/srv/irclogs.ubuntu.com/2019/06/14/#snappy.txt

sincere_foxHello00:56
sincere_foxI just installed an electron app via snap and I get this error https://pastebin.com/63MfH8NH00:58
=== epod is now known as luk3yx
mborzeckimorning05:39
mborzeckiChipaca: mvo: morning guys07:04
=== pstolowski|afk is now known as pstolowski
pstolowskimorning07:11
mborzeckipstolowski: hey07:12
mvohey pstolowski and mborzecki07:14
mborzeckihave you seen https://forum.snapcraft.io/t/new-base-snap-nix-base/11787 ?07:16
mvoyeah, its very cool07:17
mborzeckion the same note, wodner if any help is needed here: https://forum.snapcraft.io/t/base-runtime-freedesktop-sdk-runtime-19-08/11153/07:18
mborzeckiwould that enable a flatpak to be repacked as snap and vice versa?07:18
mvomborzecki: I think so07:19
mvomborzecki: or at least flat->snap it seems, not sure about the reverse but having a common base seems to be sensible07:20
Chipacamborzecki: morning07:21
mvopstolowski: mup is still down :( I pushed a tiny PR to not error if "current" is missing in "patch"07:21
mvopstolowski: please have a look, maybe its too naive07:21
mvohey Chipaca, gooood morning :)07:21
Chipaca:)07:22
mborzeckianyone up for some mkfs magic? https://github.com/snapcore/snapd/pull/699707:23
mborzeckior i shoudl say mkfs/mtools07:23
pstolowskimvo: thanks. i've been thinking about what to do about that.. looking07:28
mvopstolowski: again, maybe too naive but I think we don't want to fail07:28
mvomborzecki: I like your idea about the symlink /snap to avoid the full cost of switching in the majaro case07:32
mborzeckimvo: i'm bit afraid this would break when poeple grab snapd from AUR07:33
mborzeckimvo: which they have done in the past :/07:33
mvoyeah07:33
pstolowskimvo: okay, i think it's fine, let me comment in the PR07:34
mvopstolowski: \o/07:43
mvopstolowski: thank you07:43
pstolowskimvo: added comments07:47
mvopstolowski: thanks, makes perfect sense07:58
mvomborzecki: looking at your mkfs stuff now07:58
mborzeckimvo: ta07:58
mvomborzecki: its actually short, I like that07:59
pstolowskimvo: thanks for the change, one more remark for the test change you made07:59
mvopstolowski: thanks, looking!08:00
mvopstolowski: indeed, very good point08:00
mvopstolowski: thank you!08:00
pstolowskity!08:00
mborzeckiusual issues https://www.reddit.com/r/linux/comments/c07mp1/ubuntu_devs_testing_chromium_browser_transition/ snaps are slow, themes, auto refresh etc08:10
mvomborzecki: some ideas for your PR  looks great overall08:20
mvogentle ping about 6691, it needs a second review08:25
pstolowskimvo: #7000, what a milestone! :)08:35
mvopstolowski: indeed, 7k PRs its slightly crazy08:36
Chipacamvo: does brave work for you?09:15
mvoChipaca: it used to work09:16
mvoChipaca: I can try again, shall I use the script from my gist?09:16
mvoChipaca: or just deb/snap?09:16
mvoChipaca: what release of ubuntu?09:16
Chipacamvo: nah, just tell me if the snap's app runs09:17
Chipacahere it dies09:17
Chipacabut i'm on xenial09:17
Chipacawondering if it's universally broken or what09:17
Chipaca(it complains about the sandbox)09:17
Chipacawait, drat09:18
Chipacait does work09:18
Chipacahmm hmm09:18
mvoChipaca: aha, ok. its still installing here :)09:18
Chipacamvo: maybe i just needed to install it with assertions once for it to pick up the connections09:18
Chipaca:)09:18
mvook09:19
pstolowskigrr @go-flags09:40
Chipacapstolowski: no why?09:54
Chipacanow*09:54
Chipacamvo: grr @ snaps that do a bunch of copying stuff around on first run :-/10:16
Chipacaon an HDD, that's nearly a whole minute waiting for the window10:17
Chipacaimpressive10:17
Chipacas/nearly/more than/10:17
Chipacabrave:none:596299776:69398:3087:270710:17
mvoChipaca: ohhhh, so thats the reaon?10:17
mvoChipaca: reason?10:17
mvoChipaca: its the copying of stuff not actually anything else?10:18
Chipacathat's SNAP:MODE:SIZE:START:START2:WALK10:18
mvoChipaca: I'm very happy I asked you to look at this :)10:18
Chipacais 69 seconds for first run, 3 seconds for 2nd (from an empty cache)10:18
mvoChipaca: is that perf output?10:18
Chipacathat's my script output10:18
Chipacaone line of it10:18
mvoChipaca: *nod*10:18
Chipacaan evolution of yours :)10:18
mvoChipaca: nice find, out of curisoity, what is it copying around?10:19
Chipacarunning it on my laptop with its blazing fast ssd, this wasn't noticeable10:19
Chipacabah, a little: 6s to 4s or something like tht10:19
Chipacaon the gaming computer that has HDD homes, the above10:19
Chipacaanyway, will push the script and the results in a bit10:20
Chipacai don't have zstd though so somebody else needs to run those to get that10:20
Chipaca(where is zstd supported?)10:20
=== ricab_ is now known as ricab
mvoChipaca: I can build you a sqaushfs-tools with zstd10:21
mvoChipaca: its only available in git ./10:21
Chipacamvo: will my kernel know how to mount it?10:21
mvoChipaca: how do things compare to "on squashfs vs on ext4"? that is the most interessting one to me :)10:21
mvoChipaca: thats a xenial kernel? I can check, iirc zstd in kernel was done way before squashfs-tools supported it (which is kinda-ironic)10:22
mvoChipaca: 4.1.410:22
Chipacaxenial here, bionic on the hdd10:22
mvoChipaca: so should be in xenial, let me double check10:22
ChipacaCONFIG_SQUASHFS_ZSTD=y ← bionic10:23
Chipacanothing like that in xenial10:23
mvomeh, ok10:24
Chipacagnome-calculator:try:16560272:4577:3320:39110:24
Chipacagnome-calculator:xz:4218880:4120:2904:32810:24
Chipacamvo: ^ ext4 vs xz-compressed squashfs10:24
Chipacaabout the same10:25
Chipaca(granted g-c is small)10:25
mvoChipaca: silly me, misread - its 4.14 so no luck for you10:26
mvoChipaca: niiiiiiiiiiiiiiice10:26
mvoChipaca: I think that is a really good find10:26
mvoChipaca: well, second run is with hot cache?10:26
Chipacamvo: https://paste.ubuntu.com/p/82p3BFzwD9/10:27
Chipacamvo: no, both drop caches10:27
Chipacamvo: second run is without whatever the snap does on first run10:27
mvoChipaca: \o/10:27
mvoChipaca: you rock10:27
mvoChipaca: this data looks very encouraging10:27
Chipacamvo: https://gist.github.com/chipaca/2efec6e9d89cadada6478d8a010c5c2e10:28
pedronisChipaca: what are you actually trying?  me is a bit confused here10:29
Chipacapedronis: the effect of different compression options of squashfs on the speed of apps in a snap10:30
mvopedronis: we talked yesterday that one of the open questions is how much overhead squashfs/xz is compared to ext4. I had some concerns about this from measurements on brave (and from feedback from the advocy team)10:30
pedronismvo: talked when?10:30
Chipacaday before yesterday i think10:30
Chipacapedronis: you were off10:30
mvopedronis: do you have concerns about this ?10:32
pedronismvo: multiple, mostly process ones10:33
mvopedronis: as in "not the right time"?10:33
pedronisyes, also is untracked10:34
pedronisalso it's hard for me to help if I don't have context10:34
pedronismvo: if I look at trello I wouldn't know John is working on this10:36
mvopedronis: meh, sorry then. I updated the trello to have some info on it. also its hopefully not a long task, ~1d or so. but happy to talk some more during the standup and address concerns10:36
pedronisor that it was started at all10:36
mvopedronis: yeah, my bad, let me fix this10:36
mborzeckiChipaca: https://paste.ubuntu.com/p/wd3YjtbWbW/10:40
pedronismvo: is John also working on https://trello.com/c/mwEvN8Hk/237-squashfs-install-performance-fuse-sucks ?10:40
pedronisor just the one you moved to WIP10:40
Chipacapedronis: i am not working on that one10:41
pedronisok, I thought you were10:41
pedronisthat's why I was also confused10:41
Chipacapedronis: we did talk about that i was working on this yesterday in the standup10:41
pedronisI understood as the other one10:41
pedronisfunnily enough10:41
pedroniswe did talk about lxd, no?10:42
Chipacapedronis: i mentioned that tangentially i suggested the unpack-on-install approach as a global flag and that might be a solution to the fuse thing10:42
pedronisthat's the other thing though10:42
pedronisso my confusion10:42
mvopedronis: indeed, now it all makes sense. sorry for the confusion10:42
Chipacaheh10:42
Chipacapedronis: this was one of the things that distracted me a bit from working on the benchmarking script itself (which is now done for now)10:43
Chipacathat, and people being wrong on the internet10:43
mvoChipaca: heh :) I think we should write a forum post and maybe igor from advocacy can make it into a blog post). also would be nice to figure out what brave is doing, ideally we could suggest a fix to them (but timeboxed, lets not get dragged into a rathole :)10:44
mvoChipaca: anyway, very happy with these numbers10:44
Chipacamvo: http://paste.ubuntu.com/p/PVS5shj2QZ/10:48
Chipacamvo: that's 'find ~/snap/brave' after first run, starting from it being empty10:48
Chipacamvo: it's 37MB big10:49
Chipaca(so not big enough to justify how long it takes to do it)10:49
mvoChipaca: thanks! its probably also doing some sort of generation for this data. oh well10:51
Chipacamvo: desktop-launch10:51
mvoChipaca: oh? snap run --trace-exec brave should give us some clues about this10:52
Chipacaok let me do that on the slow box :)10:52
mvoChipaca: i need to get lunch now, will read backlog10:52
mvoChipaca: thank you!10:52
Chipacapedronis: you ok with landing #6978 then? i'm not sure, from your comments10:53
Chipacamvo: https://paste.ubuntu.com/p/sw3wGxgWzk/ (spoiler: it's update-mime-database)10:59
mborzeckioff to pick up the kids11:00
pedronisChipaca: ?11:16
pedronisChipaca: no, need to have a pass, mostly about those switch summary messages11:17
Chipacaok11:17
pedronisChipaca: made my main comments11:26
pedronisChipaca: I understand they from no channel came before, but given we are touching all of them it's a chance to improve that11:33
Chipacayep11:33
Chipacapedronis: but11:33
Chipacapedronis: can i address it in a followup?11:34
pedronisChipaca: worried about conflicts ?11:35
Chipacapedronis: wanting to get the pr that exposes all these things (with spread test) up11:35
Chipacaand it's stacked on this one11:35
pedronisah, ok11:35
pedronisChipaca: ack that in the PR11:38
pedronis*acked11:38
mborzeckire11:39
=== ricab is now known as ricab|lunch
pedronismvo: pstolowski: made a comment on #700011:58
pstolowskipedronis: thanks, thinking about it right now12:03
pedronispstolowski: also we have a bigger problem, what if snapd comes from snapd12:03
pedronisbut you have core installed12:04
mvopedronis: thank you, in a meeting right now but will look at it once that is over12:05
pstolowskipedronis: right.. this logic predates snapd snap.. the reset only happens for level 6 under these conditions, and patches should be idempotent though12:07
pstolowskipedronis: nb, we have 2 other issues i found that led to all this, we can discuss before/after standup12:07
pedronispstolowski: yes, the problem is that we wouldn't run them though12:07
pedronisbecause snapd will be updated before core maybe12:07
pedronisbut we look at core12:07
pedronis(not sure it depends on ordering issues)12:08
pstolowskipedronis: hmm i see12:09
zygahey everyone12:13
zygalast day12:13
pstolowskihey zyga! how is it going? tired?12:14
zygait's been a fantastic week12:14
zygalots of really amazing results12:15
zygaI strongly recommend reading the forum summary12:15
pstolowskizyga: yep i saw the tweets etc12:15
zygahttps://forum.snapcraft.io/t/snapcraft-summit-montreal-2019-day-1-2-3/11763/5 <- for anyone interested12:15
zygapstolowski: building firefox out of nixos packages in ~10 lines == mind blown12:15
zygaa little tired :)12:15
zygabut it's 14C here so I cannot complain12:16
zygahow was the week back home?12:17
pstolowskizyga: insanely hot.. with thunders yesterday afternoon, so a little relief12:18
zygaI'm off for breakfast, ttyl12:20
=== ricab|lunch is now known as ricab
=== ogra is now known as Guest59507
vladoskiI'm trying to install Heroku via snap, but everytime gets stuck on this and fails: error: cannot perform the following tasks:13:14
vladoski- Run configure hook of "heroku" snap if present (run hook "configure": execv failed: Permission denied)13:14
Chipacavladoski: strange13:17
Chipacavladoski: what snap is it?13:17
Chipacad'oh, "heroku"13:17
vladoskiChipaca: yep it's heroku13:18
Chipacavladoski: looks like a buggy snap13:19
Chipacavladoski: do you have /snap/bin ?13:19
vladoskiahm, so what should I do?13:19
vladoskiChipaca: yes13:20
vladoskiI have 4 snaps in it13:21
Chipacavladoski: what's the output of 'snap version'?13:21
vladoskisnap    2.39.1-1.fc3013:26
vladoskisnapd   2.39.1-1.fc3013:26
vladoskiseries  1613:26
vladoskifedora  3013:26
vladoskikernel  5.1.8-300.fc30.x86_6413:26
vladoskiChipaca: that's the output13:26
Chipacamborzecki: ^ ?13:28
vladoskiI'm using Fedora 30 if it helps13:29
vladoskioh nvm it was in the log ahah13:29
Chipacavladoski: yeah, mborzecki has been looking at issues on fc30 which is why i'm asking him to take a look13:29
Chipacajust in case it's a known issue / known workaround13:29
mborzeckivladoski: it's aknown problem, we have a fix for it and it's in fedora src repo already, we just need to update the package13:30
Chipacahuzzah13:30
vladoskiYes I know that snap is a little buggy now with fedora, one week ago I had to remove it because it was giving problems with battery status, function keys and system errors13:31
mborzeckiwondering if zyga or Eighth_Doctor could fedpkg build a new package maybe?13:31
Eighth_Doctorfor what?13:32
mborzeckivladoski: you can temporarily switch selinux to permissive, setenforce 0, install the snap, and then setenforce 113:32
mborzeckiEighth_Doctor: snapd13:32
Eighth_Doctoroh the fixes you sent in yesterday, right?13:32
mborzeckiEighth_Doctor: yup :)13:33
Eighth_DoctorI'll fedpkg build them now13:33
Eighth_Doctoronly reason I didn't was that I thought mvo was going to release 2.39.3 with that fix integrated13:33
ChipacaEighth_Doctor: hmm, yesterday mvo in the standup decided not to do a .3, and have it be distro-patched, so we could move on to .4013:34
mborzeckiEighth_Doctor: i think we'll do 2.40, let me double check with mvo13:35
ChipacaEighth_Doctor: we dropped the ball on letting you know13:35
Eighth_Doctorwelp13:35
mborzeckiah, missed that13:35
Chipacamborzecki: was that a "mvo should have told you to communicate that"?13:35
mvomeh, sorry, did I fail at doing this?13:35
zygamborzecki: next week at earliest13:35
mborzeckiChipaca: nope, i should have picked that up myself :P13:36
Eighth_DoctorI guess I'll update to 2.39.2 and push it13:36
mborzeckiso basically, /me dropping the ball :)13:36
Chipacamborzecki: yes but mvo should've double-checked :)13:36
Chipaca<--> etc13:36
* Chipaca lazy13:36
mborzeckiEighth_Doctor: that last 2 patches are not in .2, so please apply those separately13:36
Eighth_Doctorwill do13:37
mborzeckiEighth_Doctor: great, thank you!13:37
Eighth_Doctorthey're building now13:47
Eighth_Doctorzyga, mborzecki: I've been thinking that for Fedora snaps, we should look to plug into the modularity toolchain13:49
Eighth_Doctorinstead of trying to beat snapcraft into working13:49
Eighth_DoctorI'm tired of trying to figure out how to make snapcraft do what I want, and getting the modularity toolchain to do it would make it possible for Fedora contributors to officially build and submit snaps13:49
mborzeckiha13:50
mborzeckiEighth_Doctor: sounds like what the nix guys did13:50
zygaEighth_Doctor: I have the exact same feeling this week13:50
Eighth_Doctorand like the flatpaks/ namespace we have in Dist-Git, we could have a snaps/ one for people to make snap modules13:50
zygaEighth_Doctor: it is the most direct way to get the right things to happen13:50
zygaEighth_Doctor: oh, btw, I asked the godot deveoper to go to flock13:50
Eighth_Doctoryeah13:50
zygahis work on using fedora in snaps is amazing13:50
Eighth_Doctorwhich one?13:50
zygaand it  is really an eye-opening experience as to why this is done this way13:50
zygaEighth_Doctor: godot stack, including games, are all built from fedora bits13:51
zygaEighth_Doctor: but the story there is much deeper13:51
Eighth_Doctornice!13:51
Eighth_DoctorI knew that at least one godot dev liked the fedora stack :)13:51
Eighth_Doctorand that would be an awesome surprise to put into our presentation at flock: https://pagure.io/flock/issue/16313:52
Eighth_Doctorzyga, mborzecki: https://src.fedoraproject.org/projects/flatpaks/*13:53
Eighth_DoctorI want exactly the same thing for snaps13:53
mborzeckicachio: https://github.com/snapcore/snapd/pull/7004 this should work (works locally in xenial vm), i'll try to run one of nested core tests13:53
Eighth_Doctorzyga: I'm pretty close to having a new simple tool for producing a base snap13:55
Eighth_Doctorthere's a few things I'm changing up so that it works in a decently portable way across rpm distros without baking too much logic into it for that13:55
Eighth_Doctorbut also the part I don't know is uploading snaps properly without requiring snapcraft13:56
Chipacasiiiiigh13:56
Chipacahome isn't auto-connected on core, right?13:57
Chipacasi³³gh13:57
* Chipaca fixes the spread test13:57
cachiomborzecki, nice13:57
cachioi'll try it13:58
mborzeckicachio: i'm trying google-nested:ubuntu-16.04-64:tests/nested/core/hotplug13:59
zygaChipaca: correct13:59
cachiomborzecki, nice, I integrated that to the PR #6892 and I am running on 18.04 and 16.0414:01
cachioI'll tell you how it works14:01
Eighth_Doctorzyga, mborzecki: bodhi updates proposed14:02
Eighth_DoctorF30: https://bodhi.fedoraproject.org/updates/FEDORA-2019-17c0fb9ce614:02
Eighth_DoctorF29: https://bodhi.fedoraproject.org/updates/FEDORA-2019-91c0be0cc514:02
Eighth_DoctorEL7: https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2019-502ed7605314:02
Eighth_DoctorI need testing and karma14:03
mborzeckiEighth_Doctor: great, thank you!14:03
* Chipaca reckons Eighth_Doctor should have enough karma to launch a spaceship, at this point14:05
* Eighth_Doctor snorts14:05
mborzeckivladoski: you should be able to get a working version from updates-testing repo once it's there ^^14:06
vladoskimborzecki: okay thanks!14:07
* cachio afk14:09
* zyga is extremely sleepy and tired14:10
zygasorry everyone14:10
zygaEighth_Doctor: some more attention on the update, thank you for working on this !14:13
abeatojdstrand, hey, I have a couple of minor updates to interfaces waiting for review: https://github.com/snapcore/snapd/pull/6962 and https://github.com/snapcore/snapd/pull/697514:24
juliankI recently realized that snapped browsers break integration with KeePassXC, because it uses some native messaging thingy14:39
juliankbut I'm not sure what could be done about that14:39
pedronispstolowski: I reviewed #696014:39
juliankApparently this is super nasty stuff, where the browser invokes a binary on the host14:40
julianke.g. https://developer.chrome.com/extensions/nativeMessaging14:41
pstolowskipedronis: ty14:42
cachiopstolowski, when you have some time could you please take a look to #689215:11
pstolowskicachio: oh yes, looking now15:13
cachiopstolowski, thanks15:13
pstolowskicachio: so the problem with assertion disk & snapshot remains unresolved, right? no problem with the workaround, just curious, did Maciej have any other ideas?15:17
cachiopstolowski, no new ideas so fat15:18
cachiofar15:18
cachioI juest tested the change on how assertions disk is created15:18
cachioand works very well15:18
cachioI'll wait Maciej reply before merging this PR15:19
pstolowskicachio: ok, +1 from me15:23
cachiopstolowski, nice, thanks15:23
* cachio lunch15:47
=== pstolowski is now known as pstolowski|afk
=== ogra is now known as Guest19708
* cachio afk19:05

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