Skip to content

Latest commit

 

History

History
41 lines (33 loc) · 1.45 KB

README.md

File metadata and controls

41 lines (33 loc) · 1.45 KB
Raw

Toy Codegen

Toy Codegen is the example compiler for semantic-analyzer-rs.

As Codegen backend used LLVM, with inkwell.

Basic design

There is a basic implementation of code-generation. The Semantic AST structure is used as a basic data source. This means that any subset of programming languages can be represented, the main criterion for which is representability in a finite AST structure of Semantic Analyzer, which is used in the semantic-analyzer library.

Codegen implementation

The codege is focused only on those instructions that are presented as an AST sample. And it is not intended to cover all possible complex cases of code generation based on a complete set of SemanticStateContext instructions. However, the end result is a fully functional binary code.

The basic parts is:

  • Function declarations
    • Functions parameters initializations
  • Types declarations
  • Constants declarations
  • Function body implementation
    • Let-binding
      • Codegen variables state
    • Binding
    • Expressions
      • Codegen type conversions
    • If conditions flow
    • Loop conditions flow
    • Function return flow
  • Target triple initialization (LLVM backend)
  • LLVM IR source code generation (and store to file)
  • Binary code generation (and store to file)
  • Code generation error flow with custom error types

LICENCE: MIT