=== chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [03:53] Hi there [03:54] I'm checking the snappy articles on developer.ubuntu.com and I saw a few mistakes [03:54] https://developer.ubuntu.com/en/snappy/build-apps/snapcraft-advanced-features/ → references a 'examples' directory that is no longer. It has been replaced by a 'demos' directory [03:55] https://developer.ubuntu.com/en/snappy/build-apps/your-first-snap/ → there is a formatting problem down around "Some more suggestions in our docs" === chihchun_afk is now known as chihchun [04:16] i want a chicken nugger and a french fried === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [05:06] haha [07:13] ePierre: can you please open bugs on that and poke dholbah or mhall119 about this [07:13] sergiusens: looking [07:14] zyga, ok! Do we have a project for the developer portal? [07:15] ePierre: perhaps, I don't know, target snappy on launchpad and they can get re-assigned [07:20] ara: o/ [07:20] hey zyga! [07:24] ara: it was worth putting that post up :-) There's already a person interested in contributing [07:24] zyga, cool :) [08:01] zyga You know about my problem with the tmpfs dir that isn't large enough. can I resize that so I can develop further? [08:23] noizer: tmpfs is using a fraction of your ram, unmount it and use regular storage there [08:23] noizer: it's probably setup by /etc/fstabl [08:23] noizer: it's probably setup by /etc/fstab [08:45] hi [08:46] Any artik owners around ? [08:55] how to build artik5-snappy-20160317 ? [08:56] RzR: hi, sorry I don't have an artik around [08:58] didrocks, hey! Thanks for your answer on the snapcraft mailing list! [08:59] didrocks, I thought by default, when using "python3", snapcraft would try to go to PyPi and download the latest available release. This is what it seems to do, so I'm not sure what the source would be for... But I'm gonna try anyway [08:59] ePierre: no, it doesn't know about the source, you can point it to a tar (and avoid naming the package) [08:59] Is it possible to have a part which just pulls data from an external source and puts it in the snap? Like a copy plugin, which does a git clone? [09:00] I can't figure an easy way to do this, but maybe I need more coffee [09:00] ePierre: so just use source: and point it to your tar, then setup.py will be run [09:00] popey: in copy plugin, you can do it [09:00] and use source: [09:00] then files: ['*': '.'] [09:00] oooh [09:00] (yeah, I tried to argue that being useless and default should copy everything) [09:01] I tried a long time, but failed [09:01] * popey tries this [09:01] hm, it doesn't like your format of files [09:02] Issue while loading plugin: properties failed to load for fgdata: [{'*': '.'}] is not of type 'object' [09:02] oh sorry, it's rather: [09:02] files: [09:02] - '*': '.' [09:02] ah [09:02] my bad ;) [09:03] same error [09:03] oh? [09:04] yeah [09:04] popey: mind pastebining it? [09:05] sure [09:07] didrocks: lp:~popey/+junk/flightgear [09:10] sorry no "-" [09:10] files: [09:10] '*': '.' [09:10] it's a dict :/ [09:10] my fault [09:11] popey: use make [09:11] popey: git + make == bliss ;) [09:11] oh, I see didrocks gave a better example :) [09:12] ahhh [09:12] thanks chaps :) [09:13] .oO( Wish I had a local cache for github like I have a cache for apt ) [09:20] yw! [09:30] Hi guys, I am closed to have fully debugged Freecad snap [09:30] I still face some issue with locales [09:31] got GTK-Warning: locale not supported by C library [09:31] any idea on how to workaround this ? [09:31] I have connected the application to locale-control but still the same issue [09:32] zyga its not set in the /etc/fstab file [09:32] zyga ow excuse me my mistake [09:43] zyga, noizer, fstab is generated by the initrd from /etc/system-image/writable-paths at boot [09:44] ogra_: Ok then its not good to change it there. [09:44] you wont be able to [09:44] it is readonly [09:44] (inside the snap) [09:45] ogra_: I Changed it with a chrooted xenial [09:45] ogra_: But how can i fix the no space left issue for now (just for development) [09:46] i dont know, you could try to unmount /tmp or some such ... so the disk is used ... but that will likely break everything that currently uses /tmp [09:46] so not sure that can work at all [09:48] I'll give it a shot [09:49] i suspect it will just break the world ... so you will have to wait til snapd learns to use a different temp dir to unpack [09:49] noizer: or bind mount /var/tmp over /tmp [09:49] bit try it :) [09:49] **but [09:52] zyga I did that : sudo mount --bind /tmp/ /var/tmp/ [09:52] zyga but without any succes [09:53] that's the other way around :) [09:53] zyga ogra_ Does the desktop version got more memory over there? [09:53] noizer: it's not desktop vs anything [09:53] noizer: it's just that tmpfs is a ram-based filesystem [09:53] well [09:53] noizer: so if you have 1GB of ram it will allow a fraction of that to be used to store files [09:53] the desktop version might find more ram :) [09:54] since your desktop is likely to khave mmore than a rpi [09:54] ogra_: OK i will try that [09:56] ogra_: where can I find the desktop image? [09:57] there is no image [09:57] snapd runs on the desktop and installs the core snap [09:58] when you execute a snap app it gets run inside that env [10:02] ooh, managed to make a 1.2GB snap. Fun :) [10:05] popey: what's in it? [10:13] data ! [10:13] (and little planes i guess) [10:14] lt cmd data!? [10:14] haha [10:15] ogra_: huh how? I don't understand [10:16] snaps are executed inside the core snap [10:16] (this is why it gets installed automatically when you install your first snap on a desktop) [10:16] oooh okay [10:18] mvo, is there any reason for us to still have snappy_ab= inside uboot.env.in ? (i just noticed thats still there) [10:18] zyga: flightgear [10:20] zyga, I got a patch for the bite-size snappy bug and a PR, what is left to do? https://github.com/snapcore/snapd/pull/1249 [10:20] ogra_: no, that can go [10:21] yeah, thought so [10:21] thx [10:21] thibran: hey [10:22] hi [10:22] thibran: looking :-) [10:22] thibran: looks good [10:22] thibran: if you want to go all the way I can show you how to add a simple test for this [10:23] yes, why not [10:23] (I'm a IRC n00b) [10:24] thibran: look at this file: https://github.com/thibran/snapd/blob/e3b02f8807934eec904d1dd81e67608f67137c98/cmd/snap/cmd_list_test.go [10:24] thibran: copy-paste the TestList function, call it TestListEmpty [10:24] thibran: in the copy-pasted function, change the copy of this line https://github.com/thibran/snapd/blob/e3b02f8807934eec904d1dd81e67608f67137c98/cmd/snap/cmd_list_test.go#L39 [10:24] thibran: to return empty results there... [10:24] fmt.Fprintln(w, `{"type": "sync", "result": [{"name": "foo", "status": "active", "version": "4.2", "developer": "bar", "revision":17}]}`) [10:24] change it to [10:25] fmt.Fprintln(w, `{"type": "sync", "result": []}`) [10:25] finally change the copy pasted version of this line https://github.com/thibran/snapd/blob/e3b02f8807934eec904d1dd81e67608f67137c98/cmd/snap/cmd_list_test.go#L49 [10:25] to read something like c.Check(s.Stdout(), Equals, "") [10:25] then run go test [10:25] and see what it really says (the test should fail) [10:25] and paste that thing in to the "" :-) [10:26] try that, if you need anything, just ask [10:42] thibran: how's it going? [10:43] zyge, have truouble running the test in go (dont know right now why) [10:44] thibran: what happens when you run "go test" from the cmd/snap directory? [10:44] works, but it shouldn't [10:45] thibran: can you git add the code [10:45] and push it [10:45] I'll have a look [10:45] git add -p [10:45] then you see what you are adding, very useful :) [10:47] I dont understand, u want me to push the broken test? [10:47] thibran: yes [10:47] thibran: I just want to see it [10:47] thibran: don't worry, it's not a problem [10:48] done [10:49] thibran: reading [10:49] let me try to run it [10:49] thibran: btw, did you see my 2nd post about how to set up development environment? [10:49] thibran: if it passes for you it may do so because you are still running the "stock" version of snap source code in reality [10:49] yes [10:50] I moved the forked repo into the name-space of the official repo [10:50] ah [10:50] I see what happens [10:50] zyga@x200t:~/work/src/github.com/snapcore/snapd/cmd/snap$ go test [10:50] No snaps are installed yet. Try 'snap install hello-world'. [10:50] No snaps are installed yet. Try 'snap install hello-world'. [10:50] OK: 63 passed [10:50] PASS [10:50] ok github.com/snapcore/snapd/cmd/snap 0.944s [10:51] we have a small thing in place to let us "redirect" (but not really) stdout for testing [10:51] let me guess, I have to use tabWriter? [10:52] thibran: http://paste.ubuntu.com/16887305/ [10:52] no, you just have to use Fprintf and print to Stdout :) [10:52] then the test fails and you can fix it [10:56] * zyga -> lunch [10:56] thibran: push the code when ready, I think it will land shortly :-) [10:56] thanks a lot for being here and welcome :-) [10:57] thanks for the help and the kind words [11:01] I signed the CLA, do I have to do anything to connect my launchpad account with my github account? [11:22] thibran, just make sure the email address you used to commit is on your LP account === hikiko is now known as hikiko|ln [11:40] So I'm using snapcraft to create a snap; I'm using the copy plugin, and I need Python installed. [11:40] I've added `stage-packages: [ python ]`, and when I run `snapcraft snap`, it goes off and does an apt-get update. [11:41] However, it uses my host's sources.list. [11:41] Doesn't this make the build difficult to reproduce in another location with a different sources.list? [11:53] And, next question: I'm now seeing a "Bad system call" error from the app I'm snapping; is there a good way for me to identify what's causing that (debug logs etc.)? [11:54] Odd_Bloke: dmesg | grep [11:54] probably find something at the bottom of that which indicates which system call is being used/abused [11:57] Aha, it looks like Python is being denied some things. [11:58] Specifically: http://paste.ubuntu.com/16888573/ [12:09] OK, so I'm seeing: ... apparmor="DENIED" operation="exec" profile="snap.google-cloud-sdk.gcloud" name="/sbin/ldconfig" pid=3420 comm="sh" requested_mask="x" denied_mask="x" fsuid=1000 ouid=0 [12:09] But /snap/google-cloud-sdk/current/sbin/ldconfig exists which I would expect to be usable (and the thing that is used). [12:09] What am I missing here? [12:11] /sbin/ldconfig != /snap/google-cloud-sdk/current/sbin/ldconfig [12:11] you are not in a chroot ;) === hikiko|ln is now known as hikiko [12:11] use ./sbin/ldconfig or sbin/ldconfig === hikiko is now known as hikiko|ln [12:12] (or the full path) [12:14] ogra_: Hmm, I _think_ it's Python that's making that call, any ideas how to instruct it to use the correct path? [12:15] well, python doesnt do that in other snaps [12:15] at least i havent heard of it [12:16] re === hikiko|ln is now known as hikiko === luke is now known as Guest40102 [12:29] does anyone know if I can install ubuntu onto my HTC one M8#? === seb128_ is now known as seb128 [12:34] Guest40102, not with snappy :) (wrong channel, i guess you want #ubuntu-touch) [12:35] ogra_: not _yet_ ;-) [12:35] yeah :) [12:42] hi any artik5 owners in this channel ? [12:43] i think didrocks has one [12:46] launching ubuntu-core-launcher snap.freecad.FreeCAD snap.freecad.FreeCAD /usr/bin/locale -a returns me C, C.UTF-8 and POSIX [12:46] any idea on why I do not see the other locale ? [12:47] I have change the app armor configuration to get access to the locale-archive but without any success [12:49] vejmarie: did you try to add locale to your snap (in snapcraft)? [12:49] yep it is inside [12:49] hmm [12:49] it looks like more an issue regarding access to the locale archive [12:49] didrocks: here's an odd one for you. I think I found a bug in the files copy thing. It seems to sometimes erroneously copy directories to the wrong place... http://paste.ubuntu.com/16889987/ [12:50] on the host I got all my locale through locale -a [12:50] but it is currently impossible to get access to the locale within the snap [12:50] maybe ogra_ remembers our locale conversation, but I don't know how to fix that; I think everyone solves this somehow but details elude me [12:50] I am trying to find out what we are missing [12:50] i am connected to the locale-control interface [12:51] zyga, well, i guess you would need something that calls locale-gen in the snap before snapping it to get other languages [12:51] ogra_, thx I'll ping him [12:51] though even that is probably not enough [12:52] ogra_: I tried it but still no way to make it work [12:52] http://bazaar.launchpad.net/~ogra/+junk/nethack/view/head:/nethack.sh [12:52] this is what i use to make loacles work in my nethack package [12:53] (currently not buildable, it is still using the old pre-interfaces snapcraft.yaml) [12:53] let me try that [12:53] thanks will tell you if it works ! [12:54] i have libc-bin, console-setup-linux and locales in my stage packages [12:55] line 7-21 in the nethack.sh script then actually set the locale before executing the game [12:57] If I run a snap that quits with "Bad system call" in normal mode and runs successfully in devmode, I should expect some message in klog about the syscall being refused, right? [12:58] ogra_: yes I have seen that [12:59] I believe that my I18NPATH was wrong I pointed out to usr/share/il8n/locales instead of the parent dir, I am rebuilding my snap will tell you what it gives ! [12:59] good luck :) === Aria|away is now known as Aria [13:01] ogra_: thanks [13:02] sergiusens: https://github.com/ubuntu-core/snapcraft/pull/435 what's happening with autopkgtest? I can't see the "details" [13:03] Ok thanks for telling me ill wait to hear something about iut [13:09] elopio: of the edge ppa you mentioned in -tech mailing list, should its version number be behind what is in x? foo, versus foo~ppa42 [13:10] Oh, OK, I actually needed the network-bind plug to make that error go away. === chihchun is now known as chihchun_afk [13:32] zyga is it possible to install a lubuntu-desktop in a snap and get so an userinterface [13:32] ? [13:32] noizer: put all of lubuntu-desktop in a snap? [13:33] noizer: maybe but then the integration with othe snaps wouln't be there really [13:33] noizer: yes but that's a lot of work and has many prerequisites [13:35] zyga maybe you have a other idea. I want to add a screen on my raspberry pi to view some things but for this there is a UI needed for [13:36] you wuold laso need to have a working xserver (or Mir) [13:36] *would also [13:37] noizer: yeah, get a mir/x snap in place and stick it there [13:43] zyga ok I will look into mir snap [13:43] ty ogra and zyga [13:46] ogre_: I got a localdef permission denied [13:46] * zyga sees ogre [13:46] ogra_: sorry [13:47] vejmarie, hmm, probably use the relative path in your script then [13:47] ok [13:47] to make it use the localedef inside the snap [13:48] (as i siad, that nethack package comes from a time where we had diffeent security rules, it can well be that localedef isnt accessible anymore while it was back then) [13:50] ogre_: I believe yes :( [13:50] I am not sure that the current security rules allows to support locale [13:50] which might be an issue for some application and translation [13:50] it surely does ... but you cant run the localedef from the os [13:50] (which my script does) [13:55] why don't we allow to read the locale ? Where is the security issue there ? [13:55] because presenting an app in english to a chinese person or french is not a good idea [13:55] the issue is that i.e. ubuntu-core doesnt have locales installed [13:55] ogra_: yes I have seen that [13:56] (this was the first place I looked at ;)) [13:56] but snapy should probably rely on the locale settings of the user who launch the snap [13:56] your snaps run inside ubuntu-core by default ... on classic interfaces can provide you bits from the outside world though [13:57] yes, I have copied the locales into my snap, but it looks like that ubuntu-core setting is getting rid of this :( [13:57] jdstrand, tyhicks: any objection for merging https://github.com/ubuntu-core/snap-confine/pull/14 [14:06] ysionneau that is behind a vpn, you can as elopio for exact details of what you are hitting [14:07] zyga: hey - just got out of a meeting - let me have a look [14:08] tyhicks: thanks [14:10] Odd_Bloke there's `snapcraft cleanbuild` that does an isolated build [14:12] Odd_Bloke wrt python I suggest looking at the python plugin. That ldconfig is probably run due to missing library paths. Another suggestion (I've been told it was easy to do). Is just grab python from sources. [14:13] sergiusens: The python{2,3} plugins look like they expect a setup.py; I don't have one for this. :( [14:13] Odd_Bloke: add one! [14:13] it's mostly trivial [14:15] Odd_Bloke oh, launch your app in a snap [14:16] Odd_Bloke with the env vars as if it "were" created from the python plugin [14:16] zyga: I don't have an objection to it per se, but the timing isn't great. also, shouldn't it be s/snap-run/snap-confine/ for the binaries? [14:16] jdstrand: that's piled in another branch [14:16] jdstrand: can you tell me more about the timing? [14:17] zyga can a webbrowser been shown in a mir snap? [14:17] just the various branches that are in flight. people have to adjust for stuff (again) [14:17] Odd_Bloke create the py3-example snap and copy the exported vars that show up in the snap directory (command nd wrapper are in the script names) [14:17] it isn't the worst thing in the world I just find myself doing a lot of redoing things... === chihchun_afk is now known as chihchun [14:24] mvo: hey, you here? [14:25] ok I have been able to make work localdef [14:25] but still have this stupid GTK issue :( [14:26] whats that ? [14:26] Locale not supportted by C library [14:26] Gtk-WARNING: Locale not supported by C library. Using the fallback 'C' locale [14:26] might it be coming from the lack of locale in ubuntu-core [14:27] that is crazy because the localedef is also in ubuntu-core [14:28] call "env" in your wrapper script and make it log somewhere to see what is actually set [14:29] noizer: I suspect so though I don't know any details [14:29] jdstrand: I hope the changes are minimal, only to list new files in Makefile.am [14:31] I think that LC_ALL is properly set but let's try to see how it works [14:34] jdstrand: btw, can we land the argument filtering branch [14:34] jdstrand: I saw you made some changes [14:34] jdstrand: I didn't do a 2nd review but I can do one quickly and we can have that one ticked off [14:39] zyga: I'm not done with my changes. turns out tyhicks's point about the stored data affected 32 bit builds, so I'm fixing that [14:39] vejmarie create an app under `apps` called "shell" and make the command be `bash`. Make sure the `plugs` and others match whatever you want to run [14:40] ok [14:40] jdstrand: ah, I remember that, the hash table thing [14:40] ok [14:40] my LOCPATH is set to /home/vagrant/snap/freecad/100001 [14:40] zyga: like I said, it isn't a huge deal [14:40] jdstrand: in that case I'd like to merge the autotooling as is and let you rebase / merge and tweak the two affetcted makefiles [14:40] where there is the en_US.UTF-8 [14:40] properly setup by localedef [14:41] my I18NPATH is set tp /snap/freecad/1000001/usr/share/i18n [14:41] zyga: I'm not blocking the autotools merge [14:42] where the i18n data are [14:42] vejmarie, and are there locales in /home/vagrant/snap/freecad/100001 ? [14:42] so if others agree-- that's fine [14:42] jdstrand: thanks [14:42] yes en_US.UTF-8 [14:42] jdstrand: did you review it? anything broken? :) [14:42] hmm [14:44] when it does export LC_ALL=en_US.UTF-8 [14:44] it is complaining that it can't find the file [14:47] vejmarie, well http://manpages.ubuntu.com/manpages/xenial/en/man1/locale.1.html ... see at the bottom [14:47] you might need LANG too [14:47] ogra_: want to review my autotool branch? I'm really hoping someone more experienced with autoconf/make than I am will say "this looks sane" [14:48] zyga: I didn't review it-- you asked if I had objections to landing it-- that's a different question :) [14:48] zyga, and you think that one would be me ?!? [14:48] I mean, I perused it [14:48] :D [14:48] zyga, i can look though [14:48] :) [14:48] ogra_: thanks [14:48] tyhicks: are you looking at the autotools PR? if not yet, do you want me to instead? [14:49] https://github.com/ubuntu-core/snap-confine/pull/14 [14:49] tyhicks: I'm of course fine if you are doing it-- I just didn't think we both needed to do it [14:53] Trevinho: hey, yes [14:53] jdstrand: I've already left comments [14:53] #!/bin/bash [14:53] export I18NPATH=$SNAP/usr/share/i18n [14:53] export LOCPATH=$SNAP_USER_DATA [14:53] LANG=en_US [14:53] mvo: hey, about the desktop file support in snappy... for unity7, I'd like to reiterate things a little [14:53] ENC=UTF-8 [14:53] LOC="$LANG.$ENC" [14:53] # generate a locale so we get properly working charsets and graphics [14:53] if [ ! -e $SNAP_USER_DATA/$LOC ]; then [14:53] $SNAP/usr/bin/localedef --prefix=$SNAP_USER_DATA -f $ENC -i $LANG $SNAP_USER_DATA/$LOC [14:53] fi [14:53] export LC_ALL=$LOC [14:54] This is what I put into my wrapper (sorry for the whole code) [14:54] jdstrand: I'm going to do a before and after build to compare the logs and then my review will be done [14:54] the export LC_ALL fails :( [14:55] mvo: I mean, things work ok when we launch an app from dash, since we use some startup notification foo to get better results [14:55] Trevinho: ok, what do you need? [14:55] mvo: however in case of multiple installed versions or in case we launch something froma terminal (or from different sources than gio app launcher, using a desktop file) there might be matching issues [14:56] elopio: can you tell me what's going wrong here please ? 162.213.35.179:8080/job/github-snapcraft-autopkgtest-cloud/813/ [14:56] tyhicks: sounds good [14:57] mvo: so, I've prepared a bamf version where if something is launched with BAMF_DESKTOP_FILE_HINT env var, pointing to a desktop fiel, then we use it... overriding any mathcing [14:57] Trevinho: aha, so we need to put that into the generated desktop file? [14:57] Trevinho: or how do you envision this to work? [14:58] mvo: not the desktop fiel, but in the automatically generated launch script. [14:58] mvo: is that something possible? [14:58] Trevinho: I need to look at it, it will have to point to the /var/lib/snapd/application/desktop/foo_bar.desktop file (the envvar)? [14:59] mvo: yes [15:00] mvo: I mean actually all the desktop files will point to a script that initializes some SNAP_* env var... Would be possible to make sure that if a .desktop file launches that, then you also add the env var on it? [15:00] err, all the launch scripts actually [15:00] Trevinho: ok, doable but a bit of work, when we generate the env vars we will need need to iterate over the dekstop files and match there. so yes [15:02] Trevinho: I add a trello card for this [15:02] mvo: thanks, I wanted to know whether a bug or a card was better, but... well, what's work best for you [15:03] jamiebennett: how would you like to track additional requirements like the one above? trello? bugs? if trello, which board? [15:03] Trevinho: I have no strong preference, but I want to double check with jamie first :) [15:03] mvo: I could even try to help in doing this, but my go or snappy knowledge isn't the best yet... So I could do with training :-) [15:04] Trevinho: ha! deal! happy to train you - go is really fun and the codebase is quite nice [15:04] mvo: Trello is best as that is what most people are tracking on a daily basis [15:04] The Vancouver board is getting new requirements so add them there [15:05] jamiebennett: thank you! [15:06] Trevinho: https://trello.com/c/aEB6d6r0 [15:06] Trevinho: feel free to hack away! [15:08] :) [15:16] Yeees it works ! [15:16] ! [15:16] popey: yeah, I guess this happens when you don't clean up first [15:17] popey: I guess bug report for kyrofa, he loves the copy plugin :) [15:17] :'( [15:17] dont we all ? [15:18] vejmarie, what was missing ? [15:18] I had a redefine in my script of LOCPATH which was not pointing anywhere [15:18] good to see FreeCAD fully working into a snap [15:19] vejmarie \o/ [15:19] awesome [15:19] also good to know that shipping your own loacales still works :) [15:19] popey, can I see the YAML you're using? [15:20] kyrofa: lemme try and re-produce it first with a clean config [15:21] vejmarie, nice job! Still in devmode? [15:21] popey, alright, let me know [15:23] will do [15:24] no longer [15:24] I focused on fixing that normally [15:24] noce one :) [15:24] *nice even [15:28] let me validate that and push it on the store if it works [15:28] +1 [15:45] kyrofa: ok, reproduced it, http://paste.ubuntu.com/16889987/ https://code.launchpad.net/~popey/+junk/flightgear is the yaml etc === Aria is now known as Aria|away === Aria|away is now known as Aria [15:49] popey, alright, let me play with it for a sec [15:49] Oh. More than a sec. It's not small [15:49] only 1.5G [15:49] * kyrofa grabs some more coffee [15:50] You know it's going to be a long day when the coffee is gone before noon [15:51] :) [15:51] sorry, i should have warned you :) [15:54] popey, haha, it's not a problem [15:56] done published if some of you want test [15:56] popey, can you confirm that parts/fgdata/build looks okay? [15:56] Hello. I have created snap with confinement: devmode in snapcraft.yaml. After doing snap install though I get apparmor "denied" messages. What am I doing wrong? [15:57] (it is still not passing automatic test due to the need to attach home and locale ! [15:57] I don't know who is doing the manual review but the process seems long ? [15:57] sborovkov, you still need to install with --devmode [15:57] sborovkov, eventually snapd will catch that and simply say "hey, you need to install with devmode" [15:58] Ah, ok. Thanks. === Aria is now known as Aria|away [15:58] kyrofa: parts/fgdata/build looks okay, same as ../src [15:58] sborovkov, there are no plans to automatically put it into devmode (a social security issue, if you will) [15:59] popey, yeah, this looks like a bug. An odd one [15:59] kyrofa: understood. I was just surprised by what's happening [15:59] vejmarie, yeah, manual reviews can take very long [16:00] sborovkov, once the snapd side of things is done I hope it'll be more clear :) [16:00] kyrofa: is snapd going to complain/inform in some manner in that case? === Aria|away is now known as Aria === Aria is now known as Aria22 [16:00] * jdstrand is curious what it is going to say [16:01] jdstrand, the plan as I understand it is to disallow installation of snaps that say they require devmode unless --devmode is provided [16:02] jdstrand, no idea what the message will actually say. Chipaca is working on it [16:02] jdstrand, and that's assuming of course that I understand correctly [16:03] hmmm === Aria22 is now known as Aria22|away [16:04] interesting. I figured complaining/informing that --devmode wasn't specified and installing as strict made sense. refusing to install as strict... not sure what that is intended to address, but I guess we'll see :) === Aria22|away is now known as Aria === Aria is now known as Aria22 [16:05] jdstrand, I think the idea is that if the developer knows the snap doesn't work under strict confinement, the only experience that can happen is a bad one [16:05] kyrofa: yay! [16:05] popey, I'm not even sure what you'd say in that bug... but do you mind logging one? [16:06] sure thang, against what kyrofa ? [16:06] popey, against snapcraft: https://bugs.launchpad.net/snapcraft/+filebug [16:06] kyrofa: i will try and create a smaller test case :) [16:08] one that is only 1.4GB ? [16:11] So helpful ogra_ :P [16:11] * ogra_ grins === Aria22 is now known as Aria [16:14] popey, ah, I figured it out [16:14] popey, make `files` be `'*': 'fgdata/'` [16:14] (note the ending slash) [16:15] popey, still a bug obviously, but now I think I know what's happening [16:15] will do! [16:15] thanks kyrofa [16:25] zyga: can you build the autotools branch as a debian package? dpkg-buildpackage fails for me [16:25] tyhicks, how [16:26] ogra_, zyga: https://paste.ubuntu.com/16897029/ [16:27] dh_install is what fails [16:27] looks like it doesnt build anything at all [16:27] it doesn't [16:27] there isn't any compiler output on stdout [16:27] right [16:28] I'll leave a comment in the PR === jkridner|work is now known as jkridner [16:44] jdstrand: do you think you could have a look at https://github.com/snapcore/snapd/pull/1251 ? it looks like python socket using apps on i386 needs this change [16:47] mvo: that is bug 1576066 [16:47] bug 1576066 in libseccomp (Ubuntu) "32bit glibc calls old socketcall() syscall, causing seccomp problems" [High,Confirmed] https://launchpad.net/bugs/1576066 [16:47] * jdstrand steps away for a few minutes === Aria is now known as Aria|away [16:58] ysionneau: sorry, I missed your ping. Note: Bypassing https://pypi.python.org/simple/file-magic/ (disallowed host; see http://bit.ly/1dg9ijs for details). [16:59] that's likely our firewall blocking the outside world. I'll make a note to request the firewall exception. [17:00] sergiusens: kyrofa: all green here: https://github.com/ubuntu-core/snapcraft/pull/532 What else? Get a review from somebody who knows more macaroons than me? === Aria|away is now known as Aria [17:07] elopio I believe that is good already, let me have a look again [17:07] or maybe nessita wants to look :-P ^ [17:08] qengho: hello (looking for old pings). Maybe we can now change that version, but the PPA was not made for people in x, it is for our image generators. [17:08] we might actually need a ppa for people who want to follow what will come to proposed. [17:12] I like PPA over Proposed because armhf is first class in PPAs. :P [17:18] sergiusens, is python? if yes I can certainly take a look === Aria is now known as Aria22 [17:23] sergiusens, so facundo will take a look at the PR, he implemented the APIs and the macaroon work === Aria22 is now known as Aria22|away === Aria22|away is now known as Aria [17:32] tyhicks: thanks for really trying my autotools branch, I fixed the thing you found [17:32] tyhicks: now running comparative builds [17:32] nice! [17:33] nessita yes, python [17:34] tyhicks: I'd say now the flags are stronger? [17:35] sergiusens, colin already left some comments that I think should be addressed [17:35] tyhicks: http://paste.ubuntu.com/16899272/ the - is the old packaging, + is the new packaging [17:36] +gcc -DHAVE_CONFIG_H -I. -I.. -Wdate-time -D_FORTIFY_SOURCE=2 -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -c -o snap_run-main.o `test -f 'main.c' || echo './'`main.c [17:36] -cc -c -D_GNU_SOURCE -O2 -Wall -Werror -g -O2 -fstack-protector-strong -Wformat -Werror=format-security main.c -o main.o [17:36] jdstrand, tyhicks ^^ [17:37] tyhicks: is this sufficient? [17:38] zzarr: looks like we lost -Wall and -Werror [17:38] oops [17:38] zyga: ^ [17:42] tyhicks: indeed, I can add those [17:44] ls [17:46] gcc -DHAVE_CONFIG_H -I. -I.. -Wdate-time -D_FORTIFY_SOURCE=2 -Wall -Werror -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -c -o snap_run-main.o `test -f 'main.c' || echo './'`main.c [17:46] tyhicks: fixed [17:46] tyhicks: note that lintian complains about some hardening things [17:46] slangasek: around? [17:47] slangasek: did you fix some of those in ubuntu-core-launcher +1 in debian? [17:47] http://paste.ubuntu.com/16899643/ (the list I see) [17:49] tyhicks: merged! thanks for looking :) [17:51] mvo: I commented in the PR but want tyhicks' input since this has been an ongoing prioritization discussion [17:58] jdstrand, tyhicks: https://github.com/ubuntu-core/snap-confine/pull/18/files [17:58] ideas welcome, should that cover more than apparmor? [18:01] tyhicks: I'll take that ^ [18:01] zyga: commenting [18:01] thanks! [18:01] perfect timing [18:02] sergiusens, hi! who can add facundo to the proper organization so he can comment in the PR? [18:04] thanks! [18:07] nessita you shouldn't need access for comment afaik [18:10] sergiusens, ack, thanks, I blindly trusted facundo's comment [18:11] well, thanks for the comments whatever they will be === Aria is now known as Aria22 [19:12] jdstrand: The "home" interface flags snaps for manual review in the store, what's the plans for that? Is it always going to require manual review, even though it doesn't get connected automatically? [19:17] beuno: how long after a snap is published in the store before it's availble in the Software app or in "snap find"? [19:20] beuno: ignore me [19:24] mhall119: there are changes coming in that area. home will be autoconnected and under certain circumstances it will be autoapproved [19:25] zyga: so, I can't figure out how to get to a clean state for commiting after the autotools update (without manually removing files) [19:26] zyga: I used to be able to do 'make clean' and then the tree would be clean for committing [19:26] jdstrand: can we share the timeline for those changes and the circumstances where it would apply? The krita snap is hitting this issue [19:27] zyga: now I need to run 'debian/rules build' (for autoreconf), do things, etc and when ready to commit, I don't see how to get back to a clean state. I've tried 'clean', 'dist-clean', 'maintainer-clean', 'debian/rules clean', etc. there are files left over [19:28] mhall119: use 'confinement: devmode' and it shouldn't block approval [19:28] mhall119, ignore you more than usual? that'll be hard! [19:28] beuno: not in a few weeks it won't be :-P [19:28] jdstrand: um...that won't block it for approval? [19:28] mhall119: it will autoapprove when specified with unity7 and/or x11 [19:29] jdstrand: ah, good, krita uses both, so the issue should eventually go away on it's own [19:29] mhall119: no. confinement: devmode and uploading to edge is precisely to not block it for approval [19:29] (for these confinement things) [19:29] edge is a channel? [19:30] when you upload, you specify the series (ie, 16) and the channel (stable, ..., edge) [19:31] zyga: I could git ignore them, but I'm thinking this isn't intentional [19:32] mhall119: the review tools changes should be committed to trunk in the next few days and land in the store hopefully end of next week === Aria22 is now known as Aria === chihchun is now known as chihchun_afk === Aria is now known as Aria22 [21:06] jdstrand: you can try make distclean [21:06] jdstrand: which files are the problem? [21:07] jdstrand: I'll get git ignore to behave better for all that autotool cruft [21:07] jdstrand: I think some of those are intentional but I was expecting dist-clean to remove all of them really [21:07] jdstrand: anyway, point taken, I'll make it nicer [21:08] I tried distclean [21:09] debian/rules clean ; make distclean is close, but a few things remain [22:11] elopio: if pypi is not allowed, then how are the "tests" passing then with the other dependencies?