ci: more log, less timeout @K900

[K900]

Pick up new nixpkgs-mozilla with the patchelf patch.

Flake lock file updates:

• Updated input 'lib-aggregate': 'github:nix-community/lib-aggregate/d44755862cce5ba5e040ec8f7df6c6b33e47c8a0' (2023-09-17) → 'github:nix-community/lib-aggregate/cb8bfd550aaaf32a330c1c8870a3d9a5bfa00954' (2023-09-24) • Updated input 'lib-aggregate/nixpkgs-lib': 'github:nix-community/nixpkgs.lib/819180647f428a3826bfc917a54449da1e532ce0' (2023-09-17) → 'github:nix-community/nixpkgs.lib/01fc4cd75e577ac00e7c50b7e5f16cd9b6d633e8' (2023-09-24) • Updated input 'mozilla': 'github:mozilla/nixpkgs-mozilla/d8a0a26242ca12145c9a1a4e08e21efb15fc7776' (2023-09-21) → 'github:mozilla/nixpkgs-mozilla/7800b921f749d74ecb8456f35f7ef04cd49b4d24' (2023-09-25) • Updated input 'nixpkgs': 'github:nixos/nixpkgs/970a59bd19eff3752ce552935687100c46e820a5' (2023-09-17) → 'github:nixos/nixpkgs/e35dcc04a3853da485a396bdd332217d0ac9054f' (2023-09-22)

[colemickens]

cc: @Artturin any idea why the test is still failing even with this?

The update script should be updating flake lock inputs before the test, but the test seems to still fail, so it's not getting published.

[K900]

The test seems to be timing out.

[K900]

@colemickens can you try running it on this commit? I don't think 10 minutes of 4 cores is enough to run 4 VM tests.

[colemickens]

I can (I'll have to just test around on master though) but I'm still suspicious. It doesn't seem like even one test is passing here: https://github.com/nix-community/flake-firefox-nightly/actions/runs/6301320902/job/17106100984

[K900]

I wouldn't be super surprised, given it's 10 minutes to fetch + build + run the VMs on probably pretty heavily overcommitted hardware.

[colemickens]

I'm kind of ill, my brain isn't firing on all cylinders. I can of course just run the update locally and let it go and see what happens.

[colemickens]

It definitely seems stuck locally too:

many many minutes of:

vm-test-run-firefox-nightly-bin-unwrapped> machine: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d'
vm-test-run-firefox-nightly-bin-unwrapped> (finished: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d', in 0.01 seconds)
vm-test-run-firefox-nightly-bin-unwrapped> machine: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d'
vm-test-run-firefox-nightly-bin-unwrapped> (finished: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d', in 0.01 seconds)
vm-test-run-firefox-nightly-bin-unwrapped> machine: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d'
vm-test-run-firefox-nightly-bin-unwrapped> (finished: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d', in 0.01 seconds)
vm-test-run-firefox-nightly-bin-unwrapped> machine: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d'
vm-test-run-firefox-nightly-bin-unwrapped> (finished: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d', in 0.01 seconds)
vm-test-run-firefox-nightly-bin-unwrapped> machine: must succeed: xwininfo -root -tree | sed 's/.*0x[0-9a-f]* \"\([^\"]*\)\".*/\1/; t; d'

I'll take a look but, again, brain no work-y well right now.

[K900]

Weird, it worked fine for me.

[K900]

OK, it's definitely flaky, I'm looking.

[K900]

How about this?

[Artturin]

Pushed cherry-picked ci: better logs for tests to master

[Artturin]

I pushed a lock update to master, after reverting reverting of the nightly update the test still passes for me

[colemickens]

why does the test use {firefoxPackage.unwrapped.binaryName} instead of the wrapped binary?

[K900]

Because the wrapper doesn't pass through binaryName.