mirror of https://github.com/allaunthefox/SilverSight.git synced 2026-07-31 01:25:21 +00:00

allaun f6cbddcbf2 feat(tests): Python AVM port rewrite + test harness, Go test harness

Python port rewritten to match spec:
- Added Q0_16, PUSH_Q0, PUSH_BOOL as separate opcodes
- Added V6 comparison (lt_q16_v6)
- Added floor division (Lean Int.ediv)
- Added stack depth limit (AVM_MAX_STACK = 1024)
- Added type checking in exec_prim
- All 10 tests passing

Go AVM port: added test_avm_test.go with 8 test cases

Milestone: Python → ✅, Go → 🔄

2026-06-30 17:56:09 -05:00

2.7 KiB

Raw Blame History

Milestone: AVM ISA 1:1 Port Coverage

Goal: Every language with an LSP installed on the build infrastructure must have a complete, Lean-validated AVM ISA v1 port with cross-implementation test harness.

Coverage Matrix

#	Language	LSP	AVM Port	Tests	Wolfram Validated	Status
1	Lean	`lean --server`	`formal/SilverSight/AVMIsa/`	`E2E.lean`	✅	Reference
2	Rust	`rust-analyzer`	`rust/src/avm/mod.rs`	`test_add_q16`	✅	✅
3	Python	`pyright`	`python/avm.py`	`test_avm_python.py`	❌	✅
4	Coq	`coq-lsp`	`coq/AVMIsa/avm.v`	❌	❌	🔄
5	R	—	`r/AVMIsa/avm.r`	`test_avm.r`	✅	✅
6	Julia	—	`julia/AVMIsa/avm.jl`	`test_avm.jl`	✅	✅
7	Scala	`metals`	`scala/avm.scala`	❌	❌	🔄
8	Go	—	`go/avm.go`	`go/avm_test.go`	❌	🔄
9	C	`clangd`	`c/avm.c`	❌	❌	🔄
10	C++	`clangd`	`cpp/avm.hpp`	❌	❌	🔄
11	Fortran	`fortls`	`fortran/avm.f90`	❌	❌	🔄
12	Octave	—	`octave/avm.m`	❌	❌	🔄

1:1 Verification Protocol

For each port, run the same test vector through every implementation:

test_vector = [
    # (op, a, b, expected_q16)
    ("add", 5*65536, 3*65536, 8*65536),         # 5 + 3 = 8
    ("sub", 10*65536, 3*65536, 7*65536),         # 10 - 3 = 7
    ("mul", 5*65536, 3*65536, 15*65536),          # 5 * 3 = 15
    ("div", 10*65536, 2*65536, 5*65536),          # 10 / 2 = 5
    ("lt",  5*65536, 3*65536, False),             # 5 < 3 = false
    ("lt",  -5*65536, -3*65536, True),            # -5 < -3 = true
    ("add_sat", cload_max-1, 2, cload_max),       # saturation
    ("div_neg", -5*65536, 3*65536, -1*65536),     # floor division: -5/3 = -2
]

All results must match Lean #eval witnesses to pass.

Action Items

Python — Write tests/test_avm_python.py with Lean-cross-validated test vector
Coq — Write coq/AVMIsa/test_avm.v with Example witnesses
Scala — Write test harness in scala/src/test/
Go — Write go/avm_test.go
C — Write c/test_avm.c
C++ — Write cpp/test_avm.cpp
Fortran — Write fortran/test_avm.f90
Octave — Write octave/test_avm.m
CI — Add a GitHub Actions workflow that runs all port tests on every push
Wolfram Alpha — Re-audit all ports against the Lean reference

Definition of Done

All 12 ports pass the same cross-validated test vector
Test outputs match Lean #eval witnesses for every operation
CI pipeline runs all port tests on push
Audit report (docs/avm_ports_audit.md) updated with "1:1" status

2.7 KiB Raw Blame History