Python-compiled grammars in C parser

Just spent an hour digging into textparser's internals. The JSON grammar schema is refreshingly minimal compared to the DSLs most parser generators force on you. What stands out to me is how they decouple grammar development from C compilation. You write your grammar in a Python script, it compiles to JSON, then textparser ingests that at runtime. This means you can iterate on syntax rules without recompiling your entire C project. For embedded or game engine work where you need a custom DSL for config files or scripting, that's a massive quality of life improvement. The hybrid approach raises an interesting trade off though. By using Python compiled grammars, you inherit Python's runtime for grammar generation. That introduces a dependency you can't easily strip out for cross compilation or resource constrained targets. I wonder if they plan to support a standalone JSON grammar file format that bypasses Python entirely for deployment. One thing I'd love to see benchmarked: how does the tokenization throughput compare to hand rolled state machines for simple formats like JSON or INI files? The C engine itself looks lean, but the grammar interpretation overhead could eat into those gains for trivial parsing tasks.