A schema for the data and the bytes
I’m building Vexil to put a protocol’s data model and byte layout in one schema. Bit widths, field ordinals, integer encodings, bounds and evolution rules are part of that contract.
The toolchain generates codecs for Rust, TypeScript, Go and Python. I want the same value to produce the same bytes across implementations, and changes to that agreement to be visible before they ship.
What it does
- Checks schemas and generates codecs.
- Carries a canonical schema hash with each generated target.
- Classifies compatible and breaking changes through the compatibility checker.
- Provides normative language and wire-format specifications, shared conformance vectors and cross-language examples.
The wire format is deliberately not self-describing. Both peers compile the schema they intend to use and can compare its hash. That gives them an explicit contract for interpreting the bytes, with less flexibility at runtime.
Current status
Vexil is pre-1.0. Rust and TypeScript have the broadest generated-code evidence. Go and Python cover a representative shared wire matrix, with less coverage than the other two targets. The support matrix describes those differences.
The language server in current source builds reports compiler diagnostics for single-file documents. It doesn’t yet load projects or provide completion, navigation, hover or formatting. This work is newer than the published CLI.
Source and examples
The project website has the documentation and specifications. The GitHub repository contains the source, examples and conformance material.
The quickstart follows one schema through generation, exact bytes and a round trip. The cross-language example uses the same fixture in all four generated targets.
