[00:05] hi [00:07] the snappy guide starts off with the command snappy info, but in fact I get a login prompt - what is the username/password? === tyhicks` is now known as tyhicks === devil is now known as Guest11632 [06:44] good morning! [08:07] good morning [08:08] hey dholbach! [08:09] salut didrocks [08:10] didrocks, shall we talk tomorrow at 10:30 or 11? [08:10] not sure if we'll require thibautr, but we could invite him too [08:12] dholbach: not possible to have one catching up discussion today even? [08:12] good morning :) [08:12] the earlier, the better ;) [08:12] hey zyga [08:12] oh sure... maybe in the afternoon? [08:12] dholbach: good to me, pick a time [08:12] 16:30 or something? [08:13] dholbach: hum, I have to leave at 17:15 today [08:13] so that's not going to work I'm afraid :p [08:13] ok, that should work - we'll make it quick then ;-) [08:13] oh? [08:13] can't we talk for 30m? [08:13] dholbach: hum, I would prefer that we schedule it before, so that I don't have rush of last minutes email before leaving tonight [08:13] 15:30 - 16:00 then? [08:13] wfm! [08:13] will david c. be around? [08:14] (to show up what he worked on if he has anything already?) [08:14] I don't know [08:14] but you know about the progress, right? [08:14] I thought we'd also talk about the "getting feedback" bit I mentioned in my mail [08:14] I know what I worked on [08:15] dholbach: I got feedback from most of managements (olli, rick…) at mwc [08:16] right... [08:50] morning! [09:13] Good morning === zenlot1 is now known as zenlot [10:03] Hi guys, I have a question about skills [10:03] can i make my own custom skill [10:04] The skill just need to execute the wrapper executables of an other snap [10:17] Is someone active on the last day of februari xD [10:19] thanks mvo :) [10:19] noizer: not directly, you'd have to discuss this and propose a pull request to snappy [10:19] noizer: and I suspect that running an exec from one snap from another snap is a no-go [10:20] noizer: perhaps if you tell us what the goal is, we can suggest an alternative [10:20] zyga: yw! [10:20] noizer: and FYI, today skills are called interfaces :) [10:20] hahah right xD [10:21] mvo: can I do something to help land https://github.com/ubuntu-core/snappy/pull/518 [10:21] zyga I will make something how i can explain my problem to you [10:21] mvo: ? [10:21] noizer: note that you can prototype a new interface in isolation, but you won't be able to use it in snaps that go into the store really [10:21] noizer: so the earlier you know what you need the easier the road ahead is [10:23] zyga I know what i want just i will make a quick draw on my whiteboard en explain it to you [10:24] zyga: if it looks good I think the next step is to update all the integration test packages and rerun the integration test suite. it seems jenkins is a bit unhappy today so maybe not the best time? I would really like to land this only if jenkins is green [10:24] zyga: maybe you can check if https://github.com/mvo5/snappy/commit/806756730b1a04ace671aea202fb0c8517b46131 is ok? [10:25] zyga: no more git grep -i skill so we should be good in the branch :) [10:25] mvo: sure, understood [10:26] mvo: as for jenkins, let's see what happens next, if all else fails we can run integration tests locally [10:30] * zyga pushed https://github.com/ubuntu-core/snappy/pull/546 [10:31] zyga http://imgur.com/orqlhRD. Here you got a easy drawing how it needs to work. In the middle there is a main snap. This snap needs to controll all installed snaps. So he needs to start stop other snaps. So an very easy interface for users [10:31] But is would be easy that i can call the wrapper out of my own main snap [10:32] noizer: so you _can_ run other apps [10:32] noizer: in the same snap [10:33] noizer: you don't need interfaces for that [10:33] noizer: I would just suggest to call the raw executables, not the snappy-generated wrappers [10:33] noizer: try that [10:33] is that possible? [10:33] noizer: yes [10:33] hmmm that should be nicee [10:34] mvo: offtopic, upstream merged go-flags patch [10:34] mvo: it's low priority but if I prepare the debian patch, will you sponsor this for me? [10:34] but is it not a little bit better to call the wrapper? [10:35] noizer: no [10:35] noizer: in fact you cannot do that [10:35] zyga ok xD [10:35] zyga: I wonder if we should just push a new upstream release maybe? how much else did change? [10:35] zyga thx [10:35] noizer: the wrapper translates unconfined and vanilla environment to the confined snappy environment [10:35] zyga: last release in xenial is 20160218 [10:36] mvo: thanks, I was about to complain about lack of tags upstream [10:37] mvo, can we sync the version numbers among the arches ? i have actually a script i could run by cron (as interim) that could convert the tarballs daily [10:38] but with that we end up with the same version string across all snaps (which is probably wanted anyway, but currently the store has massively varyimg versions for ubuntu-core) [10:38] ogra_: yeah, I like that [10:43] mvo: not much: http://paste.ubuntu.com/15242051/ [10:43] mvo: (it took me forever to compute that, failing at the github ui first) [10:44] zyga: that looks like a new upstream is the simplest option [10:44] mvo: there's my patch and a trivial fix [10:51] so lets do this route [11:02] Chipaca: http://paste.ubuntu.com/15242182/ [11:03] Chipaca: some commands use 3rd person objective while some commands use imperative mode [11:03] Chipaca: what's the mode we want? [11:07] * zyga pushed https://github.com/ubuntu-core/snappy/pull/547 [11:38] zyga, mvo has the interfaces work landed into an image? [11:43] sergiusens: not yet, there is a branch for this but its not in yet, but almost [11:43] sergiusens: some, most notably old-security is still not merged [11:43] sergiusens: I suspect it will be all in today [11:43] zyga, mvo thanks; I'll keep track of https://bugs.launchpad.net/snapcraft/+bug/1549427 [11:43] Launchpad bug 1549427 in Snappy "migrate from skills to interfaces" [Undecided,In progress] [11:44] so please update when it makes it [11:46] sergiusens: I saw that, I'va ssigned part to myself [12:32] Good morning [12:34] Good morning xD [12:35] moaning ... [12:38] ogra_, not a good morning? [12:38] kyrofa, lots of mail from last week to wade through ... the curse of having been at a conference [12:39] ogra_, oh brutal, yeah I bet [12:39] (742 MPs in my "merges" mail folder ... 1200 in "bugs" etc etc ... ) [12:39] ogra_, hey, quick question-- I just flashed the dragon image onto an SD card, but when I boot the lights just remain solid and nothing happens. I was about to connect the serial board to see if it's complaining about anything, but have you seen that? [12:39] Yikes [12:40] kyrofa, hmm, did you use the very latest of my images ? [12:40] there was an issue with the board when no serial board was attached, that was worked around in the latest image [12:41] ogra_, dragonboard-rolling-edge-251.img.xz ? [12:41] (and will be actually fixed with the next u-boot (thanks ppisati btw !!!) ) [12:41] kyrofa, oooh, no, thats ancient [12:41] ogra_, that might be the issue :P [12:41] and my other one was moved away prior to the dragon announcement [12:42] ogra_, no longer here then? http://people.canonical.com/~ogra/snappy/dragonboard/ [12:42] i wonder if i can put it back ... sadly olli is off so i cant ask him for permission === Guest11632 is now known as devil__ [13:22] mvo, what happened to the branding of the shell prompt that we had in the 15.04 images, is that gone for good or just missing yet in all-snaps [13:23] * ogra_ finds it hard to actually know which board he is working on now that all of them say "ubuntu@localhost" ... with a handfull of boards runnning and a ssh terminal open on each it gets quite confusing [13:25] ogra_: stevebiscuit was working on the branding part recently but let me double check about the shell branding [13:27] it is indeed only an issue if you run multiple different boards ... so perhaps not that important [13:35] zyga I tried it now to call an executable from an other snap but it doesn't work [13:36] ogra_: uploaded a fix, will be part of the next ubuntu-core update [13:36] mvo, cool [13:36] * ogra_ remomembers he needs to add rfkill to writable-paths before next os snap [13:43] done :) [13:45] noizer: you cannot run apps from other snaps, just from your own [13:45] noizer: what do you want to do and why? [13:45] noizer: maybe I can help somehow [13:46] yeah, the correct answer would be to bundle everything you need into the same snap and manage it internally there [13:46] zyga I explained you earlier this day xD but here got it . I want a main snap. This snap needs to start other programs (snaps) [13:47] zyga http://imgur.com/orqlhRD [13:47] noizer: that diagram confused me [13:47] noizer: it has one snap and many apps [13:47] noizer: that _is_ allowed [13:47] noizer: running any apps from other snaps is not [13:47] noizer: if you control all the snaps just bundle everything [13:47] noizer: there are no "library snaps" [13:48] noizer: note that snap is the package [13:48] noizer: and app is something you can run in a snap [13:48] zyga ow wait you got 100 snaps with 1 main snap this main snap can start every snap on you device [13:48] noizer: snap == package, app == executable [13:48] noizer: a snap cannot have any snaps [13:48] sorry I did some mistakes for my vocabulary xD [13:49] noizer: so do you know what I mean now? [13:49] wait is that possible you have now the OS and that i build everything on my OS and then install snaps and that the build on my OS (not in a snap) can manage the snaps [13:50] zyga I know what you mean but I don't think you understand me correctly [13:51] noizer: wha you describe sounds more like ubuntu personal [13:51] noizer: where you just have a desktop and can pick apps to run [13:51] noizer: currently we don't allow this in ubuntu core [13:51] yes but the advantage of snappy that every snap is seperate from each other [13:51] noizer: I'd suggest emailing snappy-app-devel mailing list with the details of what you want to build [13:52] noizer: (that is true in personal as well) [13:52] noizer: and perhaps others can suggest what to do [13:58] zyga ok I will give it a shot [13:59] mvo, hmm, did the all-snaps u-d-f change the --install option ? i just built a dragonboard image with --install webdm.canonical ... it finished fine and i also saw it installing webdm ... but: [13:59] ubuntu@localhost:~$ snappy list|grep webdm [13:59] ubuntu@localhost:~$ [14:00] http://paste.ubuntu.com/15243621/ is the build log [14:01] ogra_: hm, no recent changes there [14:01] weird [14:01] ogra_: anything in ls /snaps? [14:02] ubuntu@localhost:~$ ls /snaps/webdm.canonical/ [14:02] 0.15 [14:02] ubuntu@localhost:~$ [14:02] no current link [14:03] GRRR ! [14:03] Rejected: [14:03] Unable to find distroseries: xernial [14:03] Further error processing not possible because of a critical previous error. [14:03] ubuntu-core-config (0.6.38) xernial; urgency=medium [14:04] * ogra_ fixes [14:08] silly xernial ... doesnt exist :P [14:20] fgimenez, seems like the jenkins thing is not completely fixed; I guess you can hand over to elopio_ now [14:20] but there was only 1 succesful run so far [14:20] ogra_, what are you pushing though? maybe push to distroseries 'ogra' ;-) [14:21] sergiusens, yeah what's going on with https://github.com/ubuntu-core/snapcraft/pull/344? [14:21] kyrofa, seems some of the backend is dead; I had one succesful run this morning though [14:21] sergiusens, I tried all weekend. Made me cry [14:21] sergiusens, indeed, i should push to omnious ogra :) [14:22] kyrofa, same thing; the error doesn't even help [14:22] sergiusens, no kidding [14:22] didrocks, here's the thing I'd mention I'd try out from SCaLE in case you want to look; it is very deep dive https://github.com/ubuntu-core/snapcraft/pull/335 [14:23] kyrofa, elopio_ since weekly is cancelled I guess we will do a regular standup [14:23] sergiusens, sounds good [14:23] sergiusens: oh nice! I'm going to give it a look (tomorrow morning probably) [14:24] sergiusens, elopio_ yep, the instances still look bad http://paste.ubuntu.com/15243858/ [14:24] didrocks, thanks; any ideas for improvement welcome [14:25] fgimenez, have you tried pinging IS directly on #is? [14:26] sergiusens, sure :) http://paste.ubuntu.com/15243874/ [14:27] fgimenez, oh, so end of week maybe? :-P but really :-( [14:29] sergiusens, hopefully it won't take too long, we have had issues like this before and they were solved quite quickly [14:30] sergiusens, dang... maybe start running locally and merging without? [14:31] sergiusens, I'll be 2 minutes late [14:57] kyrofa, oh, integration tests happening there or not? [14:57] sergiusens, yeah they will [14:58] I'll get the examples tests running locally, merge that other one, then write the integration tests and ping you [15:04] zyga: any idea how to run a snap app in gdbserver? === Saviq_ is now known as Saviq [15:04] the app segfaults and I wanna know why [15:05] ysionneau: hmmm, interesting [15:05] I already have gdbserver on the target, but since there is the sandboxing... [15:05] ysionneau: you can patch the wrapper script [15:05] and I cannot run the wrapper with gdb [15:05] ah yes [15:05] * ysionneau stupid [15:05] ysionneau: or just setup a manually crafted script that runs gdb and your app via ubuntu-core-launcher [15:06] ysionneau: or try it from the unconfined shell first, maybe you can test it better this way [15:06] my snap needs some lib [15:06] so I cannot run it without the wrapper which sets all the env [15:06] ysionneau: you can [15:07] ysionneau: the wrapper isn't magic, look at what it does [15:07] ysionneau: just set the right environment variables [15:07] ysionneau: you can try the whole wrapper with or without ubuntu-core-launcher (again, by manually hacking the wrapper script) [15:07] ack [15:07] ysionneau: though the launcher (not the wrapper, the actual launcher) does some interesting things that may be a factor too [15:08] ysionneau: I'd suggest trying without the launcher, just with altered environment [15:08] ysionneau: and do take advantage of classic dimension, apt-get install all the support tools and use them on the outside [15:10] ysionneau, or grab a core dump and analyze it with gdb separately [15:16] ysionneau, you can use the classic shell and attach to the PID as well :) [15:19] (oh, zyga said so already) [15:24] hi, I just installed snappy core for my RPi2, as it does not come with wifi support out of the box, could someone tell me what packages I need to install to get the wifi working ? [15:27] ace139, what wifi card is that ? the hw should definitely be recognized on the latest images [15:27] ace139: it should work out of the box, just edit /etc/network/interfaces.d/wlan0 [15:27] ace139: and add typical ifupdown wlan config there [15:28] ace139: quick test is to run ifconfig -a [15:28] ubuntu@localhost:~$ cat /etc/network/interfaces.d/wlan0 [15:28] allow-hotplug wlan0 [15:28] iface wlan0 inet dhcp [15:28] wpa-ssid grawert.net [15:28] wpa-psk XXXXXXXX [15:28] ace139: if you see wlan0 or something like that that will be a good sign that the hardware is supported [15:28] that is what i use here (on a dragonboard with builtin WLAN though) [15:29] (and indeed XXXXXXX is actually my wpa key) [15:29] ogra_, zyga the wifi adapter I am using, is not sensed somehow. [15:29] I have made changes to the interfaces.d/wlan0 [15:30] ace139: does it work on your regular ubuntu [15:30] zyga, yes, and on raspbian too [15:30] ace139, what chipset is that then, and what image (old 15.04 or the newer 16.04 ones) [15:31] ogra_, 15.04 [15:31] (though both latest images should ship a ton of wifi drivers) [15:31] ace139: try 16.04 [15:31] zyga, can I get a link ? [15:32] ace139: https://github.com/zyga/devtools/blob/master/ubuntu-image [15:32] ace139: try that :) [15:32] ace139, the latest and greatesd should be at http://people.canonical.com/~mvo/all-snaps/ [15:32] or try zyga's tool to build one [15:32] zyga, ogra_ thanks, let me try [15:32] :) [15:32] * zyga smells raspberry pi 3 in the mail tomorrow [15:32] snif snif [15:33] fresh hardware :) [15:33] tell us about the wlan then :) [15:33] I really want to be able to build my gadged and kernel nsaps [15:33] snaps* [15:33] its a shame it cant actually do 64bit from the start [15:33] for my work on interfaces and for doing community ports of other boards I have [15:33] ogra_: hehe, given what pi1 was I'm super positive they will get there [15:34] definitely [15:34] ogra_: remember how crappy it was in the beginning [15:34] ogra_: and it makes sense to ship hardware before the software [15:34] i dont, i only started using pi's with pi2 [15:34] ogra_: well, all the blobs, the uncertain situation of the kernel [15:34] * ogra_ doesnt care much about HW that cant run ubuntu :P [15:34] hi [15:34] ogra_: now it's all upstream and the GPU is getting fantastic work done in the open [15:34] mojo: hey [15:34] yeah === mojo is now known as Guest36587 [15:35] dragonboard actually has a free driver though :) [15:35] ogra_: is dragonboard using ardeno or something else? [15:35] * zyga lost track [15:35] (but a similar crappy closed bootloader made out of blobs) [15:35] zyga, yeah ... so freedreno should just work [15:35] ogra_: mmm, nice [15:35] it's amazing what so few people can make [15:36] yeah [15:36] all the gpu hacking, folks fit in a bigger lift together [15:36] a quick questions guys, can ubuntu core snappy run on a i386 system? [15:36] (hope they use the stairs though) [15:36] Guest36587: yes [15:36] i wish i would have been able to go to linaro connect to actually talk to the people to get this working in snappy OOTB [15:36] Guest36587: use https://github.com/zyga/devtools/blob/master/ubuntu-image to build a 16.04 based image [15:36] Guest36587: good luck [15:36] ogra_: the one in bangalore? [15:36] ogra_: when is it? I think ~week(s) from now? [15:36] yeah, i386 didnt get much attention yet ... might be a matter of luck [15:37] zyga, right [15:37] can it be installed as regular ubuntu image on a SD card? [15:37] Guest36587, you dd it to the SD and boot [15:37] Guest36587: just "dd" it to anything that your machine boots from, it can be a HDD as well [15:37] right [15:37] Guest36587: or you can use it to start a VM [15:37] ogra_: is it too late for you to go? [15:37] zyga, asac stole my ticket :P [15:38] aaahh [15:38] hehe, I knew that asac was going [15:38] ok. thank you i will give it a try. [15:38] 16:16 < ogra_> ysionneau, you can use the classic shell and attach to the PID as well :) < yep but it's not a daemon and it crashes right away [15:38] zyga, but then asac will have to implement freedreno in the dragonboard images, so all is fine :P [15:38] ysionneau: use gdb and run your program from gdb, then you can see why [15:38] ysionneau: or enable core dumps and inspect the core later [15:39] ogra_: haha, I'm sure asac will love that [15:39] for sure :) [15:39] * zyga spent all weekend staring at elf sections and program headers and x86 disassembly [15:39] but core files is something I have yet to play with [15:39] I haven't touched a core file in a decade [15:43] when running the script i get error unenxpected "(" error on line 391 [15:44] Guest36587: oh, let me see [15:44] Guest36587: hmm, the script has 214 lines [15:44] Guest36587: can you show me what you did please? [15:44] Guest36587: if I broke it I'll happily fix it :) [15:44] * zyga should add travis CI to the script [15:45] ok i did copied the script selected options i386 - then vanilla [15:45] Guest36587: can you pastebin the output of the script? [15:45] Guest36587: or just the error perha [15:45] ps [15:45] ok sec [15:47] Building image... /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d743ee368b2337f956765670f4128210dca7df71501035f7391: 1: /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d743ee368b2337f956765670f4128210dca7df71501035f7391: Syntax error: "(" unexpected [15:47] Guest36587: ahhh [15:47] * zyga looks [15:48] * zyga tries to reproduce [15:48] Guest36587: what is your host OS? [15:48] ubuntu 15.04 running on vbox [15:48] Guest36587: I was testing this on Ubuntu 16.04 (xenial) [15:48] Guest36587: I suspect 15.04 works as well though [15:49] I can check that in a moment, trying to reproduce the issue now [15:49] well i can install xenial and give it a try [15:49] Guest36587: can you check what is inside /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d743ee368b2337f956765670f4128210dca7df71501035f7391 [15:50] Guest36587: and perhaps one more thing: [15:50] Guest36587: can you pleas add 'set -x' to the top of the script (without the quotes) [15:50] seems like a data file [15:50] Guest36587: re-run it and pastebin the whole log [15:50] Guest36587: ok [15:50] Guest36587: my run will finish in a moment [15:50] ok let me set debug on it [15:50] Guest36587: so far all is good [15:51] (a few more minutes) [15:52] ogra_, since the pi3 has a 64-bit processor, can we still run 64-bit ubuntu core on there even if the standard OS is only 32-bit? [15:53] ok so running debug on script: [15:53] + udf=/root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d743ee368b2337f956765670f4128210dca7df71501035f7391 + chmod +x /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d743ee368b2337f956765670f4128210dca7df71501035f7391 + [ -n ] + [ -n ] + [ -n ] + test -z ubuntu-core.canonical + test -z canonical-pc-linux.canoni [15:53] Guest36587: thanks [15:53] Guest36587: please pastebin everything [15:53] Guest36587: (copy paste and paste it to paste.ubuntu.coM) [15:54] Building image... + exec sudo /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d743ee368b2337f956765670f4128210dca7df71501035f7391 core rolling --channel edge --kernel canonical-pc-linux.canonical --os ubuntu-core.canonical --gadget canonical-i386.canonical -o i386.img /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f1 [15:55] ca7df71501035f7391: 1: /root/.cache/ubuntu-image/blob.d8acf6b199a0a73def00dd61302afc99718cee101ec0c8d2ef845168ca9b5820ed943e90f17a6d7 [15:55] and the rest of it [15:55] he have a limited buffer here [15:56] Guest36587: it just built fine for me [15:56] Guest36587: you can try this: [15:57] Guest36587: ./ubuntu-image --no-developer-image i386 [15:57] Guest36587: ./ubuntu-image --no-developer-image i386 | pastebinit [15:57] Guest36587: ^^ this command will be non-interactive (apart from sudo) and will paste the output (though not the error streams) [15:57] Guest36587: ./ubuntu-image --no-developer-image i386 2>&1 | pastebinit [15:57] perhaps this one [15:58] Guest36587: for example this is a session I just ran to test it: [15:58] http://pastebin.ubuntu.com/15244574/ [16:00] hmmm [16:00] i'm missing something here [16:00] ubuntu-image i guess [16:00] :D [16:00] ? [16:01] i need to install it [16:01] ubuntu-image should run without any extra deps beyond what ubuntu-device-flahs (which it downloads) requires [16:01] somehow [16:01] Guest36587: the only thing that I can think of are tools like kpartx [16:02] i belive its simpler then that ;) [16:02] oh? [16:02] yup i dont have ubuntu device flash installed :O [16:02] noobie [16:02] ;) [16:03] Guest36587: that is normal [16:03] Guest36587: ubuntu-image downloads a special version [16:03] Guest36587: the version you can install won't work (though perhaps it will help to install other dependencies) [16:03] Guest36587: try installing it anyway and re-run ubuntu-image [16:04] but wait i was expecting that after installing ubuntu-device-flash, ubuntu-image would get installed [16:04] but no it seems that i need some aditional software [16:07] Guest36587: why would you expect that, ubuntu-image is a script that is not packaged anywhere [16:08] Guest36587: I'm not sure what happened or what actually caused that error to show up, try again (after you've installed ubuntu-device-flash) [16:08] Guest36587: or get a pre-made image from http://people.canonical.com/~mvo/all-snaps [16:08] (AFAIR) [16:09] thank you [16:10] Guest36587: good luck :) [16:11] ;) [16:13] zyga: if I want to allow an app to open some usr/lib/lib.so (via dlopen), I added /usr/lib/lib.so in the read-paths in security-override, is it the right way? [16:13] I cannot give the absolute path since it's in /snaps/name/some_hash/ [16:15] kyrofa, sergiusens: with the ros plugin gone, should the ros doc go as well? [16:15] ysionneau: hmmm [16:15] ysionneau: you _should_ be able to dlopen anything in your snap [16:16] dholbach, oh no-- there were two ros plugins, the useless one was removed [16:16] ah no... it's updated [16:16] fyi when the app dlopen, it segfaults [16:16] thanks kyrofa [16:16] dholbach, though the one on dev.ubuntu.com needs to be updated-- want me to do it? [16:16] ysionneau: I assume /usr/lib/lib.so is actually relative to your installed snap [16:16] kyrofa, no, I'll do it [16:16] ysionneau: what's the deny you get (tip, check syslog) [16:16] dholbach, okay :) [16:16] kyrofa, working through the whole list now :) [16:16] dholbach, heh, good luck! [16:18] I get this in syslogs zyga http://pastebin.com/szMDAht0 [16:19] I cannot see any deny, except from ulogcatd [16:19] the name of my app here is "wifid" [16:21] if in my code I do dlopen("/usr/lib/libfoo.so"); it will be able to open a lib in /snaps/wifid.sideload/INcVRgETGKeR/usr/lib/wifid/ ? [16:22] or should it be dlopen("libfoo.so") :o [16:22] ysionneau: you don't run with a chroot, you have to read appropriate $SNAP.. environment variable and dlopen stuff relative to that [16:22] oh, so the app needs to know it runs in a "snap" env [16:23] ysionneau: yes [16:23] that's not transparent [16:23] ok [16:23] ysionneau: or use a relative path [16:23] ysionneau: btw, offtopic: Feb 29 16:11:23 localhost kernel: wifid[1240]: unhandled level 1 translation fault (11) at 0x68730000, esr 0x83000005 [16:23] ysionneau: this doesn't look good [16:23] it crashes in the dlopen() [16:23] ysionneau: we've patched some common things to understand snappy better (bluez, network-manager), have a look at the approach taken there [16:24] ysionneau: basically we look at the relevant environment and change compiled-in constants to variables/functions [16:24] like env variable? [16:27]