This is my version of brkirch's afsctool

AFSC (Apple File System Compression) tool is an utility that can be used
to apply HFS+ compression to file(s), decompress HFS+ compressed file(s), or
get information about existing HFS+ compressed file(s).
Mac OS 10.6 or later is required.
(https://brkirch.wordpress.com/afsctool/)

I have made several modifications, mostly concerning the compression feature:
- improved error reporting
- an attempt to reduce memory pressure compressing large files
- support for multiple files/folders specified on the commandline
- a backup option while compressing (that comes in addition to the existing undo if
  something went wrong)
- support for files that are read-only (and/or write-only) by changing their permissions
  temporarily. No error checking is done for this feature; failure will lead to
  errors that are already caught.


The main new feature that justifies the version bump, however, is the parallel
processing feature, allowing the user to specify an arbitray (though positive :))
number of threads that will compress the specified files in parallel.
This feature has two modes that each perform best in specific conditions:
- perform only the actual compression in parallel. Disk IO (reading the file into
  memory, writing it back to disk) is done with exclusive access to prevent disk
  trashing. This mode is selected with the -j option.
- perform all file compression steps in parallel. This mode is suitable for solid-
  state disks, file collections that reside on different disks or collections with
  files of different sizes (including notably very large files). This mode is
  selected with the -J option.
The performance difference is never enormous in my testing, but YMMV.

Interestingly, the optimum performance (on large collections) is not necessarily
obtained with as many worker threads as there are CPU cores. Collections with
a significant number of very large files and many more smaller files can apparently
be processed faster when using a rather larger number of workers, with a gain
approaching the number of cores.