This is a follow-on to #3553: that places all the backup files in a single directory using the original dataset names. Unfortunately, we have no guarantee that dataset names are unique, and so this could lead to conflicts and overwritten files. This PR modifies the previous to place each backup in a subdirectory named using the dataset resource ID. Since each resource ID is unique to its dataset, this ensures that there are no conflicts. This also has the side-effect that we no longer need to store the *.md5 files, since the hash value is now embedded in the file path.
This PR consists of two commits. The second is described above; the first is a few small changes to appease the linter.
This is a follow-on to #3553: that places all the backup files in a single directory using the original dataset names. Unfortunately, we have no guarantee that dataset names are unique, and so this could lead to conflicts and overwritten files. This PR modifies the previous to place each backup in a subdirectory named using the dataset resource ID. Since each resource ID is unique to its dataset, this ensures that there are no conflicts. This also has the side-effect that we no longer need to store the
*.md5files, since the hash value is now embedded in the file path.This PR consists of two commits. The second is described above; the first is a few small changes to appease the linter.
PBENCH-1259