TODO

- More docs
- More unit tests
- More xample programs
- optimize for large systems (>1000 block devices)
  - action sorting
  - clearpart?
  - audit device setup/teardown policies to reduce unnecessary churn
- aynschronous actions
- API for managing fstab and crypttab
- improve interaction with the environment
  - disallow partitioning changes to disks with in-use partitions?
  - improve detection of active devices and associated error
    handling
- improve clarity of exceptions raised
- btrfs
  - raid10 needs four disks
  - use btrfs for resize instead of btrfsctl
  - detection of btrfs raid level, compression
- md name collisions in existing system
- add StorageConfigurationError or similar so autopart failures due to the
  sanity check can be handled by offering the option of keeping the layout to
  attempt to fix the problem instead of always resetting