=== JanC_ is now known as JanC [05:04] morning [05:05] good morning [05:06] zyga: hey [05:06] how was your weekend? [05:08] zyga: good, spent no more than an hour in front of the computer throughout the whole weekend [05:08] oh and i bought a bike, finally replacing my old one :) [05:08] mborzecki: I spent a few hours yesterday trying to put some life into and old netbook my mom gave me [05:08] ooh, nice [05:09] we managed to clean the garage to make room for the bikes [05:09] did a "garage sale" and let go some stuff [05:09] the bigger things (old kiddy beds) will likely be picked up this week so we will have all the room available [05:10] nice [05:10] btw. let me show you something [05:13] zyga: https://i.imgur.com/YM0lCsE.jpg [05:14] is that the one with folded keyboard? [05:14] ah [05:14] no [05:14] :) [05:14] no, unfortunately not [05:14] I saw one recently [05:14] what's inside? [05:14] asking price was ~30eur [05:14] pretty low, the cosmetics are nice [05:14] did you buy it? [05:15] nope, bought a road bike instead [05:15] :-) [05:16] I'm using sports tracker for my biking fun [05:16] e.g. this https://www.sports-tracker.com/workout/zygmuntkrynicki/5af879cec3d63959d624fc56 [05:16] got a used bike for ~540eur, with my frugal spending habits, another 30 felt like too much [05:18] zyga: looks like direct competition to endomondo [05:18] yeah, plenty of those apps [05:18] I've been using it for some time [05:19] it was working on my blackberry first :D [05:19] and it integrates with apple health now so I like it (it tracks my biking in the apple health app) [05:19] my wife uses endomodo as well but this one seems like what you get if you pay for endomodo [05:23] zyga: wonder if you can see this https://www.endomondo.com/users/20483467/workouts/1119982468 [05:24] yes [05:24] nice route :-) [05:25] that's the medium loop ;) got a couple of variants between 25-35km, the larger ones are 35-50, but those i can only do on weekends [05:30] do you go alone or with family? [05:31] for us the main limiting factor is how far the kids are willing to go [05:32] alone mostly [05:33] usually have like 1.5-2h of free time, so i can squeeze 25-35km loop [05:35] nice :) [05:36] we usually go on 20km loops at most, also going around Warsaw is not as fluid as there are many street lights and other choke points [05:36] I started going south from where we live, where there's fewer people and buildings but I haven't found a good route yet [05:37] my plan is to do a longer route along the river and then go back through "lasek kabacki" (a bit off to the west) so that the full loop would be close to 40km but we haven't done that y et [05:38] oh btw, I'm going to be off on Thursday, Friday, (this week) Monday, Tuesday and Wednesday (next week) [05:38] I need to file that [06:23] * zyga -> walk the dog [06:25] zyga: you entirely sure that the arrow does not go in the opposite direction? :) [06:25] haha [06:25] the truth has come out :) [06:31] Bug #1484898 opened: device tarball needs to allow setting sysctl defaults [06:49] zyga: another look at https://github.com/snapcore/snapd/pull/4504 when you're back? [06:49] PR #4504: snap, wrappers: systemd WatchdogSec support [06:49] Ack [06:49] Still walking but close to home now [06:49] Or I can just try now === pstolowski|afk is now known as pstolowski [07:07] mornings [07:07] pstolowski: hey [07:11] zyga: https://paste.ubuntu.com/p/XTPx5xJFVY/ [07:11] mborzecki: nice :) [07:12] is that the volume broken by your script? [07:12] yes [07:14] need to read more about the name checksums [07:32] github slow today? [07:37] mborzecki: I haven't noticed [07:37] mborzecki: I'm puzzled by something [07:37] I bet it is silly [07:37] if !(strings.HasPrefix(notifySocket, "/") || strings.HasPrefix(notifySocket, "@")) || strings.ContainsAny(notifySocket, `?*[]{}^"`) { [07:37] hm? [07:37] so a string must start with either / or @ [07:37] and must not have ? and other guys [07:37] right? [07:38] zyga: right [07:38] so [07:38] if s is the string [07:38] and s is "/?" [07:38] the first condition will be true, right? [07:39] so || will short-circuit and won't evaluate the rest [07:39] which is wrong [07:39] am I missing something? [07:39] if you add the test for special chars separately [07:39] zyga: the first condition will be false [07:40] wh? [07:40] why? [07:40] strings.HasPrefix(s, "/") [07:40] !(..) [07:40] that's separate [07:41] the condition inside the ( ) will be true [07:41] then the ! will negate [07:41] but that's separately evaluated [07:43] zyga: not sure what you're aiming at, you have !(true) || true for /? [07:43] oh boy [07:43] I'm blind [07:43] I didn't notice the () around the first two [07:43] zyga: or in more detail !(true || false) || true [07:43] I though it was around all three [07:44] I tweaked the code slighttly [07:44] maybe i should separate the cheks to make it more legible? [07:44] http://paste.ubuntu.com/p/zYzt3mW46f/ [07:44] is this okk? [07:44] I can push it [07:46] zyga: maybe s/AARE/illegal characters/ ? [07:46] (in the error message) [07:46] meta-characters? [07:46] invalid doesn't say much [07:47] yeaah, wrong but why? [07:47] hm [07:48] zyga: "cannot use %q as notify socket path: contains meta-characters" vs "...: contains illegal characters" [07:49] contains apparmor meta-characters [07:49] or contains one of "..." [07:49] that would be explicic [07:49] oh, the last one [07:50] zyga: `cannot use %q as notify socket path: contains one of ?*[]{}^"` [07:50] doing that [07:50] zyga: thanks! [07:59] moin moin [08:04] o/ [08:04] hey Chipaca, pstolowski [08:09] mborzecki: question about watchdog [08:09] does systemd enfroce any minimum values [08:09] mborzecki: what's the resolution there? seconds [08:09] ah, I see yes [08:09] is 1 second valid? [08:10] yeah, i think so [08:10] i mean, why wouldn't it? [08:10] I'm wondering if putting unreasonable value would be harmful to the system [08:10] if you feel like pinging every second and risking being restarted due to a missed ping, then so be it ;) [08:11] I just realized we should limit the number of apps in a snap [08:11] if we create a snap with 100M services [08:12] it would harm the system eqaisly === chihchun_afk is now known as chihchun [08:12] zyga: i think it's self regulating, if you put a value too low and miss it and malfunction you'll have the need to regulate it [08:12] zyga: the same with apps, i'd guess you'd bring down systemd if a snap has ridiculously large number of apps [08:13] s/apps/services [08:13] PR snapd#5154 closed: releases: merge 2.32.8 back into master [08:22] hmm [08:22] when selecting server snaps to install in subiquity, should the snaps actually be installed? [08:22] or should subiquity build a seed.yaml for the installed system? [08:24] pedronis: hi, this seems like something you would know? :) ^ [08:27] mwhudson: pedronis might not be with us yet [08:27] oh right [08:28] timezones how do they even [08:28] mwhudson: well, it's 10:30 in his tz [08:28] oh [08:28] heh [08:28] mwhudson: but it's a national holiday in the helvetic confederacy [08:29] ah [08:29] i guess i should email really [08:38] Chipaca: 'interdenominational-counterintelligences' lol [08:38] Trips off the tongue. [08:39] mwhudson: popey: and that's just unrepresentative-herpetology.~1~, the next version is going to be even better, i hear [08:41] I'd tell you what channel to get it from, but it won't fit in this margin [08:41] i wonder how badly this would break the snap list view i'm making for subiquity [08:42] mwhudson: answering that question is why I made it :-) [08:43] mwhudson: to be clear, package name, version, and summary are max length. Revision in particular is not, because I got lazy. [08:43] heh [08:43] what is the max revision length? [08:44] mwhudson: undocumented [08:44] len(str(sys.maxint))? [08:44] i don't know that the store team would be super happy with you uploading billions of reviions [08:44] mwhudson: probably log2((1<<31)-2) or something silly like that [08:45] mwhudson: exactly, too lazy to change country again because of being on the run from the store people [08:45] hahaha [08:48] ok now to make this stuff actually talk to snapd [08:49] mwhudson: what're you writing in? [08:49] Chipaca: python [08:50] anyone seen https://paste.ubuntu.com/p/wRXzv8vJQK/ ? [08:50] mwhudson: if you have requests, the http snap uses that + requests-unixsocket [08:50] Chipaca: that was what i had found with google [08:51] mwhudson: I'm sorry I'm only on par with, and not better than, a multi-billion corporation [08:51] Chipaca: i trust your recommendation a lot more! [08:52] * Chipaca wonders at the wisdom of that [08:54] Chipaca: I'm around now, is not a holiday for me (next Monday is) [08:54] hi! [08:55] need help understanding prime and stage [08:56] pedronis: ah! my bad :-) [08:56] pedronis: mwhudson had a question for you about becoming a farmer or something about seeds [08:57] ryan____: that's a bit vague :-) what do you need help with? [08:59] mwhudson: either is fine, depends what you can actually do, to install things the system needs to have been booted from the installed fs [09:00] pedronis: i have managed to 'boot' the installed system in a container [09:01] and then install snaps, even [09:01] understanding whats the difference in purpose between prime and stage. [09:03] ryan____: 'prime' is what the snap ends up as [09:03] mwhudson: so that boot seeds the real system I suppose, the issue there is at also starts the holding period for refreshes which we might want or not [09:03] ryan____: you can even install it using 'snap try' [09:03] What could cause snap build process to change the signature of a repackaged binary ? [09:03] case: https://github.com/snapcrafters/sublime-text/issues/16 [09:03] ryan____: it gets populated from 'stage', which is where parts install to [09:03] pedronis: hmm yes [09:04] mwhudson: I think both approaches are workable, it's a bit of a matter of what's the speed of one or the other and details like that one [09:04] also the speed of the first boot after install [09:05] om26er: differ from what? [09:05] it sounds like a small step from build to prime [09:05] Chipaca: from the tarball downloaded from the official release channel [09:06] zyga: any idea what's the status of https://github.com/snapcore/snapd/pull/4369 ? do you still plan to introduce the attribute? (i can look into that if needed) [09:06] PR #4369: add write permission to optical-drive interface [09:06] the binary sublime_text under this tar https://download.sublimetext.com/sublime_text_3_build_3176_x64.tar.bz2 is different from what we have in the snap, even though the source is the same. The upstream is concerned chipaca [09:07] om26er: snapcraft is probably patching the binary using patchelf [09:07] om26er: is this a classic snap? [09:07] Chipaca: yes, that's true [09:07] om26er: which is true [09:07] hmm, would need to add https://github.com/snapcrafters/android-studio/blob/master/snap/snapcraft.yaml#L31 ? [09:08] Chipaca: we need some kind of articulated response to the upstream on that one. [09:08] om26er: who is creating the snap? [09:08] Chipaca: we (snapcrafters) [09:09] om26er: so, you need to figure out what you're doing to the binary, and explain it to them === chihchun is now known as chihchun_afk [09:09] Wimpress: ^ [09:11] Wassup? [09:11] I kind of was not expecting the signature to change, so someone with more knowledge of why and how that is happening could produce a better response :) [09:11] popey: https://github.com/snapcrafters/sublime-text/issues/16 [09:11] om26er: if it's running patchelf, it's changing the binary; if it's changing the binary, its signature will change [09:12] and if patchelf is causing crashes on linux, it's doing it wrong [09:12] you can disable patchelf [09:13] mborzecki: that is blocked on a decision by gustavo [09:13] om26er: use the build-attributes -no-patchelf [09:13] sorry, I was afk for a moment, we're giving away baby gear and a friend came over for a summer kiddy bed [09:13] om26er: see android-studio [09:13] popey: sure, we did that for Android Studio a bit ago, it was not starting at all. [09:13] yeah [09:14] I will do a PR but would be better if you could write on that issue popey === chihchun_afk is now known as chihchun [09:20] PR snapd#5134 closed: Shrink image generated with snap prepare [09:26] om26er: ok leave it with me [09:28] popey: resurrecting #1739097 if you could give it a read [09:28] Bug #1739097: "This leaves tracking edge." [09:28] I do have a PR though https://github.com/snapcrafters/sublime-text/pull/17 but I still am not able to build it because snapcraft eats 100 cpu during build and never gets further. [09:28] PR snapcrafters/sublime-text#17: Dont use patchelf [09:30] om26er: I'd suggest checksumming that before declaring it fixed [09:30] om26er: also, dude, if you don't explain _why_ you're doing a change, how do you remember why you did things a year from now? [09:33] om26er: hang fire for now. I'm building and testing locally [09:33] I'll respond to upstream [09:34] popey: yep, that would make sense. [09:35] Chipaca: I added some description now, thanks for the reminder. [09:40] PR snapd#5134 opened: Shrink image generated with snap prepare [09:51] pstolowski: is https://github.com/snapcore/snapd/pull/4968 still blocked? [09:51] PR #4968: ifacemgr: remove stale connections on startup [09:52] mborzecki: nope, let me merge master back [09:52] ok [09:53] looks like google:ubuntu-16.04-32:project is failing, is the image broken? i'm suspecting the apt lock file was left behind [10:07] heh, when starting manually project prepare on ubuntu-16.04-32 works just fine [10:53] mborzecki: maybe unattended updates? [10:59] * zyga needs more coffee [11:09] zyga, did cachio finally get the f28 google images up? [11:09] I haven't touched that, I will ask cachio when he shows up [11:09] I had to spin new packages yesterday due to failing scriptlet: https://copr.fedorainfracloud.org/coprs/ngompa/gce-oslogin/build/753629/ [11:09] thank you again for preparing that package [11:10] my main focus this week is to go through the reviews [11:11] gce-oslogin was a crappy package [11:11] but I fixed it: https://copr-be.cloud.fedoraproject.org/results/ngompa/gce-oslogin/fedora-rawhide-x86_64/00753629-google-compute-engine-oslogin/google-compute-engine-oslogin.spec [11:12] Son_Goku: I'm not sure how google handles that, can you send your improvements back? [11:12] they require a CLA [11:12] and I hate signing those [11:13] I actually more or less rewrote the spec [11:13] so I should have dropped the ASL license header in the file [11:14] I probably will drop it actually [11:14] since my spec file shares ~2-4 lines with the original [11:14] and that's it [11:23] Hi snappy people, I have a deb that I should plave in a snap. Normaly that would go into the staged-packages but that specific deb is from a non-bleshed repository, what can I do? [11:24] *place [11:29] kjackal: you could add it as a part, plugin type dump, source, url to the deb. [11:30] thanks popey, do you happen to have an example yaml for that? [11:33] popey: something like this: https://github.com/snapcrafters/mattermost-desktop/blob/master/snap/snapcraft.yaml I guess [11:33] not off the top of my head, no. [11:48] hmm [11:48] * zyga is stuck with a test failure on code he is hacking [11:51] * cachio afk [11:55] * Chipaca lunch [11:57] pstolowski: I asked a question about #4767 in your latest forum topic [11:57] PR #4767: interfaces: disconnect hooks

[11:57] pedronis: i saw it, thanks [12:33] zyga: https://paste.ubuntu.com/p/ydn2KZBTHV/ [12:34] nice [12:34] so that's exactly what we tarballed [12:34] but created programmatically, right? [12:35] -rwxr-xr-x 1 root root 8 2018-05-14 uboot.env [12:35] -rwxr-xr-x 1 root root 8 2018-05-14 uboot.env [12:35] yes [12:36] zyga: i think that core fssck is to blame [12:37] I afgree [12:37] it looks like a bug, clearly, there [12:37] when i use fsck.vfat from host the entries are looking good [12:38] hmm [12:38] hold on, and which fsck are we using? [12:39] dang, hsot fsck does the same too [12:40] zyga: well, the one from core right? [12:41] mborzecki: zyga: that's what mvo's patch to dosfstools fixes [12:41] https://github.com/dosfstools/dosfstools/pull/83 [12:41] PR dosfstools/dosfstools#83: [RFS] check.c: do lfn_remove on auto_rename() [12:41] anyways, fsck 4.1 and 3.0.28 (from core snap) appear to do the same, i end up with 2 entries, one has only shortname (uboot.env) and the other shortname (fsck0000.000) and lfn (uboot.env) [12:42] Chipaca: right [12:42] if upstream had at least ACKed the issue we could distropatch it [12:42] if they don't, we'd have to be a lot more sure about it fixing the core issue [12:42] once mounted this appears as 2 files with identical names ;) [12:42] (should kernel at least display it differently?) [12:43] mborzecki: there are options to mount to control that [12:43] mborzecki: i think the one mvo suggested was win95 [12:43] (if you look for win95 in man mount you'll see the options) [12:43] aah, it was in mvo's change, let me try that, i'm using check=s right now [12:43] try dosbox [12:44] -rwxr-xr-x 1 root root 8 2018-05-14 uboot.env [12:44] -rwxr-xr-x 1 root root 8 2018-05-14 UBOOT.ENV [12:45] all ups is the one without lfn [12:45] this way at least we can point to one or t'other :-) [12:51] PR snapd#5155 opened: interfaces/apparmor: use strict template on openSUSE tumbleweed [12:51] mborzecki: so on dosbox things work better? [12:52] zyga: you can tell which file is which [12:52] hmm [12:52] can we mount things like that in linux? [12:52] mborzecki, Chipaca: can we land https://github.com/snapcore/snapd/pull/5142 ? [12:52] PR #5142: many: add "snap debug sandbox-features" and needed bits [12:53] zyga: why squash-merge? [12:53] junk in history [12:53] zyga: that usually means you want to backport it [12:53] and that concerns me :-) [12:53] s/backport/cherrypick/ [12:53] no, I don't need that :") [12:54] my main motivation is to use that branch [12:54] to filter tests better once we do the opensuse experiment [12:54] cachio: would it be hard to add opensuse tumbleweed to CI? [12:55] cachio: even as a manually used image [12:55] I could definitely use it [12:58] yup TW makes sense [12:59] sort of early upgrade of the image is there for arch already, tw could do the same [12:59] although iirc opensuse can do more than 1 kernel installed [13:03] Chipaca, cachio: Coming? [13:03] niemeyer: trying to [13:03] :) [13:07] mborzecki, Fedora and openSUSE support multiversioned kernel packages [13:07] Son_Goku: thanks, fedora i know about, but i didn't remember how suse does it [13:07] same exact way [13:08] well, at least some drstros as sane ;) [13:08] pretty much all RPM-based distributions use multiversion/slotted kernel packages [13:31] pstolowski: feedback con 5032 [13:33] zyga: thanks, looking [13:34] zyga: ouch, thanks, mistake in conflict resolution [13:34] W: Target Translations (main/i18n/Translation-en_US) is configured multiple times in /etc/apt/sources.list.d/google-cloud-sdk.list:1 and /etc/apt/sources.list.d/google-cloud-sdk.list:2 [13:34] apart from apt lock being held [13:35] * zyga -> lunch [13:35] https://www.youtube.com/watch?time_continue=1&v=F8nrpe0XWRg <- interesting, so far watched about 1/3rd though [14:03] off to pick up the kids [14:08] pstolowski: I did a first pass on #4767 , have also some questions there [14:08] PR #4767: interfaces: disconnect hooks

[14:09] pedronis: ty, checking [15:05] zyga: #4889 needs a j-dstrand review ? [15:05] PR #4889: cmd/snap-update-ns: don't trespass on host filesystem === cachio_ is now known as cachio [15:46] pedronis: not sure, I think it doesn't but we can add one [15:46] pedronis: jamie is a bit busy lately so I think I will wait for a +1 from gustavo first [15:54] zyga, pedronis FYI issue on xenial 32bits fixed === chihchun is now known as chihchun_afk === pstolowski is now known as pstolowski|afk [16:13] cachio: what was the issue? [16:17] zyga, sometimes when the execution was done from a fast machine [16:17] zyga, the system tried to do a release upgrade because it is a new lts [16:18] zyga, and it was locking [16:18] I updated the file /etc/update-manager/release-upgrades [16:18] and the problem is gone [16:20] thanks! [16:20] that is great [16:21] I could reproduce it from canonistack [16:21] zyga, imposible from here [16:34] is there a way to unrelease a revision from a specific track using the command line? [16:35] we can't use the web UI because the track in question doesn't appear in the web UI [16:35] kenvandine: you can either release a different revision to the channel or close it [16:39] noise][, ah, then they can release a revision to that track again later? [16:41] this is for the esr channel from mozilla, they want to test releasing the snap to esr60/stable then unpublish it until they are sure [16:42] noise][, although i'm not sure why :) [16:50] kenvandine: yes, they can release to that track/risk and then close it again [16:52] noise][, thx [16:54] * cachio afk [16:59] ttfn peeps [17:08]