Files
codebase-memory-mcp/docs
Martin Vogel 97764d48b4 docs(py_lsp): final results table through Round 9
Updates BENCHMARK_PYTHON.md to capture the full trajectory:
- Bench: 81% (65 LOC) -> 79% (136 LOC) -> 95% (136 LOC) -> 98%
  (178 LOC) -> 100% (178 LOC, 52/52)
- 22 stress-test patterns probed; 19 hard-asserted, 3 documented gaps
- Documented gaps: function-as-dict-value, match sequence pattern,
  lambda parameter inference

The hard-asserted patterns include the full LSP surface a user would
expect: NamedTuple, TypedDict, Protocol, ABC, with-as, except-as,
post-early-return narrowing, tuple unpacking, dict.items
comprehension, slice typing, decorator factories, property setters,
Self chains, diamond MRO, recursive types, closures, generator
delegation, async-gen for, long fluent chains.
2026-05-09 12:49:02 +02:00
..
2026-03-26 15:14:57 +01:00