| 27 Nov 2021 |
hexa | I don't expect any big rebuilds on release-21.11 itself, so compute is ours 🙂 | 18:32:48 |
sterni | I have https://github.com/NixOS/nixpkgs/pull/147404 | 18:33:17 |
sterni | not time critical though, I think | 18:33:41 |
hexa | just setting up my workdirs locally, but yeah unless its not time-senstitive, let's collect some more changes | 18:34:08 |
tomberek | nrdxp: looks good | 18:47:02 |
nrdxp | thanks! merged | 18:47:33 |
tomberek | I'm thinking the keymap.qwertz issue maybe a qemu test driver quirk. Does anyone have an Aaarch64 with KVM to test/reproduce/etc? | 18:48:37 |
tomberek | sterni: Haskell backports seem to need quite a bit of rebuild. Does this need to be in initial release or can it be backported later? | 18:49:38 |
sterni | as said not particularly urgent | 18:50:31 |
sterni | initially I was hoping to resolve an issue on aarch64-darwin, but seems bumping LLVM is not enough (probably upstream issue tbh) | 18:50:56 |
nrdxp | I think I can set binfmt.emulatedSystems to try and test, but I'm not sure how reliable that will be, or if qemu will even run. I'll give it a shot though | 18:53:27 |
tomberek | most qemu test drivers I can find insert a short sleep between sendkey invocations. Likely a race condition when sending keys too fast. | 19:18:11 |
nrdxp | I tried to insert a sleep statement as you suggested, but my run of the test failed:
vm-test-run-keymap-qwertz> BrokenPipeError: [Errno 32] Broken pipe
This is on an x86_64-linux system with boot.binfmt.emulatedSystem = [ "aarch64-linux" ] set, running nix run .\#legacyPackages.aarch64-linux.nixosTests.keymap.qwertz
| 19:50:13 |
hexa | In reply to @tomberek:matrix.org I'm thinking the keymap.qwertz issue maybe a qemu test driver quirk. Does anyone have an Aaarch64 with KVM to test/reproduce/etc? I can see if my pi4 wants to run tests today | 20:23:00 |
hexa | machine # [ 64.370940] systemd[1027]: Reached target Current graphical user session.
(67.39 seconds)
VT Keymap
altgr
machine: must succeed: openvt -sw -- /nix/store/z7637m80di2c64rrlajcxn9rhdrk83j7-test-input-reader 6 '@|{[]}' >&2 &
(0.18 seconds)
machine: waiting for file ‘/tmp/readerReady‘
machine # [ 65.612455] root[1074]: testReader: START: Waiting for 6 characters, expecting '@|{[]}'.
(1.31 seconds)
machine: must succeed: sleep 1
(1.17 seconds)
machine: sending monitor command: sendkey alt_r-q
machine: sending monitor command: sendkey alt_r-less
machine: sending monitor command: sendkey alt_r-7
machine: sending monitor command: sendkey alt_r-8
machine: sending monitor command: sendkey alt_r-9
machine: sending monitor command: sendkey alt_r-0
machine: waiting for file ‘/tmp/readerResult‘
machine # [ 68.260545] root[1087]: testReader: FAIL: Expected '@|{[]}' but got '@|{[9}'.
(1.30 seconds)
machine: must succeed: grep -q 'PASS:' /tmp/readerResult
machine: output:
Test "altgr" failed with error: "command `grep -q 'PASS:' /tmp/readerResult` failed (exit code 1)"
cleanup
kill machine (pid 6)
machine # qemu-system-aarch64: terminating on signal 15 from pid 4 (/nix/store/8cj0plgdv6929yx4rkv3rp6m3hcj5h00-python3-3.9.6/bin/python3)
(0.14 seconds)
| 20:57:25 |
hexa | on
commit 9c191ebcdfe917043195c54ab6ae8e934434fe7b (HEAD -> master, origin/master, origin/HEAD)
Merge: 092154c92f9 1cfecb636b1
Author: Timothy DeHerrera <tim.deh@pm.me>
Date: Sat Nov 27 11:08:43 2021 -0700
Merge pull request #147609 from primeos/revert-restarting-systemd-socket-units
Revert "Merge pull request #141192 from helsinki-systems/feat/improve…
| 21:04:40 |
hexa | ping me if you need this tested again | 21:05:03 |
jonringer | AFAIK there hasn't been too many staging PRs which need a backport. | 23:22:25 |
tomberek | Looking at https://hydra.nixos.org/build/159774271 iso_minimal for i686-linux | 23:22:57 |
tomberek | error code 5 is XZ's code for malloc failure. | 23:24:09 |
| 28 Nov 2021 |
tomberek | Probably a nix2.3 thing: https://hydra.nixos.org/build/159784106/nixlog/3/tail
‘’’
machine: must succeed: nix store verify --no-trust -r --option experimental-features nix-command /run/current-system
machine # error: 'store' is not a recognised command
machine # Try 'nix --help' for more information.
machine: output:
cleanup
‘’’ | 22:39:19 |
tomberek | Same for biosUsb and biosCdrom? | 22:40:50 |
| 29 Nov 2021 |
tomberek | https://hydra.nixos.org/build/159837962 builds! | 06:59:22 |
tomberek | nixos-21.11 channel and branch have been created. | 16:39:53 |
tomberek | https://github.com/NixOS/nixpkgs/pull/147898 | 16:40:09 |
tomberek | Who can merge https://github.com/NixOS/nixos-artwork/pull/70 ? | 16:57:27 |
hexa | samueldr | 17:02:22 |
| 30 Nov 2021 |
@grahamc:nixos.org | I hear we're basically ready to go! | 00:36:27 |
@grahamc:nixos.org | so jon asked me to look in to why the 21.11-aarch64 channel hasn't bumped yet and why status.nixos.org is borkeeeeeeeeed | 00:44:29 |
@grahamc:nixos.org | sigh, gnome-shell is at 100% cpu and causing things to be wacky | 00:45:02 |