Concatenate JWKS path with existing path on domain if present

[ltj]

In the current jwks url provider implementation any path on domain will be replaced by the "well known" jwks path. This will likely work for most providers, but some (e.g. AWS Cognito) uses a unique resource path on a shared endpoint to identify the issuer ie. https://cognito-idp.ap-southeast-2.amazonaws.com/ap-southeast-2_example

I'm aware this might have been a deliberate decision given that there is a unit test to verify this behavior. However, I think allowing for a more flexible approach will benefit this library since it's being actively used in other projects such as ktor

Changes

• Directly concatenate the domain string with the JWKS path and use URI.normalize() to ensure correct behavior regardless of trailing slash is present or not.
• Added/changed unit tests to reflect new behavior

References

Please include relevant links supporting this change such as a:

• support ticket
• community post
• StackOverflow post
• support forum thread

Testing

• This change adds test coverage
• This change has been tested on the latest version of Java or why not

Checklist

• I have read the Auth0 general contribution guidelines
• I have read the Auth0 Code of Conduct
• All existing and new tests complete without errors

[lbalmaceda]

We welcome the change! However I think it's better to documented in the javadocs that if a path is included domain string passed, it will be appended the well-known one. Mainly because the existing unit test that you've changed proved the opposite. What do you think?

Also, we rather have explicit imports for readability. Do you mind changing that back?

-import java.net.MalformedURLException;
-import java.net.URL;
-import java.net.URLConnection;
+import java.net.*;

[ltj]

We welcome the change! However I think it's better to documented in the javadocs that if a path is included domain string passed, it will be appended the well-known one. Mainly because the existing unit test that you've changed proved the opposite. What do you think?

Also, we rather have explicit imports for readability. Do you mind changing that back?

Thanks @lbalmaceda. I concur so just added a commit with explicit imports and updated javadoc to reflect actual behavior

[lbalmaceda]

Thanks!