Repository navigation
Conversation
Reviewer's guide (collapsed on small PRs)Reviewer's GuideThis structural refactor moves three self-named Rust module files to File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the
Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info
SummaryMove three directory-backed Rust modules to
Update the ChecksGitHub’s build-test, CodeQL and CodeScene checks pass on head WalkthroughTwo test module path attributes now point to different file locations. No scheduler behaviour or public declarations changed. ChangesTest module path updates
Priority: ⬇️ Low Change: Refactor Merge Risk: ⚪ Minimal · up to
|
Clippy's `self_named_module_files` requires a directory-backed module to live in `mod.rs`. Three modules already had a sibling directory, so each `foo.rs` is now `foo/mod.rs` with `git mv`, keeping the change content-preserving: - `src/metrics.rs` -> `src/metrics/mod.rs` - `src/server/runtime/server.rs` -> `src/server/runtime/server/mod.rs` - `src/client/pool/scheduler.rs` -> `src/client/pool/scheduler/mod.rs` The `#[path]` attributes pointing at test modules were recomputed against the new locations, not simply retained. `scheduler_tests.rs` stays in the parent directory and now uses `../scheduler_tests.rs`; `observation_tests.rs` already lived in `server/`, so its attribute drops the now-redundant `server/` prefix. Co-Authored-By: Claude Code <noreply@anthropic.com>
cc48b70 to
d493667
Compare
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d4936677f2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
|
||
| #[cfg(test)] | ||
| #[path = "scheduler_tests.rs"] | ||
| #[path = "../scheduler_tests.rs"] |
There was a problem hiding this comment.
Update source-of-truth docs for the moved modules
This move leaves current traceability documentation pointing to files that no longer exist: docs/adr-013-client-pool-scheduler-and-slot-ownership.md:110 names src/client/pool/scheduler.rs, docs/adr-014-server-shutdown-ownership.md:41-45 names both src/server/runtime/server.rs and src/metrics.rs, and docs/users-guide.md:2857 also names src/metrics.rs. Update these active references to their new mod.rs paths so the repository's documented source surfaces remain navigable; historical execution-plan references can remain unchanged.
AGENTS.md reference: AGENTS.md:L44-L47
Useful? React with 👍 / 👎.
Scope
Structural layer only: three
foo.rsfiles with an existingfoo/directory become
foo/mod.rs, satisfying Clippy'sself_named_module_filesahead of the final lint configuration. Content is preserved; every file below
the move is byte-identical apart from the recomputed
#[path]attributes.src/metrics.rs->src/metrics/mod.rssrc/server/runtime/server.rs->src/server/runtime/server/mod.rssrc/client/pool/scheduler.rs->src/client/pool/scheduler/mod.rsscheduler_tests.rsstays in the parent directory, so its attribute becomes../scheduler_tests.rs.observation_tests.rsalready lived insideserver/,so its attribute drops the now-redundant
server/prefix. Both were recomputedagainst the new location rather than retained blindly.
Position and dependencies
Layer 4 of 5. Stacked on the examples layer, and the parent of the census
layer that closes the stack.
harden-lint-examples-20261022(Fix example lint findings and inject the environment boundary #797)harden-lint-census-completion-20261022(Clear the helper-crate lint census #805)Reviewer entrypoints
src/client/pool/scheduler/mod.rs:395andsrc/server/runtime/server/mod.rs:398— the two recomputed#[path]attributes
git mvrenames with no content changeValidation evidence (head
d493667)This branch was restacked onto
origin/main=b194531on 2026-10-23, aftermain moved three commits including #794's toolchain bump from
nightly-2026-03-26tonightly-2026-10-22. The previous headcc48b70and itslocal gate results belong to the retired toolchain and are not current
evidence.
The
weavemerge driver was disabled for the replay(
-c core.attributesFile=/dev/null), and the replay was checked against adriver-free
merge-treeresult rather than trusted.Hosted, at this head. The
build-testlane ran at this exact commit andpassed every step. It executes the repository's own targets, so its result is
per-layer evidence for the format and lint gates rather than a result inherited
from a neighbour:
make check-fmtmake lint— rustdoc with-D warnings, Clippy with-D warnings, Whitaker--allmarkdownlint-cli2over**/*.mdmake spellingmake nixiemake test-workflow-contractsmake test-feature-gatecargo llvm-cov nextest --workspace, default features and targetshttps://github.com/leynos/wireframe/actions/runs/39959689502/job/120434257945
The instrumented run executed 1207 tests, 1207 passed, 0 skipped and
reported 87.79%, a 0.01 pp fall from the committed 87.8% baseline that is
inside the ratchet's tolerance band.
Read from the job log, not inferred from the job conclusion.
That run is narrower than the Make targets — no
--all-targets, no--all-features, no doctests — which is why the local runs below are stillrequired.
Local. Run sequentially at this exact head with
make:Known baseline failures
None observed at this head.
Deferred work
Remaining source census findings, final lint-configuration adoption, the
testkit argument-count decision and Whitaker
--workspacecoverage.Summary by Sourcery
Enhancements:
mod.rslocations while preserving their contents and updating affected test paths.References
harden-lint-census-completion-20261022)gate run and the census this layer defers to.
unsafe_code = "forbid"can apply workspace-wide #804 — isolate the unsafe allocator sounsafe_code = "forbid"can applyworkspace-wide.
/home/leynos/docs/leynos-wireframe/rust-baseline/2026-10-15/train-ledger.md🤖 Generated with Claude Code