were on beta, and I close it, you're now in stable [09:34] if I re-open beta, you flip back to beta? [09:34] yes [09:34] wow [09:34] who understands why gimp isn't picking up localisations? because I sure as chocolate don't understand it [09:34] this is not well known [09:34] popey: not even the whole snapd team was aware of it :-) [09:34] loooolz [09:34] I found this post https://forum.snapcraft.io/t/lack-of-compiled-locales-breaks-gettext-based-localisation/3758/14 but that claims it's fixed [09:35] diddledan: doesn't everybody speak esperanto already? [09:35] if they speak esperanto they'll not get the UI in their language, instead it'll be en_US [09:35] diddledan: i thought en_US was esperanto 2.0 [09:40] zyga: what will we do with the strace of the cuda program? [09:40] popey: "$trackingChannel for $snap is closed, temporarily forwarding to $snapChannel"? [09:41] better [09:41] k [09:41] We will be able to identify libraries needed for a cuda program to function, right? Then what? [09:43] PR snapd#5160 opened: snapstate: make TestDoPrereqRetryWhenBaseInFlight less brittle [09:46] kjackal: look at the ser of files being accessed [09:46] kjackal: then we check if current nvidia integration handles all of them [09:47] Moins [09:47] Hey [09:47] * zyga waits in line to file paperwork, 60 people in front, ouch [09:50] "current nvidia integration" ? [09:51] 'support all snap apps, including "classic" apps' what are classic apps? [09:51] pedronis: the above pr is to make the issue in the handlers_prereq_test.go:247: more robust [09:52] kjackal: we have special code that bridges nvidia driver to snap space [09:52] kjackal: but it doesn’t apply to classic confinement [09:53] Riddell: classic confinement is the way we describe snaps that run without a special mount namespace and with little to no confinement [09:53] Like typical classic packages [09:53] Hence the name [09:53] Riddell: i don't know what classic apps are in that context; classic snaps are what zyga just said [09:54] I'm writing the KDE Plasma 5.13 and this is what I have for Plasma Discover changes 'Snap support now follows Snap application permissions and it's possible to use classic Snap formats.' [09:55] Riddell: I literally can't even :-) [09:55] Riddell: you're going to have to ask the author for clarification i'm afraid [09:56] i don't know what "classic Snap formats" are, and I can only guess at what "Snap application permissions" are [09:57] Those are probably worded incorrectly [09:57] yes [09:57] They should be “snaps using classic confinement” and “snap interfaces” [09:57] Not permissions [09:58] zyga: _if_ that's what they meant [09:58] I guess I fall back to "refusing to guess" too easily (but it's saved me in the pasts) [09:58] past* [09:58] Well, if they meant something else I’d like to see those patches b cause we don’t have such features [09:59] man, i wish you could toggle a spread run to debug [10:04] PR snapd#5161 opened: cmd/snap: fix the message when snap.channel != snap.tracking [10:09] Chipaca reviewed [10:09] Chipaca: could you look at #5077 ? [10:09] PR #5077: overlord/snapstate,overlord/auth,store: coalesce no auth user refresh requests [10:10] zyga: the truncated bit is because the fake store replies with "" for tracking channel [10:10] zyga: (there's an ad hoc comment on the test) [10:10] Ah [10:11] zyga: but you're right about the i18n thing, will fix [10:11] Thanks [10:11] zyga: also, will reply there [10:11] why am i doing it here [10:11] Ack :-) [10:11] just to make your life harder [10:16] pedronis: nice PR [10:18] Yeah, reading too [10:20] zyga: thanks changed to 'Snap support now follows Snap interfaces and it's possible to use the Snaps using classic confinement.' [10:21] What does “follows” here mean? [10:21] Otherwise it reads ok [10:21] good question [10:22] 'uses Snap interfaces'? 'allows Snap interfaces'? [10:22] Riddell: is this talking about a GUI? [10:23] Chipaca: yes [10:23] does it mean that it now exposes toggles to connect interfaces? [10:23] Chipaca: that sounds likely [10:23] toggle is a real term? I love that [10:23] if so, then maybe saying that ("now allows toggling interface connections") might be preferably, although i might be too deep in the argot [10:24] Riddell: it's a verb and a noun :-) [10:24] you just need a place name and you can go all 'buffalo buffalo buffalo' on it [10:25] preferable* [10:25] my language is shot (toady?) :-/ [10:26] thanks, I've used that in https://www.kde.org/announcements/plasma-5.12.90.php#discover [10:27] Riddell: nice [10:28] Thanks Riddell [10:28]