Bugfix release

- Added `--threads` options which can be used to change number of runner threads
- Refactored multi-threading. When `--threads=1` is specified, everything is run in a single thread
- More operations were moved to run-time. This saves some RAM (about 50-60 Mb) and speeds compilation up

A little release which makes code shorter and makes use of druntime's unittest thing

First public release