Add model-free GBNF grammar validator

[mcharytoniuk]

Validate GBNF grammars against llama.cpp's own parser without loading a model.

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR adds a new standalone Rust crate (llama-cpp-gbnf) that validates GBNF grammars by calling into llama.cpp’s native grammar parser/engine via a new llama-cpp-bindings-sys FFI wrapper, enabling grammar validation without loading a model.

Changes:

• Introduces llama-cpp-gbnf crate with a public validate_gbnf(grammar, root) API and unit tests covering common validation outcomes.
• Adds a new C/C++ FFI surface (wrapper_gbnf.h/.cpp) to validate grammars using llama.cpp internals and exposes it via wrapper.h.
• Wires the new crate into the workspace, unit test runs, and coverage gating.

Reviewed changes

Copilot reviewed 9 out of 10 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
Makefile	Adds llama-cpp-gbnf to unit test execution and coverage gating.
Cargo.toml	Registers llama-cpp-gbnf as a workspace member and workspace dependency.
Cargo.lock	Locks the new crate and its dependencies.
llama-cpp-gbnf/Cargo.toml	Defines the new crate package metadata, dependencies, and lint configuration.
llama-cpp-gbnf/src/lib.rs	Establishes crate-level lint policy and exports modules.
llama-cpp-gbnf/src/gbnf_validation_error.rs	Adds a typed error enum for grammar validation outcomes.
llama-cpp-gbnf/src/validate_gbnf.rs	Implements the Rust validation API and unit tests.
llama-cpp-bindings-sys/wrapper.h	Includes the new GBNF wrapper header for bindgen/FFI exposure.
llama-cpp-bindings-sys/wrapper_gbnf.h	Declares the new validation status enum and validation function.
llama-cpp-bindings-sys/wrapper_gbnf.cpp	Implements model-free grammar validation against llama.cpp’s grammar parser/engine.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.