mirror of
https://github.com/leanprover/lean4.git
synced 2026-03-17 10:24:07 +00:00
mvcgen-with-grind
mvcgen' with grind
This PR adds a persistent `Sym.Simp.State` cache to VCGen and a `reassocNatAdd` simproc that rewrites `(a + m) + n → a + (m + n)` for Nat literals. Hypotheses and targets are simplified via `replaceLocalDeclDefEq` and `replaceTargetDefEq` (O(1)) before grind internalization, avoiding O(n) re-internalization of unsimplified `s + 1 + 1 + ...` chains. The persistent cache gives O(1) amortized simplification per VC since pointer-based cache hits on shared subexpressions avoid redundant work. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This is the repository for Lean 4.
About
- Quickstart
- Homepage
- Theorem Proving Tutorial
- Functional Programming in Lean
- Documentation Overview
- Language Reference
- Release notes starting at v4.0.0-m3
- Examples
- External Contribution Guidelines
Installation
See Install Lean.
Contributing
Please read our Contribution Guidelines first.
Building from Source
See Building Lean.
Languages
Lean
94.3%
C++
4.1%
Python
0.6%
Shell
0.4%
CMake
0.3%