Say an upload is interrupted for some reason (crash etc). On the next run, when Duplicacy skips over chunks that are already present, will it see the partially uploaded chunk and incorrectly skip over it too? Or is there something in place (eg. chunks being renamed after successful upload) to prevent this from happening?
For storages that behave more like a traditional file system (local disk, sftp, dropbox, etc), chunks are uploaded with temporary names and then renamed after successful transfer. For all other storages, we rely on the servers to do the right thing (not saving the incomplete file when the connection is abruptly terminated).
Are temporary names used for Google Drive?
No, renaming would incur an extra API call so is generally avoided whenever possible.