From 72b4de4d688c80b71a60ffb5b18457bd7915fca0 Mon Sep 17 00:00:00 2001
From: Dongjoon Hyun <dongjoon@apache.org>
Date: Wed, 9 Apr 2025 10:27:01 +0900
Subject: [PATCH] [SPARK-51749] Add `MacOS` integration test with Apache Spark
 4.0.0 RC3

---
 .github/workflows/build_and_test.yml   | 18 +++++++++++++++-
 Tests/SparkConnectTests/SQLTests.swift | 29 ++++++++++++++++++++++++--
 2 files changed, 44 insertions(+), 3 deletions(-)

diff --git a/.github/workflows/build_and_test.yml b/.github/workflows/build_and_test.yml
index 76f740a..0c37870 100644
--- a/.github/workflows/build_and_test.yml
+++ b/.github/workflows/build_and_test.yml
@@ -55,7 +55,7 @@ jobs:
     - name: Build
       run: swift build -v
 
-  integration-test:
+  integration-test-linux:
     runs-on: ubuntu-latest
     services:
       spark:
@@ -72,3 +72,19 @@ jobs:
         swift-version: "6"
     - name: Test
       run: swift test --no-parallel
+
+  integration-test-mac:
+    runs-on: macos-15
+    steps:
+    - uses: actions/checkout@v4
+    - uses: swift-actions/setup-swift@v2.2.0
+      with:
+        swift-version: "6"
+    - name: Test
+      run: |
+        curl -LO https://dist.apache.org/repos/dist/dev/spark/v4.0.0-rc3-bin/spark-4.0.0-bin-hadoop3.tgz
+        tar xvfz spark-4.0.0-bin-hadoop3.tgz
+        cd spark-4.0.0-bin-hadoop3/sbin
+        ./start-connect-server.sh
+        cd ../..
+        swift test --no-parallel
diff --git a/Tests/SparkConnectTests/SQLTests.swift b/Tests/SparkConnectTests/SQLTests.swift
index c875a80..997e888 100644
--- a/Tests/SparkConnectTests/SQLTests.swift
+++ b/Tests/SparkConnectTests/SQLTests.swift
@@ -30,11 +30,25 @@ struct SQLTests {
 
   let regexID = /#\d+L?/
   let regexPlanId = /plan_id=\d+/
+  let regexLocation = /file:[a-zA-Z0-9\.\-\/\\]+/
+  let regexOwner = /(runner|185)/
+
+  private func cleanUp(_ str: String) -> String {
+    return removeOwner(removeID(removeLocation(str)))
+  }
 
   private func removeID(_ str: String) -> String {
     return str.replacing(regexPlanId, with: "plan_id=").replacing(regexID, with: "#")
   }
 
+  private func removeLocation(_ str: String) -> String {
+    return str.replacing(regexLocation, with: "*")
+  }
+
+  private func removeOwner(_ str: String) -> String {
+    return str.replacing(regexOwner, with: "*")
+  }
+
   @Test
   func testRemoveID() {
     #expect(removeID("123") == "123")
@@ -44,6 +58,17 @@ struct SQLTests {
     #expect(removeID("plan_id=123") == "plan_id=")
   }
 
+  @Test
+  func removeLocation() {
+    #expect(removeLocation("file:/abc") == "*")
+  }
+
+  @Test
+  func removeOwner() {
+    #expect(removeOwner("runner") == "*")
+    #expect(removeOwner("185") == "*")
+  }
+
 #if !os(Linux)
   @Test
   func runAll() async throws {
@@ -54,8 +79,8 @@ struct SQLTests {
 
       let sql = try String(contentsOf: URL(fileURLWithPath: "\(path)/\(name)"), encoding: .utf8)
       let jsonData = try encoder.encode(try await spark.sql(sql).collect())
-      let answer = removeID(String(data: jsonData, encoding: .utf8)!)
-      let expected = removeID(try String(contentsOf: URL(fileURLWithPath: "\(path)/\(name).json"), encoding: .utf8))
+      let answer = cleanUp(String(data: jsonData, encoding: .utf8)!)
+      let expected = cleanUp(try String(contentsOf: URL(fileURLWithPath: "\(path)/\(name).json"), encoding: .utf8))
       #expect(answer == expected.trimmingCharacters(in: .whitespacesAndNewlines))
     }
     await spark.stop()