By default we should test all our code for regressions. I'm always
running the examples anyway to test new features, so we should have
that recipe run afterwards.
If a test fails, that's first priority to fix.
If an example fails, time to continue working.
Setup build system (POSIX sh), gitignore, basic C file with an
implementation of something I really wanted to setup.
It just hashes a snippet of lorem ipsum. Testing seems to indicate
it's working. That's all it does lol.
This is a really pressing matter; all my previous Lisps always just
made the strings on the fly and that irked me deeply. I want a smart
implementation that really tries to save memory on something as
intensive as symbols.
