BE-197: Implement MIR inline pass for HashQL

[indietyp]

🌟 What is the purpose of this PR?

Implements a function inlining pass for HashQL's MIR (Mid-level Intermediate Representation). The pass reduces function call overhead and enables further optimizations by replacing call sites with the callee's body. It uses a two-phase approach: heuristic-based normal inlining followed by aggressive inlining for filter closures in graph read pipelines.

🔍 What does this change?

• Adds libs/@local/hashql/mir/src/pass/transform/inline/ module with:
  • mod.rs: Main pass implementation with two-phase inlining (normal + aggressive)
  • analysis.rs: Cost estimation and body analysis (loop detection, caller counting)
  • heuristics.rs: Scoring algorithm with configurable bonuses (leaf, single-caller, unique-callsite, loop)
  • find.rs: Callsite discovery for aggressive filter closure inlining
  • rename.rs: Reference renaming for inlined code (locals, blocks, places)
  • tests.rs: Comprehensive unit tests and snapshot tests
• Processes SCCs in dependency order to ensure callees are optimized before inlining
• Never inlines recursive calls (same SCC) to prevent infinite expansion
• Respects per-caller budget (default: 120) to avoid code bloat
• Aggressive phase runs until fixpoint or cutoff (default: 16 iterations

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

This PR:

• does not modify any publishable blocks or libraries, or modifications do not need publishing

📜 Does this require a change to the docs?

The changes in this PR:

• are internal and do not require a docs change

🕸️ Does this require a change to the Turbo Graph?

The changes in this PR:

• do not affect the execution graph