Reduce noise in gix-blame tests

[cruessler]

This is a PR that reduces noise in gix-blame tests by implementing From<(Range<u32>, ObjectId)> for UnblamedHunk. This implementation allows us to replace UnblamedHunk { range_in_blamed_file: 3..5, suspects: [(suspect, 3..5)].into(), source_file_name: None, } by (3..5, suspect).into() and new_unblamed_hunk(0..5, suspect, Offset::Added(0)) by (0..5, suspect).into(). I think this change significantly increases the signal-to-noise ratio.

In addition to this, we could also implement From<(Range<u32>, (ObjectId, Range<u32>))> for UnblamedHunk or From<(Range<u32>, ObjectId, Offset)> for UnblamedHunk in order to convert places where the range associated with a suspect is not identical to range_in_blamed_file, but the benefit would be smaller. What do you think?

[Byron]

This is awesome!

Sure, it's a bit less clear what these values mean at first glance, but tests can be expert mode where a thing or two have to be learned in order to write or understand them.

In addition to this, we could also implement From<(Range<u32>, (ObjectId, Range<u32>))> for UnblamedHunk or From<(Range<u32>, ObjectId, Offset)> for UnblamedHunk in order to convert places where the range associated with a suspect is not identical to range_in_blamed_file, but the benefit would be smaller. What do you think?

Go for it :).