object list might be truncated

[martinklepsch]

s3/list-objects returns a potentially truncated list. we need to retrieve remaining if it's truncated.

Depending on the size of the bucket/number of to-be-uploaded-objects retrieving object summaries individually might be more efficient. E.g. 10k objects in bucket but only requesting to upload two files.

Getting object summaries on a per key basis would also allow for metadata to be diffed, something I initially decided to skip in #4