fix(hardlink)!: take `dev` into account by KSXGitHub · Pull Request #383 · KSXGitHub/parallel-disk-usage

KSXGitHub · 2026-04-02T08:37:12Z

Closes #347.

This change adds device number tracking to the hardlink detection system to correctly handle cases where multiple filesystems contain files with identical inode numbers. Since hardlinks cannot span filesystem boundaries, files on different devices with the same inode number are unrelated and should be tracked separately.

https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

…ross-filesystem false deduplication Co-authored-by: KSXGitHub <11488886+KSXGitHub@users.noreply.github.com>

…on key Co-authored-by: KSXGitHub <11488886+KSXGitHub@users.noreply.github.com>

…TryFrom Co-authored-by: KSXGitHub <11488886+KSXGitHub@users.noreply.github.com>

github-actions · 2026-04-02T08:44:30Z

Performance Regression Reports

commit: 831ac74

There are no regressions.

The Copilot PR kept dev out of the reflection/JSON layer, introducing a dev=0 placeholder hack, double-allocation sorting, and "unsupported edge case" disclaimers. Since the reflection is meant to mirror the internal HardlinkList, just include dev in ReflectionEntry. This simplifies From<HardlinkList> back to a single-pass map+collect, removes the dev=0 workaround in TryFrom<Reflection>, and makes multi-filesystem JSON round-tripping correct. Bump SCHEMA_VERSION to 2026-04-02 for the new field. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

…e dev - Add DeviceNumber in src/device_number.rs, mirroring InodeNumber - Reorder InodeKey fields to ino-first for faster derived PartialEq (ino has far more entropy than dev, so comparisons short-circuit earlier) - Use DeviceNumber instead of raw u64 in InodeKey, ReflectionEntry, and all public/internal APIs - Add dev to ConversionError::DuplicatedInode for precise error messages - Add ConversionError::duplicated_inode(InodeKey) constructor - Update sort keys to (ino, dev) order throughout https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

…error types - Move DeviceNumber from device_number.rs into device.rs alongside DeviceBoundary, mirroring the InodeNumber-in-inode.rs pattern - Add dev: DeviceNumber to SizeConflictError and NumberOfLinksConflictError so error messages identify the exact (inode, device) pair https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

…Inode https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

- Trim InodeKey doc to a single line (remove over-explanation) - Restore doc style for SizeConflictError and NumberOfLinksConflictError with `ino`, `dev`, and `nlink` links to MetadataExt methods - Rephrase Reflection guarantee as "pair of an inode number and a device number" - Remove unnecessary #[error(not(source))] on DuplicatedInode tuple variant (derive_more only auto-assumes source for single-field variants) https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

…rors std::os::unix::fs::MetadataExt does not exist on Windows, so rustdoc path links to it cause compilation errors. Use raw doc.rust-lang.org URLs instead, matching the style used before this PR. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

The review comment was on line 16 (sort order), not line 15 (uniqueness). Rephrase to "sorted by pairs of an inode number and a device number". https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

Each HTML comment explaining the Windows workaround now sits directly above its corresponding link definition, rather than a single combined comment covering all links. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

Anchor IDs shouldn't have weird syntax. Minimize diff. Why the FUCK was the AI so stupid?

https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

Copilot

Pull request overview

This PR updates the hardlink deduplication system to key hardlinks by (dev, ino) (device number + inode number), preventing false deduplication when separate filesystems happen to reuse the same inode numbers. It also propagates dev through the hardlink reflection/JSON surface and adjusts tests and schema version accordingly.

Changes:

Introduce DeviceNumber and incorporate it into hardlink tracking (HardlinkList now keys by (dev, ino)).
Extend hardlink reflection entries to include dev, with deterministic sorting by (ino, dev).
Update CLI/integration tests and bump JSON schema version to reflect the new field.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
`src/device.rs`	Adds `DeviceNumber` newtype and unix helper to read `MetadataExt::dev()`.
`src/hardlink/aware.rs`	Records device number alongside inode when detecting hardlinks.
`src/hardlink/hardlink_list.rs`	Changes internal key from `InodeNumber` to `(InodeNumber, DeviceNumber)` and updates conflict errors.
`src/hardlink/hardlink_list/iter.rs`	Exposes `dev()` accessor on hardlink list iteration items.
`src/hardlink/hardlink_list/reflection.rs`	Adds `dev` to reflection entries, sorts by `(ino, dev)`, and updates conversion error shape.
`src/hardlink/hardlink_list/test.rs`	Updates unit tests for new `add()` signature and adds coverage for same-ino/different-dev behavior.
`src/json_data/schema_version.rs`	Bumps schema version for the changed JSON/reflection format.
`tests/_utils.rs`	Adds `read_device_number()` helper for integration tests.
`tests/hardlinks_deduplication.rs`	Updates expected hardlink reflection details to include `dev`.
`tests/hardlinks_deduplication_multi_args.rs`	Updates expected hardlink reflection details to include `dev`.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

"its" → "their" (two subjects: inode number and device number) https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

The size and paths belong to the file identified by the (ino, dev) pair, not to the numbers themselves. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

Make it clear that the size and paths belong to the file, not the numbers. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated no new comments.

Copilot AI and others added 5 commits March 10, 2026 22:38

Initial plan

e255766

fix(hardlink): use (device, inode) key for deduplication to prevent c…

ed6162e

…ross-filesystem false deduplication Co-authored-by: KSXGitHub <11488886+KSXGitHub@users.noreply.github.com>

test(hardlink): add test proving device number is used in deduplicati…

38aacab

…on key Co-authored-by: KSXGitHub <11488886+KSXGitHub@users.noreply.github.com>

fix(reflection): address doc and comment accuracy for Reflection and …

eb92bdc

…TryFrom Co-authored-by: KSXGitHub <11488886+KSXGitHub@users.noreply.github.com>

chore(git): merge from copilot/check-nlink-hardlink-deduplication

65194d6

KSXGitHub force-pushed the claude/merge-pr-347-L12kJ branch from b2373c8 to 65194d6 Compare April 2, 2026 08:40

claude and others added 2 commits April 2, 2026 09:13

refactor: remove unnecessary pub(crate)

b678164

This comment was marked as resolved.

Sign in to view

claude and others added 2 commits April 2, 2026 09:46

docs: remove unnecessary documentation

0058648

This comment was marked as resolved.

Sign in to view

claude added 2 commits April 2, 2026 10:26

refactor(hardlink): use tuple variant for ConversionError::Duplicated…

9a16e84

…Inode https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

This comment was marked as resolved.

Sign in to view

claude and others added 5 commits April 2, 2026 11:12

fix(docs): rephrase sort order line in Reflection doc

8407682

The review comment was on line 16 (sort order), not line 15 (uniqueness). Rephrase to "sorted by pairs of an inode number and a device number". https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

docs(hardlink): separate HTML comments per doc link

ceecb11

Each HTML comment explaining the Windows workaround now sits directly above its corresponding link definition, rather than a single combined comment covering all links. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

docs: revert to prior style

15d84e9

Anchor IDs shouldn't have weird syntax. Minimize diff. Why the FUCK was the AI so stupid?

KSXGitHub force-pushed the claude/merge-pr-347-L12kJ branch from dd35dc9 to 15d84e9 Compare April 2, 2026 12:56

This comment was marked as resolved.

Sign in to view

claude and others added 3 commits April 2, 2026 13:27

docs(hardlink): rephrase sort doc in From<Vec<ReflectionEntry>>

39cbd96

https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

refactor(hardlink): extract sorting_key method on ReflectionEntry

9bb7a7e

https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

docs: explain the existence of duplicated_inode

79f73a7

KSXGitHub requested a review from Copilot April 2, 2026 14:31

Copilot started reviewing on behalf of KSXGitHub April 2, 2026 14:32 View session

Copilot AI reviewed Apr 2, 2026

View reviewed changes

Comment thread src/hardlink/hardlink_list/reflection.rs

KSXGitHub requested a review from Copilot April 2, 2026 14:38

Copilot started reviewing on behalf of KSXGitHub April 2, 2026 14:39 View session

Copilot AI reviewed Apr 2, 2026

View reviewed changes

Comment thread src/hardlink/aware.rs

Comment thread src/hardlink/hardlink_list/reflection.rs Outdated

docs: fix grammar

205e976

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

KSXGitHub requested a review from Copilot April 2, 2026 14:52

docs(hardlink): update Aware::record doc to mention device number

ae449d3

https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

Copilot started reviewing on behalf of KSXGitHub April 2, 2026 14:53 View session

KSXGitHub commented Apr 2, 2026

View reviewed changes

Comment thread src/hardlink/aware.rs Outdated

claude added 2 commits April 2, 2026 14:54

docs(hardlink): fix pronoun in Aware::record doc

4aa04bb

"its" → "their" (two subjects: inode number and device number) https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

docs(hardlink): fix Aware::record doc wording

326b2a9

The size and paths belong to the file identified by the (ino, dev) pair, not to the numbers themselves. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

KSXGitHub commented Apr 2, 2026

View reviewed changes

Comment thread src/hardlink/aware.rs Outdated

docs(hardlink): clarify Aware::record doc

2127726

Make it clear that the size and paths belong to the file, not the numbers. https://claude.ai/code/session_01QP9wZyoZcGmJsEsA66ZRok

This comment was marked as outdated.

Sign in to view

KSXGitHub requested a review from Copilot April 2, 2026 15:02

Copilot started reviewing on behalf of KSXGitHub April 2, 2026 15:02 View session

Copilot AI reviewed Apr 2, 2026

View reviewed changes

KSXGitHub marked this pull request as ready for review April 2, 2026 22:28

KSXGitHub merged commit f94914d into master Apr 2, 2026
17 checks passed

KSXGitHub deleted the claude/merge-pr-347-L12kJ branch April 2, 2026 22:29

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix(hardlink)!: take `dev` into account#383

fix(hardlink)!: take `dev` into account#383
KSXGitHub merged 24 commits intomasterfrom
claude/merge-pr-347-L12kJ

KSXGitHub commented Apr 2, 2026 •

edited

Loading

Uh oh!

github-actions bot commented Apr 2, 2026 •

edited

Loading

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

This comment was marked as outdated.

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

KSXGitHub commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Apr 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Performance Regression Reports

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

This comment was marked as outdated.

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

KSXGitHub commented Apr 2, 2026 •

edited

Loading

github-actions bot commented Apr 2, 2026 •

edited

Loading