Compilation Pipeline

This page tracks how a translation unit moves through NyxLang.

1) Lexing (Flex) — c11.l

Input: source text
Output: stream of tokens used by the parser.

Lexer responsibilities typically include:

• identifiers/keywords
• numeric literals
• operators/punctuators
• comments/whitespace handling

2) Parsing (Bison) — c11.y

Input: token stream
Output: AST.

Parser responsibilities:

• enforce grammar
• precedence/associativity
• build correct AST node kinds

3) AST — ast.zig

Input: parse result
Output: in-memory structured program representation.

AST responsibilities:

• represent syntax (expressions, statements, decls, types)
• preserve source positions if supported
• avoid semantic assumptions

4) Semantic Analysis — semanticAnalyzer.zig

Input: AST
Output: validated program + metadata (resolved symbols, types), or errors.

Common checks:

• “identifier not declared”
• redefinition in same scope
• type compatibility (assignments, calls, returns)
• correct usage of lvalues/rvalues
• function declarations/definitions consistency

5) IR generation (3AC) — 3ac.zig

Input: semantically valid AST
Output: linear IR (three-address instructions + labels).

IR goals:

• make control-flow explicit
• make temporaries explicit
• preserve enough information for efficient RV32 lowering

6) Backend — RV32 — assembler.zig

Input: 3AC
Output: RISC-V RV32 assembly.

Backend responsibilities:

• calling convention + stack frames
• register allocation strategy (even if simple)
• instruction selection + lowering
• prologue/epilogue emission