Sequential, distributed, POSIX-style job queue processing

 Filespooler is a Unix-style tool that facilitates local or remote command
 execution, complete with stdin capture, with easy integration with various
 tools. Here's a brief Filespooler feature list:
 .
  - It can easily use tools such as S3, Dropbox, Syncthing, NNCP, ssh, UUCP, USB
   drives, CDs, etc. as transport.
 .
   - Translation: you can use basically anything that is a filesystem as a
     transport
 .
  - It can use arbitrary decoder command pipelines (eg, zcat, stdcat, gpg, age,
   etc) to pre-process stored packets.
 .
  - It can send and receive packets by pipes.
 .
  - Its storage format is simple on-disk files with locking.
 .
  - It supports one-to-one and one-to-many configurations.
 .
  - Locking is unnecessary when writing new jobs to the queue, and many arbitrary
   tools (eg, Syncthing, Dropbox, etc) can safely write directly to the queue
   without any assistance.
 .
  - Queue processing is strictly ordered based on the order on the creation
   machine, even if job files are delivered out of order to the destination.
 .
  - stdin can be piped into the job creation tool, and piped to a later executor
   at process time on a remote machine.
 .
  - The file format is lightweight; less than 100 bytes overhead unless large
   extra parameters are given.
 .
  - The queue format is lightweight; having 1000 different queues on a Raspberry
   Pi would be easy.
 .
  - Processing is stream-based throughout; arbitrarily-large packets are fine and
   sizes in the TB range are no problem.
 .
  - The Filespooler command, fspl, is extremely lightweight, consuming less than
   10MB of RAM on x86_64.
 .
  - Filespooler has extensive documentation.
 .
 Filespooler consists of a command-line tool (fspl) for interacting with queues.
 It also consists of a Rust library that is used by fspl. main.rs for fspl is
 just a few lines long.