Experiments in porting the Inform type system to Rust, via a Python program that generates the relevant types.
In any case where I've got to choose between faithfulness to Inform and Rust ergonomics, I'll pick Rust ergonomics.
You can play with it if you want -- see main.py for a convincing example.
However, it's just a hack.
Edit main.py to reflect your project, then run it. Consider creating your
project with Cargo first. (it will overwrite existing files but doesn't
generate any that overlap a default cargo project.)
You should add the world module to main.rs in order to load the generated
code.
Ideally you should make all requirements available in a venv. (by installing requirements.txt) I have only tested my code on Python 3.7, but I expect my code will probably work on Python 3.6 and Python 3.8 too.
Inform has a great type system for video game content -- each instance has its own type and you can comprehense over all the items of a type. Each is effectively global.
Things I did:
- things (renamed to "entities") and kinds
- actions
- relations+initializers (minus equivalence classes)
Things I didn't do yet:
- meaningful examples
- string-generating procedures
- literally any parsing
A lot of stuff is pretty impoverished. Also, there are no tests -- my testing methodology is "run main.py, then see if the resulting Rust code compiles."
You really shouldn't use this yet.
I'll read your code if you make a pull request, but you should really contact me in advance.
Please run black on default settings to format your code.
Named for Andrew Plotkin, one of the three Inform 7 designers (the other two being Graham Nelson and Emily Short.)