docs: the Windows tests that ran nowhere now gate, and one still does not - #66
Merged
Conversation
… not `resources.rs` carried a note saying its two `#[cfg(windows)]` tests did not run in CI, and it named what would change that: "the windows CI job compiling THIS crate". #62 is that change — `browser_host.rs`'s ungated `ns_window()` is gated, so `brains-desktop` links on Windows, and the `rust` matrix runs `cargo test --workspace`, which includes this crate. So the note now says the opposite of the truth, and points at a job that no longer exists. Rewritten with the evidence rather than the inference — from the Windows leg's own log: test resources::tests::a_verbatim_resource_dir_loses_its_prefix_before_reaching_node ... ok test resources::tests::a_genuine_unc_path_is_left_alone ... ok That is the regression coverage for the verbatim-path bug finally executing on the only platform where it can fail. The same note claimed `cargo test -p brains-desktop resources` could not be run on Windows at all; it now passes locally there too (3 passed). The `#[cfg(windows)]` gate itself stays, and the note keeps saying why: `dunce::simplified` is the identity function off Windows, so the assertions cannot hold there, and making them platform-free would mean asserting against a reimplementation of dunce instead of the code that ships. Also records the gap #62 exposes rather than creates, in the recording engine's README beside the other open Windows items: `tests/fake_sidecar.rs` is `#![cfg(unix)]`, so the Windows leg runs 0 of its 4 tests while covering everything else in the crate. Worth naming because of what it leaves unpinned there — the one-shot window-peak contract, whose breakage made `micDeviceMismatch` dead on every platform, is proven today only on macOS and in jsdom. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to #62, which merged an hour ago. Two files, comments only, no behaviour change.
The note that is now false
src-tauri/src/resources.rssaid its two#[cfg(windows)]tests do not run in CI, and named what would change that: "the windows CI job compiling THIS crate". #62 is that change.browser_host.rs's ungatedns_window()is gated, sobrains-desktoplinks on Windows, and therustmatrix runscargo test --workspace— which includes this crate.The note therefore now says the opposite of the truth and points the reader at a job that no longer exists (#45's separate
Windows (recording)job, removed in #62's merge).Rewritten with evidence instead of inference, from the Windows leg's own log on #62's run:
That is the regression coverage for the verbatim-path bug finally executing on the only platform where it can fail. The same note also claimed
cargo test -p brains-desktop resourcescould not be run on Windows at all — it now passes locally there too (3 passed).The
#[cfg(windows)]gate stays, and the note keeps saying why:dunce::simplifiedis the identity function off Windows, so the assertions cannot hold there, and making them platform-free would mean asserting against a reimplementation of dunce rather than the code that ships.The gap #62 exposes rather than creates
Recorded in the recording engine's README beside the other open Windows items:
tests/fake_sidecar.rsis#![cfg(unix)], so the Windows leg runs 0 of its 4 tests while covering everything else in the crate. It was invisible while no Windows leg existed.Worth naming for what it leaves unpinned there: the one-shot window-peak contract — whose breakage is what made
micDeviceMismatchdead on every platform — is proven today only on macOS and in jsdom. Porting the fake from a shell script to Node would close it. Not done here; it is a real change, not a comment.Why bother with a comments-only PR
Because a comment that misdirects is worse than no comment, and this one tells the next person to widen a job that is gone. Rule 4 of
CLAUDE.mdtreats these headers as contracts.Gates
npm run lint:sizegreen ·cargo test -p brains-desktop resources3 passed on Windows 11 · no code touched, so nothing else can move.🤖 Generated with Claude Code