java-topology/defects/flink/unit/FlinkJobGraphJarDedupTest.java

package unit;

import java.util.ArrayList;
import java.util.LinkedHashSet;
import java.util.List;

/**
 * flink-0001: JobGraph user-jar dedup — List.contains() O(n²) vs LinkedHashSet O(n).
 *
 * Standalone unit test — no JUnit required.
 * Compile: javac -d . FlinkJobGraphJarDedupTest.java
 * Run:     java unit.FlinkJobGraphJarDedupTest
 */
public class FlinkJobGraphJarDedupTest {

    static long slowOps;
    static long fastOps;

    /** Simulates JobGraph.addJar() with ArrayList — O(n) contains per add. */
    static List<String> slowAddJars(List<String> jars) {
        slowOps = 0;
        List<String> userJars = new ArrayList<>();
        for (String jar : jars) {
            slowOps++;                          // entry into addJar()
            for (String existing : userJars) {  // ArrayList.contains() scan
                slowOps++;
                if (existing.equals(jar)) break;
            }
            if (!userJars.contains(jar)) {
                userJars.add(jar);
            }
        }
        return userJars;
    }

    /** Patched: LinkedHashSet.add() is O(1) amortised — no scan needed. */
    static List<String> fastAddJars(List<String> jars) {
        fastOps = 0;
        LinkedHashSet<String> userJars = new LinkedHashSet<>();
        for (String jar : jars) {
            fastOps++;          // O(1) hash add — single op per entry
            userJars.add(jar);
        }
        return new ArrayList<>(userJars);
    }

    static void run(int n, int expectedNx) {
        List<String> jars = new ArrayList<>();
        for (int i = 0; i < n; i++) jars.add("jar-" + i + ".jar");
        // add duplicates at the end to stress the dedup path
        for (int i = 0; i < n; i++) jars.add("jar-" + i + ".jar");

        List<String> slowResult = slowAddJars(jars);
        List<String> fastResult = fastAddJars(jars);

        boolean resultsMatch = slowResult.equals(fastResult);
        boolean quadraticWorse = slowOps > fastOps * expectedNx;

        System.out.printf("n=%-4d  slow=%6d  fast=%4d  ratio=%5.1fx  match=%b  PASS=%b%n",
            n, slowOps, fastOps, (double) slowOps / fastOps, resultsMatch,
            resultsMatch && quadraticWorse);

        if (!resultsMatch || !quadraticWorse) {
            throw new AssertionError(
                "FAIL n=" + n + " resultsMatch=" + resultsMatch +
                " slowOps=" + slowOps + " fastOps=" + fastOps +
                " needed ratio>" + expectedNx);
        }
    }

    public static void main(String[] args) {
        System.out.println("=== flink-0001: JobGraph jar dedup O(n^2) vs O(n) ===");
        run(50,  5);
        run(200, 10);
        run(500, 20);
        System.out.println("3/3 PASS");
    }
}