Adding SRPM duplicates check. (#13315)

Author: PawelWMS

.github/workflows/check-srpm-duplicates.yml

@@ -0,0 +1,51 @@
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+# This action checks that the specs in this repo
+# generate SRPMs with unique names.
+name: SRPMs duplicates check
+
+on:
+  push:
+    branches: [main, 2.0*, 3.0*, fasttrack/*]
+  pull_request:
+    branches: [main, 2.0*, 3.0*, fasttrack/*]
+
+jobs:
+  check:
+    name: SRPMs duplicates check
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        # Each group is published to a different repo, thus we only need to check
+        # for SRPM duplicates within the group.
+        specs-dirs-groups: ["SPECS SPECS-SIGNED", "SPECS-EXTENDED"]
+
+    steps:
+      # Checkout the branch of our repo that triggered this action
+      - name: Workflow trigger checkout
+        uses: actions/checkout@v4
+
+      # For consistency, we use the same major/minor version of Python that Azure Linux ships
+      - name: Setup Python 3.9
+        uses: actions/setup-python@v5
+        with:
+          python-version: 3.9
+
+      - name: Switch to stable toolkit
+        run: git fetch --all && git checkout 2.0-stable -- toolkit
+
+      # Generate the specs.json files. They are the input for the duplicates check script.
+      - name: Generate specs.json
+        run: |
+          set -euo pipefail
+
+          for spec_folder in ${{ matrix.specs-dirs-groups }}; do
+            echo "Generating specs.json for spec folder '$spec_folder'."
+
+            sudo make -C toolkit -j$(nproc) parse-specs REBUILD_TOOLS=y SPECS_DIR=../$spec_folder
+            cp -v build/pkg_artifacts/specs.json ${spec_folder}_specs.json
+          done
+
+      - name: Check for duplicate SRPMs
+        run: python3 toolkit/scripts/check_srpm_duplicates.py *_specs.json

toolkit/scripts/check_srpm_duplicates.py

@@ -0,0 +1,67 @@
+#!/usr/bin/env python3
+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+
+import argparse
+import json
+import os
+import sys
+from collections import defaultdict
+
+_REPO_KEY = "Repo"
+_SPEC_PATH_KEY = "SpecPath"
+_SRPM_PATH_KEY = "SrpmPath"
+
+
+def find_srpm_duplicates(specs_file_paths: list[str]) -> list[tuple[str, set[str]]]:
+    """
+    Analyze multiple specs JSON files to find specs producing the same SRPM.
+    """
+    srpm_to_specs = defaultdict(set)
+
+    for specs_file_path in specs_file_paths:
+        with open(specs_file_path, "r") as f:
+            data = json.load(f)
+
+        if _REPO_KEY not in data:
+            raise ValueError(
+                f"Invalid JSON format in {specs_file_path}. Expected '{_REPO_KEY}' key."
+            )
+
+        # Process each item in the repo
+        for item in data["Repo"]:
+            if _SRPM_PATH_KEY not in item or _SPEC_PATH_KEY not in item:
+                raise ValueError(
+                    f"Invalid JSON format in {specs_file_path}. Expected '{_SPEC_PATH_KEY}' and '{_SRPM_PATH_KEY}' keys in each element of '{_REPO_KEY}'."
+                )
+
+            srpm = os.path.basename(item[_SRPM_PATH_KEY])
+            srpm_to_specs[srpm].add(item[_SPEC_PATH_KEY])
+
+    return [
+        (srpm, specs_paths)
+        for srpm, specs_paths in srpm_to_specs.items()
+        if len(specs_paths) > 1
+    ]
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "specs_file_paths",
+        nargs="+",
+        help="Paths to the specs JSON files to analyze.",
+    )
+    args = parser.parse_args()
+
+    srpm_duplicates = find_srpm_duplicates(args.specs_file_paths)
+    if srpm_duplicates:
+        print("Error: detected specs building the same SRPM.", file=sys.stderr)
+        for srpm, specs_paths in srpm_duplicates:
+            print(f"{srpm}:", file=sys.stderr)
+            for spec_path in specs_paths:
+                print(f"  - {spec_path}", file=sys.stderr)
+            print(file=sys.stderr)
+        sys.exit(1)
+
+    print("No SRPM duplicates found.")