Package Details: snapd 2.32.5-3

Git Clone URL: (read-only)
Package Base: snapd
Description: Service and tools for management of snap packages.
Upstream URL:
Licenses: GPL3
Conflicts: snap-confine
Submitter: Barthalion
Maintainer: aimileus (bboozzoo, zyga)
Last Packager: bboozzoo
Votes: 26
Popularity: 8.682240
First Submitted: 2018-01-07 17:37
Last Updated: 2018-04-18 08:31

Latest Comments

lockheed commented on 2018-04-12 06:54

@bboozzoo thanks for clarification. Could you update arch wiki to reflect this change?

bboozzoo commented on 2018-04-12 06:48

@lockheed snap.refresh.timer has been removed in favor of an internal mechanism of snapd. It can be controlled by setting refresh.timer, eg. snap set core refresh.timer=0:00-24:00/4. See for more details on how to set refresh.timer. snap refresh --time shows last/next refresh times.

FWIW the syntax is more expressive and as we learned people try to do pretty crazy things with refresh times. Also, snap.refresh.timer was also acting as a workaround for bad (or nonexistent) support for monotonic timers in Go, what could cause the refresh to not happen at all (hence the separate, systemd-driven, timer which would kick the refresh anyway).

lockheed commented on 2018-04-11 20:24

Why is there no snapd.refresh.timer in this package?

systemctl start snapd.refresh.timer

Failed to start snapd.refresh.timer: Unit snapd.refresh.timer not found.

aimileus commented on 2018-03-23 16:37

@cmsigler, you are right that these files should be removed. Upstream includes a script to clean up the snapd files, but this does not include the udev rules. Since these rules are generated on the fly (one per snap), it is also not possible to include them in the package. For now I just decided to remove them manually in snapd.install.

cmsigler commented on 2018-03-21 15:47


I hope these aren't stupid questions.

A few days ago I installed snapd from AUR and tried it. The snaps I installed didn't work and that's fine, no complaint as it was purely an experiment. I removed snapd. This left behind one file I found today -- `/etc/udev/rules.d/70-snap.core.rules'. This is some kind of USB modem manager file(?). Please see:

1.) The package itself seems to generate this file? Does that mean it's not intended to go in /usr/lib/udev/rules.d/ ? Or should it? 2.) Is it possible to have the package manager clean up this file upon removal?

Just something I noticed. The answer to both may be "no." :) HTH and thank you.

Clemmitt Sigler

bboozzoo commented on 2018-03-12 06:31

@archlinux38 Can you post that to so that we don't spam the comments here?

The only think I can suggest now, having a cloned tree and while in the package directory:

  • go env and paste the output somewhere
  • export GOPATH=$PWD
  • try to build one of the components, eg. go build -x -v -o test-snapd once it fails, upload the log somewhere

archlinux38 commented on 2018-03-11 20:15

@aimileus I wish I could answer that but I can't. I don't think I have changed anything like that. Can you guide me how to look for those changes?

aimileus commented on 2018-03-10 15:50

@archlinux38, do you have a dirty go environment that might complicate things, e.g. a GOPATH/GOROOT environment variable set? Could you try building in a clean chroot, running extra-x86_64-build instead of makepkg?

archlinux38 commented on 2018-03-03 11:05

I have tried to build snapd for a couple of months in hope that it would somehow work some day but this is what I get every time I try:

Installing govendor


cannot load DWARF output from $WORK/b105//cgo.o: decoding dwarf section info at offset 0x0: too short ==> ERROR: A failure occurred in build(). Aborting...

anthraxx commented on 2018-03-01 11:56

@aimileus: thanks a lot, very much appreciated, more arch-ish now :D PS: i love your commit messages, thats how all AUR/repo message should be, explaining the why and whats. big thumbs!

All comments