Algorithm
Given any digest algorithm H (a Hash function algorithm),
a RecursiveDigest(H, path) is:
- for a file:
H("F" || file_content) - for a symlink:
H("S" || symlink_content) - for a directory:
H("D" || directory_content)
As you can see a one-letter ASCII prefix is used to make it impossible
to create a file that has the same digest as a directory,
etc. The drawback of this approach is that RecursiveDigest(H, path) of
a simple file is not the same as just a normal digest of it (H(file_content)) .
file_content is just the byte content of a file.
symlink_content is just the path the symlink is pointing to, as bytes.
directory_content is created by:
- sorting all entries of a directory by name, in ascending order, using a simple byte-sequence comparison
- for all entries concatenating pairs of:
H(entry_name)RecursiveDigest(H, entry_path)