allow test-go-pg to fail on windows

coder · hugodutka · Dec 3, 2024 · Nov 14, 2024 · Nov 14, 2024 · Nov 15, 2024
commit cd9b4d1489bd6bc3be037fc256f94884c1cdd0d3
diff --git a/.github/workflows/ci.yaml b/.github/workflows/ci.yaml
@@ -373,6 +373,10 @@ jobs:
     runs-on: ${{ matrix.os == 'ubuntu-latest' && github.repository_owner == 'coder' && 'depot-ubuntu-22.04-4' || matrix.os == 'macos-latest' && github.repository_owner == 'coder' && 'macos-latest-xlarge' || matrix.os == 'windows-2022' && github.repository_owner == 'coder' && 'windows-latest-16-cores' || matrix.os }}
     needs: changes
     if: needs.changes.outputs.go == 'true' || needs.changes.outputs.ci == 'true' || github.ref == 'refs/heads/main'
+    # This timeout must be greater than the timeout set by `go test` in
+    # `make test-postgres` to ensure we receive a trace of running
+    # goroutines. Setting this to the timeout +5m should work quite well
+    # even if some of the preceding steps are slow.
     timeout-minutes: 25
     strategy:
       fail-fast: false
@@ -428,6 +432,21 @@ jobs:
             DB=ci gotestsum --format standard-quiet -- -v -short -count=1 ./...
           fi
 
+      # This is used by the `required` job to determine if the test-go-pg job
+      # failed or not on the given OS. Matrix jobs don't support `outputs`
+      # well - the last job to run overwrites them. Instead, we write to
+      # artifacts.
+      - if: always()
+        run: echo "0" > "test-go-pg_result_${{ matrix.os }}"
+      - if: failure()
+        run: echo "1" > "test-go-pg_result_${{ matrix.os }}"
+      - name: Upload result artifact
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: "test-go-pg_result_${{ matrix.os }}"
+          path: "test-go-pg_result_${{ matrix.os }}"
+
       - name: Upload test stats to Datadog
         timeout-minutes: 1
         continue-on-error: true
@@ -808,28 +827,53 @@ jobs:
         uses: step-security/harden-runner@0080882f6c36860b6ba35c610c98ce87d4e2f26f # v2.10.2
         with:
           egress-policy: audit
-
+      - name: Download test-go-pg Artifacts
+        uses: actions/download-artifact@v4
+        with:
+          path: test-go-pg_result
+          pattern: test-go-pg_result_*
+          merge-multiple: true
       - name: Ensure required checks
+        shell: python
         run: |
-          echo "Checking required checks"
-          echo "- fmt: ${{ needs.fmt.result }}"
-          echo "- lint: ${{ needs.lint.result }}"
-          echo "- gen: ${{ needs.gen.result }}"
-          echo "- test-go: ${{ needs.test-go.result }}"
-          echo "- test-go-pg: ${{ needs.test-go-pg.result }}"
-          echo "- test-go-race: ${{ needs.test-go-race.result }}"
-          echo "- test-js: ${{ needs.test-js.result }}"
-          echo "- test-e2e: ${{ needs.test-e2e.result }}"
-          echo "- offlinedocs: ${{ needs.offlinedocs.result }}"
-          echo
-
-          # We allow skipped jobs to pass, but not failed or cancelled jobs.
-          if [[ "${{ contains(needs.*.result, 'failure') }}" == "true" || "${{ contains(needs.*.result, 'cancelled') }}" == "true" ]]; then
-            echo "One of the required checks has failed or has been cancelled"
-            exit 1
-          fi
-
-          echo "Required checks have passed"
+          import json
+          import sys
+          import os
+          from pathlib import Path
+
+          print("Checking required checks")
+
+          jobs = json.loads(os.environ["NEEDS"])
+          job_names = sorted(jobs.keys())
+          for job_name in job_names:
+              result = jobs[job_name]["result"]
+              print(f"- {job_name}: {result}")
+          print()
+
+          failed = False
+          for job_name in job_names:
+              result = jobs[job_name]["result"]
+
+              # Skip test-go-pg failures on windows
+              if job_name == "test-go-pg" and result == "failure":
+                  result_artifacts = list(Path("test-go-pg_result").glob("test-go-pg_result_*"))
+                  results = {f.name: int(f.read_text()) for f in result_artifacts}
+                  del results["test-go-pg_result_windows-2022"]
+                  if sum(results.values()) == 0:
+                      print("test-go-pg on windows-2022 failed, but we are temporarily skipping it until it's fixed")
+                      continue
+
+              if result in ["failure", "cancelled"]:
+                  failed = True
+                  break
+
+          if failed:
+              print("One of the required checks has failed or has been cancelled")
+              sys.exit(1)
+
+          print("Required checks have passed")
+        env:
+          NEEDS: ${{ toJSON(needs) }}
 
   # Builds the dylibs and upload it as an artifact so it can be embedded in the main build
   build-dylib: