/srv/irclogs.ubuntu.com/2019/06/13/#snappy.txt

mborzeckimorning05:15
mvogood morning07:03
mvomborzecki: hey, what did I miss so far :)07:03
mborzeckimvo: nothing :) i'm the only one online07:04
mborzeckimvo: unless you want to do a quick pass on https://github.com/snapcore/snapd/pull/6995 ?07:05
mborzeckimvo: as for the PR title check, what if we dumped the API and used beautifulsoup instead?07:06
mvomborzecki: oh, nice idea07:07
mvomborzecki: heh :)07:07
mvomborzecki: its old school07:07
mvomborzecki: scraping the web like its 1994 :P I like this07:08
mborzeckimvo: all we need is the title07:08
=== pstolowski|afk is now known as pstolowski
pstolowskimornings07:09
mborzeckipstolowski: hey07:09
mvopstolowski: good morning!07:10
mvomborzecki: yeah07:10
mborzeckimvo: maybe the stdlib html parser would be enough even07:10
mvomborzecki: did you look at the html already? does it look "easy" (sih)?07:15
mborzeckimvo: https://paste.ubuntu.com/p/PhwG32HG6Q/07:17
mvomborzecki: sweeeeeeeet07:18
mvomborzecki: thank you so much07:18
mborzeckipstolowski: can you take a look at https://github.com/snapcore/snapd/pull/6995 ?07:20
pstolowskisure07:21
mborzeckipstolowski: thanks!07:21
mborzeckimvo: remember the weird gadget.yaml from yesterday https://github.com/rpdeo/build-pine64-image/blob/a550cb83ba319b58042599c77548101146059386/snappy/gadget/meta/gadget.yaml ?07:38
mborzeckimvo: ubuntu-image calculates exactly the same positions as our code does07:38
mvomborzecki: good!07:59
mvomborzecki: so no risk of regression here07:59
mborzeckimvo: added a command to dump the info of a positioned volume https://paste.ubuntu.com/p/DSwjDP8zP3/ probably useful for debugging08:01
mvomborzecki: nice08:01
* mvo looks08:01
mvomborzecki: I updated the pr title check pr08:02
mborzeckimvo: thanks, looking08:03
mvomborzecki: I used the python stdlib to avoid having to worry about pulling in the right dependencies (but that is debatable I guess)08:05
mvomborzecki: actually I think I can improve my split, its a bit too silly (probably does not matter much though08:06
mvomborzecki: let me do it anyway :)08:07
mborzeckimvo: and there's ImportError due to lxml08:09
mvomborzecki: yeah, just noticed this as well, silly me08:09
mborzeckimvo: wondering if there's a command line tool for web scraping08:10
mborzeckimvo: ideally like curl .. | scrape '.head.title' | grep '<match>08:10
mborzeckimvo: check this out: https://paste.ubuntu.com/p/8R4Wkgdsxs/08:14
mborzeckimvo: this is the tool: https://github.com/ericchiang/pup08:15
mvomborzecki: interessting08:20
mborzeckimvo: and the PR is green :)08:41
mvomborzecki: yay08:42
mvonow if only the apparmor-id test would not fail :/08:42
mborzeckiappstream-id? :)08:42
mvoyeah, in my other PR that is ready to land it failed, I suspect the store is not happy but I did not look further yet08:43
mvomborzecki: https://paste.ubuntu.com/p/6f2rsHBgyg/08:43
mborzeckimvo: let me try to add some debug info there08:48
mvomborzecki: +108:48
mvomborzecki: its a bit annoying that we actually have no idea what data we got beside "computer says NO"08:49
mvomborzecki: we saw failures there a couple of times so definitely useful work to add debug there08:50
mvoa second review for 6691 would be great08:58
mvoto unblock ondra on this level08:58
mborzeckimvo: https://github.com/snapcore/snapd/pull/699609:01
mvomborzecki: thank you!09:02
mvomborzecki: do you think you can reply to "support parallel installs on desktop" in the forum? we want to do it but I am not sure where we stand right now (except that we want to do it :)09:02
mborzeckimvo: mhm, started writing a reply in the morning, but never sent it09:03
mborzeckimvo: the catch is the classic snaps & paralle installs, the guy mentions build tools, those would be classic usually09:04
mvomborzecki: yeah, eactly09:04
mvomborzecki: iirc we talked about some minimal fs namspacing around classic to support it(?)09:05
mborzeckimvo: yes, iirc zyga started looking into it at some point09:06
mvomborzecki: ok, I can reply as well in this case, I had hoped you had some more details :)09:09
mborzeckithunderstorm coming this way, might loose power for a while (hopefully not)09:10
mborzeckion the upside, it's raining now09:11
pstolowskimvo: state.json provided for yesterday's bug with current symlink. investigating09:11
mvomborzecki: good luck09:11
mvopstolowski: sweet, thank you09:11
mvopstolowski: please keep me updated09:11
pstolowskisure09:11
mborzeckiand it's gone, rained for 10 minutes :/09:16
Chipacamborzecki: just enough to keep the humidity up \o/09:31
Chipacamborzecki: for if you weren't feeling the heat properly09:31
mborzeckiChipaca: a taste of the tropics for $009:32
mborzeckianyone up for a review of this trivial PR https://github.com/snapcore/snapd/pull/6996 ?10:29
Chipacawhy would anybody include a .zip in a snap :-\10:31
Chipacamborzecki: WRT that PR, do you have an example run that failed?10:32
mborzeckiChipaca: i think mvo is the last one to see it fail10:33
Chipacamborzecki: I suspect the debug info you want is already in the log, but want to confirm10:33
mborzeckiChipaca: https://paste.ubuntu.com/p/6f2rsHBgyg/10:33
Chipacamborzecki: no, in 'debug output for …10:33
Chipacamborzecki: you should have the snapd.service logs10:34
mborzeckiChipaca: sry, that's all we have, idk which PR that was10:34
Chipacaah10:34
mvoChipaca: one sec, I find it for you10:34
Chipacata10:34
mvoChipaca: https://travis-ci.org/snapcore/snapd/builds/545102574?utm_source=github_status&utm_medium=notification10:35
mvoChipaca: in the ubuntu section10:35
Chipacamvo: mborzecki: 'context canceled'10:37
mborzeckiChipaca: timeout on nc side perhaps?10:37
mvoChipaca: oh, so it might be a real bug?10:37
Chipacamvo: mborzecki: probably10:37
Chipacawhy is it using nc for this :-(10:37
Chipacahttp would be a much better tool10:37
Chipacaor curl10:38
mborzeckiChipaca: the test does `nc -U -q 5 /run/snapd.socket`10:38
mborzeckiChipaca: heh i should fix that probably10:38
Chipacamborzecki: 1.949731ms after getting the request, nc hangs up10:38
mvota10:39
Chipacak10:39
mvoChipaca: this area of the code actually changed, yes? in this pr its about adding context to SnapInfo10:40
mvoChipaca: I wonder if there is a side-effect here ./10:40
mvoChipaca: it is using the request.Context10:40
Chipacamvo: the side effect should be that now we see the error :)10:40
sborovkovHi, is there a simple way to bypass this apparmor denial? Does not work in devmode snap as well, and the app I am debugging crashes right after/10:41
sborovkovapparmor="DENIED" operation="ptrace" profile="snap.network-manager.networkmanager" pid=1387 comm="NetworkManager" requested_mask="trace10:41
Chipacabefore snapd would've carried on waiting for the store, and tried to write the reply, to find that nc had gone away10:41
Chipacanow we see the error as soon as it happens so we get to see the log, probably10:41
Chipacamvo: that's my hypothesis anyway :)10:41
mvoChipaca: heh, ok10:41
mborzeckiChipaca: let's try with curl maybe10:41
Chipacamborzecki: k10:41
Chipacayou get the same denial in devmode?10:44
Chipacasborovkov: ^10:45
sborovkovChipaca, yes10:45
Chipacasborovkov: how do you do that?10:45
sborovkovscreenly-client      1.9.0.0         x6    candidate  -               devmode10:45
sborovkovI am not sure what actually does that10:45
sborovkovI updated Qt10:45
sborovkovnow qt webkit crashes10:45
sborovkovon core10:46
sborovkovnot on raspbian10:46
Chipacasborovkov: but that denial is not from the screenly snap, it's from networkmanager10:46
sborovkovyes, but it seems lijke it's communicating with it?10:46
sborovkovshould I try installing network manager in devmode?10:46
Chipacasborovkov: give it a try10:46
mvoChipaca: if I change the context that theStore.SnapInfo() gets to context.TODO() the issue goes away. it looks like the http.Request has an (aggressive) cancel context so probably not a good idea to reuse it(?)10:46
Chipacamvo: what happens if you use curl instead of nc?10:47
sborovkovChipaca, is there a syntax to force it to devmode with the same version of the snap?10:47
Chipacasborovkov: unfortunately there is no 'snap switch --devmode', no10:47
Chipacasborovkov: but you can 'snap refresh --devmode thesnap'10:47
Chipacadunno if it'll do what you need though10:48
mvoChipaca: heh, seems to be working10:48
Chipacaanyway, i gotta go, will bbi1h10:48
sborovkovthat won't work since it has no updates10:48
sborovkovI will just download the binary and sideload it10:48
mvoChipaca: so your theory is correct10:48
Chipacasborovkov: i'd suggest asking on forum.snapcraft.io, as zyga and/or jdstrand might be the best people to look into this10:49
Chipacasborovkov: (and they might not be around at the same time as you)10:51
* Chipaca really steps away now10:52
* pstolowski lunch11:19
mborzeckiChipaca: updated https://github.com/snapcore/snapd/pull/6996 to use curl now11:20
pedronisChipaca: could you answer this: https://forum.snapcraft.io/t/install-snaps-backup-or-export-and-relocation-to-new-pc/1177711:28
pedronisit's also probably in the wrong category11:28
=== ricab is now known as ricab|biab
* Chipaca writes <BLINK> for the first time in ages12:50
* Chipaca goes to have a quick shower before the standup because this smell will travel over hangout12:50
=== ricab|biab is now known as ricab
jdstrandpedronis: hi! I'm at the snapcraft summit and we discussed this topic and was wondering if you can comment on it soonish: https://forum.snapcraft.io/t/new-base-snap-nix-base/11787/213:07
jdstrandpedronis: I'm heading into a meeting, but can discuss via the forum or here. zyga can also speak to it13:08
kenvandinezyga: I have a snap that's failing to update13:08
kenvandinehttps://www.irccloud.com/pastebin/fWXYWFf1/error%3A%20cannot%20perform%20the%20following%20tasks%3A%20-%20Setup%20snap%20%22snap-store%22%20(111)%20security%20profiles%20(cannot%20setup%20mount%20for%20snap%20%22snap-store%22%3A%20cannot%20update%20mount%20namespace%20of%20snap%20%22snap-store%22%3A%20cannot%20update%20preserved%20namespace%20of%20snap%20%22snap-store%22%3A%20cannot%20update%20snap%20namespace%3A%20no%20such%20file%20or%20direc13:09
kenvandinetory)%20-%20Setup%20snap%20%22snap-store%22%20(111)%20security%20profiles%20(cannot%20update%20mount%20namespace%20of%20snap%20%22snap-store%22%3A%20cannot%20update%20preserved%20namespace%20of%20snap%20%22snap-store%22%3A%20cannot%20update%20snap%20namespace%3A%20no%20such%20file%20or%20directory)13:09
kenvandinethat was noisy :)13:09
zygakenvandine: hey13:09
zygakenvandine: wow, that's indeed noisy13:09
zygakenvandine: I'm at an event now and I cannot  focus on this; two small requests please: does it fail on edge core/snapd snaps and if so, can you please file a bug with the details13:10
kenvandinei have revision 117 installed and trying to refresh to 13613:10
kenvandineok13:10
zygathank you13:10
pedronisjdstrand: I'll see what I can do, I skimmed and to be honest I was a bit confused by it13:13
kenvandinezyga: fails with edge too.  See bug 183272513:21
mborzeckimkfs wrapper https://github.com/snapcore/snapd/pull/699713:32
cjp256any know of a method to fetch a snap from the store for arch different from the running system?13:38
Chipacacjp256: yes13:39
Chipacacjp256: UBUNTU_STORE_ARCH=i386 snap download core16 --edge13:39
cjp256thanks chipaca! :)13:39
Chipacacjp256: (from my bash history)13:39
Chipacawe should add an --arch, but, eh13:40
Chipacapriorities13:40
=== ricab is now known as ricab|lunch
zygakenvandine: thank you,14:08
ChipacaI've just been called to the boys' school, need to run off14:10
* Chipaca runs off14:10
Chipacaof course it was just when my tea was _perfect_14:10
=== Chipaca is now known as ChipAway
zygakenvandine: can you perhaps send the the two revisions14:11
zygamvo: hey14:12
mvohey zyga14:12
zygamvo: do you want to sync quickly or are you find with waiting for next week?14:12
kenvandinemarcustomlinson couldn't reproduce it14:12
kenvandineso it's something with my system14:12
kenvandinei even removed snap-store, installed stable and tried refreshing to candidate14:12
kenvandineblew up14:12
mvozyga: if you have anything important we can chat otherwise friday/next week is fine14:13
zygamvo: ok,14:13
zygamvo: just good news :)14:14
zygamvo: I posted a small PR https://github.com/snapcore/snapd/pull/699814:14
mvolooking14:14
zygawe're building an image with those patches for testing14:14
zygajamie also posted on the forum about nixos base snap14:14
zygaand if there's any confusion about that I'd love to clear it up (interactively)14:14
zygaso that nix can continue14:14
zygamvo: manjaro will seed snaps on the iso14:15
zygamvo: I'd like to spend a moment with someone to walk me through snap prepare-image --classic14:15
mvozyga: manjaro switches to /snap from /var/lib/snapd/snap ? do we need to do anything about this transition?14:15
mvozyga: on our side I mean, or is that just handlded in the packaging?14:16
mvozyga: I mean when users upgrade14:16
zygamvo: yes, we do14:16
zygamvo: it's in the patch14:16
zygamvo: that's more tricky, I think this kind of transition is tough in general14:17
zygamvo: I think you need to remove and reinstall snaps14:17
zygamvo: I'd like to change how we probe distributions14:17
zygabut it's not something we need this week14:17
kenvandinezyga: http://people.ubuntu.com/~ken-vandine/snap-store.tar.bz214:26
zygamvo: we're discussing the migration path14:28
zygawe don't have a good story for migration14:29
mvook14:32
zygamvo: we are working on a migration script14:41
zygakenvandine: thank you14:41
mvozyga: ok14:44
pedroniszyga: what's the question about prepare-image --classic ?14:50
zygapedronis: I'd like understand how to use it14:51
zygapedronis: so that manjaro can use it correctly to preseed the snap store stap14:51
=== ChipAway is now known as Chipaca
=== ricab|lunch is now known as ricab
pedroniszyga: it's relatively straighforward, it might not do the right thin yet if you have only core18 snaps though, and you need to specify all bases and all content providers and a model14:59
zygapedronis: we only want to seed the snap-store snap14:59
zygaso if you walk us through that that would be amazing15:00
Chipacamvo: d'oh, did /v2/download _just_ miss out on having context from day 115:00
zygaright now we (I learned) use a shell implementation of prepare-image that Wimpress shared15:00
zygaand I'd be much more comfortable with the proper approach15:00
pedroniszyga: something like this (ATM):  https://pastebin.ubuntu.com/p/9fm4p6Rvy9/15:09
zygalooking15:10
pedronisas I said atm is fetching core instead of snapd15:10
mvoChipaca: :/ I think so15:10
zygapedronis: does it fetch generic.model?15:10
Chipacamvo: (d'oh)³15:10
zygaor do we need to get it somehow15:10
pedroniszyga: no15:10
pedronisthat is like for core15:10
zygahow can we get it?15:11
pedroniszyga: snap known model on a ubuntu system for example15:11
pedronisor they can make a model15:11
pedronisbut for that they need brand account and keys15:11
zygapedronis: do you recommend that  they make a model?15:11
zygawe could do that15:11
pedronisif they want and are comfortable managing the keys15:12
zygacould you or someone else walk us through the process?15:12
pedroniszyga: they need a brand account if they really want this, I would chat with your store person there15:13
zygathat's what we will do15:13
zygaonce we have that, what else do we need15:13
noise][+1 for them to have a distinct model15:14
pedroniscreate a key,  register it, backup ~/.snap/gnupg15:14
zygapedronis: we'll work with daniel15:15
zygapedronis: and then we need to craft a model file?15:15
zygaI'm not familiar with that process15:15
pedronisthere should be instructions around15:16
pedronisgiven that is for classic it needs the bare minimum15:16
pedronislike our generic one15:16
Chipacazyga: you there?15:19
zygaChipaca: yes15:19
Chipacazyga: https://forum.snapcraft.io/t/request-for-a-base-snap-for-godot/11808/5?u=chipaca15:19
Chipacazyga: i don't understand15:19
Chipacazyga: maybe i'm missing something15:19
Chipacazyga: you were there, maybe there's more context than is stated in the post, that i'm missing?15:20
Chipacaor maybe there's something i'm ignorant of15:20
zygaChipaca: I can explain over a call?15:21
Chipacazyga: wfm15:21
zygapedronis: do we need a vault?15:21
zygapedronis: or can we keep authority-id as generic?15:21
pedronisah15:21
Chipacazyga: https://meet.google.com/fbb-gvae-mda?authuser=115:21
pedronisforgot about15:21
pedronisthat15:21
pedronisso no15:21
pedronisthey should just use generic15:21
pedronisthey would need a vault otherwise15:21
zygaoh15:22
zygahmmm15:22
pedroniswe cannot sign serial for them15:22
pedronisby definition15:22
zygaI understand15:22
Chipacazyga: i'll go make tea then15:24
zygaChipaca: stanndup15:24
zygaor meet15:24
zygahttps://meet.google.com/wmp-dfwj-jdt?authuser=015:25
cmatsuokamvo: I have a few unpushed recovery fixes (written in airport & plane), will test and push tonight after the event15:25
mvocmatsuoka: please do, I get errors that partition 4 cannot be created right now15:36
cmatsuokamvo: hmm, current HEAD was supposed to be functional (at least able to create everything and install), I'm running prepare on a new directory to see what's happening15:42
zygapedronis: quick question: can we today create a snap that can be content shared by anyone but is not published by canonical?15:47
zygapedronis: the use case is the godot runtime support snap for anyone publishing games via godot15:48
pedroniszyga: yes, it's a matter of setting up the right snap-declaration15:48
pedronisthat's how we do it also for canonical snaps15:48
zygaexcellent, thank you15:48
zygaCC Chipaca ^15:48
zygawe were just discussing that and wanted to double check with you15:48
Chipacazyga: pedronis: ah! good, phew15:48
Chipacazyga: pedronis: for some reason i was still stuck in the 'publisher is the same, or is "canonical"' world15:49
Chipacaglad that's moved on :)15:49
mvocmatsuoka: thanks! its a bit cryptic, the error just says "Could not create partition 4 from 23606048 to 3920448" no reason why is given :/15:55
* cachio afk15:56
pedronisChipaca: for content, canonical is not special I think, it's really same publisher or overriden by the declaration15:56
cmatsuokamvo: will check that and report back as soon as I have everything downloaded and built15:56
* Chipaca goes for a walk16:01
mvocmatsuoka: thank you, I will have dinner and check back16:01
mvocmatsuoka: looks like my core-build checkout was not in the writable-ramdisk PR16:13
mvocmatsuoka: eh, branch, let me try to figure out why16:13
cmatsuokamvo: oh ok that explains the problem16:13
mvocmatsuoka: yeah, sorry for the noise16:13
cmatsuokamvo: still building here, slow network and slow cpu16:14
cmatsuokamvo: the recent fixes are related to extracting the kernel and initrd from the snap16:14
mvocmatsuoka: thanks16:15
cmatsuokamvo: so recovering using a different kernel version was still booting with the original kernel (I didn't notice because in my tests I used the same kernel in both recoveries)16:16
cmatsuokamvo: to go through the entire recovery cycle selecting a different recovery version you'll also need the pc gadget on writable-ramdisk to recognize a new "recovery_reboot" snap mode16:17
mvocmatsuoka: I merged/uploaded your pc-gadget changes to 20/edge16:17
cmatsuokathis can be handled in a different way but that was the least intrusive one (but still required some changes in the grub configuration)16:17
cmatsuokamvo: ah nice, thanks16:18
mvocmatsuoka: I'm cleaning my dir now to ensure I have all the right checkouts now16:18
pedronishave the re-reg managers_test test passing, need some cleanups though (and a likely a couple follow ups)16:19
jdstrandpedronis: do you have a moment to talk about nix-base?16:32
pedronisjdstrand: I read the description more carefully, it seems ok if a bit of a degenerate (as used in math terminology) case16:36
zygamvo: do we use snapd snap by default nonw?16:36
jdstrandpedronis: nix is quite unique in that nix packages ship only the libs they need and everything is mounted under /nix16:37
jdstrandpedronis: so there is nothing needed in there. hence the analogy to static binaries and the bare base snap16:37
jdstrandpedronis: anyhoo. can you comment in the forum that you are ok with it?16:38
cmatsuokamvo: the re-prepared version worked here to install, didn't try recover yet16:40
pedronisjdstrand: do you know why it needs the /nix directory in it, we don't let layout create directories at that level?16:40
jdstrandpedronis: the ayouts feature does not allow creating toplevel directories under /16:41
pedronisok16:41
jdstrandpedronis: it is a limitation of the implementation16:41
pedronisjdstrand: it would be good if the topic mentioned this for clarity16:42
jdstrandI can add that detail16:42
jdstrandpedronis: done16:44
jdstrandpedronis: it isn't clear if you approve this (I haven't approved the snap due to this discussion)16:49
=== pstolowski is now known as pstolowski|afk
pedronisjdstrand: sorry, it took a bit to understand where the comment was added, I added a comment myself16:56
jdstrandpedronis: thanks!16:57
* Chipaca EODs17:11
mvozyga: we don't auto-transition people yet but if you only have core18 we install it18:01
mvocmatsuoka: yeah looks good18:02
albertosottilejdstrand: hi, any chance you could review our answer in the Syncplay classic confinement request topic?20:07
jdstrandalbertosottile: I commented in the topic. It is now moving forward and there is one more step (publisher vetting; cc popey_ and Wimpress)20:13
jdstrandhttps://forum.snapcraft.io/t/classic-confinement-request-syncplay/11189/1220:13
albertosottilejdstrand: understood, thanks. Please let me and Et0h know if we can assist in any way during the vetting20:14
popey_albertosottile: done20:36
albertosottilepopey_: thanks!20:37
popey_no problemo20:38
jdstrandalbertosottile: granted. please see the forum comment20:40
albertosottilejdstrand: thank you :) We will submit another revision asap20:41
jdstrandalbertosottile: cool. sorry it took so long. it was a bit of a different use case for classic and we wanted to make sure we considered everything20:41
jdstrandalbertosottile: thank you for your patience :)20:42
albertosottilejdstrand: I totally understand20:42
albertosottileWe also wanted to ask for an alias for a syncplay-server command, is now the right time to do so?20:42
jdstrandalbertosottile: yes. that will go quickly. please make a request in the forum20:43
albertosottileSure20:43
albertosottilejdstrand: https://forum.snapcraft.io/t/alias-request-syncplay-server/1182120:48

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!