)]}'
{
  "commit": "dfeeb16e37bac6ec28ef0236cff00d53bbc9c0f0",
  "tree": "1f296b374f81017585dfac0a2215aa00a7d06101",
  "parents": [
    "069ead80c522c9909bc2762b757f7987df10cb99"
  ],
  "author": {
    "name": "Zhen Wang",
    "email": "643348094@qq.com",
    "time": "Tue Sep 01 17:18:54 2026 +0800"
  },
  "committer": {
    "name": "GitHub",
    "email": "noreply@github.com",
    "time": "Tue Sep 01 17:18:54 2026 +0800"
  },
  "message": "[#2769] improvement(ci): clean build outputs between sequential Maven profiles (#2770)\n\n### What changes were proposed in this pull request?\n\nAdd the Maven `clean` goal to every profile-specific invocation in\n`.github/workflows/sequential.yml`, including:\n\n- Spark 4, 4.1, 3, and 2\n- Hadoop MR 2.8 and 3.2\n- Tez\n- Dashboard\n\n### Why are the changes needed?\n\nThe sequential workflow runs multiple incompatible Maven profiles in the same\nworkspace. Reusing existing `target` directories can cause Maven or\n`scala-maven-plugin` incremental compilation to treat stale classes as\nup-to-date.\n\nCleaning before each invocation ensures that every profile compiles and\npackages only its selected sources and dependencies, making the workflow\nindependent of execution order.\n\nCloses #2769.\n\n### How was this patch tested?\n\nexisting CI\n\n----\nGenerated by Codex (GPT-5.6)",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "a7a3f1d0e337b9ba61334be092d2731aaea5fa61",
      "old_mode": 33188,
      "old_path": ".github/workflows/sequential.yml",
      "new_id": "6e574b3466c3265fb0e3a013a0a99955a54dbbe6",
      "new_mode": 33188,
      "new_path": ".github/workflows/sequential.yml"
    }
  ]
}