Cross-Layer Optimizer - Sprint 4

This module implements cross-layer optimization passes for the Unified HIR. The optimizer eliminates Python→C boundaries and performs optimizations that are only possible with unified Python+C knowledge.

Architecture

The optimizer is built around a pass-based architecture:

UnifiedHIR → Pass 1 → Pass 2 → Pass N → Optimized HIR

Core Passes

• Boundary Elimination: Removes Python→C FFI boundaries
• Dead Code Elimination: Removes unreachable code
• Inlining (future): Inlines small functions
• Constant Folding (future): Evaluates constants at compile time

Usage

use spydecy_optimizer::{OptimizationPipeline, BoundaryEliminationPass};
use spydecy_hir::unified::UnifiedHIR;

# fn example(unified_hir: UnifiedHIR) -> anyhow::Result<()> {
let mut pipeline = OptimizationPipeline::new();
pipeline.add_pass(Box::new(BoundaryEliminationPass::new()));

let optimized = pipeline.run(unified_hir)?;
# Ok(())
# }