dynamodb-replicator uses / as delimiter to hierarchically structure
the data in S3. Up to now it did a naive join of all path parts when
constructing paths and key names for S3. That produced different paths
in S3 whether or not a trailing slash was provided when specifying the
target. For example the following two commands produced different output
files in S3:
As that's different from other tools which interact with S3, like the
official awscli this commit changes that behavior to work like
path.join does, but by using an external package to ensure that also
under Windows paths are generated with forward slashes.
dynamodb-replicatoruses/as delimiter to hierarchically structure the data in S3. Up to now it did a naive join of all path parts when constructing paths and key names for S3. That produced different paths in S3 whether or not a trailing slash was provided when specifying the target. For example the following two commands produced different output files in S3:As that's different from other tools which interact with S3, like the official
awsclithis commit changes that behavior to work likepath.joindoes, but by using an external package to ensure that also under Windows paths are generated with forward slashes.