Skip to content

fix(tests): repair third stale CARGO_MANIFEST_DIR path missed by #1857#1860

Merged
noahgift merged 1 commit into
mainfrom
fix/stale-path-3-aprender-gpu
May 21, 2026
Merged

fix(tests): repair third stale CARGO_MANIFEST_DIR path missed by #1857#1860
noahgift merged 1 commit into
mainfrom
fix/stale-path-3-aprender-gpu

Conversation

@noahgift

Copy link
Copy Markdown
Contributor

Summary

Third stale `../../aprender/contracts/` path that #1857 missed.

`crates/aprender-gpu/src/kernels/tests/fusion_contract_falsify.rs` builds CONTRACT_PATH via:

```rust
concat!(env!("CARGO_MANIFEST_DIR"), "/../../aprender/contracts/kernel-fusion-v1.yaml")
```

Post-monorepo, `CARGO_MANIFEST_DIR` for `aprender-gpu` is `crates/aprender-gpu/`, so the old suffix resolves to `crates/aprender/contracts/` (doesn't exist). Correct suffix is `/../../contracts/`.

Comment updated to reflect post-consolidation reality too.

Verification

```bash
$ cargo test -p aprender-gpu --features cuda --lib --release falsify_fusion_001
```

The test now finds the contract and runs. It still surfaces the SAME `FusedQKVKernel`-class contract drift covered by #1858 — that's the falsifier working as intended, not a regression here.

Test plan

  • `cargo check -p aprender-gpu --lib --tests` clean
  • Test now compiles and reaches the assertion (used to fail at `include_str!` resolution)
  • Audit complete — no other stale `../../aprender/contracts/` paths remain anywhere in `crates/`

🤖 Generated with Claude Code

…on contract test

Third stale `../../aprender/contracts/` path missed by #1857.

`crates/aprender-gpu/src/kernels/tests/fusion_contract_falsify.rs`
constructed CONTRACT_PATH via:

```
concat!(env!("CARGO_MANIFEST_DIR"), "/../../aprender/contracts/kernel-fusion-v1.yaml")
```

Post-monorepo-consolidation, CARGO_MANIFEST_DIR for `aprender-gpu` is
`crates/aprender-gpu/`, so the old suffix resolved to `crates/aprender/contracts/`
which doesn't exist. The correct suffix is `/../../contracts/`.

Comment also updated to reflect post-monorepo reality.

## Verification

```
cargo test -p aprender-gpu --features cuda --lib --release falsify_fusion_001
```

Now finds the contract and runs the falsifier. The test still fails on
the SAME contract drift surfaced by #1858 (FusedQKVKernel and 4 sibling
kernels missing from `kernel-fusion-v1.yaml`) — that's the falsifier
doing its job, not a regression from this fix.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
@noahgift noahgift enabled auto-merge (squash) May 21, 2026 21:44
@noahgift noahgift merged commit daea88f into main May 21, 2026
11 checks passed
@noahgift noahgift deleted the fix/stale-path-3-aprender-gpu branch May 21, 2026 22:03
noahgift added a commit that referenced this pull request May 22, 2026
…yaml (closes #1858) (#1861)

The QA-002 falsifier (`enforcement.kernel_registry_complete`) detected
five fused kernels that exist in the codebase but lack `fusion_decisions`
entries. Now registered as FUSION-006..010, all status ACTIVE with
verified call sites.

## Why this matters

Without entries, these kernels are invisible to:
- `pv lint` / contract-coverage audits
- The BLOCKED-requires-benchmark rule (QA-003)
- The Poka-Yoke registry check (kernel_registry_complete)

The drift was previously masked because the falsifier tests couldn't
compile (broken `include_str!` paths repaired in #1857 / #1860).
Repairing those paths surfaced the drift in `fusion_gate_contract_falsify`
(aprender-serve) and `falsify_fusion_001_every_fused_kernel_has_yaml_entry`
(aprender-gpu).

## Five entries added

| ID         | Kernel                            | Replaces                          | Call site |
|------------|-----------------------------------|-----------------------------------|-----------|
| FUSION-006 | FusedQKVKernel                    | Q/K/V GEMVs                       | crates/aprender-serve/src/cuda/generate.rs:267 |
| FUSION-007 | FusedGateUpKernel                 | Gate+Up GEMV + SwiGLU (FP32 path) | crates/aprender-serve/src/cuda/generate.rs:270 |
| FUSION-008 | FusedGemmBiasGeluKernel           | GEMM + Bias + GELU                | crates/aprender-gpu/src/memory/resident/ops/composite/linear_bias.rs:232 |
| FUSION-009 | FusedRmsNormQ4KGemvKernel         | RMSNorm + Q4K GEMV                | crates/aprender-serve/src/cuda/generate.rs:202 |
| FUSION-010 | FusedGateUpQ4KGemvKernel          | Gate+Up Q4K GEMVs (shared input)  | crates/aprender-serve/src/cuda/generate.rs:273 |

Each kernel was confirmed via `grep -rn 'pub struct <Name>\b'` and
each call site verified by inspecting the dispatching file.

FUSION-007 vs FUSION-010 distinction documented inline:
- FUSION-007 = FP32-input gate+up+SwiGLU fused (FusedGateUpKernel)
- FUSION-010 = Q4K-input gate+up sharing input vector load
  (FusedGateUpQ4KGemvKernel); SwiGLU applied separately downstream.

## Verification

```
$ cargo run -p aprender-contracts-cli --bin pv -- validate \
    contracts/kernel-fusion-v1.yaml
Contract is valid.

$ cargo test -p aprender-serve --test fusion_gate_contract_falsify --release
test result: ok. 5 passed; 0 failed     ← was 4/5 with QA-002 failing

$ cargo test -p aprender-gpu --features cuda --lib --release \
    falsify_fusion_001_every_fused_kernel_has_yaml_entry
test result: ok. 1 passed; 0 failed     ← was failing on FusedQKVKernel
```

Closes #1858.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant