Language-agnostic test suite for SWHID implementations

[andrew]

Currently when implementing SWHID in a new language, you need to either:

• Extract test cases from the Python swh-model implementation
• Manually verify against known hashes
• Hope your implementation matches the spec correctly

A language-agnostic test format (JSON, YAML, or similar) with input data and expected SWHID hashes for all object types would make it much
easier to:

• Verify new implementations produce correct hashes
• Catch edge cases during development
• Ensure consistency across implementations

This came up while working on the Ruby implementation (https://github.com/andrew/swhid) where I had to manually extract test cases from the
Python library to validate correctness.

Example test cases could include:

{
  "type": "content",
  "data": "Hello, World!",
  "expected_hash": "96898574d1b88e619be24fd90bb4cd399acbc5ca"
}

{
  "type": "directory",
  "entries": [
    {
      "name": "README.md",
      "type": "file",
      "perms": "100644",
      "target": "37ec8ea2110c0b7a32fbb0e872f6e7debbf95e21"
    }
  ],
  "expected_hash": "d7ed3d2c31d608823be58b1cbe57605310615231"
}

{
  "type": "revision",
  "directory": "85a74718d377195e1efd0843ba4f3260bad4fe07",
  "parents": ["01e2d0627a9a6edb24c37db45db5ecb31e9de808"],
  "author": "Linus Torvalds <torvalds@linux-foundation.org>",
  "author_timestamp": 1436739030,
  "author_timezone": "-0700",
  "committer": "Linus Torvalds <torvalds@linux-foundation.org>",
  "committer_timestamp": 1436739030,
  "committer_timezone": "-0700",
  "message": "Linux 4.2-rc2\n",
  "expected_hash": "bc0195aad0daa2ad5b0d76cce22b167bc3435590"
}

Would be happy to help contribute test cases if this sounds useful.