G30B1N/paq

Hash file or directory recursively.

paq

Hash file or directory recursively.

Powered by blake3 cryptographic hashing algorithm.

Performance

The go programming language repository was used as a test case (478 MB / 12,540 files).

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
paq ./go	116.4 ± 2.6	111.4	120.9	1.00
shell b3sum	132.4 ± 1.5	129.6	135.9	1.14 ± 0.03
dirhash -a sha256 ./go	642.5 ± 5.8	634.7	649.8	5.52 ± 0.13
shell sha256sum	1583.0 ± 16.3	1568.6	1606.8	13.60 ± 0.33

Performance benchmark uses hyperfine.

Commands with shell use the following command with various <hashsum> implementations:

find ./go -type f -print0 | LC_ALL=C sort -z | xargs -0 <hashsum> | <hashsum>

Installation

Cargo Install

Installation requires cargo.

Install From Crates.io

cargo install paq

Install From Repository Clone

1. Clone this repository.
2. Run cargo install --path . from repository root.

Pre-Built Binary Package

1. Find Latest Release .zip archive for computer Operating System and Architecture.
2. Download and extract .zip.
3. Modify permissions of the extracted paq binary to allow execution.
4. Move paq to system path.

Usage

Command Line Interface executable or Crate library.

Included in this repository is an example directory containing some sample files, a subdirectory and a symlink to test paq functionality.

Executable

Run paq [src] to hash source file or directory.

For help, run paq --help.

Hash Example Directory

paq ./example

Path to example directory can be relative or absolute.

Expect different results if -i or --ignore-hidden flag argument is used.

Crate Library

Add paq to project dependencies in Cargo.toml.

Use Library

use paq;

let source = std::path::PathBuf::from("/path/to/source");
let ignore_hidden = true; // .dir or .file
let source_hash: paq::ArrayString<64> = paq::hash_source(&source, ignore_hidden);

println!("{}", source_hash);

Hash Example Directory

use paq;

let source = std::path::PathBuf::from("example");
let ignore_hidden = true;
let source_hash: paq::ArrayString<64> = paq::hash_source(&source, ignore_hidden);

assert_eq!(&source_hash[..], "a593d18de8b696c153df9079c662346fafbb555cc4b2bbf5c7e6747e23a24d74");

Expect different results if ignore_hidden is set to false.

Content Limitations

Hashes are generated using file system content as input data to the blake3 hashing algorithm.

By design, paq does NOT include file system metadata in hash input such as:

• File modes
• File ownership
• File modification and access times
• File ACLs and extended attributes
• Hard links
• Symlink target contents (target path is hashed)

Additionally, files or directory contents starting with dot or full stop can optionally be ignored.

How it Works

1. Recursively get path(s) for a given source argument.
2. Hash each path and file content if path is for a file.
3. Sort the list of hashes to maintain consistent ordering.
4. Compute the final hash by hashing the sorted list of hashes.

License

MIT