Float Data Type

[MelbourneDeveloper]

TLDR

Float type support and arithmetic operations now return Result types for compile-time overflow safety. All arithmetic expressions (+, -, *, /, %) return Result<T, String> requiring explicit pattern matching.

What Was Added?

• Float type: FLOAT token in grammar, FloatLiteral AST node, float type in Hindley-Milner type system
• Result-wrapped arithmetic: All arithmetic operators return Result<int, String> or Result<float, String>
• Type inference helpers: isNumericType(), isFloatType(), isIntOrResultInt(), unwrapResultType() for Result type handling
• Numeric unification: unifyConcreteWithGeneric() to unify Result with Result[float] notation
• Overflow detection: LLVM intrinsics (sadd_with_overflow, etc.) for runtime overflow checking

What Was Changed / Deleted?

• Arithmetic operations: Changed from direct int/float returns to Result<T, String> for overflow safety (core_functions.go)
• Type inference: Enhanced to handle Result types in arithmetic and auto-unwrap where safe (type_inference.go: +320 lines)
• All test files: 50+ test files updated to pattern match arithmetic results
• Fiber generation: Updated fiber message passing to handle Result types (fiber_generation.go)
• Function signatures: Extended to support float parameters and Result returns (function_signatures.go)

How Do The Automated Tests Prove It Works?

• make test passes all integration tests with Result-wrapped arithmetic
• examples/tested/basics/* demonstrate correct pattern matching on arithmetic results
• Fiber tests verify Result types work with concurrent operations
• Type inference tests verify float/int unification and Result type handling
• Error case tests (failscompilation/) verify overflow detection and proper error messages

Summarise Changes To The Spec Here

• 0002-LexicalStructure.md: Added FLOAT literal syntax (3.14, 0.5, 100.0)
• 0004-TypeSystem.md: Added float primitive, documented Result<T, E> wrapping for arithmetic
• 0007-PatternMatching.md: Added requirement to match arithmetic results due to Result returns
• 0013-ErrorHandling.md: Documented arithmetic overflow detection via Result types