Add setting `parallelListingRequests` to an S3 publish endpoint. If set
to a value greater than 1, aptly will use up to that number of parallel
HTTP requests to list the S3 bucket.

Fix the test S3 server: in case on multiple CommonPrefixes being
returned, S3 specs require multiple CommonPrefix XML tags (and not a
single CommonPrefix with multiple Prefix tags).

When a publishing uses a publish prefix, instead of listing the contents
of the whole bucket under the storage prefix, only list the contents of
the bucket under the storage prefix and publish prefix, and cache it by
publish prefix.
This speeds up publish operations under a prefix.