=== chihchun_afk is now known as chihchun [05:55] good morning [06:01] zyga: morning! [06:02] zyga: did you find any solution for the "cannot find package "golang.org/x/crypto/ssh/testdata"" error on travis? === ubott2 is now known as ubottu [06:59] ogra_: ping [07:16] morphis: hey [07:16] hey ara :) === zyga_ is now known as zyga [07:16] morning zyga :) [07:16] zyga: hey :-) [07:16] morphis: I commented on https://github.com/snapcore/snapd/pull/3271 - trivial stuff if you want to apply that [07:16] PR snapd#3271: cmd/snap-confine: use /etc/ssl from the core snap [07:16] * zyga goes back to reviews [07:17] zyga: what is the difference between details and description? [07:17] ah looks like I did that wrong until now [07:18] and spread doesn't complain about details: [07:18] s/details:/description:/ [07:22] morphis: it's not used at all [07:22] zyga: yeah, just saw that I used description: in all my previous MPs [07:22] morphis: but you didn't give any details or description so I wrote some :) [07:22] morphis: it's a purely human-readable thing [07:22] will clean that up [07:23] morphis: thanks! [07:37] zyga: fixed https://github.com/snapcore/snapd/pull/3271 [07:37] PR snapd#3271: cmd/snap-confine: use /etc/ssl from the core snap [07:37] morphis: thank you [07:58] zyga: also updated https://github.com/snapcore/snapd/pull/3274 [07:58] PR snapd#3274: cmd/snap,tests/main: add force-devmode switch instead of spread system blacklisting [08:00] morphis: great, thank you [08:01] that one still needs conversion of all other tests not converted yet but once I get a pass from travis I will do that [08:02] morphis: replied [08:02] morphis: one thing about ! [08:03] fixed [08:03] morphis: great! [08:19] * zyga -> coffee [08:19] brb [08:23] morning. can i get some help finishing my sigrok snap and getting it into the store? [08:23] ali1234: hey, what kind of help do you need [08:24] it works with strict confinement, i just need to finish off the house keeping: build dependencies, version number etc [08:24] and get it uploaded [08:24] cleanbuild still does not work for me [08:24] also i have a weird bug where Qt apps are not themed correctly, but that affects every snap afaict [08:25] every snap on my system i mean [08:25] https://github.com/ali1234/mysnaps/blob/master/sigrok/snapcraft.yaml is what i have [08:27] oh, also i'm not sure why i need the network plug. libusb uses netlink and raw-usb alone isn't enough to make it work [08:27] ali1234: themes are hard :) https://forum.snapcraft.io/t/use-the-system-gtk-theme/496/4 [08:27] oh, so its because i don't use the default theme? [08:28] okay, well i guess that's the least important issue [08:37] ali1234: yes, because the theme in the snap is not the same as the themes in the system [08:37] ali1234: can you share dmesg | grep DENIED [08:38] for what? the netlink thing? [08:38] ali1234: and also dmesg | grep type=1326 [08:38] ali1234: this will show any apparmor and seccomp denials [08:38] if i have the raw-usb plug but not the network plug i get: type=1400 audit(1494097705.817:107): apparmor="DENIED" operation="create" profile="snap.sigrok.sigrok-cli" pid=8274 comm="sigrok-cli" family="netlink" sock_type="raw" protocol=15 requested_mask="create" denied_mask="create" [08:38] so we can figure out what sigrok is doing that needs more interfaces [08:39] aha so NETLINK_KOBJECT_UEVENT [08:39] it may be that sigrok supports analyzers on ethernet and is trying to scan for them [08:39] i am not sure, i only have a usb one to test with [08:39] the good news is that we have some nice netlink interface lately [08:41] even if sigrok requires a dedicated interface one can be crafted very quickly [08:41] as i said, it works with the network plug [08:42] and sigrok does support ethernet devices according to the hardware page [08:43] i think it can use rs232 devices as well [08:50] morphis, yes ? [08:53] ogra_: I saw https://paste.ubuntu.com/24535422/ in a spread test this morning [08:54] ogra_: any idea? [08:54] morphis, ouch ... looks like there was a newer shadow version uploaded to the archive ... overriding the PPA [08:54] uups [08:57] zyga: looks like you referenced the wrong link in https://github.com/snapcore/snapd/pull/3277#issuecomment-299792600 [08:57] PR snapd#3277: cmd/snap, client: add "whoami" command [08:58] aha, thanks [08:58] ogra_: if you want more details https://travis-ci.org/snapcore/snapd/builds/229850432?utm_source=github_status&utm_medium=notification [08:58] morphis, i dont need any details ... i need to merge a package :P [08:58] :-) [08:58] corrected :) [09:02] * ogra_ sighs ... and indeed its a manual merge, the CVE fix touches the same file [09:11] ogra_, hey, is there any way I can avoid starting console-conf on first run? [09:13] abeato_, i think you can touch some file but not sure what the path was ... perhaps mwhudson can help [09:15] abeato_: there is, have a look at our tests-extras repo [09:15] (not sure how the getty setup reacts on that though ... there is some special setup to print the IP and ssh keys) [09:15] abeato_: https://git.launchpad.net/~snappy-hwe-team/snappy-hwe-snaps/+git/tests-extras/tree/create-image-scripts/create-image.sh#n135 [09:15] (and to prevent spawning a login prompt) [09:16] morphis, I see, thanks [09:16] ogra_, well, provided I get a shell I can manage :) [09:17] if you would that would be a bug :) [09:17] you shouldnt get a shell login offered without configured user [09:18] yeah, I will try to create one with a system-user assertion [09:18] yeah, that shoudl work [09:26] abeato_, ogra_: /var/lib/console-conf/complete i think [09:27] mwhudson, great... how can I some debugging for console-conf, anyway? [09:31] what approach should i take to automatic building and versioning? [09:31] upstream has this repo that contains build scripts for various platforms: http://sigrok.org/gitweb/?p=sigrok-util.git;a=tree;f=cross-compile;h=e3013fd578b6d07b55ca32a4c43d082e36cc5e24;hb=HEAD [09:32] as you can see, they already have an appimage [09:32] it would make sense to put the snapcraft.yaml in here i think [09:32] but it builds from several other repos [09:36] ali1234: you can look at build.snapcraft.io [09:36] ali1234: it's a free service that's very useful :-) [09:36] upstream does not use github though [09:37] Chipaca: hey [09:37] Chipaca: do you know if any of the sprint-going people will be around today? [09:37] i assume it does the same thing as https://snapcraft.io/docs/build-snaps/ci-integration [09:38] zyga: I don't [09:38] ali1234: ci integration is sligthly different but I'm not actively using it myself so I could be worng [09:38] zyga: I do know I'm about to go off to a doctor's appointment :-) [09:40] Chipaca: take care! Look after yourself! [09:40] zyga: this is the physio \o/ [09:40] anyway, should be back for the standup [09:40] sounds good [09:41] is it still a thing to prefix your name on packages to prevent collisions? [09:41] abeato_: painfully [09:42] lol [09:42] abeato_: ui things get debugged in dry-run mode [09:42] yes, that is something I suspected... [09:42] more complicated things mostly by staring at the log file [09:42] mwhudson, it is more about some issue with handling interfaces [09:42] ali1234: if this will be the official snap I'd skip the prefix [09:42] it wont be, i am not upstream [09:42] ali1234: especially if it comes from the upstream repository [09:42] abeato_: as in NICs? [09:42] ali1234: aha [09:42] ali1234: maybe try talking to them [09:42] ali1234: it's just packaging :) [09:42] mwhudson, yes [09:42] i have done [09:42] ali1234: and they may like it [09:43] they did like it [09:43] ali1234: if they commit the snapcraft.yaml into their tree [09:43] they want to see it working [09:43] abeato_: fun times [09:43] mwhudson, where is the log stored for console-conf? [09:43] ali1234: I don't see why it should be prefixed [09:43] abeato_: /var/log/console-conf [09:43] i added my repo on build.snapcraft.io and it says it does not have a snapcraft.yaml [09:43] this is not true, it actually has three [09:43] mwhudson, ok, thanks [09:44] ali1234: it looks at a specific location, where are your snapcraft.yaml files? [09:44] ali1234: perhaps ask in the forum (forum.snapcraft.io) as the people working on this are in other timezones [09:44] well the sigrok one is in sigrok/snapcraft.yaml [09:44] if it goes in upstream it will be in cross-compile/snap/snapcraft.yaml [09:45] abeato_: you could try using a system user assertion to have an account created on first boot and then poke around more interactively [09:45] ali1234: AFAIR it can be in (paths) snap/snapcraft.yaml or snapcraft.yaml relative to the root of the project [09:45] but that's a bit hairy too [09:45] ali1234: it cannot be in an arbitrary location [09:45] okay i will fork upstream repo and try with that [09:45] mwhudson, yes, that is something I want to do too [09:45] abeato_: what sort of problem are you having? [09:46] mwhudson, the one signing the system-user assertion should be the same that signed the gadget snap, right? [09:46] mwhudson, some issue with interface configuration, it looks apparently configured but I cannot reach the store [09:46] abeato_: s/gadget snap/model assertion/ i think, but yes [09:46] mwhudson, yes, model :) [09:47] abeato_: ugggh [09:47] although it won't work for upstream since it isn;t on github [09:47] well, unconfigured interfaces still run a dhclient [09:47] even before cvonsole-conf ever comes up [09:48] mwhudson, not sure if maybe the issue comes from some p2p0 interface that is around [09:48] mwhudson, ogra_ http://paste.ubuntu.com/24535879/ [09:48] looks like you cant reach your nameserver [09:49] yes i guess you should check for broken / bogus dhcp servers :) [09:49] ali1234: where is upstream? [09:49] ogra_, in fact I can ping the device when I am in the "network connections" screen, but after "configuring" the network, the IP gets lost [09:49] http://sigrok.org/gitweb/?p=sigrok-util.git;a=tree;f=cross-compile;h=e3013fd578b6d07b55ca32a4c43d082e36cc5e24;hb=HEAD [09:50] it actually removes connectivity instead of providing it [09:50] abeato_, yeah, consiole-conf changes it [09:51] i am importing that repo to github just for testing purposes [09:51] abeato_: anything incriminating in syslog? [09:51] mwhudson, it is pretty difficult for me to grab logs, will change initramfs to get those [09:52] ali1234: perhaps you could create a launchpad project, setup git mirroring and just build on launchpad with a snap packaging recipe [09:52] ali1234: build.snapcraft.io is just a fancy front end [09:52] yes, that's the thing i linked to before :) [09:52] ali1234: aha :-) [09:52] abeato_: system user assertion approach probably best then [09:53] i wonder if that is documented anywhere yet... [09:53] but the thing is that repo gets pushes all the time for other things [09:53] ok [09:53] also it doesn't get pushes when the actual source repos are updated [09:53] so idk how to make it rebuild properly [09:53] or i dunno, unpack core snap, chroot into it, add a user, repack, reimage, reflash, would that work? [09:54] maybe not actually, you wouldn't get a getty, could always hack up systemd config to make sure you do but that gets a bit involved again [09:55] * mwhudson is rambling, it's a bit late here [09:55] ali1234: it pulls automatically every 6 hours [09:55] ali1234: it's not amazing but not terrible either [09:55] ali1234: and the recipe can build on each push [09:55] ali1234: look at this: [09:55] yes, i understand that [09:56] https://github.com/snapcore/pi2-gadget [09:56] this is mirrored on launchpad and built to a snap [09:56] exactly the way I described [09:56] and it is documented in the readme [09:56] the problem is that the repo gets pushes eg when the appimage changes [09:56] but it doesn't get pushes when the source code of the app itself changes [09:56] look at the links and you can do the same thing [09:56] aha [09:56] wait, so where is the source code? [09:56] mwhudson, it is an idea... will go first for the system-user assertion though ;) [09:56] in several other repos [09:56] I see [09:56] look at the snapcraft.yaml :) [09:56] * zyga looks [09:57] https://github.com/ali1234/mysnaps/blob/master/sigrok/snapcraft.yaml [09:57] well [09:57] no idea, ask on the forum :) [09:57] it's an interesting problem for sure [09:57] so far everything i tried to snap ended up as an interesting problem :) [09:58] that's why my repo is called "museum of broken snaps" [09:58] ali1234: this is how exploring the unknown feels like :D [09:58] none of them actually work [09:58] zyga: any idea why `snap info core` doesn't print out the snap type on Fedora? [09:58] sigrok does now though so i want to promote it :) [09:59] ali1234: i just have a launchpadlib script in cron that triggers a build every 24 hours [09:59] mwhudson: doesn't that mean users have to download an update every day, regardless of if anything changed? [09:59] ali1234: http://paste.ubuntu.com/24535900/ [09:59] the sigrok snap comes out at 74MB [10:00] morphis: no, can you share the output? [10:00] ali1234: ah yes, well nothing stops you being slightly cleverer in the script i guess :) [10:00] ali1234: i agree this is a problem, i don't have a great solution [10:00] zyga: https://paste.ubuntu.com/24535903/ [10:00] zyga: on ubuntu that gives me type: core [10:01] yeah [10:01] I see [10:01] * zyga looks [10:01] so we have maybePrintType [10:02] zyga: yes, but that should print the type for the core snap [10:02] um... why doesn't the snapcraft forum have SSO? [10:03] its in the works [10:03] zyga: ok, I know why [10:03] zyga: `snap list` marks it as broken [10:03] morphis, new core with new shadow building ... [10:03] morphis: aaah [10:03] morphis: :) [10:03] morphis: there you go [10:03] ogra_: nice [10:03] morphis: /var/lib/snapd/snap vs /snap issue somewhere [10:03] morphis: broken meant that snapd didn't find the meta/snap.yaml file [10:04] * zyga would really like for snapd to cache snap.info per revision [10:04] then we could even lazily mount snaps or keep only one revision mounted instead of three [10:04] less memory use [10:04] and we could keep a finite set mounted even if the user has many times more snaps installed (just not running) [10:04] part of the lifecycle idea we had earlier [10:09] does the name: in snapcraft.yaml have to match the registered name on build.snapcraft.io? [10:09] ali1234: yes [10:09] ali1234: well [10:09] ali1234: any name you want really but you need to be able to push snaps there [10:10] which? [10:12] okay it still can't see the snapcraft.yaml [10:12] https://github.com/ali1234/sigrok-util/tree/master/cross-compile/snap [10:14] zyga: debian should run in force devmode true, right? [10:23] morphis: yes [10:23] morphis: unless with ubuntu apparmor patches, then it will be automatically not be devmode [10:24] ok [10:31] zyga: hm .. qemu:debian-9-64 .../tests/main/forced-devmode# snap forced-devmode [10:31] false [10:31] ali1234: one of snap/snapcraft.yaml, snapcraft.yaml, or .snapcraft.yaml has to exist at the *top level* of your branch [10:31] ali1234: (I believe symlinks work) [10:31] yeah that's not going to happen [10:31] not my choice, sorry [10:31] upstream won't accept it, and that's not my choice [10:32] probably best bring it up on the forum or similar - it's not going to change without user protest [10:32] i have done https://forum.snapcraft.io/t/trying-to-build-sigrok-snap/503 [10:36] morphis: interesting [10:36] morphis: I wonder how is this possible [10:36] morphis: is it a boolean flip somewhere/ [10:36] zyga: it loosk like there is a problem in my code === olive` is now known as olive [10:54] hello, is there a snap with the same sort of functionality as plymouth for Ubuntu core? [11:03] zyga: ok, was again my fault .. [11:06] AdamH_, nope ... not yet (and it would be a bit tricky sinc eplymouth needs initrd inclusion for some use cases) [11:07] AdamH_, i'm locally playing with uboot spash screen but dont have anything to look at yet [11:07] *splash [11:07] that wont help much with the rest of the boot though ... [11:08] ogra_: the new core (1870) fixes the ubuntu-core-create-user error https://travis-ci.org/snapcore/spread-cron/builds/229908080 all green now [11:08] * ogra_ hugs fgimenez ... thanks for the feedback ! [11:08] :-) [11:08] I could use a simple review for https://github.com/snapcore/snapd/pull/3262/files [11:08] PR snapd#3262: cmd/snap-confine: aggregate operations holding global lock [11:09] * zyga goes back to other reviews [11:09] * ogra_ ponders to add a check to the core build scripts with a blacklist of packages we dont want from the archive [11:09] so it would fail the next time a newer shadow is there [11:10] ogra_: :) thank you! morphis ubuntu-core-create-user should pass with the new core [11:11] fgimenez: great! [11:22] ogra_: Thanks, If you need any help testing let me know. I will come back to this task later. Still plenty to develop! [11:22] will do, thanks for the offer [11:31] * zyga afk for a moment [11:34] * Chipaca returns [11:34] fgimenez: do you know what's up with ubuntu-core-16-64:tests/main/ubuntu-core-create-user ? [11:35] ogra_: or maybe you? [11:35] /usr/bin/chfn: unrecognized option '--extrausers' [11:35] ^ [11:36] Chipaca, fixed already [11:36] re-run wih a newer core [11:37] woot [11:38] ogra_: what had happened? [11:38] Chipaca, CVE upload in the archive with a higher version [11:38] so the PPA package wasnt used anymore [11:40] ogra_: in an ideal world we'd get alerted when that happened, yes? [11:41] Chipaca, yes. see what i wrote above ... [11:41] gra_ ponders to add a check to the core build scripts with a blacklist of packages we dont want from the archive [11:41] so it would fail the next time a newer shadow is there [11:41] ogra_: yeah! [11:42] must be blind because i scanned up several times and didn't spot you saying that there [11:42] Is mvo not around today? [11:42] sborovkov: I wouldn't be surprised if he took the day off [11:42] (he's been travelling back from the sprint) [11:42] JamieBennett might know more [11:43] that's fine. it's not urgent anyway [11:43] just wanted to ask him to add another rpi config option [11:43] sborovkov: he is on vacation today [11:43] sborovkov, just provide a PR ;) [11:45] sborovkov, just provide a PR ;) [11:45] oope [11:45] *oops [11:45] sorry [11:45] oh right I can do that [11:45] and it seems it should be one liner [11:46] * ogra_ notes he just killed his pi3 boot ... adding a unit to multi-user-target.wants that has a Before=network.service inside seems to not be such a clever thing ... [11:46] ogra_: it shouldn't kill it though [11:46] ogra_: systemd will break the cycle by ignoring a unit at random [11:47] Chipaca, well, ssh doesnt come up, i'm in the living room having some lunch, board is upstairs in the office ... [11:47] ogra_: guess which one it chose to break the cycle [11:47] hehe [11:47] ogra_: easy to fix without going upstairs: power cycle the house [11:48] * ogra_ plugs two fingers in the socket [11:48] * Chipaca sends flowers [11:48] ah, damn, you said cycle ... means i need to go to the basement after the fuse blew [11:49] so that ends up in being the same effort ... i'll just finish lunch :P [11:50] Any pointers on where to look to rotate the screen to portrait and hide the cursor? I am using mir-kiosk but can't see any config files or options? [11:50] PR core#38 opened: Add another pi-config option [11:50] ogra_: who should I ask to take a look at that pr - https://github.com/snapcore/core/pull/38 [11:50] PR core#38: Add another pi-config option [11:51] sborovkov, you are aware that enabling the usb mode isnt reversible ? ... not sure thats a clever thing to have in the default options [11:52] (or did that bevaiour change) [11:52] *behaviour [11:53] ogra_: I am not sure about that behavior, but how can I set it otherwise? Configure hook will comment out everything else in config.txt and well without it I can't even set it [11:53] at all [11:54] and we want to use that option eventually because SD cards are dying very often and that's actually super cool feature [11:54] sborovkov: question, have you signed the CLA? [11:54] sborovkov, oh, why are your SD cards dying ? [11:54] (couldn't figure out your lp nick from your irc nick, if you have an lp account, which is the easiest way of me checking CLA status) [12:02] ogra_: well we have around 10000 rpis in production... that's the most frequent case of failure [12:02] Chipaca: not sure - https://launchpad.net/~serge-borovkov [12:03] sborovkov: that's a nope [12:05] Chipaca: where do i sign it actually? here? https://www.ubuntu.com/legal/contributors/submit [12:06] sborovkov, and adding 10000 USB keys is a cost effective solution ? :) [12:07] sborovkov: yep [12:07] ogra_: it actually is. Because for the bigger clients especially they can have a bunch of devices across the country. Sending technician to replace SD card for them is not very convenient and would cost more [12:07] Chipaca: Please add the Canonical Project Manager or contact. Who should I list here? [12:08] sborovkov: Jamie Bennett [12:22] re [12:22] mi [12:24] fa [12:30] so [12:30] Chipaca: I think we want to do a small experiment on exposing interface meta-data [12:30] Chipaca: e.g. an optional method [12:30] Chipaca: that returns a struct [12:30] Chipaca: that has various stuff [12:30] Chipaca: it could replace AutoConnect [12:30] Chipaca: Name [12:30] Chipaca: (well, if we do Name it could be mandatory) [12:30] Chipaca: it could return description [12:31] Chipaca: and one-to-many/many-to-many [12:31] Chipaca: what do you think? [12:31] zyga: you mean so the client could make smarter decisions for tab completion? [12:32] Chipaca: yes [12:32] Chipaca: and we could show nice stuff in CLI [12:32] Chipaca: e.g. (what is this interface) [12:32] Chipaca: is that interface restricted [12:32] Chipaca: whatever we can come up with [12:32] zyga: could do. Or we could add something on the daemon like "what can connect to " [12:32] Chipaca: it's just a struct [12:32] ah, yeah, that'd be nice to have [12:32] Chipaca: yes, that's possible too but I think we want the meta-data anyway; I agree that the daemon may make better decisions about what to suggest as an option [12:33] zyga: that would make “snap interface foo:bar” make sense [12:33] Chipaca: especially if we could get rid of some of the optional boring copy pasted methods now [12:33] Chipaca: yes! [12:33] Chipaca: I just +1'd 3214 [12:33] shall I merge? [12:34] oh! that old thing [12:34] zyga: sure! :-) [12:35] PR snapd#3214 closed: cmd/snap: iterate interface tab completion [12:35] Chipaca: if you review https://github.com/snapcore/snapd/pull/3272 quickly we could land it [12:35] PR snapd#3272: add interfaces-shutdown-introspection spread test [12:35] it has 1.9 +1s [12:40] zyga: on it [12:43] jdstrand: snapd#3272 is GTG, unless you want to remove the restore section RSN we should land it [12:43] PR snapd#3272: add interfaces-shutdown-introspection spread test [12:43] bitrot an' all that [12:49] Chipaca: hey, I can remove it [12:49] jdstrand: your call :-) [12:49] you've got the +1s either way [12:51] * Chipaca goes in search of tea [12:51] Chipaca: I can remove it [12:51] jdstrand: oh [12:51] jdstrand: hey :) [12:51] zyga: hey :) [12:51] * zyga read backlog in the wrong order :D [12:51] jdstrand: go ahead then, I'll gladly merge it when tests show green [12:53] fgimenez: did you see this: [12:53] __all__: The name test-snapd-kernel-module-control-consumer should not be longer than 40 characters. [12:54]