feat!: hardlink deduplications

[KSXGitHub]

TODO:

• Fix the algorithm.
• Report the number of unrecorded links: stats.nlink() - links.len().
• More efficient data structure to insert detected hardlinks (Vec randomly pauses).
• Resolve the TODO comments.
• Add detected hardlinks to the JSON output.
• Print detected hardlinks from JSON input.
• Test.
• Multi arguments.
• Fix: duplicated arguments.
• Update the README.
• Add examples to --help.
• Add competing benchmarks.
• ci(benchmark): update dust to include --no-progress #297
• Refactor: Rename Hook to RecordHardlinks, DoNothing to Ignorant.

[github-actions]

Performance Regression Reports

commit: 4d4ceea

There are no regressions.

[Copilot]

Pull Request Overview

This PR implements hardlink deduplication functionality to accurately calculate disk usage when hardlinks are present. When enabled via --deduplicate-hardlinks, the tool detects files with multiple links and adjusts the reported sizes to avoid double-counting the same data.

Key changes include:

• Detection and recording of hardlinks during filesystem traversal
• Deduplication logic that reduces sizes in the data tree for hardlinked files
• JSON output enhancement to include hardlink information with summary and detailed reporting

Reviewed Changes

Copilot reviewed 48 out of 49 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/usual_cli.rs	Updated test constructors to use new FsTreeBuilder signature with hardlink recorder
tests/json.rs	Modified JSON tests to handle new JsonTree structure with hardlink information
tests/hardlinks_*.rs	Added comprehensive test suites for hardlink functionality with and without deduplication
tests/cli_errors.rs	Updated error test to use new FsTreeBuilder signature
tests/_utils.rs	Added utilities for creating test workspaces with hardlinks and symlinks
src/runtime_error.rs	Added error types for unsupported features on non-POSIX platforms
src/reporter/	Enhanced progress reporting to include hardlink detection statistics
src/lib.rs	Added new modules for hardlink and inode handling
src/json_data.rs	Extended JSON data structure to include hardlink information
src/inode.rs	New module for inode number handling
src/hardlink/	Complete hardlink detection, recording, and deduplication system
src/fs_tree_builder.rs	Updated to support hardlink recording during tree construction
src/data_tree/	Added hardlink deduplication logic to data tree
src/args.rs	Added CLI flags for hardlink deduplication and JSON output control
src/app/	Updated application logic to handle hardlink processing and reporting

Comments suppressed due to low confidence (1)

src/hardlink/aware.rs:105

• Silently ignoring errors from hardlink recording could mask important issues and make debugging difficult. Consider logging the error or handling it more explicitly.

            .map(|values| (*values.size(), values.paths().clone()))