Commit Graph
3089 Commits
Author SHA1 Message Date
LGUG2Z 865ffcebda docs(ci): what is guaranteed about KVM on hosted runners, what is drawn again, and what is not covered 2026-11-06 09:18:43 -08:00
LGUG2Z 996330f19b fix(ci): draw the host of a VM suite again when it has no KVM, in the same run
A hosted runner of the ubuntu-24.04 pool is not promised nested virtualization, and
3 of the 16 suites of run 42182554717 failed the check of /dev/kvm. rerun-no-kvm.yaml
could not help: a workflow_run workflow runs only from the default branch, where the
file is not, and a run cannot rerun its own failed jobs while it is in progress.

The step "Enable KVM" of the nix action is now the only one that enables KVM
(enable_kvm of install-nix-action is off), runs before the disk and the install,
waits for udev to settle and polls /dev/kvm for 30 seconds instead of testing it
once, 42 ms after an asynchronous udevadm trigger, and prints what tells a missing
device from a rule that was not applied yet. With defer-no-kvm a host without KVM
is an output, not a failure.

_vm.yaml calls _vm-attempt.yaml up to three times: the suites that found no KVM are
listed through the artifacts of the run and run again on new machines, and the last
attempt fails them. vm-fast and vm-slow are calls of it. rerun-no-kvm.yaml is
removed.
2026-11-06 09:18:43 -08:00
LGUG2Z d57b2d7ed3 fix(ci): build from source when a download from the cache fails
A NAR of 460 MB was cut by the edge of the cache ("Stream error in the HTTP/2 framing layer") and the
resumed request failed the same way, so the job kwin-events failed before its VM test started. With
fallback = true Nix builds the derivation from source when a substitute cannot be downloaded: one dropped
connection costs minutes of compiling in that job, not the job. HTTP/2 is left at its default: nothing
shows that turning it off prevents the error.
2026-11-06 09:17:08 -08:00
LGUG2Z 97e1637393 fix(test): do not check the pages of the cli for macOS on a Windows host
ApplicationIdentifier::Class is described per platform ("Class" on
Windows, "Role or subrole" elsewhere), and the description is taken from
the host the test is built for. every_macos_page_is_as_checked_in renders
the pages of komorebic for macOS from the command line, so on the Windows
runner six of them (ignore-rule, manage-rule, workspace-rule,
named-workspace-rule, initial-workspace-rule and
initial-named-workspace-rule) came out with "- class: Class" and were
reported as stale, though the checked-in pages are right.

The pages are the same on the hosts that have the macOS wording, and the
linux test run checks them, so skip the test on Windows.
2026-11-05 23:41:05 -08:00
LGUG2Z 8642f7b184 fix(ci): quote the description of the push input of the nix action
An unquoted ": " inside the description made action.yaml invalid YAML ("mapping values are not allowed in
this context"), so every job that uses the action would have failed at its first step.
2026-11-05 16:48:38 -08:00
LGUG2Z 82366f7134 fix(nix): evaluation no longer reads store paths that builtins.path makes
nix flake check --no-build failed in the nix-eval job on the runner with "path ...-Cargo.lock is not
valid": crane's vendorCargoDeps runs pathExists on the lock at evaluation time, and `repoFile` is a
store path that builtins.path makes during evaluation. Nix 2.35 with an empty store, as on a fresh
runner, does not have it yet under --no-build; Lix, and a store that had seen the file, do.

The parsed lock now comes from the flake source (cargoLockParsed in nix/lib.nix), and the Windows guest
reads tools/test-partition.json the same way. The derivations are unchanged (same drv paths), so no
cache key moves. Reproduced and fixed with upstream Nix and an empty store:
nix shell nixpkgs#nixVersions.latest -c nix flake check --no-build --all-systems --store <empty dir>
2026-11-05 16:48:10 -08:00
LGUG2Z 166e5f353a ci(cachix): write to the cache on every push, read it in every job that runs nix
The write token goes to cachix-action on a push (master and the temporary komomono
trigger, without naming either), a manual run and a schedule; a pull request, a fork
and the merge queue only read. The flake-lock update goes through the nix action too,
and the nixpkgs canary reads the cache but never pushes (push: "false" replaces
cache: "false").

docs(ci): the write rule, the read-only path and a checklist for the first push to
komomono.
2026-11-05 16:40:29 -08:00
LGUG2Z 17fb726ecf ci(hygiene): print the tool versions of the flake, and write down that stage 0 runs the latest
The tools of stage 0 are the latest releases, taken from the flake, and the code is changed to satisfy
them, not the other way round: every run says which versions it had, and docs/dev/ci.md says how they
are kept current (the weekly lock update, and nix flake update rust-overlay or nixpkgs in between).
2026-11-05 16:37:59 -08:00
LGUG2Z cb2f49dc17 build(nix): update rust-overlay, so the nightly rustfmt of the checks is the latest
The pin was a week old: the hygiene shell had the nightly of 2026-10-27. The nightly of
2026-11-04 formats the whole tree unchanged. Every other tool of the shell was already the latest
release (actionlint 1.7.12, shellcheck 0.11.0, cargo-deny 0.20.2, ripgrep 15.2.0, jq 1.8.2, cargo 1.99.1).
2026-11-05 16:37:59 -08:00
LGUG2Z d5d5fe7fe7 ci(hygiene): stage 0 runs the tools the flake pins, not the ones of the runner
The first stage ran shellcheck, rustfmt, cargo-deny and actionlint as the runner image, the
rust-toolchain rustup and two actions shipped them, and the runner's older shellcheck reported
SC2317 on a trap handler that the flake's does not. hygiene, fmt and deny now run every step in
`nix develop .#hygiene`, so CI and a local run use the same versions.
2026-11-05 16:31:20 -08:00
LGUG2Z 194fbd4281 build(nix): a hygiene dev shell with the pinned tools of the first CI stage
devShells.hygiene is the toolchain with the nightly rustfmt, cargo-deny, shellcheck, actionlint, jq,
ripgrep and git, without the Windows SDK, the VM tooling and the rest of the full shell, so that the
first stage sets up in seconds and a local `nix develop .#hygiene` runs what CI runs.
2026-11-05 16:31:20 -08:00
LGUG2Z 15c8a04459 ci(actionlint): run actionlint 1.7.12 in the hygiene job
1.7.7 does not know the hosted Arm Windows runner and reported windows-11-arm as an unknown label.
1.7.12 is the latest release and the version nixpkgs gives `just lint`, so the hygiene job and a
local run now agree.
2026-11-05 16:21:14 -08:00
LGUG2Z 658827fcd5 fix(lint): build errors with Report::msg, not the eyre! macro the workspace bans
The ban (`just lint`, and now the hygiene job) came from the macOS and KDE repositories and was
never applied to the Windows code, so the unified tree had sixteen uses of the macro: seven in
the Windows daemon, six in komorebi-core (written during the unification), one in the KDE
backend and two in the recorder. Each becomes eyre::Report::msg, with format! where the message
has arguments.
2026-11-05 16:13:37 -08:00
LGUG2Z 416cda802d docs(ci): the ladder, the gaps, the settings the owner changes and the komomono checklist 2026-11-05 15:51:50 -08:00
LGUG2Z 2e649ef8e5 ci(windows-guest): an experimental, manual workflow builds the Windows guest and runs the Windows tests in it 2026-11-05 15:51:50 -08:00
LGUG2Z 04c79adefd ci(sweeps): the slow VM suites, the NixOS evaluation and the nixpkgs canary run outside the ladder; the flake lock is updated weekly 2026-11-05 15:51:50 -08:00
LGUG2Z e727f6b14d ci(macos): the macOS jobs run on the self-hosted Mac when the repository enables it 2026-11-05 15:51:50 -08:00
LGUG2Z 3a7d3d9e1f ci: the ladder of stages, with one required aggregator, and the reusable macOS jobs 2026-11-05 15:51:50 -08:00
LGUG2Z 21e86824e4 ci(paths): .github/ci-paths.yaml says which files concern which platform 2026-11-05 15:51:50 -08:00
LGUG2Z d49b155045 ci(actions): a composite action installs nix, enables KVM and reads the Cachix cache 2026-11-05 15:51:50 -08:00
LGUG2Z e23bfcfb36 build(nix): the macOS options and nix-darwin module validations are checks on every system
They take a plain pkgs and evaluate against any nixpkgs, so a Linux runner can build them. They stay in the macos-options set.
2026-11-05 15:51:50 -08:00
LGUG2Z f05d7fe737 build(just): test-partition-check, test-windows (the partitioned subset) and test-windows-full (the whole Windows suite by hand) 2026-11-05 15:51:50 -08:00
LGUG2Z 56afe5b286 build(nix): the tests of komorebi-macos-core run in linux-nextest and macos-nextest 2026-11-05 15:51:50 -08:00
LGUG2Z 98295e0065 build(nix): windows-nextest runs the windows and os_sensitive packages of the partition, windows-nextest-full all of them 2026-11-05 15:51:50 -08:00
LGUG2Z 8e1a2b261c refactor(macos): the animation maths, the arrival record and the enhanced-UI rules move into komorebi-macos-core, which needs no macOS 2026-11-05 15:51:50 -08:00
LGUG2Z b83cf06cc6 feat(tools): test-partition says which packages are tested where, and fails when the classification and the workspace disagree 2026-11-05 15:51:50 -08:00
LGUG2Z 3464fb0c58 test(transport): the listener tests run on Windows too, against the uds_windows listener the Windows daemon uses 2026-11-05 15:51:50 -08:00
LGUG2Z a81a1120cd docs: update paths for crates/ and the komorebi directory, add a contributor path map 2026-11-05 15:03:27 -08:00
LGUG2Z c234b1d884 build(tools): route the justfile and tools to crates/ 2026-11-05 15:03:27 -08:00
LGUG2Z 279ac0f759 style: format the files the path edits touched with rustfmt and nixfmt 2026-11-05 15:03:27 -08:00
LGUG2Z 259575a1b3 build: route the workspace to crates/
The workspace members, the path dependencies, the include_str! and test
fixture paths, the recorder's indexed source keys, and the Nix source
filter and path literals follow the new layout. Cargo.lock is unchanged.
2026-11-05 15:03:27 -08:00
LGUG2Z 6def6bdfbf refactor(layout): move library and dev-tool crates under crates/, rename komorebi-daemon to komorebi
Pure renames, no content changes. This revision does not build on its own;
the next one routes the workspace to the new paths.
2026-11-05 15:03:27 -08:00
LGUG2Z fdfb0cef73 chore(deps): dependencies.json lists this workspace's crates by name and version
cargo deny list names a workspace crate by its path+file:// source, which is a path on the machine that
ran it (C:/Users/... before, /Users/... now). depgen strips it, and the file is regenerated for all three
platforms' crates.
2026-11-05 15:03:27 -08:00
LGUG2Z 5b3e3f598a build(deny): ignore RUSTSEC-2025-0059, servo-fontconfig is unmaintained
It is reached only through font-loader, which komorebi-bar looks up system fonts with; it was
already in the lock before the recorder and schemars-nixgen merged.
2026-11-05 15:03:27 -08:00
LGUG2Z 112e624c5e build(nix): regenerate the Nix options from the schemas 2026-11-05 15:03:27 -08:00
LGUG2Z b16fedebe7 docs(nixgen): schemars-nixgen has a README, and the docs say how to run it 2026-11-05 15:03:27 -08:00
LGUG2Z a342dc39f5 ci(windows): the Windows tests leave out schemars-nixgen 2026-11-05 15:03:27 -08:00
LGUG2Z 58ecb37f38 build(just): nixgen-linux and nixgen-macos run schemars-nixgen from the tree 2026-11-05 15:03:27 -08:00
LGUG2Z c6d29a9268 build(nix): schemars-nixgen is built and checked by the flake, and in the dev shell
nix/nixgen has the package and the clippy and nextest checks (in the default check sets), from a
"nixgen" platform of nix/workspace-src.nix that is the tool alone; it is a stub in every other
platform's source, so an edit of it rebuilds none of theirs.
2026-11-05 15:03:27 -08:00
LGUG2Z 622646c3f7 build(workspace): schemars-nixgen is a member of the workspace, not a default member
It is a development tool: a bare cargo build leaves it out, publish is off, and no release
ships it. Its manifest uses the workspace dependencies it shares and keeps its own version.
2026-11-05 15:03:27 -08:00
LGUG2Z 3d33fac0cd merge: the history of schemars-nixgen 2026-11-05 15:03:27 -08:00
LGUG2Z e499ebb774 docs(recorder): the recorder is documented as part of the komorebi repository 2026-11-05 15:03:27 -08:00
LGUG2Z a0f7d871e5 ci(windows): the recorder is built with the rest, its tests informational until they have run on Windows 2026-11-05 15:03:27 -08:00
LGUG2Z e803212de2 build(just): the recorder is a target of every platform, with record, icon and build-windows-recorder 2026-11-05 15:03:27 -08:00
LGUG2Z 1db6d7d3fd build(nix): the recorder is built, checked and shipped by the flake
nix/recorder holds the native packages and checks of macOS and Linux (the recorder and the
sources it indexes are the "recorder" platform of nix/workspace-src.nix), the Windows
build has it as one of its members, and the dev shell gets ffmpeg, resvg and GStreamer. The
flake inputs komorebi-mac and komorebi-windows are gone: the sources are in the tree.
2026-11-05 15:03:27 -08:00
LGUG2Z 170a6f51c0 feat(catalog): index the komorebi source in this repository, not a checkout of each komorebi
The recorder is a member of the repository it indexes: each flavor is built from the
directory of its own daemon (komorebi-macos, komorebi-windows, komorebi-linux) and the shared
crates that log. The statements of the container, workspace and monitor moved into
komorebi-core, so the targets of the real log sample in the fixture are komorebi_core's.
KOMOREBI_SRC and KOMOREBI_COMMIT replace the three KOMOREBI_*_SRC variables, for a build
that is not in a checkout.
2026-11-05 15:03:27 -08:00
LGUG2Z 9854f90db7 build(wix): the komorebi installer ships komorebi-recorder, which has no installer of its own 2026-11-05 15:03:27 -08:00
LGUG2Z c64c71541b build(workspace): komorebi-recorder is a member of the workspace, at 0.1.0-dev 2026-11-05 15:03:27 -08:00
LGUG2Z 38a56fd544 merge: the history of komorebi-recorder 2026-11-05 15:03:27 -08:00
LGUG2Z c69fb806b3 build(just): the justfile does the right thing on Windows, macOS and Linux 2026-11-05 15:03:27 -08:00