=== verterok` is now known as verterok === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [04:01] Are there any Classic RPi kernel images that support the apparmor snapd requires? [05:36] qengho: ogra_ will know the answer when he's around === devil is now known as Guest57113 [06:15] hello [06:16] anybody is there? === Guest57113 is now known as devil_ === devil_ is now known as devil__ === devil__ is now known as devil_ [06:16] Girish: Why? [06:17] i have question to ask. [06:17] what about asking it then? [06:17] instead of asking if somebody is interested in you to ask [06:18] how to download source codes for kernel version 3.19.0-15-generic? [06:19] that's not really a snappy question [06:20] Girish, unsure if that's the one you want but https://launchpad.net/ubuntu/+source/linux/3.19.0-15.15 [06:24] hey, good morning! [06:25] hey dholbach [06:33] hey seb128 [06:38] hey dholbach! [06:38] salut didrocks [06:47] Girish: You should not expect anyone to commit to answering your unasked question. When on IRC, just ask, and say what you tried already to find the answer. [06:55] Morning dholbach, seb128 [06:55] hey niemeyer [06:56] qengho: Probably a good life advice too, even if a bit raw.. ;) [07:07] !help | Girish [07:07] Girish: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience [07:08] qengho: ^ that's a thing :) [07:13] Folks.. that's enough do-not-ask-to-ask advice for a life time :) [07:14] Girish: Sorry about that.. [07:16] yes, I apologize Girish :) [07:30] hey niemeyer [07:30] seb128: o/ [07:49] good morning [07:50] hey zyga, how are you today? [07:50] o/ zyga [07:51] hey guys, I'm good :-) [07:51] that's good :) [08:38] qengho, http://cdimage.ubuntu.com/releases/16.04/release/ubuntu-16.04-preinstalled-server-armhf+raspi2.img.xz === chihchun is now known as chihchun_afk [09:02] ara: coming [09:02] zyga, I am running late as well === cpaelzer is now known as cpaelzer_away === hikiko is now known as hikiko|bbl [09:38] hello anybody is there? [09:39] how to make a kernel module from a source code ? [09:40] hi [09:40] Girish: you have to build the whole kernel snap, you cannot ship one module [09:41] what does it mean? actually i am new in linux environment [09:46] actually i have to modify the source code in order to detect a LTE modem in its USB interface but the linux (ubuntu 15.10, kernel version 3.19.0-15-generic) that i m using have option.ko file instead of option.c file so i download the option.c file and modify it now i am confused how do i make its make file? [09:47] zyag are you there? [09:48] Girish: yes [09:48] Girish: are you doing this for a desktop system or an Ubuntu Core system? [09:48] Girish: I think the question nhaines asked is crucial [09:49] actually i m using lenovo desktop which is boot into linux through UEFI/LEGACY [09:50] Then this is the wrong channel, although I'm not sure which would be the right one. [09:51] Girish: then you just want to fix the issue in the ubuntu kernel, snappy is not a part of this [09:51] how to do this? [09:52] Grtry asking in the #ubuntu or #ubuntu-kernel channel then === chihchun_afk is now known as chihchun [10:15] jdstrand: do you think we can allow sched_setparam or switch seccomp to EPERM instead of killing the process? === dpm_ is now known as dpm [11:01] kyrofa: is this something we still want? https://code.launchpad.net/~kyrofa/snap-confine/create_user_common_data/+merge/293555 === cpaelzer_away is now known as cpaelzer === cpaelzer is now known as cpaelzer_away === hikiko|bbl is now known as hikiko [12:21] Hi. From what I understand currently systemd logs on snappy are written to SD card (in case of RPI at least). Is it possible to configure it to use tmpfs for logs? [12:22] zyga (cc tyhicks): we can probably allow sched_setparam with seccomp arg filtering (sched_setparam 0). as for EPERM, that is on the security team's todo list [12:26] zyga, no. It's no longer necessary with snap run [12:31] hmm, do we have any interface that allows setpriority ? [12:32] jdstrand: thanks [12:40] ...probably from chromium. [12:42] kyrofa: changed [12:42] qengho, heh, well, i'm packahing nw.js .. so chromium is kind of involved (in case you talked to me above) [12:43] runs fine in --devmode ... and makes an awesome webapp container ;) [12:46] niemeyer, zyga: hey, I'm having some trouble with go accessing the data in []interface{}. Can you look at: http://paste.ubuntu.com/18023741/ [12:47] jdstrand: looking [12:47] niemeyer, zyga: I feel like I am being dense or missing a critical piece of info [12:48] zyga, niemeyer: dang it, that had a typo. use this: http://paste.ubuntu.com/18023866/ [12:48] jdstrand: interface{} is "anything" but the real object behind is has its true type, what you want to do is to progressively check that the type is of a given kind until you reach the data you want [12:49] looking again [12:49] zyga: yes, I have no problem with interface{}. that is easy. eg: something.(string) [12:49] jdstrand: "baz" is a []interface{} probably [12:50] zyga: the problem is []interface{}. I can't .([]string) it and I can't range or iterate on it [12:50] zyga: it is [12:50] jdstrand: so you want baz, ok := slot.Attrs.([]interface{}) [12:50] I say that in the paste :) [12:50] and then access items in baz the same way [12:50] ok, let me try that [12:50] it's not a []string! [12:50] it's a []interface{} [12:50] and baz[0] may be a .(string) [12:50] right? [12:50] that's distinct from []string [12:51] because baz[1] may be .(int) [12:51] Go is still martian to me. [12:51] baz[0] is a .(string), yes. I understand why .([]string) doesn't work. I just couldn't figure out what *would* work [12:51] let me try the thing you mentioned [12:54] zyga: baz, ok := slot.Attrs.([]interface{}) [12:54] invalid type assertion: slot.SlotInfo.Attrs.([]) (non-interface type map[string]interface {} on left) [12:55] for baz, ok := range slot.Attrs.([]interface{}) doesn't work either [13:00] zyga: sorry, I've looked at this for far longer than I care to admit but I just can't seem to pull out the data. it seems like I need to iterate on slot.Attrs["baz"] so I can .(string) on baz[0], baz[1], but I can't figure out how to do that and I'm not sure if I should be doing something else [13:02] jdstrand: let me help you after the standup [13:23] jdstrand: can you push something more concrete? do you have a branch with a new interface? === cpaelzer_away is now known as cpaelzer [13:35] zyga: well, I have a branch that I am converting to using the yaml as described and it is severely broken because I'm blocked on this, but give me a few minutes and I can push that somewhere [13:35] jdstrand: ok [13:35] jdstrand: I'll help you out [14:00] zyga: https://github.com/jdstrand/snapd/tree/dbus-bind-broken [14:01] jdstrand: thanks [14:01] zyga: I'm using: go build -tags=excludeintegration -v github.com/snapcore/snapd/... && go test -tags=excludeintegration -v github.com/snapcore/snapd/interfaces/builtin [14:01] FYI, test implies build [14:02] zyga: you'll see soon enough, the problem is in SanitizeSlot [14:02] jdstrand: looking :) [14:09] zyga, hi [14:09] sabdfl: hey :-) [14:10] o/ sabdfl, how are you? [14:11] well thank you, sad at brexit [14:11] understandable frustrations, incomprehensible response [14:11] on the bright side... SNAPS! :) [14:12] yes! :) [14:12] sabdfl: have you seen the increasing amout of snaps in the playpen? https://github.com/ubuntu/snappy-playpen [14:12] sabdfl: lets not mention the England football result either [14:12] *amount [14:13] jamiebennett, there are some games where one side HAS to lose [14:13] other gae where everybody CAN win [14:13] games, that is [14:14] let's just say this was not a game where penalty shootouts were called for :) [14:14] tsimonq2, yes, great traction, and with distros too [14:15] sabdfl: I'm curious, speaking of other distros, what's your opinion on Flatpak? [14:16] tsimonq2, it will end up duplicating snaps, judging from the roadmap they announced, unfortunately [14:16] cant really prevent it, the nature of competitive instincts [14:17] sure, I get it :) [14:17] just wondering your opinion [14:17] i wish they wouldn't say nasty things about us to try to get support for their own efforts [14:17] it's fine to NIH stuff if you feel you have to do that [14:17] i uspect most fedora / rhel user will choose to use snaps if they are high quality [14:17] yeah, I recently started contributing to Snapcraft, and now I love snaps. I used to find them confusing, but I became pleasantly surprised [14:18] but it's fine that to have other options too [14:18] tsimonq2, :D [14:18] we just have to blog loudly about the great thinking that underpins the snap architecture [14:18] technology and quality should win these races, and we should challenge the other guys to leave the politics at home [14:19] yes, agreed [14:19] it's sort of ironic how we are supposed to have a universal package format now [14:19] but there is multiple formatsd [14:19] *formats [14:20] Snappy, Flatpak, and I hear of more [14:20] appimage ? [14:20] yeah I think that :) [14:21] sabdfl: on the note of me contributing, I am really glad I decided to work with the people in the Snapcraft team to get som code in [14:21] *some [14:21] I've learned a lot [14:22] pay it back by teaching someone else in the community later ;) [14:22] that's my plan :) [14:22] :D [14:23] tsimonq2, did you make a plugin? [14:24] sabdfl: well no, I added Subversion support [14:25] just as useful :) [14:25] \o/ [14:25] sabdfl: but I have a couple bugs assigned to me right now [14:25] for example, I'm adding checksum support for tar and zip sources [14:26] that's nice, will help all the different user communities [14:26] gtg [14:27] o/ [14:40] zyga: any luck? [14:52] jdstrand: I was pulled into a few meetings [14:52] jdstrand: I have vim open but I didn't touch it much [14:52] heh, ok, no worries [14:53] I have a meeting myself soon [14:59] jdstrand: ok, I'm more available now [15:41] jdstrand: sorry, this is a pretty interesting day [15:41] jdstrand: I'll be with you shortly [15:42] hi, about socket activated services in Snappy. Are they automatically started at bootup? [15:42] Or only when socket is accessed? [15:42] zyga: no worries. out of meeting and have a bunch of other stuff I can work on [15:42] The behaviour I am seeing here is : it's started at bootup. is this normal? [15:43] zyga: I am quite curious what I'm missing though === chihchun is now known as chihchun_afk [16:07] jdstrand: hacking on it now [16:08] jdstrand: quick note, pulsaudio 9.0.1 changes APIs a lot, we could drop the shm stuff entirely [16:08] jdstrand: we should have a backlog item for it [16:08] jdstrand: arch has it today [16:11] elopio hey, thinking about migrating the sources while we wait; is it just a matter of movinf the organization? [16:11] sergiusens: yes, afaik. You can ask niemeyer if he had to do something else. [16:11] sergiusens: hey, i'm trying to use the python3 plugin to pull in plainbox to my snaps, as per the plugin discussion, however i don't get entry points for the programs in the package created like i do if just run pip3 on my host - do you know why that might be? [16:12] sergiusens: oh, and reset the jenkins hook, set the team permissions, that too. [16:13] re-enable coveralls, update the badges in the README. [16:13] Chipaca, Hi! I've been trying to make a snap of httpie [16:14] Chipaca, I just realised that you did one a few weeks ago [16:14] Chipaca, But with a very different approach. [16:15] Chipaca, I am struggling with the relative modules and this snapcraft.yaml: https://github.com/pedrococa/httpie_snap/blob/master/snapcraft.yaml [16:15] sergiusens: Moving the repo? Yeah, it's straightforward, and GH internally maps the old links to the new place [16:18] Chipaca, Do you know how to overcome the relative modules issues with httpie, when you point to the repo and use the python3 plugin? [16:24] niemeyer yup. already done ;-) Sent an email and it is kind of cool that redirects happen automatically and all PRs just DTRT [16:25] elopio ok, I can take care of badges et.al. but no idea how to reset the jenkins stuff [16:25] wooooooooooooah [16:25] https://github.com/snapcore/snapcraft !!! [16:28] sergiusens: In setting of the project > hooks. Set http://162.213.35.179:8081/ghprbhook/, with permission to Pull Request and Issue comment. [16:28] sergiusens: I'm leaving now, I can help when I'm back. [16:28] elopio sure; all good [16:32] elopio heh, lost my admin :-P [16:32] niemeyer can you make me admin of snapcraft? [16:33] @sergiusens Of course [16:33] niemeyer: No such command! [16:33] joc_ they all have those script entries in setup.py, right? [16:33] Oops :) [16:33] sergiusens: Of course :) [16:35] thanks niemeyer [16:35] sergiusens: not too familiar with the format of setup.py but i think so, there are entry points declared like this: http://bazaar.launchpad.net/~checkbox-dev/checkbox/trunk/view/head:/plainbox/setup.py#L95 [16:36] joc_ yeah, console_scripts should just work, let me try something simpler to see if there's a hidden bug (later today though) [16:36] sergiusens: You and kyrofa are admins, and added the snapcraft-contributors team as write access [16:36] thanks! [16:37] np [16:37] Thanks niemeyer [16:37] sergiusens: ok thanks [16:38] sergiusens, kyrofa: I've also created a snapcraft-bots team.. not sure what level of access you'll want to give it, but let's please keep real contribs and bots separated [16:38] We've been doing that already on the other teams [16:50] kyrofa lool https://github.com/snapcore/snapcraft/pull/611 [16:51] sergiusens: LGTM, do you want me to do something to land it? [16:51] I dont think I'm in the snapcore org [16:54] lool no, just sending your way since you complained ;-) [16:55] lool I will wait for elopio on this one [16:55] sergiusens: I didn't complain! [16:55] I sent you a heads up in case you didn't about it [16:55] elopio btw, the hook is already configured [16:55] lool ah, well, a complaint I took in a positive way ;-) [16:55] sergiusens: FYI snapcraft/tests/test_sources.py mentions the old branches, if real checkouts are made and we ever remove the old branch, you would want to update them to the new one [16:55] lool, stop complaining that you didnt complain ! [16:55] :) [16:56] I'm being harassed by people who pretend I'm complaining, I'll file a complaint [17:00] lool those never reach the checkout phase :-) [17:00] sergiusens: ok, I didn't check [17:01] hence the conditional [17:01] and caveats === dieb_ is now known as Guest32438 [17:50] Hi, guys! [17:51] hello on http://snapcraft.io/ there is a little mistake in section Release channels. the channel option is mentioned with a single dash (-) yet in the section beneath it is mentioned with a equal (=) . just happened to see this and thought to inform you. have a nice day! [17:57] jdstrand: hey [17:57] jdstrand: sorry it took me so long, really unpredictable day [17:58] jdstrand: if you pull from my remote you will see how to use this [17:58] jdstrand: I can expand this anyway you like so please just have a look and tell me if that answers your question [17:58] jdstrand: and if not let's work on this some more [17:59] * zyga coffee and back to snap-confine [18:01] zyga: thanks! I see what you are doing. man, I was so close [18:01] I almost tried something like that. thanks! [18:03] jdstrand: syntax can be confusing :) [18:08] zyga: yes, and go's interface{} isn't the most intuitive thing in the world [18:19] jdstrand: it helps to read how it is implemented [18:20] jdstrand: it is essentially a proxy that knows how to dispatch each of the methods defined inside (note that {} doesn't have to be empty!) [18:24] Any eta. on a release image for Rpi3? [18:42] zyga: yeah, I did read how it was implemented. there was a nice article on it. it was the '[]' part of '[]interface{}' that threw me since I couldn't range it and it seemed like I should be able to [18:47] kyrofa su? [18:47] sergiusens, ah sorry, mentioned on telegram but should have put it here-- I'm not feeling well, I need to lay down [18:48] kyrofa: I hope you feel better [18:48] kyrofa ah, get better [18:48] Thanks guys [18:51] jdstrand: you can range it [18:51] jdstrand: you can range it and each element is a interface{} [18:57] slangasek: hey, I'm just doing this not to forget [18:58] slangasek: I'd love if you could commit debian packaging back to snap-confine master, this would help us in tetsing debina [18:58] slangasek: and then merge the one change back to debian, the change in debian/rules, specifically commit 2b130c575010a6f35a9f05d8683a761c8c02e2a0 [18:59] slangasek: this will become critical when we switch to snap-exec and when snap-confine will be started from a relative path [18:59] slangasek: which will be essential for core snap updates [19:01] zyga: I'm pretty sure I don't have commit rights on snap-confine master fwiw, but I will take time this week to review and raise pull requests [19:01] zyga: that is what I tried to do, but I didn't end up with the right incantation [19:01] slangasek: just send me a pull request or a branch on alioth that can be merged easily [19:02] slangasek: I looked at doing this but I gave up because master's weren't compatible [19:02] jdstrand: ack [19:02] hence the 'so close' comment [19:02] :) [19:02] anyhoo, thanks! [19:02] jdstrand: my pleasure [19:02] slangasek: I'd like to upload a sid vm to linode so that we can do per-commit tests on debian too [19:02] slangasek: and then having dpkg-vendor patches would be essential [19:14] josepht https://bugs.launchpad.net/snapcraft/+bug/1596044 [19:14] Launchpad bug 1596044 in Snapcraft "Parts loader needs refactoring" [Undecided,New] [19:15] sergiusens: is that different from the ordering refactoring you mentioned in the SU? [19:16] josepht nope [19:16] sergiusens: ah, what's missing is to update the jobs. I'm on it. [19:16] sergiusens: I see. I was reading it as the "parser's parts loader". I'll assign to me. [19:19] sergiusens: the tests can be triggered now. But the user can't update the status on the PR, it says: http://162.213.35.179:8080/job/github-snapcraft-examples-tests-cloud/1082/ [19:20] so you need to give snappy-m-o write access. [19:28] elopio ok, write access as in to comment? [19:29] the webhook or the user/bot? [19:32] sergiusens: the user. Write access as in pushing changes to the repo. [19:36] niemeyer hey, can you add elopio's bot to https://github.com/orgs/snapcore/teams/snapcraft-bots ? [19:37] niemeyer also, can you enable coveralls for snapcore/snapcraft, seems it needs an org admin [19:38] niemeyer: and please send us the token in case it's needed. [19:41] mhall119 give this a try https://github.com/snapcore/snapcraft/pull/612 you will get further but stuck on the fact you need to export alternate paths for GIR [19:50] attente hey you around? Can you tell me what would be needed to be able to open "resources" using a snap app; like https://telegram.me/joinchat/BQHZRAiN_pnOeE4mT31YCQ in the telegram snap? [19:51] sergiusens: what kind of resources? is it just a url opened via xdg-app? [19:52] attente yeah, probably, going to that link offers tg://join?invite=BQHZRAiN_pnOeE4mT31YCQ [19:52] sergiusens: the snapd-xdg-open helper only allows http, https and mailto [19:52] attente so it is the other way around than from the bug you were working on [19:52] attente right that is from snap to world [19:52] attente I face a much simpler problem now, from world to snap [19:55] sergiusens: that doesn't sound easier. wouldn't we have to hack the default xdg-open script that comes with xdg-utils? [19:56] attente I don't know, hence I am asking you :-) [19:57] attente does xdg-open look for desktop files to figure out uri handlers or are they registered at install time? [19:58] sergiusens: just glancing at the script right now, it looks like it examines the desktop files, so maybe there's some way to prioritize the ones in the snaps [19:58] attente right, right now I click and nothing happens :-) [20:00] sergiusens: maybe that's an apturl problem actually [20:01] sergiusens: er. sorry, never mind [20:03] sergiusens: is there an open bug for this? [20:12] attente no, just brought to my attention by pmcgowan :-) [20:14] sergiusens: is there a general directory where the desktop files of snaps is located? the same way that there's /snap/bin? [20:14] attente yup [20:14] attente /var/lib/snapd/desktop/applications/ [20:15] mvo or zyga can correct me if I am wrong here ^ [20:15] sergiusens: looks right here [20:17] sergiusens: maybe adding /var/lib/snapd to XDG_DATA_HOME in the session might make the xdg-open script look there first [20:17] attente so as it is right now it should be looking? [20:18] in my case, it is not opening anything at all [20:18] attente what desktop key would I need for this? [20:19] desktop key? [20:20] attente yeah, in the desktop file [20:21] i would've thought MimeType=x-scheme-handler/tg, but your telegram snap already has it [20:27] sergiusens, elopio: Still out for some exercising, but will do [20:53] sergiusens: can you try manually editing ~/.config/mimeapps.list to add x-scheme-handler/tg=telegram-sergiusens_telegram.desktop [20:54] this seemed to launch the snapped telegram for me [21:13] sergiusens: that didn't seem to help me any [21:14] sergiusens: does snapcraft cleanbuild install snapcraft in the lxc, or does it copy the host's snapcraft install into it? [21:15] attente, that worked [21:16] attente, although it seems to always open a new instance [21:17] hmmm, looks to install it via apt [21:20] yeah, same here [21:24] sergiusens: I'm cowboy-patching snapcraft in the lxc container created by cleanbuild [21:24] will see how it goes [21:24] * genii hands mhall119 the lariat [21:33] sergiusens: so I patched /root/parts/mail/bin/pkg-config and re-ran "snapcraft snap" and it did indeed get further, but it spit out a whole bunch of cmake warnings about link directories or something [21:34] oh wait, it might be something I did [21:37] ah ha, yes, I threw an extra print(env) line in there, and evidently something was consuming that [21:43] thanks sergiusens, that unblocks me on snapcraft now I think, I'll try and get vala/gir help from the elementary devs tomorrow