DRY: generate the code and documentation from a data-spec - Githubissues

rrthomas / mit

A simple stack-based VM

Other

10 stars 3 forks source link

DRY: generate the code and documentation from a data-spec #52

Open rrthomas opened 5 years ago

rrthomas commented 5 years ago

Literate Racket combining mit.tex and src/spec.py, which contains a complete data description that can be turned into YAML for easy parsing. It's not a general-purpose VM description DSL, but should contain:

Types (really only 1 for 0.1 VM)
Registers
Stack (including frame stack and error handling)
Memory
Primitive operations (like pForth's primitives) that "everyone knows the meaning of" to describe the semantics of the instructions (logic, arithmetic, memory access etc.)
Execution cycle

With this it should be possible to:

run the description as an interpreter, so it can pass the test suite.
Ideally, generate all C stack/frame code from it, including stack effects of all extra instructions.

rrthomas commented 4 years ago

This issue is not properly fixed: the spec document is still not generated. Divergence from the code has recently been noted! Reopening.