compare the way we store symlinks in our link metadata to the reference implementation

[shibumi]

Please fill in the fields below to submit an issue or feature request. The
more information that is provided, the better.

Description of issue or feature request:
We still need to decide if link metadata stores the path of the source or the target of a symlink.
The python reference implementation is might be useful for this.

Current behavior:

unknown

Expected behavior:

see python reference implementation

[shibumi]

See PR #37 and #55

[adityasaky]

This came up in #191. There was an artifact path collision when it should have recorded the link paths (both would still have the same hash). The python implementation does this as well and I think it makes the most sense.

[adityasaky]

@shibumi do you think we can consider this done with #194 merged?

[shibumi]

Good question. I would suggest running in-toto-golang and the python implementation and testing if they really create the same output.

We really need some sort of automated tests that regularly run the Go and Python implementations and make sure that they produce consistent and equal output.

[adityasaky]

This should be a test case for in-toto/in-toto#563