v1.0.2 - Performance Improvements

• Significant performance improvements for multithreaded usage (~30-35%) due to
  the counter now using AtomicU32 again rather than Arc<Mutex<32>>, enabled
  thanks to the stabilization of fetch_update
  (3a5cc2e)
• Generating CUIDs and slugs is now ~15-20% faster, thanks to reductions in the
  number of allocated strings and improvements to how space is pre-allocated
  during the generation of radix strings (a09bfad)
• Updated non-v1 requirements to require minor version compatibility
• Updated rand requirement from ~0.7 to ~0.8.0
• Switched from Travis to GH Actions