feat/try compat - #108
Conversation
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Sorry @tonythethompson, your pull request is larger than the review limit of 150000 diff characters
|
Note: |
|
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:
📝 WalkthroughWalkthroughThe PR adds per-Nu-minor activation profiles, lifecycle-aware version switching, explicit package resolution for ChangesActivation and compatibility workflows
Estimated code review effort: 5 (Critical) | ~120 minutes Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 8✅ Passed checks (8 passed)
✨ Finishing Touches 💡 1🛠️ Fix failing CI checks 💡
📝 Generate docstrings
🧪 Generate unit tests (beta)
✨ Simplify code
Warning Review ran into problems🔥 ProblemsLinked repositories: Public OSS repositories can only analyze public repositories installed in this organization. Analyzed Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
PR Summary by QodoMake
AI Description
Diagram
High-Level Assessment
Files changed (16)
|
Greptile SummaryThe PR introduces compatibility-first
Confidence Score: 5/5The PR appears safe to merge because no blocking failure remains in the previously reported paths. The nested-lock failure, pinned-release compatibility leak, and concurrent install-to-activate handoff race are all resolved in the current code.
|
| Filename | Overview |
|---|---|
| src/cmd/try_cmd.rs | Implements package-specific compatibility resolution and retains the root mutation lock continuously across install and activation; the previously reported pinned-scope and handoff issues are fixed. |
| src/cmd/activate.rs | Adds an under-lock activation entry point that skips nested lock acquisition while preserving the existing activation lifecycle. |
| src/cmd/activation_switch.rs | Adds activation-profile leave and restore orchestration for managed Nu version switches without introducing a finding eligible for this follow-up. |
| src/state/activation_profile.rs | Persists desired plugin and module activation sets by Nu minor and supports synchronization and rollback. |
| src/cmd/use_cmd.rs | Integrates managed Nu selection with activation-profile teardown and restoration. |
| src/core/resolve.rs | Adds compatibility helpers that correctly evaluate candidate Nu versions against a specific pinned release when requested. |
Reviews (16): Last reviewed commit: "fix: classify modules by on-disk state a..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
Code Review by Qodo
1.
|
![qodo-code-review[bot]](https://avatars.githubusercontent.com/in/484649?s=60&v=4){kind=small}
Qodo Fixer✅ Committed (1) · ☑ Fixed (1) Commits pushed directly to this PR — no separate fix PR opened. Process — 1 fixed
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 18
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/cmd/try_cmd.rs (1)
294-342: 🩺 Stability & Availability | 🟠 Major | ⚡ Quick winDrop the install lock before activating in
numan try.
try_packageholds_lockfor the whole function, then callsactivate::execute, which re-acquires the same root mutation lock.try_write()fails immediately on the second acquisition, so non---no-activateinstalls fail withActivated "<id>" but activation failed: Another Numan mutation is already in progress for this root.Scope_lockto install only, and update the lock-held tests if they check the outer lock fromtry_cmd.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/try_cmd.rs` around lines 294 - 342, In try_package, scope the acquire_mutation_lock(root) guard only around the install operation so it is dropped before the activate call. Keep installation setup and result handling under the lock, then release it before invoking activate; update any try_cmd tests that assert the outer lock remains held.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@AGENTS.md`:
- Line 240: Insert a blank line immediately before the “Learned Workspace Facts”
heading in AGENTS.md, separating it from the preceding list so the markdownlint
MD022 requirement is satisfied.
In `@README.md`:
- Line 215: The compatibility retry guidance must preserve an explicitly
supplied package version. Update README.md lines 215-215 to instruct users to
retry the same <owner/name[`@version`]> specification, and update AGENTS.md lines
252-252 so the learned workspace guidance preserves the explicit `@version`;
ensure both references cover versioned and unversioned inputs.
- Around line 357-358: Update the documentation for numan use restoration
semantics: in README.md lines 357-358, state that activations are restored after
successful restoration and that restoration can fail; in CHANGELOG.md line 14,
remove any implication that target restoration is guaranteed; and in AGENTS.md
line 152, document that post-marker restoration may be partial and can fail
while leaving the activation profile unchanged.
In `@src/cmd/activate.rs`:
- Around line 244-247: Filter the packages passed to sync_user_activate_profile
in the post-activation flow so only packages with successfully written
activation records are included, matching the deactivate path’s liveness check.
Apply this to the explicit-packages branch before syncing, while preserving the
existing no-op sync behavior and any_failed handling.
In `@src/cmd/activation_switch.rs`:
- Around line 383-386: Update the comment above the plugin-only
has_compatible_version fallback in the restore classification logic to state
that modules with installed versions absent from the index are intentionally
classified as Ok and restored, while plugins must still have a compatible
version. Keep the condition and behavior unchanged.
- Around line 157-163: Update the Lockfile::load handling in leave_current_nu so
load failures propagate as errors instead of returning an empty report and
success. Preserve Lockfile::load’s existing behavior for a missing lockfile,
while ensuring malformed or unreadable lockfiles stop the switch before
active-version markers or paths are updated.
In `@src/cmd/doctor.rs`:
- Around line 530-534: Remove the behavior-narrating helper comment in
src/cmd/registry.rs lines 210-230 unless it documents a non-obvious output
constraint. In src/cmd/doctor.rs lines 530-534, replace the summary with a brief
comment explaining that the legacy-path fallback is required for pre-migration
roots, or remove the comment entirely.
- Around line 535-552: Update resolve_managed_nu_binary to add descriptive
context to errors from version_manager::active_nu_binary(root) and
version_manager::latest_installed_version(root), using .context(...) that
identifies the failed managed Nu resolution operation while preserving the
existing error propagation and fallback behavior.
In `@src/cmd/try_cmd.rs`:
- Around line 668-809: Add tests around the explicit-version handling in
try_package: cover an unknown requested version and verify the error lists
available versions, and cover an incompatible requested version while asserting
the retry hint preserves the `@version` suffix. Also add coverage for an
unparsable explicit version returning the expected Invalid version error; ensure
install and activation callbacks are not invoked on these failures.
- Around line 178-181: Update report_incompatible and its retry-hint formatting
to accept and use the original package_spec instead of package_id for the numan
try command. Modify both report_incompatible call sites in try_package (the
indicated branches) to pass package_spec, preserving the explicitly requested
version in retry instructions.
- Around line 159-182: Update the recommendation flow around
select_recommended_nu so an explicit package version only considers Nu
candidates compatible with that requested version entry, or suppresses the
recommendation when the selected rec equals current_nu. Ensure the output never
labels the current Nu as a recommendation or emits a no-op numan use command,
while preserving recommendations for other compatible Nu versions.
In `@src/cmd/use_cmd.rs`:
- Around line 646-651: Remove the vacuous comparison in the ordering test around
order and unreg_pos. Instrument the observable path_refresh boundary used by
switch_active_nu_version after the active-version marker write, record its
position in order, and assert that the unregister hook position precedes that
boundary while preserving the existing presence assertion and message intent.
In `@src/core/resolve.rs`:
- Around line 526-528: Update src/core/resolve.rs lines 526-528 in version_score
to avoid packed u64 arithmetic by returning a non-overflowing version
representation, then update lines 594-621 to extract one shared nearest-version
comparator used by both min_by sites. Compare version components with abs_diff
and remove the duplicated closures and unsafe i64 casts.
- Around line 464-468: Update the public doc comment for the candidate-version
function in resolve.rs to remove the claim that it includes verified-with
versions; describe only candidates derived from the nu_version constraint, while
preserving the existing lower-bound and exclusive-upper-bound behavior.
- Around line 941-984: Remove the unused verified parameter from
plugin_with_verified and update both callers to construct the fixture without
it, unless the tests are intended to validate verified_with; in that case, add
an assertion whose result depends on the fixture’s verified_with value.
- Around line 594-621: Extract the duplicated nearest-version comparison logic
from the installed_all and compatible min_by calls into one reusable comparator
near the surrounding resolution logic. Define its parameter signature to
accommodate the differing reference depth at both call sites, use u64::abs_diff
instead of casting scores through i64, and invoke the shared comparator in both
places while preserving the existing tie-breaker favoring the higher version
score.
In `@src/state/activation_profile.rs`:
- Around line 190-225: Introduce a batch profile-update API alongside
ensure_contains_for_paths and ensure_absent_for_paths that loads the profile
once, applies the requested operation to all package IDs, and saves at most once
when changes occur. Update sync_profile_after_user_activate and
sync_profile_after_user_deactivate to classify all IDs first and call the batch
API instead of invoking the per-ID helpers in a loop, preserving the existing
minor/version and ProfileKind behavior.
- Around line 85-130: Update ActivationProfile::profile_path to store
activation-profile.json under the state directory so create_snapshot rollback
captures it with the related state. Keep load, load_or_default, and save using
this centralized path helper, preserving their existing behavior.
---
Outside diff comments:
In `@src/cmd/try_cmd.rs`:
- Around line 294-342: In try_package, scope the acquire_mutation_lock(root)
guard only around the install operation so it is dropped before the activate
call. Keep installation setup and result handling under the lock, then release
it before invoking activate; update any try_cmd tests that assert the outer lock
remains held.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: 96daf7ba-7c48-4cac-b5a8-0ee2fad1ade7
📒 Files selected for processing (16)
AGENTS.mdCHANGELOG.mdREADME.mdsrc/cli.rssrc/cmd/activate.rssrc/cmd/activation_switch.rssrc/cmd/deactivate.rssrc/cmd/doctor.rssrc/cmd/mod.rssrc/cmd/registry.rssrc/cmd/remove.rssrc/cmd/try_cmd.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/state/activation_profile.rssrc/state/mod.rs
🔗 Linked repositories identified
CodeRabbit considers these linked repositories for cross-repo context during reviews:
tonythethompson/QuickShell(manual)tonythethompson/numan(manual)tonythethompson/dependency-chain-substrate(manual)
📜 Review details
⏰ Context from checks skipped due to timeout. (1)
- GitHub Check: Greptile Review
🧰 Additional context used
📓 Path-based instructions (15)
**/*
📄 CodeRabbit inference engine (CLAUDE.md)
**/*: Use Serena's semantic, symbol-aware tools as the primary tools for reading and editing code; use built-in Read, Glob, Grep, and Edit only under the stated exceptions.
Before editing a code file, inspect its symbol overview, read the specific symbols being changed, and edit them with Serena's symbol-aware tools.
Understand the existing code before changing it and make the smallest change that satisfies the request; avoid unrelated cleanup, premature abstractions, unnecessary error handling, feature flags, and compatibility shims.
Prefer editing existing files over creating new ones, and never create Markdown or README files unless explicitly requested.
For exploratory questions, provide a 2–3 sentence recommendation with the main tradeoff and do not implement until the user agrees.
For UI or frontend changes that cannot be tested in a browser, explicitly state that browser testing was not performed rather than claiming success.
Address security issues when discovered, including injection, XSS, SQL injection, path traversal, and secret leaks.
Pause and obtain confirmation before destructive, hard-to-reverse, externally visible, or third-party-upload actions, including deleting files or branches, dropping tables, force-pushing, modifying CI/CD, posting externally, or uploading content.
When blocked, investigate the root cause instead of bypassing it with --no-verify, --force, or deletion; investigate unfamiliar files, branches, and configuration before deleting them.
Only commit when explicitly asked; do not proactively update git configuration or push changes.
Do not skip Git hooks unless explicitly asked; if a pre-commit hook fails, fix the issue, re-stage, and create a new commit rather than amending.
Stage files by name rather than using git add -A or git add ., to avoid accidentally including secrets or large binaries.
Use a HEREDOC for commit messages, and add co-author attribution only when the user explicitly requests the exact trailer.
Do not force-pus...
Files:
src/cmd/mod.rssrc/state/mod.rsCHANGELOG.mdsrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rsREADME.mdAGENTS.mdsrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
**/*.{js,jsx,ts,tsx,py,java,go,rs,rb,php,c,cpp,h,hpp,cs,swift,kt,kts}
📄 CodeRabbit inference engine (CLAUDE.md)
Add comments only when the WHY is non-obvious; do not narrate what the code does, reference the current task, or reference the PR in comments.
Files:
src/cmd/mod.rssrc/state/mod.rssrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rssrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
!**/.env,!**/credentials.json,!**/*.pem
📄 CodeRabbit inference engine (CLAUDE.md)
Do not commit files that appear to contain secrets, including .env, credentials.json, and PEM files; warn before doing so even if explicitly requested.
Files:
src/cmd/mod.rssrc/state/mod.rsCHANGELOG.mdsrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rsREADME.mdAGENTS.mdsrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
**/*.rs
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.rs: Use the Rust 2021 edition.
Useanyhow::Resultwith.context(...)in application code; usethiserrorfor library error types that callers match on.
Useclapderive macros for CLI definitions.
Useserdewithserde_jsonortomlfor serialization.
Function parameters must use&Path, not&PathBuf.
Library code must not panic; error paths should returnanyhow::Resultwith context where appropriate.
Add or update tests for behavior changes, including relevant failure paths.
New mutating code paths must acquire the mutation lock viaacquire_mutation_lock(root)and snapshot the lockfile before writes.
Lockfile, journal, and state-file JSON writes must usewrite_json_atomic.
numan installmust write only to$NUMAN_ROOT; it must not invoke Nu or register plugins/autoloads.
Onlyactivateanddeactivatemay modify Nu integration state.
Treat the lockfile as the authoritative source of truth; derived projections such as autoload state must not be authoritative.
Install payloads under versioned, content-addressed paths and never overwrite them in place.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass plugin paths through environment variables only; do not use runtime interpolation in Nu program strings.
**/*.rs: All CI gates must pass:cargo test,cargo clippy -- -D warnings, andcargo fmt --check.
Every mutating command—includinginstall,remove,update,gc, and futurenupm import—must callacquire_mutation_lock(root).
Lockfiles, journals, and state files must usewrite_json_atomic; partial writes are not allowed.
Pending activation, autoload, and lifecycle journals must be stored under$NUMAN_ROOT/state/.
Module autoload identity must match all four fields: Nu executable hash, Nu version, vendor autoload directory, and managed file path; the lockfilemodule_activationvalue is authoritative.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass paths to Nu only throu...
Files:
src/cmd/mod.rssrc/state/mod.rssrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rssrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
**/*.{rs,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Run and keep
cargo fmt/rustfmtclean, and ensurecargo clippy -- -D warningspasses.Use
clapderive macros for the CLI andserde/serde_json/tomlfor serialization according to the established crate conventions.
Files:
src/cmd/mod.rssrc/state/mod.rssrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rssrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
**/*.{rs,nu}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.{rs,nu}: Real-Nu acceptance tests must be marked#[ignore]and should be run when changes affect activation or nupm import; unit tests must not spawn realnuand should use injectable seams such asFakeCandidateRunneror registrars.
The nupm integration must be read-only towardNUPM_HOME, must not executebuild.nu, and must not perform bidirectional synchronization.Unit tests must use
FakeCandidateRunneror injectable registrars and must not spawn a realnuprocess.
Files:
src/cmd/mod.rssrc/state/mod.rssrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rssrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
**/*.{rs,md}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Match existing naming, module layout, and documentation level in the file being edited; update
AGENTS.md,docs/, or command help when structure, conventions, or user-visible behavior changes.Tests must cover failure modes, not only successful execution.
Files:
src/cmd/mod.rssrc/state/mod.rsCHANGELOG.mdsrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rsREADME.mdAGENTS.mdsrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
**/*.{rs,md,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Use the repository's established serialization and module conventions rather than introducing unrelated refactors.
Files:
src/cmd/mod.rssrc/state/mod.rsCHANGELOG.mdsrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rsREADME.mdAGENTS.mdsrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
src/**/*.rs
📄 CodeRabbit inference engine (AGENTS.md)
Platform detection must use compile-time
#[cfg(target_env)]information from the build target, notstd::env::consts.
Files:
src/cmd/mod.rssrc/state/mod.rssrc/cmd/registry.rssrc/cmd/remove.rssrc/cli.rssrc/cmd/doctor.rssrc/cmd/activate.rssrc/state/activation_profile.rssrc/cmd/deactivate.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/cmd/activation_switch.rssrc/cmd/try_cmd.rs
src/state/**/*.rs
📄 CodeRabbit inference engine (AGENTS.md)
src/state/**/*.rs: The lockfile is authoritative for module activation;autoload-state.jsonis only a derived fast-check projection.
All JSON state files, including lockfiles, journals, andnu_state/paths.json, must be written atomically withwrite_json_atomicusing a temporary file in the same directory.
Acquireacquire_mutation_lock(root)before mutations; a second acquisition for the same root must fail immediately rather than block.
Create a snapshot before install, update, remove, activate, deactivate, nupm import, andinit --refreshmutations; snapshots must capturenu_state/paths.jsonwhen present.
Files:
src/state/mod.rssrc/state/activation_profile.rs
**/*.md
📄 CodeRabbit inference engine (REVIEW.md)
Update documentation and
AGENTS.mdwhen project structure or conventions change.
Files:
CHANGELOG.mdREADME.mdAGENTS.md
src/cmd/{doctor,use_cmd,init}.rs
📄 CodeRabbit inference engine (AGENTS.md)
Migration journals must be reconciled only when valid and safe; refuse reconciliation when
tools/nushellis a symlink or reparse point, and retain journals when repair fails.
Files:
src/cmd/doctor.rssrc/cmd/use_cmd.rs
src/cmd/{activate,deactivate,plugin_lifecycle,update}.rs
📄 CodeRabbit inference engine (AGENTS.md)
Nu plugin register/unregister callbacks are owned by the activate/deactivate lifecycle boundary; only explicitly opted-in
updateorchestration may coordinate that boundary and must not invoke callbacks directly.
Files:
src/cmd/activate.rssrc/cmd/deactivate.rs
src/cmd/{activate,deactivate}.rs
📄 CodeRabbit inference engine (AGENTS.md)
Keep activation and deactivation testable through injected registrar, unregistrar, and candidate-runner seams; unit tests must not spawn a real Nu binary.
Files:
src/cmd/activate.rssrc/cmd/deactivate.rs
src/state/activation_profile.rs
📄 CodeRabbit inference engine (AGENTS.md)
Activation profiles store desired plugin/module sets per Nu minor; cross-minor switching tears down modules then plugins and restores plugins then modules. Same-target switching is restore-only.
Files:
src/state/activation_profile.rs
🪛 markdownlint-cli2 (0.23.2)
AGENTS.md
[warning] 240-240: Headings should be surrounded by blank lines
Expected: 1; Actual: 0; Below
(MD022, blanks-around-headings)
🔍 Remote MCP DeepWiki, GitHub Copilot
Additional review context
- The repository treats
Lockfileactivation records as authoritative; plugin/module activity is tied to exact Nu identity fields, while journals support crash recovery. `` - PR
#108adds a 1,284-line activation-switch module coordinating teardown, marker writes, path refresh, and restoration. `` - Potential issue:
leave_current_nutreats allLockfile::loaderrors as “no lockfile,” although malformed/unreadable lockfiles are otherwise propagated. This could permit switching with corrupted state. - Potential issue: activation-profile synchronization after
activateuses every requested ID even when some activations failed, potentially recording failed packages as desired restorations. `` - Potential issue: an incompatible explicit
try owner/name@versionreports a retry command without@version, so the suggested retry may target a different package version. `` - Restore failures caused by missing/drifted cached paths are emitted as warnings after the active-version marker has been written; the switch can therefore complete without restoring the desired profile. ``
- CI currently reports Rust analysis, CodeQL, CodeFactor, and other analysis checks successful; Greptile remains in progress. Sourcery skipped review because the diff exceeded its 150,000-character limit. ``
🔇 Additional comments (51)
src/cmd/doctor.rs (2)
493-528: LGTM!
2343-2410: LGTM!src/cmd/registry.rs (3)
183-207: LGTM!
231-255: LGTM!
257-323: LGTM!AGENTS.md (1)
68-70: LGTM!Also applies to: 88-88, 235-239, 241-251
CHANGELOG.md (1)
18-20: LGTM!Also applies to: 24-24
README.md (1)
202-202: LGTM!Also applies to: 246-246, 263-263, 337-337
src/cli.rs (1)
62-62: LGTM!src/cmd/try_cmd.rs (10)
4-25: LGTM!
27-37: LGTM!
39-90: LGTM!
194-232: LGTM!
253-290: LGTM!
388-479: LGTM!
481-506: LGTM!
508-666: LGTM!
811-839: LGTM!
188-192: 📐 Maintainability & Code QualityRun
cargo fmt --all -- --checkbefore merge.This expression has long call-chains and a long string literal; the CI formatter job runs
cargo fmt --all -- --check, so manual formatter status is needed before merge.src/core/resolve.rs (5)
469-502: LGTM!
504-524: LGTM!
530-590: LGTM!
924-939: LGTM!
986-1070: LGTM!src/state/activation_profile.rs (4)
1-68: LGTM!
132-158: LGTM!
161-169: LGTM!
227-338: LGTM!src/state/mod.rs (1)
1-1: LGTM!src/cmd/activate.rs (2)
140-145: LGTM!Also applies to: 181-185
601-638: LGTM!src/cmd/deactivate.rs (6)
128-128: LGTM!Also applies to: 166-166, 266-281
189-200: LGTM!
231-256: LGTM!
405-458: LGTM!
481-482: LGTM!Also applies to: 570-574
1556-1574: LGTM!src/cmd/remove.rs (1)
164-166: LGTM!src/cmd/activation_switch.rs (8)
1-82: LGTM!
124-146: LGTM!
184-221: LGTM!
264-344: LGTM!
426-514: LGTM!Also applies to: 517-559, 562-593
595-1284: LGTM!
86-91: 🎯 Functional CorrectnessNo change needed: managed autoload paths use the normalized
/numan.nuform.
236-243: 🎯 Functional CorrectnessNo change needed.
numan use 0.114parses and normalizes the minor-only argument, and the subsequent minor-key reconciliation uses that same parsed value.src/cmd/mod.rs (1)
2-2: LGTM!src/cmd/use_cmd.rs (4)
7-19: LGTM!
31-98: LGTM!
122-144: LGTM!
515-645: LGTM!Also applies to: 652-657
|
All
|
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
src/cmd/try_cmd.rs (1)
700-704: 🩺 Stability & Availability | 🔵 Trivial | ⚡ Quick winAdd a regression test for the lock handoff.
The activation callback only sets
activated. The test passes even if Line 333 is removed. Add a timeout-controlled test seam that attempts to acquire the same root mutation lock during activation.As per coding guidelines: “Tests must cover failure modes, not only successful execution.”
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/try_cmd.rs` around lines 700 - 704, Add a regression test around the activation callback in the try command test to verify lock handoff: during activation, attempt to acquire the same root mutation lock using a timeout-controlled test seam, and assert the acquisition succeeds. Keep the existing package assertion and activated tracking, while ensuring the test would fail if the lock remains held through activation.Source: Coding guidelines
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Outside diff comments:
In `@src/cmd/try_cmd.rs`:
- Around line 700-704: Add a regression test around the activation callback in
the try command test to verify lock handoff: during activation, attempt to
acquire the same root mutation lock using a timeout-controlled test seam, and
assert the acquisition succeeds. Keep the existing package assertion and
activated tracking, while ensuring the test would fail if the lock remains held
through activation.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: 9d6d0249-586d-4af7-96f9-be5a9f52fd2b
📒 Files selected for processing (1)
src/cmd/try_cmd.rs
🔗 Linked repositories identified
CodeRabbit considers these linked repositories for cross-repo context during reviews:
tonythethompson/QuickShell(manual)tonythethompson/numan(manual)tonythethompson/dependency-chain-substrate(manual)
📜 Review details
⏰ Context from checks skipped due to timeout. (1)
- GitHub Check: Greptile Review
🧰 Additional context used
📓 Path-based instructions (10)
**/*
📄 CodeRabbit inference engine (CLAUDE.md)
**/*: Use Serena's semantic, symbol-aware tools as the primary tools for reading and editing code; use built-in Read, Glob, Grep, and Edit only under the stated exceptions.
Before editing a code file, inspect its symbol overview, read the specific symbols being changed, and edit them with Serena's symbol-aware tools.
Understand the existing code before changing it and make the smallest change that satisfies the request; avoid unrelated cleanup, premature abstractions, unnecessary error handling, feature flags, and compatibility shims.
Prefer editing existing files over creating new ones, and never create Markdown or README files unless explicitly requested.
For exploratory questions, provide a 2–3 sentence recommendation with the main tradeoff and do not implement until the user agrees.
For UI or frontend changes that cannot be tested in a browser, explicitly state that browser testing was not performed rather than claiming success.
Address security issues when discovered, including injection, XSS, SQL injection, path traversal, and secret leaks.
Pause and obtain confirmation before destructive, hard-to-reverse, externally visible, or third-party-upload actions, including deleting files or branches, dropping tables, force-pushing, modifying CI/CD, posting externally, or uploading content.
When blocked, investigate the root cause instead of bypassing it with --no-verify, --force, or deletion; investigate unfamiliar files, branches, and configuration before deleting them.
Only commit when explicitly asked; do not proactively update git configuration or push changes.
Do not skip Git hooks unless explicitly asked; if a pre-commit hook fails, fix the issue, re-stage, and create a new commit rather than amending.
Stage files by name rather than using git add -A or git add ., to avoid accidentally including secrets or large binaries.
Use a HEREDOC for commit messages, and add co-author attribution only when the user explicitly requests the exact trailer.
Do not force-pus...
Files:
src/cmd/try_cmd.rs
**/*.{js,jsx,ts,tsx,py,java,go,rs,rb,php,c,cpp,h,hpp,cs,swift,kt,kts}
📄 CodeRabbit inference engine (CLAUDE.md)
Add comments only when the WHY is non-obvious; do not narrate what the code does, reference the current task, or reference the PR in comments.
Files:
src/cmd/try_cmd.rs
!**/.env,!**/credentials.json,!**/*.pem
📄 CodeRabbit inference engine (CLAUDE.md)
Do not commit files that appear to contain secrets, including .env, credentials.json, and PEM files; warn before doing so even if explicitly requested.
Files:
src/cmd/try_cmd.rs
**/*.rs
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.rs: Use the Rust 2021 edition.
Useanyhow::Resultwith.context(...)in application code; usethiserrorfor library error types that callers match on.
Useclapderive macros for CLI definitions.
Useserdewithserde_jsonortomlfor serialization.
Function parameters must use&Path, not&PathBuf.
Library code must not panic; error paths should returnanyhow::Resultwith context where appropriate.
Add or update tests for behavior changes, including relevant failure paths.
New mutating code paths must acquire the mutation lock viaacquire_mutation_lock(root)and snapshot the lockfile before writes.
Lockfile, journal, and state-file JSON writes must usewrite_json_atomic.
numan installmust write only to$NUMAN_ROOT; it must not invoke Nu or register plugins/autoloads.
Onlyactivateanddeactivatemay modify Nu integration state.
Treat the lockfile as the authoritative source of truth; derived projections such as autoload state must not be authoritative.
Install payloads under versioned, content-addressed paths and never overwrite them in place.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass plugin paths through environment variables only; do not use runtime interpolation in Nu program strings.
**/*.rs: All CI gates must pass:cargo test,cargo clippy -- -D warnings, andcargo fmt --check.
Every mutating command—includinginstall,remove,update,gc, and futurenupm import—must callacquire_mutation_lock(root).
Lockfiles, journals, and state files must usewrite_json_atomic; partial writes are not allowed.
Pending activation, autoload, and lifecycle journals must be stored under$NUMAN_ROOT/state/.
Module autoload identity must match all four fields: Nu executable hash, Nu version, vendor autoload directory, and managed file path; the lockfilemodule_activationvalue is authoritative.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass paths to Nu only throu...
Files:
src/cmd/try_cmd.rs
**/*.{rs,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Run and keep
cargo fmt/rustfmtclean, and ensurecargo clippy -- -D warningspasses.
Files:
src/cmd/try_cmd.rs
**/*.{rs,nu}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.{rs,nu}: Real-Nu acceptance tests must be marked#[ignore]and should be run when changes affect activation or nupm import; unit tests must not spawn realnuand should use injectable seams such asFakeCandidateRunneror registrars.
The nupm integration must be read-only towardNUPM_HOME, must not executebuild.nu, and must not perform bidirectional synchronization.Unit tests must use
FakeCandidateRunneror injectable registrars and must not spawn a realnuprocess.
Files:
src/cmd/try_cmd.rs
**/*.{rs,md}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Match existing naming, module layout, and documentation level in the file being edited; update
AGENTS.md,docs/, or command help when structure, conventions, or user-visible behavior changes.Tests must cover failure modes, not only successful execution.
Files:
src/cmd/try_cmd.rs
**/*.{rs,md,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Use the repository's established serialization and module conventions rather than introducing unrelated refactors.
Files:
src/cmd/try_cmd.rs
src/{cmd,state,config,nu}/**/*.rs
📄 CodeRabbit inference engine (AGENTS.md)
Create a snapshot before mutations performed by install, update, remove, activate, deactivate, nupm import, or
init --refresh.
Files:
src/cmd/try_cmd.rs
**/*.{md,rs,toml,yml,yaml}
📄 CodeRabbit inference engine (AGENTS.md)
Keep user-facing fix hints aligned with
docs/numan-doctor.md, and keep review conventions aligned withREVIEW.mdand the GitHub review instructions.
Files:
src/cmd/try_cmd.rs
🧠 Learnings (1)
📓 Common learnings
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-07T23:17:09.897Z
Learning: Do not silently auto-switch the user's Nushell version; use compatibility-aware UX such as `numan try`, honest search/install results, and offer-based managed Nu pinning.
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-07T23:17:09.897Z
Learning: Use imperative commit messages shorter than 72 characters, feature/fix branch naming, no force-pushes to `master`, and squash merges for features.
🔍 Remote MCP DeepWiki, GitHub Copilot
Relevant review context
-
Confirmed critical issue:
leave_current_nutreats everyLockfile::loaderror as “no lockfile.” Since missing lockfiles already load as empty, this swallows malformed/unreadable lockfiles and allows version switching without tearing down active packages. DeepWiki confirms malformed lockfiles must block automated operations. -
Confirmed
numan trylock behavior: the PR explicitly drops the install lock before invokingactivate, addressing the non-reentrant lock failure identified in review. -
Unresolved pinned-version correctness issue: incompatible
owner/name@versionrequests pass the unversioned package ID into compatibility reporting. Recommendations are computed across all releases and the retry hint omits@version, potentially switching to or installing a different release. -
Unresolved activation-profile integrity issue: explicit activation requests are synchronized into the profile without checking that activation records were successfully written; failed activations may become desired restore state.
-
Unresolved same-target path issue:
numan use latestcan repair a dangling marker path, then enter the same-target reconciliation path without refreshingpaths.json, leaving cached Nu identity potentially stale. -
Unresolved restore fail-open issue: restore classification treats missing/unavailable registry information as compatible, allowing restoration without validating Nu compatibility.
-
Test gap: the ordering assertion compares the unregister position with
order.len(), so it cannot prove teardown occurred before marker/path switching. -
CI observations: CodeQL, CodeFactor, and language analysis checks reported success; Greptile review and Rust analysis were still in progress, while the CodeRabbit status was pending.
🔇 Additional comments (5)
src/cmd/try_cmd.rs (5)
287-289: Preserve explicit-version context in incompatibility reporting.At Line 289,
target_entryis already classified, butreport_incompatiblereceives onlypackage_id. Forowner/name@version, the report can use compatibility data from another release and print a retry command without@version. Pass the originalpackage_specand scope compatibility reporting to the selected version. This repeats the previous review finding.
673-709: Cover explicit-version failure paths.The supplied
try_packagetests passNoneforversion. They do not cover invalid, unavailable, or incompatible explicit versions. Add these cases and verify that installation and activation callbacks are not called. This repeats the previous review finding.As per coding guidelines: “Tests must cover failure modes, not only successful execution.”
Source: Coding guidelines
292-349: LGTM!
399-451: LGTM!
513-541: LGTM!
- Move activation-profile.json from nu_state/ to state/ so snapshots and rollback capture it alongside other state files. Added activation_profile sidecar to Snapshot/SidecarDigests, capture in create_snapshot, restore in rollback_to_snapshot, new ActivationProfileCommitted lifecycle stage. - Extract duplicated nearest-version comparator in resolve.rs into nearest_comparator helper using u64::abs_diff. - Remove unused verified parameter from plugin_with_verified test fixture (renamed to windows_plugin). - Preserve explicit-version context in try_cmd incompatibility reporting: report_incompatible now accepts package_spec and optional target_entry so pinned owner/name@version scopes compatibility to that release and the retry hint includes @Version. - Add compatible_nu_versions_for_entry to resolve.rs for scoped compatibility checks against a single VersionEntry. - Add tests: invalid version format, unavailable version, incompatible pinned version with scoped compat, lock handoff regression. Generated with [Devin](https://devin.ai) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
- Critical: propagate Lockfile::load errors in leave_current_nu instead of swallowing them. A missing lockfile already returns Ok(empty), so the Err arm only fires on corrupt state -- switching must stop there. - Major: filter args.packages by post-activation liveness in sync_user_activate_profile so failed activation lanes are not recorded as desired state for numan use restore. - Major: fix tautological ordering assertion in use_cmd test to assert against the observable path_refresh boundary (unreg < refresh < reg). - Minor: add .context() to active_nu_binary and latest_installed_version in resolve_managed_nu_binary for actionable doctor diagnostics. - Minor: remove behavior-narrating doc comments from registry.rs helpers. - Minor: fix incorrect doc comment on candidate_nu_versions_from_constraint (it does not include verified-with versions; caller adds those). - Trivial: add rationale comment for plugin-only fallback in classify_restore_target (plugins need target match; modules are scripts). Generated with [Devin](https://devin.ai) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Actionable comments posted: 7
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
src/cmd/use_cmd.rs (1)
532-541: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick winBoth fake Nu versions hash to the same value, so the identity dimension is inert.
Line 534 hashes
b"fake". Line 539 hashes"fake".as_bytes(). Those are the same bytes, sohash_113 == hash_114.The lockfile activation at Line 583 records
nu_executable_sha256: hash_113. Afterpath_refreshruns,nu_paths.nu_executable_hashishash_114, which equalshash_113. Any switch-path check that compares the recorded executable hash against the live one therefore cannot distinguish the two Nu installations in this test. The test name claims a cross-minor switch, but the executable-identity dimension is degenerate.Write distinct bytes per version, and hash the same bytes the test writes to disk.
💚 Proposed fix — give each version a distinct hash
fn create_fake_version(root: &Path, version: &str) { let binary_name = if cfg!(windows) { "nu.exe" } else { "nu" }; let dir = version_manager::version_install_dir(root, version); std::fs::create_dir_all(&dir).unwrap(); - std::fs::write(dir.join(binary_name), "fake").unwrap(); + std::fs::write(dir.join(binary_name), format!("fake-{version}")).unwrap(); }let (nu_exe_113, hash_113) = { let binary = version_manager::version_binary(root, "0.113.1"); - let hash = integrity::compute_sha256(b"fake"); + let hash = integrity::compute_sha256(b"fake-0.113.1"); (binary.to_string_lossy().into_owned(), hash) }; let (nu_exe_114, hash_114) = { let binary = version_manager::version_binary(root, "0.114.0"); - let hash = integrity::compute_sha256("fake".as_bytes()); + let hash = integrity::compute_sha256(b"fake-0.114.0"); (binary.to_string_lossy().into_owned(), hash) }; + assert_ne!(hash_113, hash_114, "fake Nu binaries must have distinct hashes");Check the other tests in this module that rely on
create_fake_version;test_use_switch_clears_stale_paths_cache_when_probe_failsuses a literal"deadbeef"hash and is unaffected.As per coding guidelines: "Module autoload identity must match all four fields: Nu executable hash, Nu version, vendor autoload directory, and managed file path."
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/use_cmd.rs` around lines 532 - 541, Update the fake Nu version setup in the test around nu_exe_113 and nu_exe_114 to write distinct per-version bytes to disk and compute each hash from the exact bytes written. Keep the resulting hash_113 and hash_114 values different so the cross-minor switch validates executable identity along with the other autoload fields; do not change unrelated tests such as test_use_switch_clears_stale_paths_cache_when_probe_fails.Source: Coding guidelines
README.md (1)
262-262: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winQualify
numan tryactivation by package type.The documentation treats compatibility as sufficient for activation. Scripts and completion packages remain install-only, so the documentation must distinguish installation from activation.
README.md#L262-L262: state thatnuman tryactivates only activatable package types.README.md#L214-L214: qualify the compatible-package activation statement.CHANGELOG.md#L18-L18: describe activation as conditional on the package type.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@README.md` at line 262, Update the `numan try` documentation to distinguish installation from activation: in README.md lines 262-262 and 214-214, state that activation occurs only for compatible, activatable package types while scripts and completion packages remain install-only; update CHANGELOG.md lines 18-18 with the same conditional activation wording.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@AGENTS.md`:
- Line 91: Document that activation-profile.json is captured, digested, and
restored only for new snapshots; legacy snapshots omit it and leave the live
activation profile unchanged during rollback. Update AGENTS.md lines 91 and 156
to qualify the project-structure summary and activation-profile convention, and
update CHANGELOG.md line 14 to state that the capture and rollback behavior
applies only to new snapshots.
In `@CHANGELOG.md`:
- Line 14: Update the changelog entry describing activation-profile snapshot
capture and restoration to qualify that only new snapshots include and digest
activation-profile.json; specify that legacy snapshots omit it and leave the
live profile unchanged during rollback, while preserving the existing behavior
descriptions for new snapshots.
In `@src/cmd/deactivate.rs`:
- Around line 393-417: Keep the runner validation in the terminal deactivation
flow, but avoid calling run_partial_deactivation when remaining_ids is empty;
route that case through the existing journaled run_full_deactivation path so
deletion of the managed file and AutoloadState is journaled and the journal is
cleared last. Preserve the non-terminal partial-deactivation path, and add
coverage verifying runner-driven final deactivation leaves no managed autoload
file, no autoload-state.json, and no journal.
In `@src/core/resolve.rs`:
- Around line 526-551: Extract the shared candidate iteration, version
deduplication, and sorting logic from compatible_nu_versions and
compatible_nu_versions_for_entry into a private helper that accepts the
compatibility predicate. Update both public functions to call this helper with
their respective predicates, preserving the existing deduplication key,
ascending sort order, and compatibility behavior.
In `@src/nu/bootstrap.rs`:
- Around line 1462-1468: Move PathRestoreGuard::new() to before the PATH
inspection/source scan so the process-global environment is protected throughout
the test. Update every PATH-mutating test, including
prepend_process_path_adds_canonical_dir, to acquire and retain the same guard
before reading or modifying PATH.
- Line 1459: Update the set_permissions call in the bootstrap
permission-handling code to pass perms by value instead of borrowing it.
Preserve the existing existing path and permission-setting behavior.
In `@src/state/snapshot.rs`:
- Around line 692-708: Update src/state/snapshot.rs:692-708 in
capture_activation_profile_sidecar to return SnapshotSidecar::Absent or Present
rather than Option, always compute activation_profile_sha256 with ?, and always
write the sidecar so missing profiles are represented. Ensure the documentation
at src/state/snapshot.rs:163-175 states that new snapshots record Absent or
Present. In src/state/rollback.rs:269-275, retain the SnapshotSidecar::Absent
handling, verify it is reachable, and add a test confirming rollback removes
state/activation-profile.json when the snapshot recorded no profile.
---
Outside diff comments:
In `@README.md`:
- Line 262: Update the `numan try` documentation to distinguish installation
from activation: in README.md lines 262-262 and 214-214, state that activation
occurs only for compatible, activatable package types while scripts and
completion packages remain install-only; update CHANGELOG.md lines 18-18 with
the same conditional activation wording.
In `@src/cmd/use_cmd.rs`:
- Around line 532-541: Update the fake Nu version setup in the test around
nu_exe_113 and nu_exe_114 to write distinct per-version bytes to disk and
compute each hash from the exact bytes written. Keep the resulting hash_113 and
hash_114 values different so the cross-minor switch validates executable
identity along with the other autoload fields; do not change unrelated tests
such as test_use_switch_clears_stale_paths_cache_when_probe_fails.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: d22b821b-5008-4a4b-bb62-9cd9b942d25f
📒 Files selected for processing (13)
.markdownlint.jsonAGENTS.mdCHANGELOG.mdREADME.mdsrc/cmd/deactivate.rssrc/cmd/try_cmd.rssrc/cmd/use_cmd.rssrc/core/resolve.rssrc/nu/bootstrap.rssrc/state/activation_profile.rssrc/state/lifecycle_journal.rssrc/state/rollback.rssrc/state/snapshot.rs
🔗 Linked repositories identified
CodeRabbit considers these linked repositories for cross-repo context during reviews:
tonythethompson/QuickShell(manual)tonythethompson/numan(manual)tonythethompson/dependency-chain-substrate(manual)
📜 Review details
⏰ Context from checks skipped due to timeout. (2)
- GitHub Check: Greptile Review
- GitHub Check: Analyze (rust)
🧰 Additional context used
📓 Path-based instructions (15)
!**/.env,!**/credentials.json,!**/*.pem
📄 CodeRabbit inference engine (CLAUDE.md)
Do not commit files that appear to contain secrets, including .env, credentials.json, and PEM files; warn before doing so even if explicitly requested.
Files:
.markdownlint.jsonsrc/state/lifecycle_journal.rsCHANGELOG.mdsrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rsAGENTS.mdsrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rsREADME.mdsrc/cmd/try_cmd.rs
**/*
📄 CodeRabbit inference engine (CLAUDE.md)
**/*: Use Serena's semantic, symbol-aware tools as the primary tools for reading and editing code; use built-in Read, Glob, Grep, and Edit only under the stated exceptions.
Before editing a code file, inspect its symbol overview, read the specific symbols being changed, and edit them with Serena's symbol-aware tools.
Understand the existing code before changing it and make the smallest change that satisfies the request; avoid unrelated cleanup, premature abstractions, unnecessary error handling, feature flags, and compatibility shims.
Prefer editing existing files over creating new ones, and never create Markdown or README files unless explicitly requested.
For exploratory questions, provide a 2–3 sentence recommendation with the main tradeoff and do not implement until the user agrees.
For UI or frontend changes that cannot be tested in a browser, explicitly state that browser testing was not performed rather than claiming success.
Address security issues when discovered, including injection, XSS, SQL injection, path traversal, and secret leaks.
Pause and obtain confirmation before destructive, hard-to-reverse, externally visible, or third-party-upload actions, including deleting files or branches, dropping tables, force-pushing, modifying CI/CD, posting externally, or uploading content.
When blocked, investigate the root cause instead of bypassing it with --no-verify, --force, or deletion; investigate unfamiliar files, branches, and configuration before deleting them.
Only commit when explicitly asked; do not proactively update git configuration or push changes.
Do not skip Git hooks unless explicitly asked; if a pre-commit hook fails, fix the issue, re-stage, and create a new commit rather than amending.
Stage files by name rather than using git add -A or git add ., to avoid accidentally including secrets or large binaries.
Use a HEREDOC for commit messages, and add co-author attribution only when the user explicitly requests the exact trailer.
Do not force-pus...
Files:
src/state/lifecycle_journal.rsCHANGELOG.mdsrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rsAGENTS.mdsrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rsREADME.mdsrc/cmd/try_cmd.rs
**/*.{js,jsx,ts,tsx,py,java,go,rs,rb,php,c,cpp,h,hpp,cs,swift,kt,kts}
📄 CodeRabbit inference engine (CLAUDE.md)
Add comments only when the WHY is non-obvious; do not narrate what the code does, reference the current task, or reference the PR in comments.
Files:
src/state/lifecycle_journal.rssrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rssrc/cmd/try_cmd.rs
**/*.rs
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.rs: Use the Rust 2021 edition.
Useanyhow::Resultwith.context(...)in application code; usethiserrorfor library error types that callers match on.
Useclapderive macros for CLI definitions.
Useserdewithserde_jsonortomlfor serialization.
Function parameters must use&Path, not&PathBuf.
Library code must not panic; error paths should returnanyhow::Resultwith context where appropriate.
Add or update tests for behavior changes, including relevant failure paths.
New mutating code paths must acquire the mutation lock viaacquire_mutation_lock(root)and snapshot the lockfile before writes.
Lockfile, journal, and state-file JSON writes must usewrite_json_atomic.
numan installmust write only to$NUMAN_ROOT; it must not invoke Nu or register plugins/autoloads.
Onlyactivateanddeactivatemay modify Nu integration state.
Treat the lockfile as the authoritative source of truth; derived projections such as autoload state must not be authoritative.
Install payloads under versioned, content-addressed paths and never overwrite them in place.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass plugin paths through environment variables only; do not use runtime interpolation in Nu program strings.
**/*.rs: All CI gates must pass:cargo test,cargo clippy -- -D warnings, andcargo fmt --check.
Every mutating command—includinginstall,remove,update,gc, and futurenupm import—must callacquire_mutation_lock(root).
Lockfiles, journals, and state files must usewrite_json_atomic; partial writes are not allowed.
Pending activation, autoload, and lifecycle journals must be stored under$NUMAN_ROOT/state/.
Module autoload identity must match all four fields: Nu executable hash, Nu version, vendor autoload directory, and managed file path; the lockfilemodule_activationvalue is authoritative.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass paths to Nu only throu...
Files:
src/state/lifecycle_journal.rssrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rssrc/cmd/try_cmd.rs
**/*.{rs,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Run and keep
cargo fmt/rustfmtclean, and ensurecargo clippy -- -D warningspasses.
Files:
src/state/lifecycle_journal.rssrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rssrc/cmd/try_cmd.rs
**/*.{rs,nu}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.{rs,nu}: Real-Nu acceptance tests must be marked#[ignore]and should be run when changes affect activation or nupm import; unit tests must not spawn realnuand should use injectable seams such asFakeCandidateRunneror registrars.
The nupm integration must be read-only towardNUPM_HOME, must not executebuild.nu, and must not perform bidirectional synchronization.Unit tests must use
FakeCandidateRunneror injectable registrars and must not spawn a realnuprocess.
Files:
src/state/lifecycle_journal.rssrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rssrc/cmd/try_cmd.rs
**/*.{rs,md}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Match existing naming, module layout, and documentation level in the file being edited; update
AGENTS.md,docs/, or command help when structure, conventions, or user-visible behavior changes.Tests must cover failure modes, not only successful execution.
Files:
src/state/lifecycle_journal.rsCHANGELOG.mdsrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rsAGENTS.mdsrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rsREADME.mdsrc/cmd/try_cmd.rs
**/*.{rs,md,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Use the repository's established serialization and module conventions rather than introducing unrelated refactors.
Files:
src/state/lifecycle_journal.rsCHANGELOG.mdsrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rsAGENTS.mdsrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rsREADME.mdsrc/cmd/try_cmd.rs
src/state/{lockfile,journal,plugin_deactivate_journal,migration_journal,autoload_journal,lifecycle_journal}.rs
📄 CodeRabbit inference engine (AGENTS.md)
src/state/{lockfile,journal,plugin_deactivate_journal,migration_journal,autoload_journal,lifecycle_journal}.rs: Write journal state atomically and reconcile pending journals on the next relevant operation; journal transitions must not leave inconsistent lockfile state.
Usewrite_json_atomicfor JSON state writes to prevent partial-write corruption.
Files:
src/state/lifecycle_journal.rs
src/**/*.rs
📄 CodeRabbit inference engine (AGENTS.md)
Use
serde/serde_jsonfor JSON serialization,tomlfor configuration, andclapderive macros for CLI definitions.
Files:
src/state/lifecycle_journal.rssrc/nu/bootstrap.rssrc/state/rollback.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/core/resolve.rssrc/state/activation_profile.rssrc/state/snapshot.rssrc/cmd/try_cmd.rs
**/*.md
📄 CodeRabbit inference engine (REVIEW.md)
Update documentation and
AGENTS.mdwhen project structure or conventions change.
Files:
CHANGELOG.mdAGENTS.mdREADME.md
src/cmd/{activate,deactivate,plugin_lifecycle,update}.rs
📄 CodeRabbit inference engine (AGENTS.md)
Nu plugin registration and unregistration belong exclusively to the activate/deactivate lifecycle boundary; opted-in update orchestration may coordinate that boundary but must not invoke Nu callbacks directly.
Files:
src/cmd/deactivate.rs
src/cmd/{activate,deactivate}.rs
📄 CodeRabbit inference engine (AGENTS.md)
Expose dependency-injection seams such as
execute_with_registrar,execute_with_unregistrar, and candidate-runner variants; unit tests must use fakes rather than spawning real Nu.
Files:
src/cmd/deactivate.rs
src/state/activation_profile.rs
📄 CodeRabbit inference engine (AGENTS.md)
Maintain desired activation sets per Nu minor; cross-minor switching tears down modules before plugins and restores plugins before modules, while same-target switching only restores missing desired activations.
Files:
src/state/activation_profile.rs
src/state/snapshot.rs
📄 CodeRabbit inference engine (AGENTS.md)
Create a snapshot before
install,update,remove,activate,deactivate, nupm import, andinit --refreshmutations; snapshots must capturenu_state/paths.jsonand keep referenced payloads live for GC.
Files:
src/state/snapshot.rs
🧠 Learnings (1)
📓 Common learnings
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-08T01:02:39.785Z
Learning: Run `cargo test`, `cargo clippy -- -D warnings`, and `cargo fmt --check`; CI also runs ignored real-Nu acceptance tests across Ubuntu, Windows, and macOS.
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-08T01:02:39.785Z
Learning: Source builds require explicit user consent before cloning or building, with separate consent scopes.
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-08T01:02:39.785Z
Learning: Create a feature branch, implement with tests, update AGENTS.md when conventions change, and submit a descriptive PR.
🪛 LanguageTool
AGENTS.md
[grammar] ~5-~5: Ensure spelling is correct
Context: ...es, rollback, and interoperability with nupm. ## Build & Test ```bash # Build cargo buil...
(QB_NEW_EN_ORTHOGRAPHY_ERROR_IDS_1)
🔍 Remote MCP DeepWiki, GitHub Copilot
Additional review context
-
leave_current_nutreats allLockfile::loaderrors as “no lockfile” and proceeds. This can allow switching past malformed/unreadable lockfiles instead of failing closed. The repository’s lockfile contains authoritative activation records, and lifecycle operations depend on those records matching the current Nu identity. -
Teardown ordering is implemented as modules first, then plugins; restoration is plugins first, then modules. The test verifies lifecycle callback ordering, but the
use_cmdintegration test’s marker-order assertion compares unregister position againstorder.len(), so it does not actually establish that teardown preceded marker mutation. -
Restore compatibility classification is fail-open when registry lookup fails or the installed version is absent from the registry:
classify_restore_targetonly returnsIncompatiblefor explicit incompatible entries (or missing compatible plugin versions), otherwise it returnsOk. This may restore packages without verified compatibility. -
Activation-profile persistence is now part of snapshot/rollback state. New snapshots record and digest
activation-profile.json; legacy snapshots omit it and leave the live profile unchanged during rollback. Rollback records a dedicatedActivationProfileCommittedlifecycle stage. -
numan trycorrectly scopes compatibility recommendations to the pinnedVersionEntrywhenowner/name@versionis supplied, and explicitly drops the install lock before invoking activation so activation can reacquire the mutation lock.
🔇 Additional comments (16)
src/cmd/use_cmd.rs (1)
663-668: 🎯 Functional Correctness | ⚡ Quick winThe ordering assertion still cannot fail.
version_changed_posisorder.len().unreg_posis an index intoorder, and Line 657 already asserted that"unregister"is present. Any valid index is less than the length, so the comparison at Line 665-668 is always true.The message claims the assertion verifies that the unregistrar runs before the active-version marker changes. The marker write is never pushed into
order, so its position is not observable from this vector. This is fake readiness for the ordering invariant the cross-minor switch depends on.
path_refreshruns immediately after the marker write, so use it as the observable boundary.💚 Proposed fix — assert against an observable boundary
let unreg_pos = order.iter().position(|&x| x == "unregister").unwrap(); - let version_changed_pos = order.len(); - assert!( - unreg_pos < version_changed_pos, - "unregistrar must execute before active-version marker changes" - ); + // `path_refresh` runs immediately after the active-version marker + // write, so it marks the switch boundary observably. + let refresh_pos = order.iter().position(|&x| x == "path_refresh").unwrap(); + let reg_pos = order.iter().position(|&x| x == "register").unwrap(); + assert!( + unreg_pos < refresh_pos, + "unregistrar must execute before the active-version marker changes: {order:?}" + ); + assert!( + refresh_pos < reg_pos, + "restore must run after the marker write and paths refresh: {order:?}" + );src/state/activation_profile.rs (1)
1-1: LGTM!Also applies to: 85-130, 132-169
src/cmd/deactivate.rs (1)
1572-1589: LGTM!src/state/lifecycle_journal.rs (1)
62-63: LGTM!src/state/snapshot.rs (2)
102-105: LGTM!
359-395: LGTM!src/core/resolve.rs (1)
621-644: LGTM!Also applies to: 963-963, 991-991, 1016-1016, 1036-1036
src/cmd/try_cmd.rs (1)
3-3: LGTM!Also applies to: 96-105, 184-184, 285-302, 843-843, 934-1078
AGENTS.md (3)
156-156: Document that restoration can fail after the marker write.
numan useswitches the active marker before restoring activations. The restore phase can fail or remain partial, so this convention should describe an attempted restore and its failure behavior.Source: MCP tools
247-247: Preserve the explicit package version in retry guidance.When the user supplies
owner/name@version, the retry must repeat that full specification. Dropping@versioncan select a different package version afternuman use <version>.Also applies to: 263-263
Source: MCP tools
4-8: LGTM!Also applies to: 38-39, 71-73, 132-132, 172-172, 184-190, 213-227, 245-246, 248-251, 252-262, 264-265
README.md (2)
356-357: Document restore failure after the marker write.Cross-minor
numan usecan switch the marker successfully and then fail during plugin or module restoration. Document successful restoration and the possibility of a non-zero, partial restore result.Source: MCP tools
7-8: LGTM!Also applies to: 26-26, 61-61, 94-108, 170-188, 201-201, 245-245, 308-308, 334-336, 370-370, 457-457, 468-468, 479-479, 489-489
CHANGELOG.md (1)
19-20: LGTM!Also applies to: 24-24
.markdownlint.json (1)
1-5: LGTM!src/nu/bootstrap.rs (1)
1436-1437: 🎯 Functional CorrectnessNo change needed.
validate_nushell_binary()only checks the executable bit before probing, andregister_existing_nu()returns before PATH mutation when TTY is false without--yes;probe_nu_config_path()is not called in this code path.
Doctor still looked for the legacy tools/nushell/nu path, so versioned setup nu installs were reported as missing. Co-authored-by: Cursor <cursoragent@cursor.com>
Style package ids, soft-wrap dim descriptions, and separate entries with blank lines. Co-authored-by: Cursor <cursoragent@cursor.com>
Remember per-Nu-minor activation desire so cross-minor switches leave and restore plugins/modules without shrinking leave profiles. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Fixed 7 file(s) based on 9 unresolved review comments. Co-authored-by: CodeRabbit <noreply@coderabbit.ai>
- `numan try <owner/name[@Version]>` now attempts the specified package for the current Nu and platform. - If incompatible, it lists compatible managed Nu versions, recommends the nearest, and suggests `numan use <version>` followed by `numan try <package>` without auto-switching Nu or installing alternatives. - Removed the curated-starter logic and the obsolete `try --yes` flow. - Reuses existing Resolver, transaction, and activate machinery. - Added resolve helpers for candidate Nu version discovery and recommendation. - Updated README, CHANGELOG, AGENTS, and CLI docs; added 16 unit tests. - Also fixed compile/test blockers in activation_switch and use_cmd. Generated with [Devin](https://devin.ai)
- activation_switch: use partial module deactivation when a runner is provided so teardown-order test can observe the module lane. - use_cmd: seed 0.114 activation profile and refresh fake Nu paths in the path_refresh hook so the integration test exercises plugin restore. - nu/bootstrap: isolate PATH with PathRestoreGuard in register_existing_nu_refuses_non_tty_without_yes_before_path_mutation so concurrent tests cannot race on the process-global PATH.
- Move activation-profile.json from nu_state/ to state/ so snapshots and rollback capture it alongside other state files. Added activation_profile sidecar to Snapshot/SidecarDigests, capture in create_snapshot, restore in rollback_to_snapshot, new ActivationProfileCommitted lifecycle stage. - Extract duplicated nearest-version comparator in resolve.rs into nearest_comparator helper using u64::abs_diff. - Remove unused verified parameter from plugin_with_verified test fixture (renamed to windows_plugin). - Preserve explicit-version context in try_cmd incompatibility reporting: report_incompatible now accepts package_spec and optional target_entry so pinned owner/name@version scopes compatibility to that release and the retry hint includes @Version. - Add compatible_nu_versions_for_entry to resolve.rs for scoped compatibility checks against a single VersionEntry. - Add tests: invalid version format, unavailable version, incompatible pinned version with scoped compat, lock handoff regression. Generated with [Devin](https://devin.ai) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Forward the use PreMutation snapshot into unlocked activate/deactivate module helpers, print leave summary before restore, prove restore order with TrackingRunner, and harden HOME/PATH test isolation around register. Co-authored-by: Cursor <cursoragent@cursor.com>
There was a problem hiding this comment.
Actionable comments posted: 2
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (4)
src/cmd/activation_switch.rs (1)
340-368: 🎯 Functional Correctness | 🟠 Major | ⚡ Quick winPreserve committed module results when journal cleanup fails.
activate_modules_unlockedcan returnErrafter it writes module activation records andautoload-state.json. For example,PendingAutoload::delete(root)?runs after those commits.The
Errbranch marks everymodule_idas failed without reloading actual activation state.numan usecan then report successful module activations as failures and return an incorrect lifecycle failure.Reload and classify module activation state in the
Errbranch, as theOk(true)branch already does. Record the journal-cleanup error as a separate lifecycle error. Add a failure-path test that injects an error after module state commits.As per coding guidelines: “Tests must cover failure modes, not only successful execution.”
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/activation_switch.rs` around lines 340 - 368, Update the Err branch handling the result of activate_modules_unlocked to reload the lockfile and classify each module using collect_currently_active, matching the Ok(true) branch instead of marking every module failed. Preserve successfully committed modules in report.restored_modules, record only inactive modules as failed, and retain the journal-cleanup error as a separate lifecycle error. Add a failure-path test that injects an error after activation records and autoload-state.json are committed.Sources: Coding guidelines, MCP tools
src/cmd/try_cmd.rs (3)
1063-1116: 📐 Maintainability & Code Quality | 🔵 TrivialConfirm that all Rust CI gates are green at the current head.
The observed CI state reports an Ubuntu Test failure while other jobs passed or were still running. Confirm that the failure is resolved before merge.
As per coding guidelines: “All CI gates must pass:
cargo test,cargo clippy -- -D warnings, andcargo fmt --check.”🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/try_cmd.rs` around lines 1063 - 1116, Confirm the current head passes all required Rust CI gates: cargo test, cargo clippy -- -D warnings, and cargo fmt --check. Resolve any failures, including the Ubuntu test failure, before considering the change complete.Sources: Coding guidelines, MCP tools
1110-1116: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winMake the malformed-lockfile test corrupt the file that production reads.
The test writes
{not-jsontolockfile, but the authoritative state file islockfile.json. The assertion also acceptsFailed to load lockfile, so the test can pass because the file is missing and never reach JSON parsing.Write malformed content to the exact path used by
Lockfile::load. Assert a parse-specific error.The repository contract uses
lockfile.jsonas the authoritative lockfile. As per coding guidelines: “Tests must cover failure modes, not only successful execution.”🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/try_cmd.rs` around lines 1110 - 1116, Update the malformed-lockfile test around print_install_only_hint to write invalid JSON to the authoritative lockfile.json path consumed by Lockfile::load. Replace the broad error assertion with one requiring a JSON parse-specific error, ensuring the test exercises malformed-content handling rather than a missing-file path.Sources: Coding guidelines, MCP tools
200-210: 🎯 Functional Correctness | 🟡 Minor | ⚡ Quick winRemove the duplicate error prefix and show the complete package syntax.
This error returns through the standard
anyhow::Resultpath. Do not embederror:in the message. The CLI renderer can add that prefix already.Use
numan try owner/name[@Version]in the usage text. Add assertions for both missing and empty package arguments.As per coding guidelines: update command help when user-visible behavior changes, and use lowercase
numanfor the CLI.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/cmd/try_cmd.rs` around lines 200 - 210, Update execute’s missing-package validation to remove any embedded “error:” prefix, use the complete lowercase syntax “numan try owner/name[`@version`]” in the message, and retain the example. Add assertions covering both absent and empty package arguments, and update the command help to match the revised user-facing syntax.Sources: Coding guidelines, MCP tools
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@src/nu/bootstrap.rs`:
- Around line 559-568: Update the temporary-directory validation around path
registration to run unconditionally, before the consent check and any PATH
mutation. Remove the !options.skip_path guard from
path_is_under_temp_dir(&parent), while retaining skip_path solely to suppress
the durable/user PATH update.
In `@src/util/test_paths.rs`:
- Around line 194-207: Update home_restore_guard_restores_previous_value so the
baseline HOME value is captured only after HomeRestoreGuard::new() acquires
HOME_MUTEX, and reacquire HOME_MUTEX before the post-guard restoration
assertion. Match the locking pattern used by the existing PATH restoration test
while preserving the current HOME mutation and expected-value checks.
---
Outside diff comments:
In `@src/cmd/activation_switch.rs`:
- Around line 340-368: Update the Err branch handling the result of
activate_modules_unlocked to reload the lockfile and classify each module using
collect_currently_active, matching the Ok(true) branch instead of marking every
module failed. Preserve successfully committed modules in
report.restored_modules, record only inactive modules as failed, and retain the
journal-cleanup error as a separate lifecycle error. Add a failure-path test
that injects an error after activation records and autoload-state.json are
committed.
In `@src/cmd/try_cmd.rs`:
- Around line 1063-1116: Confirm the current head passes all required Rust CI
gates: cargo test, cargo clippy -- -D warnings, and cargo fmt --check. Resolve
any failures, including the Ubuntu test failure, before considering the change
complete.
- Around line 1110-1116: Update the malformed-lockfile test around
print_install_only_hint to write invalid JSON to the authoritative lockfile.json
path consumed by Lockfile::load. Replace the broad error assertion with one
requiring a JSON parse-specific error, ensuring the test exercises
malformed-content handling rather than a missing-file path.
- Around line 200-210: Update execute’s missing-package validation to remove any
embedded “error:” prefix, use the complete lowercase syntax “numan try
owner/name[`@version`]” in the message, and retain the example. Add assertions
covering both absent and empty package arguments, and update the command help to
match the revised user-facing syntax.
🪄 Autofix
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: ASSERTIVE
Plan: Pro Plus
Run ID: e1890613-f688-464e-bfae-276cab67a9f3
📒 Files selected for processing (8)
src/cmd/activate.rssrc/cmd/activation_switch.rssrc/cmd/deactivate.rssrc/cmd/try_cmd.rssrc/cmd/use_cmd.rssrc/nu/bootstrap.rssrc/state/activation_profile.rssrc/util/test_paths.rs
🔗 Linked repositories identified
CodeRabbit considers these linked repositories for cross-repo context during reviews:
tonythethompson/QuickShell(manual)tonythethompson/numan(manual)tonythethompson/dependency-chain-substrate(manual)
📜 Review details
⏰ Context from checks skipped due to timeout. (3)
- GitHub Check: Greptile Review
- GitHub Check: Analyze (rust)
- GitHub Check: Analyze (rust)
⚠️ CI failures not shown inline (17)
GitHub Actions: CI / Real-Nu acceptance (windows-latest): feat/try compat
Conclusion: failure
##[group]Run cargo test --verbose -- --ignored --skip acceptance_process_helper --skip stage1_official_registry --skip real_nu_active_update_
�[36;1mcargo test --verbose -- --ignored --skip acceptance_process_helper --skip stage1_official_registry --skip real_nu_active_update_�[0m
shell: C:\Program Files\PowerShell\7\pwsh.EXE -command ". '{0}'"
env:
CARGO_TERM_COLOR: always
CARGO_HOME: C:\Users\runneradmin\.cargo
CARGO_INCREMENTAL: 0
CACHE_ON_FAILURE: false
##[endgroup]
##[error]The operation was canceled.
GitHub Actions: CI / Test (windows-latest): feat/try compat
Conclusion: failure
##[group]Cache Configuration
Cache Provider:
github
Workspaces:
D:\a\numan\numan
Cache Paths:
C:\Users\runneradmin\.cargo\bin
C:\Users\runneradmin\.cargo\.crates.toml
C:\Users\runneradmin\.cargo\.crates2.json
C:\Users\runneradmin\.cargo\registry
C:\Users\runneradmin\.cargo\git
D:\a\numan\numan\target
Restore Key:
v0-rust-test-Windows_NT-x64-8af1e26a
Cache Key:
v0-rust-test-Windows_NT-x64-8af1e26a-2799deb7
.. Prefix:
- v0-rust-test-Windows_NT-x64
.. Environment considered:
- Rust Versions:
- 1.97.1 x86_64-pc-windows-msvc 8bab26f4f68e0e26f0bb7960be334d5b520ea452
- 1.97.1 x86_64-pc-windows-msvc 8bab26f4f68e0e26f0bb7960be334d5b520ea452
- CARGO_HOME
- CARGO_INCREMENTAL
- CARGO_TERM_COLOR
.. Lockfiles considered:
- D:\a\numan\numan\Cargo.lock
- D:\a\numan\numan\Cargo.toml
##[endgroup]
... Restoring cache ...
Cache hit for: v0-rust-test-Windows_NT-x64-8af1e26a-2799deb7
Received 176160768 of 231634064 (76.1%), 167.0 MBs/sec
Received 231634064 of 231634064 (100.0%), 170.6 MBs/sec
Cache Size: ~221 MB (231634064 B)
[command]"C:\Program Files\Git\usr\bin\tar.exe" -xf D:/a/_temp/703466e8-814b-4022-a618-6ac5707e54d0/cache.tzst -P -C D:/a/numan/numan --force-local --use-compress-program "zstd -d"
##[error]The operation was canceled.
GitHub Actions: CI / Roadmap drift: feat/try compat
Conclusion: failure
##[group]Run curl --fail-with-body --silent --show-error --location \
�[36;1mcurl --fail-with-body --silent --show-error --location \�[0m
�[36;1m https://raw.githubusercontent.com/tonythethompson/numan/$CONTRACT_SHA/scripts/check-roadmap-drift.py \�[0m
�[36;1m -o /tmp/check-roadmap-drift.py�[0m
�[36;1mdiff -u scripts/check-roadmap-drift.py /tmp/check-roadmap-drift.py \�[0m
�[36;1m || (echo "::error::Local drift script diverges from $CONTRACT_SHA."; \�[0m
GitHub Actions: CI / Roadmap drift: feat/try compat
Conclusion: failure
##[group]Run curl --fail-with-body --silent --show-error --location \
�[36;1mcurl --fail-with-body --silent --show-error --location \�[0m
�[36;1m https://raw.githubusercontent.com/tonythethompson/numan/$CONTRACT_SHA/docs/contracts/roadmap-v1.md \�[0m
�[36;1m -o /tmp/roadmap-v1.md�[0m
�[36;1mdiff -u docs/contracts/roadmap-v1.md /tmp/roadmap-v1.md \�[0m
�[36;1m || (echo "::error::Local contract doc diverges from $CONTRACT_SHA."; \�[0m
GitHub Actions: CI / 5_Real-Nu acceptance (windows-latest).txt: feat/try compat
Conclusion: failure
##[group]Run cargo test --verbose -- --ignored --skip acceptance_process_helper --skip stage1_official_registry --skip real_nu_active_update_
�[36;1mcargo test --verbose -- --ignored --skip acceptance_process_helper --skip stage1_official_registry --skip real_nu_active_update_�[0m
shell: C:\Program Files\PowerShell\7\pwsh.EXE -command ". '{0}'"
env:
CARGO_TERM_COLOR: always
CARGO_HOME: C:\Users\runneradmin\.cargo
CARGO_INCREMENTAL: 0
CACHE_ON_FAILURE: false
##[endgroup]
##[error]The operation was canceled.
GitHub Actions: CI / 7_Test (windows-latest).txt: feat/try compat
Conclusion: failure
##[group]Cache Configuration
Cache Provider:
github
Workspaces:
D:\a\numan\numan
Cache Paths:
C:\Users\runneradmin\.cargo\bin
C:\Users\runneradmin\.cargo\.crates.toml
C:\Users\runneradmin\.cargo\.crates2.json
C:\Users\runneradmin\.cargo\registry
C:\Users\runneradmin\.cargo\git
D:\a\numan\numan\target
Restore Key:
v0-rust-test-Windows_NT-x64-8af1e26a
Cache Key:
v0-rust-test-Windows_NT-x64-8af1e26a-2799deb7
.. Prefix:
- v0-rust-test-Windows_NT-x64
.. Environment considered:
- Rust Versions:
- 1.97.1 x86_64-pc-windows-msvc 8bab26f4f68e0e26f0bb7960be334d5b520ea452
- 1.97.1 x86_64-pc-windows-msvc 8bab26f4f68e0e26f0bb7960be334d5b520ea452
- CARGO_HOME
- CARGO_INCREMENTAL
- CARGO_TERM_COLOR
.. Lockfiles considered:
- D:\a\numan\numan\Cargo.lock
- D:\a\numan\numan\Cargo.toml
##[endgroup]
... Restoring cache ...
Cache hit for: v0-rust-test-Windows_NT-x64-8af1e26a-2799deb7
Received 176160768 of 231634064 (76.1%), 167.0 MBs/sec
Received 231634064 of 231634064 (100.0%), 170.6 MBs/sec
Cache Size: ~221 MB (231634064 B)
[command]"C:\Program Files\Git\usr\bin\tar.exe" -xf D:/a/_temp/703466e8-814b-4022-a618-6ac5707e54d0/cache.tzst -P -C D:/a/numan/numan --force-local --use-compress-program "zstd -d"
##[error]The operation was canceled.
GitHub Actions: CI / 8_Roadmap drift.txt: feat/try compat
Conclusion: failure
##[group]Run ref_file=$(mktemp)
�[36;1mref_file=$(mktemp)�[0m
�[36;1mstatus=$(curl --silent --show-error --output "$ref_file" --write-out "%{http_code}" \�[0m
�[36;1m -H "Authorization: ***" \�[0m
�[36;1m -H "Accept: application/vnd.github+json" \�[0m
�[36;1m -H "User-Agent: numan-ci" \�[0m
�[36;1m "https://api.github.com/repos/tonythethompson/numan/git/refs/tags/${CONTRACT_TAG}")�[0m
�[36;1mif [ "$status" = "404" ]; then�[0m
�[36;1m echo "::warning::Tag $CONTRACT_TAG does not exist yet; skipping SHA resolution check."�[0m
�[36;1m exit 0�[0m
�[36;1mfi�[0m
�[36;1mif [ "$status" != "200" ]; then�[0m
�[36;1m echo "::error::GitHub API returned HTTP $status for tag $CONTRACT_TAG."�[0m
GitHub Actions: CI / Roadmap drift: feat/try compat
Conclusion: failure
##[group]Run curl --fail-with-body --silent --show-error --location \
�[36;1mcurl --fail-with-body --silent --show-error --location \�[0m
�[36;1m https://raw.githubusercontent.com/tonythethompson/numan/$CONTRACT_SHA/docs/plans/consolidated-multi-repo-roadmap.md \�[0m
�[36;1m -o /tmp/consolidated-roadmap.md�[0m
�[36;1mdiff -u docs/plans/consolidated-multi-repo-roadmap.md /tmp/consolidated-roadmap.md \�[0m
�[36;1m || (echo "::error::Local consolidated roadmap diverges from $CONTRACT_SHA."; \�[0m
GitHub Actions: CI / Roadmap drift: feat/try compat
Conclusion: failure
##[group]Run ref_file=$(mktemp)
�[36;1mref_file=$(mktemp)�[0m
�[36;1mstatus=$(curl --silent --show-error --output "$ref_file" --write-out "%{http_code}" \�[0m
�[36;1m -H "Authorization: ***" \�[0m
�[36;1m -H "Accept: application/vnd.github+json" \�[0m
�[36;1m -H "User-Agent: numan-ci" \�[0m
�[36;1m "https://api.github.com/repos/tonythethompson/numan/git/refs/tags/${CONTRACT_TAG}")�[0m
�[36;1mif [ "$status" = "404" ]; then�[0m
�[36;1m echo "::warning::Tag $CONTRACT_TAG does not exist yet; skipping SHA resolution check."�[0m
�[36;1m exit 0�[0m
�[36;1mfi�[0m
�[36;1mif [ "$status" != "200" ]; then�[0m
�[36;1m echo "::error::GitHub API returned HTTP $status for tag $CONTRACT_TAG."�[0m
GitHub Actions: CI / Real-Nu acceptance (macos-latest): feat/try compat
Conclusion: failure
emver=/Users/runner/work/numan/numan/target/debug/deps/libsemver-8e618bac00eed7e7.rlib --extern serde=/Users/runner/work/numan/numan/target/debug/deps/libserde-b685fc42c5c8fcbf.rlib --extern serde_json=/Users/runner/work/numan/numan/target/debug/deps/libserde_json-f87702b5f448b39c.rlib --extern sha2=/Users/runner/work/numan/numan/target/debug/deps/libsha2-2b174ae6d50727fb.rlib --extern tar=/Users/runner/work/numan/numan/target/debug/deps/libtar-48850b7ab1cdb196.rlib --extern tempfile=/Users/runner/work/numan/numan/target/debug/deps/libtempfile-a5e5a50385bf5ed2.rlib --extern thiserror=/Users/runner/work/numan/numan/target/debug/deps/libthiserror-90085ed5d82d1fba.rlib --extern toml=/Users/runner/work/numan/numan/target/debug/deps/libtoml-b3944a01ff08379e.rlib --extern uuid=/Users/runner/work/numan/numan/target/debug/deps/libuuid-57fb4db30512c65c.rlib --extern wait_timeout=/Users/runner/work/numan/numan/target/debug/deps/libwait_timeout-e98bd672c49c60a0.rlib --extern xz2=/Users/runner/work/numan/numan/target/debug/deps/libxz2-8e0ead227871d7a8.rlib --extern zip=/Users/runner/work/numan/numan/target/debug/deps/libzip-9778315c125c94c2.rlib -L native=/Users/runner/work/numan/numan/target/debug/build/libgit2-sys-b5e1493a1ae1720e/out/build -L native=/Users/runner/work/numan/numan/target/debug/build/zstd-sys-5b52c2a9924f88fa/out -L native=/opt/homebrew/Cellar/xz/5.8.3/lib -L native=/usr/lib`
�[1m�[92m Running�[0m `/Users/runner/.rustup/toolchains/stable-aarch64-apple-darwin/bin/rustc --crate-name plugin_lifecycle_real_nu --edition=2021 tests/plugin_lifecycle_real_nu.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 -C split-debuginfo=unpacked --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=43756f88bb3caab8 -C extra-filename=-c8ff62241e125e61 --out-dir /Users/runner/work/numan/numan/target/debug/deps -L dependency=/Users/runner/work/numan/numan/...
GitHub Actions: CI / 0_Real-Nu acceptance (macos-latest).txt: feat/try compat
Conclusion: failure
emver=/Users/runner/work/numan/numan/target/debug/deps/libsemver-8e618bac00eed7e7.rlib --extern serde=/Users/runner/work/numan/numan/target/debug/deps/libserde-b685fc42c5c8fcbf.rlib --extern serde_json=/Users/runner/work/numan/numan/target/debug/deps/libserde_json-f87702b5f448b39c.rlib --extern sha2=/Users/runner/work/numan/numan/target/debug/deps/libsha2-2b174ae6d50727fb.rlib --extern tar=/Users/runner/work/numan/numan/target/debug/deps/libtar-48850b7ab1cdb196.rlib --extern tempfile=/Users/runner/work/numan/numan/target/debug/deps/libtempfile-a5e5a50385bf5ed2.rlib --extern thiserror=/Users/runner/work/numan/numan/target/debug/deps/libthiserror-90085ed5d82d1fba.rlib --extern toml=/Users/runner/work/numan/numan/target/debug/deps/libtoml-b3944a01ff08379e.rlib --extern uuid=/Users/runner/work/numan/numan/target/debug/deps/libuuid-57fb4db30512c65c.rlib --extern wait_timeout=/Users/runner/work/numan/numan/target/debug/deps/libwait_timeout-e98bd672c49c60a0.rlib --extern xz2=/Users/runner/work/numan/numan/target/debug/deps/libxz2-8e0ead227871d7a8.rlib --extern zip=/Users/runner/work/numan/numan/target/debug/deps/libzip-9778315c125c94c2.rlib -L native=/Users/runner/work/numan/numan/target/debug/build/libgit2-sys-b5e1493a1ae1720e/out/build -L native=/Users/runner/work/numan/numan/target/debug/build/zstd-sys-5b52c2a9924f88fa/out -L native=/opt/homebrew/Cellar/xz/5.8.3/lib -L native=/usr/lib`
�[1m�[92m Running�[0m `/Users/runner/.rustup/toolchains/stable-aarch64-apple-darwin/bin/rustc --crate-name plugin_lifecycle_real_nu --edition=2021 tests/plugin_lifecycle_real_nu.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 -C split-debuginfo=unpacked --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=43756f88bb3caab8 -C extra-filename=-c8ff62241e125e61 --out-dir /Users/runner/work/numan/numan/target/debug/deps -L dependency=/Users/runner/work/numan/numan/...
GitHub Actions: CI / Real-Nu acceptance (ubuntu-latest): feat/try compat
Conclusion: failure
nner/work/numan/numan/target/debug/deps/libthiserror-586dfbc5650686d0.rlib --extern toml=/home/runner/work/numan/numan/target/debug/deps/libtoml-3a01d174782f9120.rlib --extern uuid=/home/runner/work/numan/numan/target/debug/deps/libuuid-1068b4ee51a57a2f.rlib --extern wait_timeout=/home/runner/work/numan/numan/target/debug/deps/libwait_timeout-fa4fc4237a4e496f.rlib --extern xz2=/home/runner/work/numan/numan/target/debug/deps/libxz2-24b073cd3f366c2c.rlib --extern zip=/home/runner/work/numan/numan/target/debug/deps/libzip-7d9f25bc64201460.rlib -L native=/home/runner/work/numan/numan/target/debug/build/libgit2-sys-8489ee2672c70f74/out/build -L native=/home/runner/work/numan/numan/target/debug/build/lzma-sys-49b2a8bf09b7cbbd/out -L native=/home/runner/work/numan/numan/target/debug/build/bzip2-sys-387644da4e3ca1ec/out/lib -L native=/home/runner/work/numan/numan/target/debug/build/zstd-sys-84e663e874f8a180/out`
�[1m�[92m Running�[0m `/home/runner/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/bin/rustc --crate-name plugin_active_update_real_nu --edition=2021 tests/plugin_active_update_real_nu.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=17687d9e7a3811a8 -C extra-filename=-0883b0d2f61c1de5 --out-dir /home/runner/work/numan/numan/target/debug/deps -L dependency=/home/runner/work/numan/numan/target/debug/deps --extern anyhow=/home/runner/work/numan/numan/target/debug/deps/libanyhow-a5ec7606a567b1e8.rlib --extern base64=/home/runner/work/numan/numan/target/debug/deps/libbase64-ef7c71042959ea5a.rlib --extern clap=/home/runner/work/numan/numan/target/debug/deps/libclap-ac5abf46c593b52c.rlib --extern clap_complete=/home/runner/work/numan/numan/target/debug/deps/libclap_complete-73c9ebc1ee885a35.rlib --extern clap_complete_nushell=/home/runner/work/numan/numan/target/debug/deps/libclap_complet...
GitHub Actions: CI / 2_Real-Nu acceptance (ubuntu-latest).txt: feat/try compat
Conclusion: failure
nner/work/numan/numan/target/debug/deps/libthiserror-586dfbc5650686d0.rlib --extern toml=/home/runner/work/numan/numan/target/debug/deps/libtoml-3a01d174782f9120.rlib --extern uuid=/home/runner/work/numan/numan/target/debug/deps/libuuid-1068b4ee51a57a2f.rlib --extern wait_timeout=/home/runner/work/numan/numan/target/debug/deps/libwait_timeout-fa4fc4237a4e496f.rlib --extern xz2=/home/runner/work/numan/numan/target/debug/deps/libxz2-24b073cd3f366c2c.rlib --extern zip=/home/runner/work/numan/numan/target/debug/deps/libzip-7d9f25bc64201460.rlib -L native=/home/runner/work/numan/numan/target/debug/build/libgit2-sys-8489ee2672c70f74/out/build -L native=/home/runner/work/numan/numan/target/debug/build/lzma-sys-49b2a8bf09b7cbbd/out -L native=/home/runner/work/numan/numan/target/debug/build/bzip2-sys-387644da4e3ca1ec/out/lib -L native=/home/runner/work/numan/numan/target/debug/build/zstd-sys-84e663e874f8a180/out`
�[1m�[92m Running�[0m `/home/runner/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/bin/rustc --crate-name plugin_active_update_real_nu --edition=2021 tests/plugin_active_update_real_nu.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=17687d9e7a3811a8 -C extra-filename=-0883b0d2f61c1de5 --out-dir /home/runner/work/numan/numan/target/debug/deps -L dependency=/home/runner/work/numan/numan/target/debug/deps --extern anyhow=/home/runner/work/numan/numan/target/debug/deps/libanyhow-a5ec7606a567b1e8.rlib --extern base64=/home/runner/work/numan/numan/target/debug/deps/libbase64-ef7c71042959ea5a.rlib --extern clap=/home/runner/work/numan/numan/target/debug/deps/libclap-ac5abf46c593b52c.rlib --extern clap_complete=/home/runner/work/numan/numan/target/debug/deps/libclap_complete-73c9ebc1ee885a35.rlib --extern clap_complete_nushell=/home/runner/work/numan/numan/target/debug/deps/libclap_complet...
GitHub Actions: CI / 4_Test (ubuntu-latest).txt: feat/try compat
Conclusion: failure
rlib --extern thiserror=/home/runner/work/numan/numan/target/debug/deps/libthiserror-586dfbc5650686d0.rlib --extern toml=/home/runner/work/numan/numan/target/debug/deps/libtoml-3a01d174782f9120.rlib --extern uuid=/home/runner/work/numan/numan/target/debug/deps/libuuid-1068b4ee51a57a2f.rlib --extern wait_timeout=/home/runner/work/numan/numan/target/debug/deps/libwait_timeout-fa4fc4237a4e496f.rlib --extern xz2=/home/runner/work/numan/numan/target/debug/deps/libxz2-24b073cd3f366c2c.rlib --extern zip=/home/runner/work/numan/numan/target/debug/deps/libzip-7d9f25bc64201460.rlib -L native=/home/runner/work/numan/numan/target/debug/build/libgit2-sys-8489ee2672c70f74/out/build -L native=/home/runner/work/numan/numan/target/debug/build/lzma-sys-49b2a8bf09b7cbbd/out -L native=/home/runner/work/numan/numan/target/debug/build/bzip2-sys-387644da4e3ca1ec/out/lib -L native=/home/runner/work/numan/numan/target/debug/build/zstd-sys-84e663e874f8a180/out`
�[1m�[92m Running�[0m `/home/runner/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/bin/rustc --crate-name plugin_deactivate_real_nu --edition=2021 tests/plugin_deactivate_real_nu.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=5f18f17a8368e0b5 -C extra-filename=-f27447e092ad8230 --out-dir /home/runner/work/numan/numan/target/debug/deps -L dependency=/home/runner/work/numan/numan/target/debug/deps --extern anyhow=/home/runner/work/numan/numan/target/debug/deps/libanyhow-a5ec7606a567b1e8.rlib --extern base64=/home/runner/work/numan/numan/target/debug/deps/libbase64-ef7c71042959ea5a.rlib --extern clap=/home/runner/work/numan/numan/target/debug/deps/libclap-ac5abf46c593b52c.rlib --extern clap_complete=/home/runner/work/numan/numan/target/debug/deps/libclap_complete-73c9ebc1ee885a35.rlib --extern clap_complete_nushell=/home/runner/work/numan/numan/target/...
GitHub Actions: CI / Test (ubuntu-latest): feat/try compat
Conclusion: failure
rlib --extern thiserror=/home/runner/work/numan/numan/target/debug/deps/libthiserror-586dfbc5650686d0.rlib --extern toml=/home/runner/work/numan/numan/target/debug/deps/libtoml-3a01d174782f9120.rlib --extern uuid=/home/runner/work/numan/numan/target/debug/deps/libuuid-1068b4ee51a57a2f.rlib --extern wait_timeout=/home/runner/work/numan/numan/target/debug/deps/libwait_timeout-fa4fc4237a4e496f.rlib --extern xz2=/home/runner/work/numan/numan/target/debug/deps/libxz2-24b073cd3f366c2c.rlib --extern zip=/home/runner/work/numan/numan/target/debug/deps/libzip-7d9f25bc64201460.rlib -L native=/home/runner/work/numan/numan/target/debug/build/libgit2-sys-8489ee2672c70f74/out/build -L native=/home/runner/work/numan/numan/target/debug/build/lzma-sys-49b2a8bf09b7cbbd/out -L native=/home/runner/work/numan/numan/target/debug/build/bzip2-sys-387644da4e3ca1ec/out/lib -L native=/home/runner/work/numan/numan/target/debug/build/zstd-sys-84e663e874f8a180/out`
�[1m�[92m Running�[0m `/home/runner/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/bin/rustc --crate-name plugin_deactivate_real_nu --edition=2021 tests/plugin_deactivate_real_nu.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=5f18f17a8368e0b5 -C extra-filename=-f27447e092ad8230 --out-dir /home/runner/work/numan/numan/target/debug/deps -L dependency=/home/runner/work/numan/numan/target/debug/deps --extern anyhow=/home/runner/work/numan/numan/target/debug/deps/libanyhow-a5ec7606a567b1e8.rlib --extern base64=/home/runner/work/numan/numan/target/debug/deps/libbase64-ef7c71042959ea5a.rlib --extern clap=/home/runner/work/numan/numan/target/debug/deps/libclap-ac5abf46c593b52c.rlib --extern clap_complete=/home/runner/work/numan/numan/target/debug/deps/libclap_complete-73c9ebc1ee885a35.rlib --extern clap_complete_nushell=/home/runner/work/numan/numan/target/...
GitHub Actions: CI / 10_Test (macos-latest).txt: feat/try compat
Conclusion: failure
n/numan/target/debug/deps/libxz2-8e0ead227871d7a8.rlib --extern zip=/Users/runner/work/numan/numan/target/debug/deps/libzip-9778315c125c94c2.rlib -L native=/Users/runner/work/numan/numan/target/debug/build/libgit2-sys-b5e1493a1ae1720e/out/build -L native=/Users/runner/work/numan/numan/target/debug/build/zstd-sys-5b52c2a9924f88fa/out -L native=/opt/homebrew/Cellar/xz/5.8.3/lib -L native=/usr/lib`
�[1m�[92m Running�[0m `/Users/runner/.rustup/toolchains/stable-aarch64-apple-darwin/bin/rustc --crate-name module_autoload_test --edition=2021 tests/module_autoload_test.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 -C split-debuginfo=unpacked --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=a4ad8e058f525c6a -C extra-filename=-81871ac1746f3cc1 --out-dir /Users/runner/work/numan/numan/target/debug/deps -L dependency=/Users/runner/work/numan/numan/target/debug/deps --extern anyhow=/Users/runner/work/numan/numan/target/debug/deps/libanyhow-a7c2b97cced6d62c.rlib --extern base64=/Users/runner/work/numan/numan/target/debug/deps/libbase64-35589c29a679fc09.rlib --extern clap=/Users/runner/work/numan/numan/target/debug/deps/libclap-91a01e217dc65eca.rlib --extern clap_complete=/Users/runner/work/numan/numan/target/debug/deps/libclap_complete-4bbb4721d590a39c.rlib --extern clap_complete_nushell=/Users/runner/work/numan/numan/target/debug/deps/libclap_complete_nushell-4c7ee6de86e61196.rlib --extern console=/Users/runner/work/numan/numan/target/debug/deps/libconsole-8e2f795b55b9c710.rlib --extern dirs=/Users/runner/work/numan/numan/target/debug/deps/libdirs-98b40dc56d3c28d7.rlib --extern ed25519_dalek=/Users/runner/work/numan/numan/target/debug/deps/libed25519_dalek-89cc653197a6cb66.rlib --extern fd_lock=/Users/runner/work/numan/numan/target/debug/deps/libfd_lock-019c225f1e06d1e9.rlib --extern flate2=/Users/runner/work/numan/numan/target/debug/deps/libf...
GitHub Actions: CI / Test (macos-latest): feat/try compat
Conclusion: failure
n/numan/target/debug/deps/libxz2-8e0ead227871d7a8.rlib --extern zip=/Users/runner/work/numan/numan/target/debug/deps/libzip-9778315c125c94c2.rlib -L native=/Users/runner/work/numan/numan/target/debug/build/libgit2-sys-b5e1493a1ae1720e/out/build -L native=/Users/runner/work/numan/numan/target/debug/build/zstd-sys-5b52c2a9924f88fa/out -L native=/opt/homebrew/Cellar/xz/5.8.3/lib -L native=/usr/lib`
�[1m�[92m Running�[0m `/Users/runner/.rustup/toolchains/stable-aarch64-apple-darwin/bin/rustc --crate-name module_autoload_test --edition=2021 tests/module_autoload_test.rs --error-format=json --json=diagnostic-rendered-ansi,artifacts,future-incompat --emit=dep-info,link -C embed-bitcode=no -C debuginfo=2 -C split-debuginfo=unpacked --test --check-cfg 'cfg(docsrs,test)' --check-cfg 'cfg(feature, values())' -C metadata=a4ad8e058f525c6a -C extra-filename=-81871ac1746f3cc1 --out-dir /Users/runner/work/numan/numan/target/debug/deps -L dependency=/Users/runner/work/numan/numan/target/debug/deps --extern anyhow=/Users/runner/work/numan/numan/target/debug/deps/libanyhow-a7c2b97cced6d62c.rlib --extern base64=/Users/runner/work/numan/numan/target/debug/deps/libbase64-35589c29a679fc09.rlib --extern clap=/Users/runner/work/numan/numan/target/debug/deps/libclap-91a01e217dc65eca.rlib --extern clap_complete=/Users/runner/work/numan/numan/target/debug/deps/libclap_complete-4bbb4721d590a39c.rlib --extern clap_complete_nushell=/Users/runner/work/numan/numan/target/debug/deps/libclap_complete_nushell-4c7ee6de86e61196.rlib --extern console=/Users/runner/work/numan/numan/target/debug/deps/libconsole-8e2f795b55b9c710.rlib --extern dirs=/Users/runner/work/numan/numan/target/debug/deps/libdirs-98b40dc56d3c28d7.rlib --extern ed25519_dalek=/Users/runner/work/numan/numan/target/debug/deps/libed25519_dalek-89cc653197a6cb66.rlib --extern fd_lock=/Users/runner/work/numan/numan/target/debug/deps/libfd_lock-019c225f1e06d1e9.rlib --extern flate2=/Users/runner/work/numan/numan/target/debug/deps/libf...
🧰 Additional context used
📓 Path-based instructions (13)
**/*
📄 CodeRabbit inference engine (CLAUDE.md)
**/*: Use Serena's semantic, symbol-aware tools as the primary tools for reading and editing code; use built-in Read, Glob, Grep, and Edit only under the stated exceptions.
Before editing a code file, inspect its symbol overview, read the specific symbols being changed, and edit them with Serena's symbol-aware tools.
Understand the existing code before changing it and make the smallest change that satisfies the request; avoid unrelated cleanup, premature abstractions, unnecessary error handling, feature flags, and compatibility shims.
Prefer editing existing files over creating new ones, and never create Markdown or README files unless explicitly requested.
For exploratory questions, provide a 2–3 sentence recommendation with the main tradeoff and do not implement until the user agrees.
For UI or frontend changes that cannot be tested in a browser, explicitly state that browser testing was not performed rather than claiming success.
Address security issues when discovered, including injection, XSS, SQL injection, path traversal, and secret leaks.
Pause and obtain confirmation before destructive, hard-to-reverse, externally visible, or third-party-upload actions, including deleting files or branches, dropping tables, force-pushing, modifying CI/CD, posting externally, or uploading content.
When blocked, investigate the root cause instead of bypassing it with --no-verify, --force, or deletion; investigate unfamiliar files, branches, and configuration before deleting them.
Only commit when explicitly asked; do not proactively update git configuration or push changes.
Do not skip Git hooks unless explicitly asked; if a pre-commit hook fails, fix the issue, re-stage, and create a new commit rather than amending.
Stage files by name rather than using git add -A or git add ., to avoid accidentally including secrets or large binaries.
Use a HEREDOC for commit messages, and add co-author attribution only when the user explicitly requests the exact trailer.
Do not force-pus...
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
**/*.{js,jsx,ts,tsx,py,java,go,rs,rb,php,c,cpp,h,hpp,cs,swift,kt,kts}
📄 CodeRabbit inference engine (CLAUDE.md)
Add comments only when the WHY is non-obvious; do not narrate what the code does, reference the current task, or reference the PR in comments.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
!**/.env,!**/credentials.json,!**/*.pem
📄 CodeRabbit inference engine (CLAUDE.md)
Do not commit files that appear to contain secrets, including .env, credentials.json, and PEM files; warn before doing so even if explicitly requested.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
**/*.rs
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.rs: Use the Rust 2021 edition.
Useanyhow::Resultwith.context(...)in application code; usethiserrorfor library error types that callers match on.
Useclapderive macros for CLI definitions.
Useserdewithserde_jsonortomlfor serialization.
Function parameters must use&Path, not&PathBuf.
Library code must not panic; error paths should returnanyhow::Resultwith context where appropriate.
Add or update tests for behavior changes, including relevant failure paths.
New mutating code paths must acquire the mutation lock viaacquire_mutation_lock(root)and snapshot the lockfile before writes.
Lockfile, journal, and state-file JSON writes must usewrite_json_atomic.
numan installmust write only to$NUMAN_ROOT; it must not invoke Nu or register plugins/autoloads.
Onlyactivateanddeactivatemay modify Nu integration state.
Treat the lockfile as the authoritative source of truth; derived projections such as autoload state must not be authoritative.
Install payloads under versioned, content-addressed paths and never overwrite them in place.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass plugin paths through environment variables only; do not use runtime interpolation in Nu program strings.
**/*.rs: All CI gates must pass:cargo test,cargo clippy -- -D warnings, andcargo fmt --check.
Every mutating command—includinginstall,remove,update,gc, and futurenupm import—must callacquire_mutation_lock(root).
Lockfiles, journals, and state files must usewrite_json_atomic; partial writes are not allowed.
Pending activation, autoload, and lifecycle journals must be stored under$NUMAN_ROOT/state/.
Module autoload identity must match all four fields: Nu executable hash, Nu version, vendor autoload directory, and managed file path; the lockfilemodule_activationvalue is authoritative.
Never overwrite foreign autoload files; respectOWNERSHIP_MARKER.
Pass paths to Nu only throu...
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
**/*.{rs,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Run and keep
cargo fmt/rustfmtclean, and ensurecargo clippy -- -D warningspasses.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
**/*.{rs,nu}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
**/*.{rs,nu}: Real-Nu acceptance tests must be marked#[ignore]and should be run when changes affect activation or nupm import; unit tests must not spawn realnuand should use injectable seams such asFakeCandidateRunneror registrars.
The nupm integration must be read-only towardNUPM_HOME, must not executebuild.nu, and must not perform bidirectional synchronization.Unit tests must use
FakeCandidateRunneror injectable registrars and must not spawn a realnuprocess.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
**/*.{rs,md}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Match existing naming, module layout, and documentation level in the file being edited; update
AGENTS.md,docs/, or command help when structure, conventions, or user-visible behavior changes.Tests must cover failure modes, not only successful execution.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
**/*.{rs,md,toml}
📄 CodeRabbit inference engine (CONTRIBUTING.md)
Use the repository's established serialization and module conventions rather than introducing unrelated refactors.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
src/**/*.rs
📄 CodeRabbit inference engine (AGENTS.md)
src/**/*.rs: Serialize mutations per root withacquire_mutation_lock(root); the second acquisition must fail immediately rather than block.
Pass Nu paths and names throughNUMAN_PLUGIN_BINARY,NUMAN_PLUGIN_CONFIG, andNUMAN_PLUGIN_NAME; the Nu program string must be a compile-time constant with no runtime interpolation.
Files:
src/util/test_paths.rssrc/nu/bootstrap.rssrc/cmd/use_cmd.rssrc/cmd/deactivate.rssrc/cmd/activation_switch.rssrc/cmd/activate.rssrc/cmd/try_cmd.rssrc/state/activation_profile.rs
src/cmd/{activate,deactivate,plugin_lifecycle,update}.rs
📄 CodeRabbit inference engine (AGENTS.md)
Nu plugin register/unregister callbacks belong exclusively to the activate/deactivate lifecycle boundary; opted-in update orchestration may coordinate that boundary but must not invoke callbacks directly.
Files:
src/cmd/deactivate.rssrc/cmd/activate.rs
src/cmd/deactivate.rs
📄 CodeRabbit inference engine (AGENTS.md)
Use
execute_with_unregistrarfor plugin tests andexecute_with_candidate_runner_and_unregistrarwhen both module and plugin lanes require fakes.
Files:
src/cmd/deactivate.rs
src/cmd/activate.rs
📄 CodeRabbit inference engine (AGENTS.md)
Use
execute_with_registrarfor plugin tests andexecute_with_candidate_runnerfor module tests; inject fakes instead of spawning a real Nu binary in unit tests.
Files:
src/cmd/activate.rs
src/state/**/*.rs
📄 CodeRabbit inference engine (AGENTS.md)
src/state/**/*.rs: Write JSON state files, including lockfiles, journals, andnu_state/paths.json, withwrite_json_atomicusing a same-directory temporary file and persist operation.
Create a snapshot before mutating install, update, remove, activate, deactivate, nupm-import, orinit --refreshstate; snapshots must preserve referenced payloads as GC live roots.
Files:
src/state/activation_profile.rs
🧠 Learnings (1)
📓 Common learnings
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-08T10:55:43.714Z
Learning: Source builds require explicit user consent before cloning or building, with separate consent scopes.
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-08T10:55:43.714Z
Learning: Create a feature branch from `master`, implement with tests, run the full test suite, update AGENTS.md when conventions change, and open a described PR.
Learnt from: CR
Repo: tonythethompson/numan
Timestamp: 2026-08-08T10:55:43.714Z
Learning: Commits must use imperative mood and be under 72 characters; use `feature/description` or `fix/description` branches; never force-push `master`; squash feature merges.
🔍 Remote MCP DeepWiki, GitHub Copilot
Relevant review context
- The repository treats
lockfile.jsonas authoritative activation state; autoload state is derived, and activation records are tied to Nu executable identity/version. Cross-minor switching must preserve lifecycle ordering and journal/snapshot consistency. - At the current PR head (
d9d37d7), three fresh review concerns remain:capture_activation_profile_sidecarreads the profile twice—once forcontent/digest and again for parsedvalue—so concurrent modification can produce an internally inconsistent snapshot.restore_desiredcan classify every module as failed whenactivate_modules_unlockedreturnsOk(true), even if some modules were committed successfully; this may incorrectly failnuman useand misreport restored state.- The missing-package error in
numan tryincludes anerror:prefix despite being returned throughanyhow::Result, potentially producing duplicated error text; its usage text also omits the documentedowner/name[@Version]shape.
- Earlier major findings were addressed: the install-to-activation lock is now intentionally held through
activate::execute_under_lock; pinned compatibility recommendations preserve the requested version; snapshot absence is represented explicitly; and module restore ordering is now tested with a tracking runner. - CI is not fully green at the observed state: Ubuntu
Testfailed, while Format, MSRV, CodeFactor, Package, Deny, and several analysis checks passed; other jobs were still running.
🔇 Additional comments (14)
src/state/activation_profile.rs (1)
376-387: LGTM!src/cmd/deactivate.rs (1)
474-525: LGTM!Also applies to: 1625-1641
src/cmd/activation_switch.rs (1)
124-153: LGTM!Also applies to: 158-229, 233-270, 458-539, 632-655, 832-859, 918-919, 1003-1003, 1072-1094, 1152-1152, 1291-1314
src/cmd/use_cmd.rs (2)
81-100: 🩺 Stability & AvailabilityResolve the failing Ubuntu Test job before merge.
Review context reports that the Ubuntu Test job failed at PR head
d9d37d7. Identify the failing test and rerun the complete Rust gate set on the final head before merge.As per coding guidelines: “All CI gates must pass:
cargo test,cargo clippy -- -D warnings, andcargo fmt --check.”Sources: Coding guidelines, MCP tools
127-152: LGTM!src/cmd/activate.rs (1)
73-74: LGTM!Also applies to: 734-767
src/cmd/try_cmd.rs (2)
316-316: LGTM!Also applies to: 353-365
328-328: 🗄️ Data Integrity & IntegrationNo change needed for unpinned installs.
transaction::install_packagetreatsNoneasResolver::resolve, which returns the same latest compatible entry used bytry_package, so passingNonedoes not install a different release.> Likely an incorrect or invalid review comment.src/nu/bootstrap.rs (4)
526-526: LGTM!
1195-1197: LGTM!
1595-1601: LGTM!
1626-1630: LGTM!src/util/test_paths.rs (2)
6-7: LGTM!
93-133: 🩺 Stability & AvailabilityNo lock-order issue found in current test call sites.
No combined
HomeRestoreGuard/PathRestoreGuardcall site acquires Home before Path.
The early temp-directory refusal in register_existing_nu fired regardless of NUMAN_TEST_NO_PERSIST_USER_PATH, so ignored acceptance tests and the #[cfg(unix)] unit tests in cmd::setup that stage binaries under tempfile roots with PathRestoreGuard hit the guard and bailed. persist_path_dir already short-circuits under that flag (no durable write, no refusal), so the half-update risk the guard prevents does not apply in test mode. Gate the early refusal on the flag being absent to align with persist_path_dir. Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Per CodeRabbit security review, drop the `!options.skip_path` gate so a session-only registration can never put a temp dir on the process PATH. In production skip_path is always false here, so this is defense-in-depth for future callers; the test harness bypass via NUMAN_TEST_NO_PERSIST_USER_PATH is preserved. Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
|
Applied in 28c04e3: removed the !options.skip_path\ gate so the temp-dir refusal is unconditional. The \NUMAN_TEST_NO_PERSIST_USER_PATH\ bypass (mirroring \persist_path_dir_*) is preserved so acceptance tests can stage binaries in TempDir under PathRestoreGuard. Verified locally: |
…rrors restore_desired's Err branch marked every requested module failed even when activate_modules_unlocked committed activation records and autoload-state.json before bailing (e.g. a journal-cleanup failure in run_module_lane step 12). Reload the lockfile and classify each module via collect_currently_active, matching the Ok(failed) branch: committed modules go to restored_modules, only inactive ones to failed, and the journal-cleanup error is retained as a separate lifecycle_errors warning rather than a per-module failure. Also address try_cmd review feedback: drop the angle-bracket placeholder and embedded "error:" prefix from the missing-package message, use the lowercase "numan try owner/name[@Version]" syntax, add an empty-package test, and tighten the malformed-lockfile test to assert a serde_json parse error against the authoritative lockfile path. Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
c74d1fe
P1 (use_cmd.rs): refresh cached paths.json after the dangling off-tree binary_path repair in execute_latest. Without this, the same-target reconcile loads the stale cache and validate_drift fails after the marker has already been rewritten. Added regression test test_use_latest_self_heal_clears_stale_paths_cache. P2 (remove.rs): move activation-profile cleanup to a preflight before lockfile/payload removal so a profile-write failure aborts the remove while state is still intact and retryable. Removed the duplicate dead warning-only call. P1 (activation_profile snapshots/rollback): already addressed by #108; snapshot capture and rollback restore are in place. No code change. Also applies cargo fmt to doctor.rs (Format CI was failing).
…#106) * Report versioned managed Nu installs in doctor. Doctor still looked for the legacy tools/nushell/nu path, so versioned setup nu installs were reported as missing. Co-authored-by: Cursor <cursoragent@cursor.com> * Improve registry packages listing readability. Style package ids, soft-wrap dim descriptions, and separate entries with blank lines. Co-authored-by: Cursor <cursoragent@cursor.com> * Auto-deactivate and restore packages on numan use. Remember per-Nu-minor activation desire so cross-minor switches leave and restore plugins/modules without shrinking leave profiles. Co-authored-by: Cursor <cursoragent@cursor.com> * Update src/cmd/activation_switch.rs Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * Update src/cmd/activation_switch.rs Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * Update src/state/activation_profile.rs Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * Update src/cmd/remove.rs Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> * fix: apply CodeRabbit auto-fixes Fixed 7 file(s) based on 9 unresolved review comments. Co-authored-by: CodeRabbit <noreply@coderabbit.ai> * fix: address review feedback and rebase onto master - Rebased onto latest master to resolve merge conflicts - Cleaned up activation_switch.rs per review (simplified logic) - Fixed use_cmd.rs per review comments - Cleaned up activation_profile.rs per review Generated with [Devin](https://devin.ai) Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com> * fix: address remaining valid PR 106 review findings Filter activate profile sync to packages that are active, retain profile desire when deactivate fails, tighten use switch ordering assertions, and document ensure_absent_for_paths error behavior. Co-authored-by: Cursor <cursoragent@cursor.com> * fix: map managed Nu doctor errors by VersionManagerError Reserve the numan use hint for dangling active markers; filesystem and marker parse failures get a permissions/reinstall message instead. Co-authored-by: Cursor <cursoragent@cursor.com> * fix: address Codex review findings on PR #106 P1 (use_cmd.rs): refresh cached paths.json after the dangling off-tree binary_path repair in execute_latest. Without this, the same-target reconcile loads the stale cache and validate_drift fails after the marker has already been rewritten. Added regression test test_use_latest_self_heal_clears_stale_paths_cache. P2 (remove.rs): move activation-profile cleanup to a preflight before lockfile/payload removal so a profile-write failure aborts the remove while state is still intact and retryable. Removed the duplicate dead warning-only call. P1 (activation_profile snapshots/rollback): already addressed by #108; snapshot capture and rollback restore are in place. No code change. Also applies cargo fmt to doctor.rs (Format CI was failing). --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> Co-authored-by: CodeRabbit <noreply@coderabbit.ai> Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Uh oh!
There was an error while loading. Please reload this page.