skol

Architecture

Component Breakdown

Component	Completeness	Source Code	Purpose
CLI	Incomplete	cli paclage	Parses command line-flags and prepares files for compilation or analysis.
Lexer	Complete	lexer package	Breaks meaningless plain text into a sequence of meaningful tokens.
Parser	Incomplete	parser package	Consumes tokens from the lexer and constructs an AST.
AST	Incomplete	ast package	Represents the structure of a source code file.
Typechecker	Incomplete	typecheck package	Ensures that everything in the AST has the correct typing. (ensures a int variable isn’t set to a string, etc.)
IR	Incomplete	N/A	Breaks a program down into simple instructions to allow for easier compilation to binary formats.
Codegen	Incomplete	codegen package	a) Transpiles Skol code into another language from the AST. b) Compiles into executables from IR.

Usual flow for compilation

1. The CLI reads the source code.
2. A parser is created, which contains it’s own lexer for the previously read source code.
3. The parser consumes tokens, creates the adequate nodes and constructs an AST out of them.
4. The typechecker ensures type correctness in the program.

Component Completeness Breakdown

CLI

• Is able to parse actions and arguments separately.
• Is able to build a file using any engine.
• Is able to start a REPL using any engine.
• Has a way to access additional tools (e.g. linter).

Lexer

• Ignores comments.
• Reads punctuators.
• Reads identifiers.
• Reads integer literals.
• Reads float literals.
• Reads character literals.
• Reads string literals.

Parser

• Parses basic constructs:
  • Variable definition/assignment and both.
  • Function/extern definition.
  • Basic control flow.
  • Structured types.
  • Array types.
• Properly handles expected lexer errors (e.g. EOF).
• 100% proven waterproof.

AST

• Is properly constructed by the parser.
• Can correctly reflect the structure of a Skol program.
  • Global variables
  • Global functons/externs
  • Global types
  • Multi-file compilation.
  • Top-level code.

Typechecker

• Can check variables.
• Can check functions.
• Can check structure types.
• Can check array types.
• Can determine value types.
• Supports built-in functions.
• Supports generic functions.

IR

• Can be constructed from any valid AST.
• Can be cached as a file.

Codegen

• Distingushes between AST-based codegen (for transpilers) and IR-based codegen (for compilers).
• Transpiles valid code from AST.
• Compiles valid code from IR.

This site is open source. Improve this page.