In progress: snapd 2.27.5


#1

The snapd 2.27.5 starts now.

Highlights of this release:

  • updated interfaces: default, mir, optical-observe, system-observe, screen-inhibit-control, unity, network-control
  • new interfaces: greengrass-support
  • snapctl work inside the snap context too (outside of hooks)
  • forced-devmode information is available via the sysinfo API now
  • seccomp argument filtering re-enabled
  • add --listswitch to snap abort and snap watch
  • add new snap commandline aliass: snap search and snap change
  • many test improvements
  • many bugfixes
  • shellcheck everywhere
  • snap list now shows the snap type in the notes (if it is not type: app)
  • support “title” via the store/rest-api
  • syslogIdentifier is now written to the generated systemd unti files
  • make snapd itself a “type=notify” daemon
  • opensuse,fedora spec files are part of the git tree now
  • auto-import system-user assertion only from ext4,vfat partitions
  • show sha3-384 hash in snap info --verbose snapfile
  • show snap-id in snap info
  • make config defaults from gadget work also at first boot
  • add support for android-boot
  • implement snap-update-ns
  • use /etc/ssl from the core snap
  • implement shortcut snap install --unaliased to install a snap without its automatic aliases snapd#3375

Snapcraft.io out of date
#2

This one should be considered too:

Support for androidboot.

(this is still WIP)


#3

And this one too:


#4

When is the cut off date for this?


#5

A first 2.27~rc3 candidate is in the “beta” channel now.


#6

@mvo it would be great to have https://github.com/snapcore/snapd/pull/3596 added to the release branch too once merged, it will prevent the snapd-notify error we were hitting last week.

Thanks!


#7

@fgimenez, I reproduced that error during beta validation, would be nice if for next RC we can have this fix.


#9

The current results for beta validations are being tracked in:


#10

snapd-2.27~rc4 currently does not build on Fedora for 32-bit architectures.

cd /builddir/build/BUILD/snapd-2.27~rc4/src/github.com/snapcore/snapd/cmd/snap-seccomp
CGO_LDFLAGS="-g" "-O2" "-Wl,-Bstatic" "-lseccomp" "-Wl,-Bdynamic" /usr/lib/golang/pkg/tool/linux_386/cgo -
objdir $WORK/github.com/snapcore/snapd/cmd/snap-seccomp/_obj/ -importpath 
github.com/snapcore/snapd/cmd/snap-seccomp -- -I $WORK/github.com/snapcore/snapd/cmd/snap-
seccomp/_obj/ -g -O2 main.go
# github.com/snapcore/snapd/cmd/snap-seccomp
In file included from /usr/include/xfs/xqm.h:21:0,
                 from src/github.com/snapcore/snapd/cmd/snap-seccomp/main.go:49:
/usr/include/xfs/xfs.h:51:12: error: size of array 'xfs_assert_largefile' is too large
 extern int xfs_assert_largefile[sizeof(off_t)-8];
            ^~~~~~~~~~~~~~~~~~~~

This issue has come up before in the distant past with snap-confine, and @zyga and @morphis fixed it before. It looks like the new snap-seccomp will need similar treatment.


#11

I’ll try to fix it now.


#12

What is current target date for 2.27? This Thursday?


#13

Hey Sergio, thanks a lot for the validation. It seems we have a few failures in there (it would be nice if the scenarios that have failures and need looking at would be marked somehow).

Looks like we have quite a few failures, where do I get e.g. the console-conf automated tests? Did you run the tests from the release/2.27 branch? I will try to reproduce the failures now.


#14

The target date is (tentatively) Monday, we need to get to the bottom of those test failures first though.


#15

Today or next Monday?


#16

Sorry, next Monday, we need to sort out the issues that Sergio reported first :slight_smile:


#17

If I could get comment rights to the gdoc, that would be great. I am going over the failed tests now (from top to bottom):

  • db res1: https://paste.ubuntu.com/25190950/: error: cannot copy request into temporary file: write /tmp/snapd-sideload-pkg-453861150: no space left on device
  • amd64: https://paste.ubuntu.com/25190256/:
  • MATCH 'activation done in' error: pattern not found, got: - I cherry picked 405d0e6 to fix this
  • 2017-07-28 02:40:50 Error executing external:ubuntu-core-16-64:tests/main/prepare-image-uboot : ... Fetching core core 48.58 MB / 69.93 MB 69.47% 49.25 KB/s 7m23ss6s <kill-timeout reached>
  • i386: https://paste.ubuntu.com/25190576/: `
  • + echo 'Remove hook was executed. It shouldn'\''t.': fixed with #3661
  • Fetching pi2-kernel pi2-kernel 92.91 MB / 96.31 MB 96.46% 132.82 KB/s 26ssss5s <kill-timeout reached>
  • snapd-notify test, same as above for amd64
  • pi2: https://paste.ubuntu.com/25223576/
  • snapd-notify: same as amd64
  • install-hook: same as amd64
  • prepare-image-uboot: same as amd64
  • pi3: https://paste.ubuntu.com/25223117/
  • 2017-08-01 20:03:07 Error executing external:ubuntu-core-16-32:tests/main/failover:rclocalcrash : kill-timeout reached, cannot reconnect to external:ubuntu-core-16-32 after reboot: dial tcp 192.168.1.10:22: getsockopt: no route to host

#18

The failure on refresh in:


#19

The failure: 2017-08-01 00:22:49 Error executing external:ubuntu-core-16-64:tests/main/interfaces-account-control : should be fixed once we cherry-pick http://github.com/snapcore/snapd/pull/3664 to release/2.27


#20

I released a new 2.27~rc7 into “beta” that should fix most of the above failures.


#21

The 2.27 release is now in the candidate channel. Please help testing by snap refresh --candidate core and let us know if everything works as expected. If we don’t find any regressions we plan to release 2.27 (-final) to stable early next week.