fix(tests): stabilize unreadable-source CLI assertion across runners by checking combined stdout/stderr output; fix README.md; fix matrics error in GAW

Author: orenlab

.github/workflows/benchmark.yml

@@ -64,22 +64,52 @@ jobs:
             memory: ""
             timeout_minutes: 60
 
-    if: >
-      (github.event_name != 'workflow_dispatch' && matrix.profile == 'smoke') ||
-      (github.event_name == 'workflow_dispatch' && matrix.profile == inputs.profile)
-
     steps:
+      - name: Resolve run profile gate
+        shell: bash
+        run: |
+          enabled=0
+          if [ "${{ github.event_name }}" != "workflow_dispatch" ]; then
+            if [ "${{ matrix.profile }}" = "smoke" ]; then
+              enabled=1
+            fi
+          else
+            if [ "${{ matrix.profile }}" = "${{ inputs.profile }}" ]; then
+              enabled=1
+            fi
+          fi
+          echo "BENCH_ENABLED=$enabled" >> "$GITHUB_ENV"
+
       - name: Checkout
-        uses: actions/checkout@v6
+        if: env.BENCH_ENABLED == '1'
+        uses: actions/checkout@v6.0.2
+
+      - name: Set up Python (macOS local benchmark)
+        if: env.BENCH_ENABLED == '1' && runner.os == 'macOS'
+        uses: actions/setup-python@v6.2.0
+        with:
+          python-version: "3.13"
+          allow-prereleases: true
+
+      - name: Set up uv (macOS local benchmark)
+        if: env.BENCH_ENABLED == '1' && runner.os == 'macOS'
+        uses: astral-sh/setup-uv@v5
+        with:
+          enable-cache: true
+
+      - name: Install dependencies (macOS local benchmark)
+        if: env.BENCH_ENABLED == '1' && runner.os == 'macOS'
+        run: uv sync --all-extras --dev
 
       - name: Set benchmark output path
+        if: env.BENCH_ENABLED == '1'
         shell: bash
         run: |
           mkdir -p .cache/benchmarks
           echo "BENCH_JSON=.cache/benchmarks/codeclone-benchmark-${{ matrix.label }}.json" >> "$GITHUB_ENV"
 
       - name: Build and run Docker benchmark (Linux)
-        if: runner.os == 'Linux'
+        if: env.BENCH_ENABLED == '1' && runner.os == 'Linux'
         env:
           RUNS: ${{ matrix.runs }}
           WARMUPS: ${{ matrix.warmups }}
@@ -90,7 +120,7 @@ jobs:
           cp .cache/benchmarks/codeclone-benchmark.json "$BENCH_JSON"
 
       - name: Run local benchmark (macOS)
-        if: runner.os == 'macOS'
+        if: env.BENCH_ENABLED == '1' && runner.os == 'macOS'
         run: |
           uv run python benchmarks/run_benchmark.py \
             --target . \
@@ -100,7 +130,7 @@ jobs:
             --output "$BENCH_JSON"
 
       - name: Print benchmark summary
-        if: always()
+        if: env.BENCH_ENABLED == '1'
         shell: bash
         run: |
           python - <<'PY'
@@ -184,9 +214,13 @@ jobs:
               fh.write("\n".join(lines) + "\n")
           PY
 
+      - name: Skip non-selected profile
+        if: env.BENCH_ENABLED != '1'
+        run: echo "Skipping matrix profile '${{ matrix.profile }}' for event '${{ github.event_name }}'"
+
       - name: Upload benchmark artifact
-        if: always()
-        uses: actions/upload-artifact@v7
+        if: env.BENCH_ENABLED == '1'
+        uses: actions/upload-artifact@v4
         with:
           name: codeclone-benchmark-${{ matrix.label }}
           path: ${{ env.BENCH_JSON }}

README.md

@@ -29,7 +29,7 @@ all with baseline-aware governance that separates **known** technical debt from
 - **Baseline governance** — known debt stays accepted; CI blocks only new clones and metric regressions
 - **Reports** — interactive HTML, deterministic JSON/TXT plus Markdown and SARIF projections from one canonical report
 - **CI-first** — deterministic output, stable ordering, exit code contract, pre-commit support
-- **Fast** — incremental caching, parallel processing, warm-run optimization
+- **Fast*** — incremental caching, parallel processing, warm-run optimization, and reproducible benchmark coverage
 
 ## Quick Start
 
@@ -42,21 +42,6 @@ codeclone . --json --md --sarif --text   # generate machine-readable reports
 codeclone . --ci             # CI mode (--fail-on-new --no-color --quiet)
 ```
 
-## Reproducible Docker Benchmark
-
-```bash
-./benchmarks/run_docker_benchmark.sh
-```
-
-The wrapper builds `benchmarks/Dockerfile`, runs isolated container benchmarks, and
-writes deterministic results to `.cache/benchmarks/codeclone-benchmark.json`.
-Use environment overrides to pin benchmark envelope:
-
-```bash
-CPUSET=0 CPUS=1.0 MEMORY=2g RUNS=16 WARMUPS=4 \
-  ./benchmarks/run_docker_benchmark.sh
-```
-
 <details>
 <summary>Run without install</summary>
 
@@ -273,10 +258,10 @@ Architecture: [`docs/architecture.md`](docs/architecture.md) · CFG semantics: [
 | Docker benchmark contract  | [`docs/book/18-benchmarking.md`](docs/book/18-benchmarking.md)                           |
 | Determinism                | [`docs/book/12-determinism.md`](docs/book/12-determinism.md)                             |
 
-<details>
-<summary>Benchmarking</summary>
+## Benchmarking
 
-## Reproducible Docker Benchmark
+<details>
+<summary>Reproducible Docker Benchmark</summary>
 
 ```bash
 ./benchmarks/run_docker_benchmark.sh
@@ -292,7 +277,8 @@ CPUSET=0 CPUS=1.0 MEMORY=2g RUNS=16 WARMUPS=4 \
   ./benchmarks/run_docker_benchmark.sh
 ```
 
-Benchmark contract: [docs/book/18-benchmarking.md](docs/book/18-benchmarking.md)
+* Performance claims are backed by the reproducible benchmark workflow documented
+  in [docs/book/18-benchmarking.md](docs/book/18-benchmarking.md)
 
 </details>
 

tests/test_cli_inprocess.py

@@ -2930,10 +2930,11 @@ def _source_read_error(
     monkeypatch.setattr(pipeline, "process_file", _source_read_error)
     _patch_parallel(monkeypatch)
     _run_main(monkeypatch, [str(tmp_path), "--no-progress"])
-    out = capsys.readouterr().out
-    assert "Cannot read file" in out
-    assert "CONTRACT ERROR:" not in out
-    assert _summary_metric(out, "Files skipped") == 1
+    captured = capsys.readouterr()
+    combined = captured.out + captured.err
+    assert "Cannot read file" in combined
+    assert "CONTRACT ERROR:" not in combined
+    assert _summary_metric(captured.out, "Files skipped") == 1
 
 
 def test_cli_unreadable_source_fails_in_ci_with_contract_error(