Repository navigation
Conversation
C6 / Beads mdo-6w5.2 (epic mdo-6w5, PR order 2 of 6; stacked on the C1
branch).
- tests/cli.rs re-implemented temp_output_for as temp_output_for_test and
had already lost stem sanitization (diverged from production). The four
call sites now call mdo_cli::temp_output_for directly, so tests compute
the same per-source private paths production does, sanitization included.
The mirror is deleted, along with the Hash/Hasher imports it alone needed.
- The one-line derive_output alias had a single caller; main.rs now calls
the real target, input.with_extension("html"), and lib no longer exposes
the alias.
- SETUP_SAMPLE_FILE_NAME / SETUP_SAMPLE_MARKDOWN had zero external callers;
both are now private. (Windows SetupSample? not referenced: mdo-setup and
mdo-open never touch them.)
- wrapper.rs's own empty_path_dir duplication is intentionally untouched:
recorded in the issue as having an honest reason.
Gates: fmt --check, clippy -D warnings, cargo test --all-targets
--all-features --locked (90 tests) all green, including the Linux --open
tests that exercise the corrected temp-output path contract.
Beads-Ref: mdo-6w5.2
Agent-Signature: kilo-kilo-auto-efficient on behalf of matt
| use std::process::Command; | ||
| use std::time::{SystemTime, UNIX_EPOCH}; | ||
|
|
||
| use mdo_cli::temp_output_for; |
There was a problem hiding this comment.
WARNING: Unused import on non-Linux test builds - cargo clippy --all-targets -D warnings fails on Windows/macOS
Every caller of temp_output_for in this integration-test file (open_output_refuses_precreated_symlink, open_launch_failure_exits_nonzero_and_reports_path, open_success_exits_zero, open_reports_failure_when_opener_exits_nonzero) is #[cfg(target_os = "linux")], so on any non-Linux target this import is unused and -D warnings turns the warning into an error. The deleted Hash/Hasher import directly below this line was cfg-gated the same way, which is this file's convention (lines 8-12).
| use mdo_cli::temp_output_for; | |
| #[cfg(target_os = "linux")] | |
| use mdo_cli::temp_output_for; |
Reply with @kilocode-bot fix it to have Kilo Code address this issue.
Code Review SummaryStatus: No Issues Found | Recommendation: Merge Files Reviewed (3 files)
Previous Review Summary (commit 9ca6598)Current summary above is authoritative. Previous snapshots are kept for context only. Previous review (commit 9ca6598)Status: 1 Issue Found | Recommendation: Address before merge Fix these issues in Kilo Cloud Overview
Issue Details (click to expand)WARNING
Files Reviewed (3 files)
Reviewed by efficient · Input: 24.7K · Output: 2.8K · Cached: 112.6K |
The import was only used in #[cfg(target_os = linux)] tests but was unconditional, causing clippy -D warnings to fail on macOS/Windows with unused_imports. This is a follow-up to Beads mdo-6w5.2 (C6). Agent-Signature: kilo-kilo-auto-efficient on behalf of matt
C6: Retire the divergent test mirror and dead public surface
Beads mdo-6w5.2 · parent epic mdo-6w5 (PR order 2 of 6; stacked on the C1 branch
agent/6w5-c1-convert-request)What
tests/cli.rs:618-633re-implementedtemp_output_foras a mirror that had already lost stem sanitization. The four--opencall sites now callmdo_cli::temp_output_fordirectly, so the tests compute precisely the paths production computes. Mirror deleted (plus theHash/Hasherimports only it needed).derived_outputalias (src/lib.rs:137, one caller) deleted;main.rscalls the real targetinput.with_extension("html").SETUP_SAMPLE_FILE_NAME/SETUP_SAMPLE_MARKDOWN(zero external callers) made private.wrapper.rs'sempty_path_dirduplication deliberately untouched — recorded in the issue as having an honest reason.Gates
fmt --check ✓ · clippy -D warnings ✓ ·
cargo test --all-targets --all-features --locked✓ (90 tests, including the Linux--opensuite that now exercises the real path contract)Recorded byte-identity work from C1 is preserved: no CLI-visible strings changed in this PR.
Signed: kilo-kilo-auto-efficient on behalf of matt (Beads mdo-6w5.2)