Yantra 0.1 was released on GitHub as a native C++ LALR(1) parser generator. The tool provides an integrated lexer, Unicode/UTF‑8 support, a built‑in AST builder, and an optional amalgamated mode that emits a single self‑contained hello.cpp file with a ready‑to‑run main().
Features and quick start
Yantra bundles a lexer that can handle multi‑mode token streams, making nested comments possible out of the box. It generates a full parser executable called ycc via a standard CMake build (cmake .. && cmake --build). After cloning the repo, a grammar like hello.y can be turned into hello.cpp with bin/ycc -c ascii -f hello.y -a. The resulting source compiles with any C++23 compiler (clang, gcc, MSVC) and can be fed strings directly via -s or files via -f Yantra.
Comparison to other generators
Unlike Bison/Lemon, which run semantic actions during bottom‑up reductions, Yantra always builds the whole AST first and then walks it top‑down, allowing parent rules to execute before children. Compared to ANTLR, Yantra stays in the LALR(1) family, preserving its time‑ and space‑efficiency while still offering a top‑down walk, but it only targets C++ and avoids a JVM dependency. Tree‑sitter focuses on incremental editing support, a feature Yantra explicitly does not provide. The project notes these trade‑offs in its documentation and lists a “Known Limitations” section for missing capabilities.
Plug‑and‑play in a codebase
Yantra has no external dependencies beyond the C++ standard library, so it can be dropped into an existing project either as separate .hpp/.cpp files or as a single amalgamated source file. A sample application lives in the companion repo lingo, and a language‑server extension for IDEs is available at yantra‑language‑server.
Caveats
The repository does not publish any pricing information, but the MIT license makes it free to use. The project is maintained by a single contributor and is newer than alternatives like Bison or ANTLR, so community support and maturity may be limited. Additionally, the docs do not cover incremental parsing or error‑tolerant recovery, which could be a deal‑breaker for editor integrations.
When to try it – If you need a lightweight, C++‑only parser generator with built‑in lexer and AST support and can accept a smaller community footprint, pull Yantra into a sandbox project and generate a simple grammar to validate the build flow before committing to production code.
