[00:46] PR snapcraft#2877 closed: [3.9 backport] meta: handle plug & slot string objects [00:46] PR snapcraft#2879 closed: [3.9-backport] meta: ensure Snap's `assumes` is initialized as a set [02:43] PR snapcraft#2881 opened: meta: fix missing provider case for get_provider_content_directories() [06:29] morning [06:32] hey mborzecki [06:33] mvo: hey [06:40] PR snapd#7998 closed: httputil: use shorter timeout in TestRetryRequestTimeoutHandling [06:52] o/ [06:52] mborzecki hey [06:52] I will be back shortly, the office is still super cold after the night [06:52] zyga-laptop: hey [07:06] I need to catch up with things from yesterday [07:08] oh, iCloud is back! [07:08] ircloud [07:17] * zyga builds 2.43.1 update [07:24] mborzecki: we have a new dependency on Suse? [07:24] we need fake root for tests [07:24] zyga: hm? [07:26] mborzecki: without fake root dependency [07:26] https://www.irccloud.com/pastebin/08hHDzfx/ [07:26] I'll send a patch shortly [07:28] https://github.com/snapcore/snapd/pull/8009 [07:28] zyga: yeah, we shuld porbably skip those tests [07:29] PR #8009: packaging/opensuse: build-depend on fakeroot [07:29] PR snapd#8009 opened: packaging/opensuse: build-depend on fakeroot [07:31] zyga: otoh, fakeroot is mocked in those tests, the test will fail if it's called and not mocked by some useful code [07:32] hmm? [07:32] what? [07:32] zyga: 2 lines up from the first point it failed [07:33] zyga: https://github.com/snapcore/snapd/blob/master/gadget/mkfs_test.go#L60 [07:34] dunno [07:35] zyga: was this on obs? [07:36] no, in "osc build" [07:37] zyga: can you run those tests with -check.work and inspect the test dir afterwards? [07:37] mborzecki: in osc build or manually? [07:39] (in either case, later, updating TW now) [07:46] I'll grab some food, see you after breakfast [08:02] morning [08:02] mvo: side note, does it make sense to start console-conf in install mode at all? [08:09] mborzecki: it does not [08:09] pstolowski: good morning [08:09] pstolowski: hey ;) [08:10] o/ [08:14] good morning, [08:14] zyga: yesterday, i solved it. i found in glibc, that dlvsym is like dlsym with version equal null. So i add a PR for snapcraft-preload. And it works for sway :-) [08:19] mborzecki: do you remember if we had a forum top to the parallel seccomp setup? [08:21] mvo: let me check [08:22] mborzecki: no need to create it if it's not there, I'm just updating the snapd roadmap with interessting bits of 2.43 [08:22] mvo: no, i don't think we had a topic for that [08:23] that's fine, thanks for checking [08:24] hm wonder whether we could build an image with systemd.debug-shell=1 out of the box [08:41] back now [08:41] sdhd-sascha: cool, I'm glad you solved it :) [08:41] hey mvo :) [08:41] hey pawel :) [08:42] zyga: thank you. Just talk about the problem with you, was a great help to choose a solution ;-) [08:42] mvo: I published the render code and will try to sync with chipaca about it [08:42] mvo: I need to send some tweaks that I thought about while sleeping but it's mostly enough what I need for explain [08:43] mvo: I published 2.43.1 in suse, shall I do fedora/debian before jumping into regular PR work? [08:44] zyga: your implementation of RuneWidth is rather optimistic :) [08:44] Chipaca: I was sure you'd be interested in that one :) [08:45] Chipaca: yeah, that's why it is in the heuristic package [08:45] Chipaca: happy to take feedback on every part [08:46] Chipaca: if you are reading it already, the two changes I was thinking about are: relative positioning (d'oh, why did I do absolute) and changing Pre widget to Paragraph that I also implemented but left out - paragraph does text folding [08:47] I think it's just generally better than a fixed text thing [08:47] hey zyga [08:47] hey pawel :) [08:47] hey zyga, good morning [08:47] zyga: does it support truncation / elision of text? [08:48] Chipaca: no, it supports folding into available width - the whole design of this is like printer paper [08:48] you have a fixed with [08:48] but all the height you want [08:48] Chipaca: it supports clipping though, so you can force it to occupy some space [08:48] zyga: thinking of things we are doing that would benefit from something like this, find and list come directly to mind [08:48] but both of them require elision [08:49] Chipaca: I think the thing you should look at is the signature of Widget.Pack [08:49] zyga: release> not super urgent to release 2.43.1, I think it's fine to release when it hits our candidate channel [08:49] if we give it information about what the constraints are (currently it's just the remaining width) we could do what you just described [08:49] mvo: ok [08:52] Chipaca: do you have time now? we could jump into a HO and talk about this [08:52] zyga: I'm having breakfast, now :) [08:52] Chipaca: ah, sorry [08:53] zyga: for 'snap list' one problem you have is that almost every field needs shrinking, and some are more 'important' than others [08:53] indeed, I was thinking about tabular widget [08:53] doing that in a generic way is doable, but you might need to be knuth [08:53] right now vbox and hbox are insufficient to perfectly replicate a table [08:54] Chipaca: heh, I remember reading that paper a few years ago - about table layout problems [08:54] it's a surprisingly hard problem [08:54] PR snapd#8010 opened: snap-bootstrap: add support for "recover" mode [08:55] zyga: the good news is fixed-width makes it a little simpler :) [08:55] the bad news is double-width characters make it hard again [09:09] Bug #1655376 changed: Add support for android/touch type images [09:22] PR snapd#7997 closed: overlord: increase settle timeout for slow machines [09:37] zyga: heh, pstolowski correctly spotted that those tests are in snap-bootstrap [09:38] mborzecki: hmm? [09:38] zyga: https://github.com/snapcore/snapd/pull/8009#pullrequestreview-343774644 [09:38] PR #8009: packaging/opensuse: build-depend on fakeroot [09:38] pstolowski: thanks pawel! [09:39] What is you opinion. What should happened if a symlink couldn't created. I think the snap shouldn't abort installation. [09:39] `cannot update snap namespace: cannot create symlink in "/etc/sway": existing file in the way` [09:39] hmm [09:39] so do we use fakeroot in the real code? [09:39] sdhd-sascha: layouts are required, if it cannot be created and there are install hooks or configure hooks then installation is aborted [09:40] yw, good! [09:40] I have to handle two case's. Case 1: no desktop packages are installed on the host. Case 2: a desktop installation exists on the host. [09:41] E. g. openGl drivers are on the host. Then the snap don't need to deliver them. [09:41] ... [09:42] Would it make sense, to add "try-symlink" to layouts ? [09:43] And/Or "try-bind" mount [09:44] zyga: yes, we do, but that's relevant only for uc20 atm [09:46] zyga: got a fix, can open a PR [09:46] it's a single patch, so you should be able to cherry pick it without trouble [09:46] mborzecki: cool, thank you! [09:47] sdhd-sascha: I don't know, can you give me an example please? [09:47] sdhd-sascha: but it feels like unrelated to layouts [09:47] sdhd-sascha: layouts cannot grab data from the host [09:48] sdhd-sascha: what you are describing seems more like interface connections [09:50] zyga: currently i used this: [09:50] layout: [09:50] /etc/sway: [09:50] symlink: $SNAP/etc/sway [09:50] PR snapd#8009 closed: packaging/opensuse: build-depend on fakeroot [09:50] PR snapd#8011 opened: cmd/snap-boostrap: add mocking for fakeroot [09:50] zyga: but this is not urgent. [09:51] mvo: posted a little guide for working with vm cloud images (incl. uc20) and feedimg some quick setup via cloud-init https://gist.github.com/bboozzoo/a1303a1e4deb516c2ad9a1e64835c341 [09:51] i had another case in the past. But can't remember, so i can give more examples in the next days. Current i test, sway on both case's. [09:52] mborzecki: maybe mention the gist on the forum [09:52] mborzecki: it will be easier to find [09:53] mborzecki: nice! [09:54] PR snapd#8012 opened: o/devicestate: do not create perfTimings if not needed inside ensureSeed/Operational [09:56] mborzecki: did you end up with network in the qemu vm? [09:56] Chipaca: hey, conflict in #8004 [09:56] PR #8004: cmd/snap: print full channel in 'snap list', 'snap info' [09:56] mborzecki: I remember there was an issue yesterday? [09:57] mvo: we got it resolved once the system booted into the run mode and kernel module were mounted at the right location [10:01] pedronis: boot speed ho? [10:03] PR snapcraft#2881 closed: [backport] meta: fix missing provider case for get_provider_content_directories() [10:13] brb [10:22] re [10:24] going to take a break in a few minutes. Nothing particularly noteworthy in triage today. [10:28] PR snapd#8002 closed: tests/core18/snapd-failover: collect more debug info [10:41] mvo: this will help a little https://i.imgur.com/86JdDRU.png [10:42] not sure why the quantity got badly formatter, maybe it's not bytes (?) [10:44] heh, the conversion is wrong ;) [10:45] mborzecki: nice! [10:45] mborzecki: you rock [10:45] btw. u-i should probably create an image large enough to fit the layout too [10:46] mborzecki: yes, probably something for sil2100 to make ubuntu-image smarter so that it knows how much space to reserve when building a uc20 image (right now it creates an image that is too smalll to fit the partitions) [10:46] * Chipaca afk [10:52] mvo: I'm looking at #7992, yes I think the issue is that all the fixture code is recovery oriented but shouldn't for this new stuff [10:52] PR #7992: bootloader: add ExtractedRunKernelImageBootloader interface, implement in grub [10:53] mvo: I'll see if I can untangle that a bit [11:00] pedronis: nice, thank you! [11:23] Chipaca: hey, is this a good time? [11:30] mvo: done, but there is probabably a bit more to do there [11:32] Chipaca: there's no code for format MiB/GiB quantities in the tree, is there? [11:33] pedronis: I'd like to merge https://github.com/snapcore/snapd/pull/7875 as-is [11:33] PR #7875: interfaces: refactor path() from raw-volume into utils with comments for old [11:34] do you think it's possible to get your review today? [11:39] * zyga jumps to another PR [11:53] ijohnson: hi, thanks for the insight into the root cause of https://github.com/snapcore/snapd/pull/7871 failures! i got it working (i don't know yet if it passes on all systems, waiting for travis, but as far as this denial is concerned, it should be fine now) [11:53] PR #7871: tests: add spread test for hook permissions <⛔ Blocked> [11:56] I have asked this before, sorry. When are we getting the "don't update me please, I'm running" for applications? [11:57] i think zyga worked on this ^ [11:59] popey: it's stuck in review for ... months? [12:00] popey: I just resolved a merge conflict https://github.com/snapcore/snapd/pull/7825 [12:00] PR #7825: many: use transient scope for tracking apps and hooks [12:00] PR snapd#8013 opened: cmd/snap-bootstrap: check device size before boostrapping and produce a meaningful error [12:01] zyga: i'll take a look at it later today [12:01] pstolowski: super, thanks! [12:01] I'll push a small tweak for older merge breaking after test rename [12:15] PR snapd#8011 closed: cmd/snap-boostrap: add mocking for fakeroot [12:16] zyga: you can cherry-pick the patch now ^^ [12:16] mborzecki: I released a variant with fakeroot build-dep [12:16] mborzecki: I'll get this for 2.44 :) [12:16] ah, ok [12:16] mborzecki: or I can do a -2 if there's some other things to cherry pick [12:19] * pstolowski lunch+errand, bbl [12:25] zyga, should all layouts be set up when i use snap run --shell for a daemon ? [12:26] ogra: yes [12:26] hmm, weird [12:27] ls prime/etc/dbus-1/system.d/ [12:27] org.freedesktop.DisplayManager.conf [12:27] root@localhost:/home/ogra# ls /etc/dbus-1/system.d/ [12:27] wpa_supplicant.conf [12:27] the above is from the packaging dir ... the below one from snap run --shell ... i have a bind layout defined for /etc/dbus-1 [12:28] i dont get any error when installing or anything ... so i'd expect to find that org.freedesktop.DisplayManager.conf file somehow ... [12:28] Bug #1859974 opened: failing to release devmode revisions to edge/beta channels [12:30] ogra: can you show me the relevant parts of snap.yaml please [12:31] * zyga jumps to setgid branch [12:31] looking at those failures now [12:36] mvo, mborzecki: on it! (soon) [12:36] sil2100: thank you [12:36] sil2100: not a blocker but certainly will improve the experience :) (all of us ran into this problem) [12:38] mvo: yeah, when I switched to the 'cleaner' way of handling system-seed, I guess I forgot about the image-size calculation [12:42] sil2100: thanks! [12:42] mvo: uc20 rite of passage [12:43] * zyga sighs at inability to stop spread [12:48] zyga: ^\ doesn't work? [12:48] zylop: i mean sigquit [12:49] zyga: ^ :) [12:51] mborzecki: yeah but that leaves garbage [12:51] zyga: you need to spread -discard -spread-pid=<..> later [12:51] yeah but ... why do I need to :/ [12:55] cmatsuoka: mvo: do you have any notes on runnnig qemu with tpm? === ricab is now known as ricab|lunch [13:14] mborzecki: once you have the tpm simulator installed, you must run qemu the the appropriate parameters [13:14] mborzecki: see them here: https://github.com/snapcore/spike-tools/blob/master/run-test.sh [13:14] cmatsuoka: thanks [13:14] mborzecki: it's the line with tpms in it [13:56] brb [13:56] morning folks [13:56] hi, I'm seeing an issue when using snapcraft candidate. it seems the command-*.wrapper files are not being created. did something change there? [13:58] sergiusens: ^ [14:00] oh I'm not on the latest candidate, I guess it just got udpated [14:13] degville: please ping me when you look at the content interface docs, there's some rarely known features there [14:13] s/'s/are [14:14] Bug #1859974 changed: failing to release devmode revisions to edge/beta channels [14:14] zyga: will do, thank you! [14:24] ackk: wrappers are now gone, instead what used to be in the wrapper is now part of the command-chain, this allows for "snap run --shell " to have the right command from the get go === ricab|lunch is now known as ricab [14:25] PR pc-amd64-gadget#33 closed: grub.cfg-boot: use $prefix/kernel instead of $ubuntu-boot/kernel [14:26] sergiusens, is there a way from inside the snap to run stuff as if it were from the outside? we currently call the wrapper script explicitly since we need to call stuff under snapcraft-runner for the LD_PRELOAD/PATH exports [14:26] sergiusens, I guess we can wrap everything with our wrapper otherwise [14:26] PR core20#18 closed: static: try using /run/mnt/snapd first in run-snapd-from-snap [14:27] ackk: hmm, that wrapper was always an implementation detail, that said, "$snap/command-chain/snapcraft-runner" will have most of what you need [14:28] PR core20#17 closed: Drop encoding digits in units and code paths. [14:28] sergiusens, I see, thanks [14:30] sergiusens, is snapcraft-runner used both for apps and daemons? [14:31] ah yeah it is, cool [14:32] * zyga goes to grab lunch === mborzeck1 is now known as mborzecki [14:37] sil2100: hi. Is there a tool to use, to generate the changelog entry's for ubuntu-image ? [14:43] pedronis: thanks for the patches to 7992, I guess I never quite understood what you meant by NoSlashBoot option when you mentioned it in our meetings but I see what you meant now. [14:44] np [14:48] PR snapd#8007 closed: tests: remove execution of ubuntu 19.04 from google backend [14:52] PR snapd#8014 opened: tests: run `uc20-snap-recovery-encrypt` test on 20.04-64 as well [14:53] mvo, pedronis should I enable ubuntu 20.04 tests? [14:53] as 19.04 is gone [14:53] I already have an image [14:54] or should I make it once I am back from my vacactions [14:54] ? [14:58] sdhd-sascha: hey! I usually just add one with dch for that - but I can also create those for you, as long as everything else is good to go [15:00] when I try to start snap-store voa terminal, I get no protocol specified and Unable to init server: Could not connect: Connection refused - what's the issue? [15:02] sil2100: i cleaned up, rootfs_size ( 182 ) and added one test. [15:07] Girtablulu: and if you try to launch it in a different way? [15:07] * cachio lunch [15:08] Start menu and terminal both dont work [15:08] Girtablulu: sounds like a bug to me [15:10] mmmh okay [15:11] Girtablulu: https://launchpad.net/snap-store to report it maybe [15:11] yea will do [15:11] cachio: yeah, I think we should enable 20.04-64, maybe to unstable at first [15:13] Girtablulu: please include the output of 'snap version' there [15:13] will do [15:13] mvo, sure, I'll add this after lunch [15:20] Girtablulu: thanks [16:13] fd [16:13] re :) [16:17] mvo: can we merge https://github.com/snapcore/snapd/pull/7238 ? [16:17] PR #7238: usersession: add systemd user instance service control to user session agent

[16:18] mvo: +3 and green [16:18] I can merge master into it to ensure tests passed recentlty [16:18] *recently [16:24] zyga: it can be merged, but yes best to merge master first [16:24] ok [16:25] jd [16:33] zyga: looks like samuele and jamie are happy so I see no reason not to merge it [16:33] cool, I'll merge it after it passes one round of tests [16:33] ok [17:06] zyga: i did one pass over #7825, i'll continue tomorrow morning [17:06] PR #7825: many: use transient scope for tracking apps and hooks [17:08] cachio, ijohnson can you take another look at #7871 if/when it gets green (just had to restart the job, it seemed to be stuck) [17:08] PR #7871: tests: add spread test for hook permissions [17:09] pstolowski: sure [17:09] thanks [17:09] eod, cu [17:37] PR snapd#8015 opened: [RFC] daemon, store: support download resume from /v2/download [18:04] zyga: I pushed a tweak to #7875 along the last comments, it can go in if it gets green [18:04] PR #7875: interfaces: refactor path() from raw-volume into utils with comments for old [19:09] PR snapcraft#2882 opened: extensions: change extension merge-strategy to fix build-environment === blackboxsw_ is now known as blackboxsw [20:43] PR snapd#8016 opened: gitignore: ignore snap files [20:51] PR snapd#8004 closed: cmd/snap: print full channel in 'snap list', 'snap info' [20:57] on a 16.04 system using the snapd deb (2.34.2), should one be able to just run `sudo snap install snapd` to get the latest? [20:58] or would there other steps to take? [21:05] PR snapd#8017 opened: tests: add ubuntu 20.04 to the tests execution and remove tumbleweed from unstable [21:09] cjp256: I'm not sure about 16.04, but I know on 18.04 if you don't have snaps, it will use the snapd snap when you first install a snap [21:10] cjp256: if you wanted to get the latest on 16.04 I'd just say do `sudo snap refresh core` [21:13] cjp256: though I did just check on 16.04 and it should be fine to install snapd like you have it [21:17] thanks ijohnson. i did a similar test and was OK, but recalled a conversation between sergiusens and Chipaca about some flag I can't recall, wanted to make sure it wasn't generally relevant :D [21:17] my test is on a fresh lxd image, so it's probably much newer base than theirs [21:18] cjp256: as long as you are doing this fresh without any snaps installed it is okay to install snapd, otherwise you would need to migrate from core -> snapd which does require a flag [21:18] err "okay to install the snapd snap" is what I meant [21:18] oh, what flag is that? [21:18] let me find it, it's an experimental flag [21:20] ijohnson: https://github.com/snapcore/snapcraft/blob/master/snapcraft/internal/build_providers/_snap.py#L332 ? [21:20] cjp256: ah yep there it is [21:21] tried searching for it earlier and couldn't find it because I was checked out on an old tag 🤦 [21:21] that's the flag you need to use if you have the core snap installed, and want to use the snapd snap instead === Girtablulu is now known as Girtablulu_ [22:33] ijohnson: cjp256: snapcraft needs to work on the un-updated snapd that didn't know about the snapd snap [22:38] Chipaca: thanks! ijohnson gave me a lot of background info and I just confirmed running an old version of snapd (2.34.2): `error: cannot install "snapd": cannot install snapd snap on a model without a base snap yet` so the two options they have are `apt update; apt install snapd` or `sudo snap install core` to get working. [22:39] cjp256: that's why snapcraft sets an experimental flag [22:39] cjp256: but even that needs a new enough snapd [22:39] if eventually installing core snap anyways, is the apt update/upgrade wasting cycles? [22:39]