chore: use priority ordering in java repo finder

[benmss]

This PR makes the configuration property repo_pom_paths a priority list. When the Repo Finder is resolving URLs to repositories, the first valid URL (based on the ordering of the repo_pom_paths) will be returned. By default, the highest priority SCM entry in a POM should be url, as this most reliably represents the URL of the source repository.

Closes #640

[tromai]

I have tested your implementation and it seems that because calling defaults.get_list method (implementation here) without duplicated_ok=True will first convert the list specified in defaults.ini to a set before returning. A set in Python doesn't preserve the inserting order of elements. This will result in the order of elements not following the original order as specified in defaults.ini and impact the behavior of JavaRepoFinder._read_pom.
@behnazh-w It sounds to me that we need to improve defaults.get_list regarding the order of the elements in the list. defaults.get_list altering the order of the list as the default behavior is not what we expected. However, improving that function might be out of scope for this PR. For now I think @benmss can address it by calling

tags = defaults.get_list("repofinder.java", "repo_pom_paths", duplicated_ok=True)

At this place.
I don't have a strong opinion on whether we fix it in this PR or in a separated PR.

[tromai]

Thanks for the changes. LGTM (given the assumption that the issue discussed in #641 (comment) is going to be addressed in a separated PR).

[behnazh-w]

@behnazh-w It sounds to me that we need to improve defaults.get_list regarding the order of the elements in the list. defaults.get_list altering the order of the list as the default behavior is not what we expected.

Yes, agreed. @tromai can you please work on this issue in a separate PR?