test: seeded goldens for the robustness seeds (#826)

tests/robustness/goldens.json records what each of seeds 0-95 builds: its steps, how it ended (ok, or the kind and the start of the error it was refused with) and, for the design it left, the bodies' total volume and area (within a relative 1e-6) and their face, edge and shell counts. tests::robustness_goldens fails on any change; ROBUSTNESS_GOLDENS_BLESS="why" records one, with the reason and the seeds it moved. Pull requests that touch crates/kernel run every 8th seed; pushes to main and the nightly run every seed.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

.github/workflows/ci.yml

@@ -47,6 +47,7 @@ jobs:
     runs-on: ubuntu-24.04
     outputs:
       native: ${{ steps.paths.outputs.native }}
+      kernel: ${{ steps.paths.outputs.kernel }}
     steps:
       - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
         with:
@@ -59,12 +60,14 @@ jobs:
           HEAD_SHA: ${{ github.event.pull_request.head.sha }}
         run: |
           native=false
+          kernel=false
           if [ "$EVENT_NAME" != pull_request ]; then
             native=true
           else
             # Compare both old and new paths of renames; unknown paths run native checks.
             git diff --no-renames --name-only -z "$BASE_SHA...$HEAD_SHA" > "$RUNNER_TEMP/ci-paths"
             while IFS= read -r -d '' path; do
+              case "$path" in crates/kernel/*|vendor/truck-*) kernel=true ;; esac
               case "$path" in
                 .github/workflows/ci.yml) native=true ;;
                 *.rs|Cargo.toml|Cargo.lock|*/Cargo.toml|.cargo/*|rust-toolchain*) native=true ;;
@@ -74,6 +77,7 @@ jobs:
             done < "$RUNNER_TEMP/ci-paths"
           fi
           echo "native=$native" >> "$GITHUB_OUTPUT"
+          echo "kernel=$kernel" >> "$GITHUB_OUTPUT"
 
   checks:
     name: Format and repository checks
@@ -150,7 +154,7 @@ jobs:
 
   tests:
     name: Native Linux tests (${{ matrix.shard }}/4)
-    needs: build-tests
+    needs: [changes, build-tests]
     runs-on: ubuntu-24.04
     timeout-minutes: 60
     strategy:
@@ -177,6 +181,16 @@ jobs:
           # the build job's path, which is the same checkout path here.
           cargo nextest run --archive-file tests.tar.zst --workspace-remap . --extract-to . --extract-overwrite \
             --profile ci --partition "slice:$SHARD/4" "${all[@]}"
+      - name: Robustness goldens sample (pull requests that touch the kernel)
+        # Every 8th seed of tests/robustness/goldens.json, a quarter per shard; pushes to main and
+        # the nightly run every seed in the step above.
+        if: github.event_name == 'pull_request' && needs.changes.outputs.kernel == 'true'
+        env:
+          SHARD: ${{ matrix.shard }}
+        run: |
+          cargo nextest run --archive-file tests.tar.zst --workspace-remap . --extract-to . --extract-overwrite \
+            --profile ci --ignore-default-filter \
+            -E "package(solidcraft-engine) & test(=tests::robustness_goldens::sample_$((SHARD - 1)))"
       - name: Upload test timings
         if: always()
         uses: actions/upload-artifact@cf430e030ddbb5b0abf93d22962f4752f3646cd9 # v7.0.2