=== chihchun_afk is now known as chihchun [06:13] Can i package a wxpython based app as a snap? [06:27] PR snapd#2635 closed: SNAPD_DEBUG is a boolean (we use GetenvBool() for it) [07:12] PR snapd#2397 closed: interfaces: add iio [07:15] PR snapd#2529 closed: interfaces: mm: permissions for protocol proxies [08:03] tyhicks: woot, I just read the thead about complain mode for seccomp [08:03] nice :) === chihchun is now known as chihchun_afk [08:08] Can i package a wxpython based app using snap? [08:16] Tcab: hey [08:16] Tcab: it should be doable, just enough massaging to fit all of the bits in /snap [08:18] hello, is this an appropriate place to write feedback on Ubuntu Core tutorials that are on ubuntu.com? [08:27] eduardas_m: I would also recommend https://rocket.ubuntu.com/channel/snapcraft or the snapcraft mailing list [08:29] zyga, I do not know if this is appropriate, but I filed my minor gripe with documentation here: https://github.com/ubuntu/codelabs/issues/17 [08:29] Because that is where the "Did you find a mistake? Please file a bug." link directs me to from the tutorials page [08:32] AlbertA: hey, I'd like to discuss https://github.com/snapcore/snapd/pull/2256/files [08:32] AlbertA: can you please ping me when you have the chance === chihchun_afk is now known as chihchun [08:37] Ok thanks [08:39] Would be great if there was a simple examp,e / template for a wxpython app. It could thenbe used by many others [08:40] Wxpthon probably means compiling wxpython ? As part of the build process? [08:47] PR snapd#2648 closed: cmd/snap-confine: use flags rather than magic bool constants [09:09] PR snapd#2650 opened: also include system-shutdown helper in snapd.install [09:51] PR snapd#2600 closed: tests: remove the snapd dirs last (should fix random test errors) [10:27] Join [10:38] PR snapd#2651 opened: interfaces,overlord/ifacestate: small refactor around reference methods [11:26] Hi all [11:26] pedronis: thanks for the yesterday's link to the integration tests [11:27] man. that was a trip... [11:27] I used systemctl edit snapd.service to inject http_proxy for snapd daemon [11:28] it occured that this file, has to contain Service section before the Environment declarations [11:28] I've lost some time on it :) maybe it is worth to mention in a official documentations? [11:29] anyway, here is how /etc/systemd/system/snapd.service.d/override.conf should look like: [11:30] cat /etc/systemd/system/snapd.service.d/override.conf [11:30] [Service] [11:30] Environment=HTTPS_PROXY=bla:port [11:30] Environment=HTTP_PROXY=bla:port [11:30] Environment=http_proxy=bla:port [11:30] Environment=https_proxy=bla:port [11:31] without the [Service] section , does not work. [11:35] PR snapd#2256 closed: overlord/ifacestate: fix missing security setup for connected slot [12:11] PR snapd#2651 closed: interfaces,overlord/ifacestate: small refactor around reference methods [12:14] PR snapd#2328 closed: Add download manager interface [12:14] PR snapd#2652 opened: overlord/ifacestate: setup security of snaps affected by auto-connection [12:16] PR snapd#2653 opened: tests: skip i18n test when no "snappy.mo" file is available [12:21] ogra_: where is the right location of our reference gadget snaps these days? I see https://docs.ubuntu.com/core/en/reference/gadget#examples-of-production-ready-gagdet-snaps pointing to launchpad but I also see them on github.com/snapcore [12:24] morphis, https://github.com/snapcore/pi3-gadget and https://github.com/snapcore/pi2-gadget [12:24] updating the latter atm) [12:24] ogra_: then documentation really needs to be updated [12:24] yeah [12:24] otherwise we're pointing people to old stuff :-) [12:25] who owns that nowadays ? [12:25] davidcalle, ^^^ is that still your area ? [12:27] ogra_: I guess you can just do a PR against https://github.com/CanonicalLtd/ubuntu-core-docs or file a bug there [12:33] @ogra_ yes, PR is fine :) [12:33] davidcalle: No such command! [12:34] ogra_: please have a look at other links you find in the gadget area, in case there is more to update [12:34] will do [12:34] Thanks [12:51] hello, is the image ubuntu-core-16-amd64.img outdated when it comes to snapcraft tutorials in any way? [12:51] eduardas_m: where do you have it from? [12:52] morphis, http://releases.ubuntu.com/ubuntu-core/16/ubuntu-core-16-amd64.img.xz [12:52] I believe I am using this one [12:53] eduardas_m: should be ok [12:54] eduardas_m: worth running $ snap refresh inside to get the latest updates [12:55] the thing is I am trying to follow the snapcraft tutorials word by word... and I found something that works differently [12:56] eduardas_m: what is it? [12:56] this might be nitpicking, but I do not get the snap build failure described in https://tutorials.ubuntu.com/tutorial/create-first-snap#3 === ben_r_ is now known as ben_r [12:56] bash installs on first go without adding configflags: ["--infodir=/var/bash/info"] [12:57] so I thought either the tutorials are outdated or the image is [12:58] checked the sources of both GNU hello and bash [12:58] both use the same directory for infodir [12:58] but no conflict described in tutorial [13:00] Quick snapcraft question, which package sources are used for stage-packages - is it the /etc/apt/sources.list.. from the system where snapcraft is run? [13:02] jdstrand: comments changed as requested, thank you for the capabilities angle, I didn't consider this https://github.com/snapcore/snapd/pull/2630 [13:02] PR snapd#2630: many: detect potentially insecure use of snap-confine [13:03] PR snapd#2654 opened: i18n: look into core snaps when checking for translations [13:05] if possible, I would like an explanation: Ubuntu Core is for headless systems only or a GUI on a TFT LCD can be added for handheld device? [13:06] are there pre-build snaps that enable X server + desktop on Ubuntu Core or somehing similar? [13:11] longsleep, yea its from your hosts sources.list [13:11] stokachu: ok great thanks! [13:12] i'd prefer it had its own sources though so i could stage packages from zesty et [13:12] etc* [13:12] longsleep, np [13:14] Bug #1616629 changed: could not unmarshal state entry "snap-type" [13:14] PR snapd#2230 opened: Add an interface that allows clients to use media-hub over dbus [13:48] PR snapd#2448 closed: interfaces,snap: account for trusty-specifics in interfaces [14:05] PR snapd#2653 closed: tests: skip i18n test when no "snappy.mo" file is available [14:09] PR snapd#2575 closed: cmd: move snap-discard-ns to dedicated directory [14:11] Bug #1649934 changed: USB Auto-mount for assertion import fails [14:20] zyga: I should be done with version 2 of the seccomp patches this week so everything is still looking good [14:25] tyhicks: will you also work on snap-confine changes? [14:25] tyhicks: do you know if those patches will find their way back to the xenial kernel? [14:25] zyga: yeah, I should be able to do the snap-confine changes [14:25] zyga: I'll be backporting the kernel and libseccomp changes all the way back to xenial [14:26] tyhicks: and trusty I assume [14:26] zyga: ah, right... libseccomp changes will go back to trusty [14:36] PR snapd#2542 closed: many: use snap-confine to save cost of repackaging core snap for testing [14:38] PR snapd#2528 closed: tests: speed up update_core_snap_with_snap_exec_snapctl [14:45] jdstrand: is the new 'dbus' interface in snapd 2.20? [14:46] and is it provided by ubuntu-core, or only core? [14:46] mhall119: yes: https://github.com/snapcore/snapd/wiki/Interfaces#dbus [14:47] mhall119: I don't know the status of ubuntu-core vs core. it is in snapd 2.20. snap --version on an ubuntu-core system should tell you what it has [14:47] I think they are in sync again, but I don't know when that happened [14:47] jdstrand: I have 2.20.1ubuntu1, but "snap interfaces" doesn't show dbus [14:48] mhall119: no, it won't because it isn't an implicit slot. it is for providing snaps to use [14:48] popey: can you run "snap interfaces" and see if you have "dbus"? IIRC you switched from ubuntu-core to core [14:48] oh, nvm popey [14:48] it won't be there [14:49] install corebird-diddledan or ktuberling [14:49] I have ktuberling [14:49] ktuberling [14:49] Qt: Session management error: None of the authentication protocols specified are supported [14:49] "Couldn't register name 'org.kde.ktuberling-19990' with DBUS - another process owns it already!" [14:49] mhall119: refresh it [14:49] it's brand new from upstream [14:49] mhall119: this morning it was approved [14:50] mhall119: like, less than 1 hour ago [14:50] jdstrand: I got it from KDE's build servers directly [14:50] I have no idea what they have. I can tell you that what is in the store is implementing the dbus interface [14:51] jdstrand: which store channel? [14:51] edge r5 [14:51] * diddledan pricks his ears up [14:52] mhall119: also, 'another process owns it already' suggests that you have multiple ktuberlings running that are allowed to use the interface [14:52] so I think it is working as expected [14:53] I don't have other instances of ktuberling running [14:53] I suspect it takes a failure to secure the dbus service name as an indication of another being there [14:53] I don't know why it would say what it is saying then [14:53] perhaps something in the backgoround>? > [14:53] is there a way to list dbus session services? [14:54] install d-feet [14:54] there are other commands, but I don't know the invocation otoh [14:55] regardless, if you install r5 from the store then do 'snap interfaces' you should see an interface called ktuberling:session-dbus-interface [14:57] jdstrand: yup, and I can confirm that the versions in the store do in fact work and it secures the dbus name, so it's something wrong with the new builds, sorry to bother you [14:57] np [14:57] youtube believes this is relevant to my interests: https://www.youtube.com/watch?v=vPBM0g9usMs [14:57] (it isn't *wrong*, per se) [14:57] mhall119: what is the name of the content snap? [14:58] kde-frameworks-5 [14:58] mhall119: it doesn't seem to be installed when I do 'snap install ktuberling --edge' fwiw [14:58] ah, in --edge [14:58] IIRC, snapd doesn't auto-install needed content snaps yet [15:00] mhall119: ok, after installing, connecting the content snap and doing 'sudo /usr/lib/snapd/snap-discard-ns ktuberling' I can confirm that 2.20.1ubuntu1 works as advertised [15:01] that isn't surprising; ktuberling was one of the test snaps for the interface :) [15:01] jdstrand: right, I just unpacked the one from their build server and it doesn't have the dbus session interface in snap.yaml, I'll work with them to fix it [15:02] cool [15:04] zyga: ping [15:08] tedg_ or mterry, coudl one of you trigger a rebuild of the unity8-session snap ? theer were mir fixes in the overlay PPA that should make it work on the pi, i'd like to try them [15:08] ogra_: you got it [15:16] sergiusens, not sure if I'm asking the right person but here it goes: what's the difference between the ubuntu-core snap and the core snap? [15:16] PR snapd#2639 closed: snap: add {Plug,Slot}Info.SecurityTags [15:16] Kaleo, one has ubuntu- in the name [15:16] sergiusens, installing a local snap with classic confinement if you don't have the core snap installed (but the ubuntu-core snap instead): should it work? [15:16] ogra_, :) [15:16] Kaleo, ubuntu-core is deprecated and the snapd team is working on a migration path away from it [15:16] (they are identical ) === JanC_ is now known as JanC [15:17] ogra_, ah ok [15:17] but snapcraft requires the core one for classic builds i think [15:17] sergiusens, oh sweet, so will people eventually get core installed automatically? [15:17] (it checks for the name) [15:17] thats the plan [15:17] Kaleo, new installs already do, people with older installs will transition [15:18] sergiusens, sweet; I guess I'll follow https://bugs.launchpad.net/snappy/+bug/1655599 [15:18] Bug #1655599: No migration of ubuntu-core to core === souther_ is now known as souther [15:19] sergiusens, ok and so, installing a local snap with classic confinement if you don't have the core snap installed (but the ubuntu-core snap instead): should it work? [15:20] Kaleo, it won't [15:20] Kaleo, if you are not invested in snaps yet apt remove --purge snapd [15:20] if not, wait [15:23] PR snapd#2655 opened: cmd: switch to non-recursive make [15:25] sergiusens, I did not understand what you meant with being invested [15:25] sergiusens, thanks for the answers [15:25] Kaleo, using snaps a lot, as puring snapd makes you lose all your snap data [15:25] sergiusens, ah yes [15:25] sergiusens, ok [15:31] mterry, doesnt feel like it ... i just end up with a black screen (mir-kiosk at least gives me a cursor and turned on backlight on black screen, unity8 doesnt, backlight of the monitor goes off, no cursor) [15:33] http://paste.ubuntu.com/23822572/ is the unity8 log [15:34] looks like it misses the gallium driver [15:35] sadly i need to re-flash the SD so i cant test further ... [15:46] ogra_: well all I did was rebuild it after you mentioned -- does https://launchpad.net/~unity-team/+snap/unity8-session-silo/+build/17398 have the versions of packages that you expect? === cmiller_ is now known as qengho [15:49] mterry, i guess it does ... but i think we are missing a mesa driver that kgunn provides in his mir-kiosk snap [15:50] i cant imagine mesa-kms or mesa-x11 to work on a pi [15:51] hmm, if you know of a package we ought to include, I can add it. But drivers seem low-level for the unity8 snap, long term. Maybe we're including them now as a stop-gap tho [15:55] mterry, well, i'm not sure what mir-kiosk includes that you dont ... else i'D simply tell you :) [15:55] ogra_: sure :) I'll make a note to ask kgunn for his snapcraft.yaml === chihchun is now known as chihchun_afk === sjn__ is now known as sjn [16:25] mterry, hmm, it might actually be something different, i see the snap ships its own cgmanager service and tried to start it ... which clashes with the system cgmanager that is already running [16:26] ogra_: oh is that provided by a different snap now? [16:26] * ogra_ sees complaints in the logs [16:26] that was just to get us going [16:26] mterry, it is provided by the OS on the ubuntu-core pi2/3 images [16:27] though that shouldnt be any different to the ubuntu-core kvm images ... [16:27] perhaps its a red herring [16:27] i only noticed it complaining in the logs [16:27] sure, but that's a good bit of cleanup we can do regardless [16:28] not sure if snaps actually have access to it though ... when not running in --devmode [16:28] ah no interface yet [16:28] yep [16:28] PR snapd#2545 closed: overlord: allow max 500 changes in "ready" state to avoid growing changes for 24h === stgraber_ is now known as stgraber [16:49] PR snapd#2656 opened: spread: refresh apt cache before first install [16:50] PR snapd#2657 opened: snapmgr, ifacemr: add automatic ubuntu-core -> core transition === gbisson_ is now known as gbisson [17:14] hello folks. I want to build a classic-confinement snap. How do I set the "core_dynamic_linker"? [17:18] roadmr, you need the core snap [17:18] kyrofa: AH [17:18] kyrofa: thanks :) [17:18] roadmr, dreadfully unhelpful error, we know-- it's fixed in the next release [17:19] kyrofa: yes, I just found it by grepping around on snapcraft's source :) installing core now... yay [17:22] PR snapd#2656 closed: spread: refresh apt cache before first install [17:24] kyrofa: can't install core while ubuntu-core is installed, but can't remove ubuntu-core :( [17:25] roadmr, indeed-- you have to purge remove snapd I'm afraid. The snapd team are working on a migration to make that bette [17:25] r [17:26] kyrofa: mvo just proposed a pull request for this [17:26] zyga, nice! [17:26] kyrofa: it will be likely fixed next week [17:26] (next release) [17:26] zyga, thanks for the heads up [17:27] kyrofa: /o\ not ideal :) but it'll do. Purging! [17:32] kyrofa: my shiny classic snap is ready :) thanks! [17:37] roadmr, any time, sorry for the inconvenience! [17:45] ogra_: Thanks for your comment on bug 1576282. Could you perhaps try to summarize exactly how data behaves? I might not have my details right. My assumption was, that snaps are compressed, but at one point they will be uncompressed in memory or possibly swap space - so eventually 3.5M becoming 35M, or 5M becoming 125M if we wanted more locales, would still weigh down on the system [17:45] Bug #1576282: Snaps built from deb can't be gettext translated

[17:45] [17:46] kalikiana, they are just loop mounted as-is ... run "mount" on your pc ;) [17:46] zyga, this question could use you: https://askubuntu.com/questions/872991/trying-to-install-via-snap?noredirect=1#comment1355694_872991 [17:47] ogra_: Well, surely apps have to access them uncompressed since they're not aware of the filesystem details? [17:47] kalikiana, indeed if you copy anything out of the snap into a common dir or the user dir it gets uncompressed [17:47] kyrofa: looking [17:47] Maybe I'm just thinking about it the wrong way [17:48] kalikiana, they are uncompressed on the fly when accessed ... any only the bits that go into ram [17:48] by the kernels squashfs driver [17:48] that is how we fit a 2GB filesystem onto a 600MB CD since years already ;) [17:48] ogra_: Hmmmm so maybe I should look at the data, locales in the particular case, that the app actively uses, to see what memory it will require? [17:49] there is a loop mount and the inodes are exposed (so you have file references) on access data is decompressed on the fly [17:49] Right, I know the disc can be smaller, but those CDs usually require a lot of memory [17:50] well, but not 2GB ... it is really only what is actively open that occupies RAM [17:50] for your locales only the files in use will be in ram [17:50] but they wont need any diskspace or have to be uncompressed anywhere [17:50] thats the beauty of squashfs [17:52] you can safley measure a snaps size by its actual size ... only if you have any scripts that unpack stuff to a userdata dir or into the common dir your snap will occupy more [17:52] I guess I'm used to thinking worst case. But indeed the memory of a single locale would be small if all strings were loaded [17:52] yeah [17:52] and the unudes locales just sit in the compressed snap [17:53] if adding locales adds 3.5MB thats about it ... it wont grow in any way [17:54] this is why the ubuntu-core image effectively only occupies ~150MB for everything even though it would be closer to 500 if you actually unpacked all the files [17:54] Maybe we can get all locales in then - I was making a snap with only the subset used on the phone as the original data is quite big [17:57] ogra@localhost:~$ sudo du -hcs /snap/core/922 [17:57] 186M /snap/core/922 [17:57] 186M total [17:57] ogra@localhost:~$ df -h | grep core/922 [17:57] that might give you an impression ;) [17:57] err [17:58] ogra@localhost:~$ df -h | grep core/922 [17:58] /dev/loop3 65M 65M 0 100% /snap/core/922 [17:58] better :) [17:58] effectively it only occupies 65M ... even though the content is actually 3x the size [17:58] Anyone seen a snap fail to run with accessing /run/user/1000/snap. ? [17:58] http://paste.ubuntu.com/23823245/ getting that, maybe jdstrand knows? :) [17:58] it's installed in devmode [17:59] popey: "No such file or directory" [17:59] popey: mkdir -p /run/user/1000/snap.servo [17:59] there is a bug on that. let me find it [18:00] make the dir as part of the launcher? [18:00] as a workaround for the bug, yes [18:00] that'll do for now, thanks [18:00] I guess I should parameterise it? $UID or whatever for 1000 ? [18:01] ogra_: Impressive indeed. Thanks for enlightening me [18:02] popey: https://bugs.launchpad.net/snappy/+bug/1656340 [18:02] Bug #1656340: XDG_RUNTIME_DIR is not created on app startup [18:03] popey: no, XDG_RUNTIME_DIR is correctly set. just do 'mkdir $XDG_RUNTIME_DIR' [18:03] jdstrand: thanks [18:04] popey: there is an alternative way to do it in comment #1 [18:04] will subscribe to bug and workaround for now. [18:09] Ihow do I determine the key associated with my "brand", where I created my sso account a long time account. (myapps.developer.ubuntu.com lists my Account-Id but I am not sure which (if any) key that is associated with. [18:10] a long time AGO (I meant) [18:13] none of my local snap keys (snapcraft list-keys) have the same fingerprint as the Account_id listed in the store. [18:16] pedronis, is the relation between account id and keys documented anywhere? [18:28] Bug #1657552 opened: [spread] install-sideload:reexec0 failure [18:34] kyleN: are you logged in as that SSO account ? [18:42] jjohansen: hey, in case you are around, I had a look at the new kernel and I've collected some data (attached to the bug) but perhaps it would be better if you had a look and told me what to look for [18:42] jjohansen: it seems that the failure I get is different than before [18:42] jjohansen: (before == stock kernel) [18:42] jjohansen: in any case, I will be around for a few more moments, please ping me with instructions if you can [18:43] zyga: yep, just getting in, I will have a look at it [18:43] oh hi :) [18:43] great [18:43] pedronis, yes [18:46] kyleN: snapcraft can get for the store enabled keys that are not local, but doesn't have a way to print them afaik atm [18:46] s/for the store/from the store/ [18:47] pedronis, ok. I'll keep looking around. thanks [18:49] PR snapcraft#1024 closed: gradle plugin: update gradle plugin to support both gradle and gradlew [18:56] kyleN: this might print relevant info: python3 -c "import snapcraft.storeapi; print(snapcraft.storeapi.StoreClient().get_account_information().get('account_keys'))" [19:00] stgraber, have you seen LP: #1657252 ? [19:00] Bug #1657252: lxd no longer works with 2.21 [19:01] sergiusens: I believe I commented in it [19:01] ah, sorry, you have [19:01] I didn't refresh! [19:04] jdstrand, just so I'm clear, the serial port interface attributes you mentioned here does not currently exist? https://bugs.launchpad.net/snappy/+bug/1645445/comments/2 [19:04] Bug #1645445: Turtlebot needs /dev/kobuki [19:14] jdstrand, is there a way to manually connect? [19:27] PR snapd#2652 closed: overlord/ifacestate: setup security of snaps affected by auto-connection [19:35] sergiusens: is snap connect not working for you? [19:36] kyrofa: they all exist. see https://github.com/snapcore/snapd/wiki/Interfaces#serial-port. note this is slot side in the gadget snap [19:41] jdstrand, ah excellent, I misunderstood then [19:52] PR snapcraft#1055 opened: store: proper error colors for login failures [19:52] sergiusens: I responded in the bug [20:05] PR snapd#2655 closed: cmd: switch to non-recursive make