repository debug dump refactor: support versioned metadata and dumping artifacts

[jku]

--repository-dump-dir currently dumps current metadata that repository is serving.

If it is seen useful, we could also output the targets we are serving:

• could create the directory structure that the web server is using
• the artifact filename should be hash prefixed I suppose

[jku]

I'm not 100% sure this is worth the amount of work but the use case I'm thinking of is manual debugging

• Run test, dump repo
• the dump content tries to accurately match the http server contents
• user can run python -m http.server or similar and the manually run their client against that server to see what happens

(this sequence could even be automated into a debug tool but that's step 2)

[jku]

the dump content tries to accurately match the http server contents

this does not happen currently for the metadata dump: current dump tries to be as readable and diffable as possible so e.g. version numbers are not used in filenames

[jku]

This might be something we want as it makes it possible to run other clients against a repository version (say, example client from python-tuf) for debugging purposes.

What would need to happen for metadata:

• change the dump method so that every other metadata file is prefixed with version except timestamp
• there should be an additional metadata directory, so e.g. if dump dir is /tmp/tuf-conformance-dump then test_basic_init_and_refresh metadata goes in /tmp/tuf-conformance-dump/test_basic_init_and_refresh/refresh-1/metadata/

and then artifacts

• artifacts for the same test go in /tmp/tuf-conformance-dump/test_basic_init_and_refresh/refresh-1/targets/
• artifact targetpath subdirectories should be created and hash prefix should be used. so "a/artifact" would go in
  /tmp/tuf-conformance-dump/test_basic_init_and_refresh/refresh-1/targets/a/<HASH>.artifact