java-topology/defects/sinatra/unit/SinatraTest.java

package unit;

import java.util.*;
import java.util.regex.*;

/**
 * SinatraTest — CWE-407 benchmarks for Sinatra base.rb defects.
 *
 * sinatra-0001: content_type calls add_charset.all? {|p| !(p === mime_type)} on every response.
 *   Slow: Array#all? iterates all entries (Strings + Regexps) — O(k) per content_type call.
 *   Fast: Set#include? for string entries (O(1)), Regexp only on miss.
 *
 * sinatra-0002: provides condition block calls types.include?(response_content_type) per request.
 *   Slow: Array#include? — O(n) where n = number of types in the provides() call.
 *   Fast: Set#include? built once at route-definition time — O(1) per request.
 *
 * Ruby Array#include? and Java List#contains are both O(n) linear scans.
 * Ruby Set#include? and Java HashSet#contains are both O(1) hash lookups.
 * The Java model faithfully represents the algorithmic complexity.
 */
public class SinatraTest {

    static void bench(String label, Runnable slow, Runnable fast, long sOps, long fOps) {
        // warmup
        slow.run();
        fast.run();
        long t0 = System.nanoTime();
        slow.run();
        long sMs = (System.nanoTime() - t0) / 1_000_000;
        long t1 = System.nanoTime();
        fast.run();
        long fMs = (System.nanoTime() - t1) / 1_000_000;
        double r = fMs > 0 ? (double) sMs / fMs : 0;
        System.out.printf("  %-52s  slow:%4dms (%,d ops)  fast:%4dms (%,d ops)  speedup:%.1fx%n",
                label, sMs, sOps, fMs, fOps, r);
    }

    // -----------------------------------------------------------------------
    // sinatra-0001: content_type add_charset scan
    // Simulates: settings.add_charset.all? { |p| !(p === mime_type) }
    // add_charset contains both String exact-match entries and Regexp patterns.
    // -----------------------------------------------------------------------

    // Models Ruby's === for mixed String/Regexp array
    static boolean addCharsetMatchSlow(List<Object> addCharset, String mimeType) {
        for (Object p : addCharset) {
            if (p instanceof String && ((String) p).equals(mimeType)) return true;
            if (p instanceof Pattern && ((Pattern) p).matcher(mimeType).find()) return true;
        }
        return false;
    }

    // Fast: check Set<String> first (O(1)), then Regexp array only on miss
    static boolean addCharsetMatchFast(Set<String> strings, List<Pattern> patterns, String mimeType) {
        if (strings.contains(mimeType)) return true;
        for (Pattern p : patterns) {
            if (p.matcher(mimeType).find()) return true;
        }
        return false;
    }

    // -----------------------------------------------------------------------
    // sinatra-0002: provides condition — types.include? per request
    // -----------------------------------------------------------------------

    static boolean providesCheckSlow(List<String> types, String responseContentType) {
        // Ruby: types.include?(response_content_type) || types.include?(base_ct)
        if (types.contains(responseContentType)) return true;
        int semi = responseContentType.indexOf(';');
        String base = semi >= 0 ? responseContentType.substring(0, semi) : responseContentType;
        return types.contains(base);
    }

    static boolean providesCheckFast(Set<String> typesSet, String responseContentType) {
        // Built once at route-definition time
        if (typesSet.contains(responseContentType)) return true;
        int semi = responseContentType.indexOf(';');
        String base = semi >= 0 ? responseContentType.substring(0, semi) : responseContentType;
        return typesSet.contains(base);
    }

    public static void main(String[] args) {
        System.out.println("Sinatra CWE-407 Benchmarks");
        System.out.println("==========================");
        System.out.println();

        // --- sinatra-0001: add_charset scan ---
        // Default Sinatra add_charset: ["application/javascript", "application/xml",
        //   "application/xhtml+xml", "application/json"] + /^text\//
        // We extend it to demonstrate scaling with larger add_charset arrays.

        // Default (k=5)
        List<Object> addCharsetDefault = Arrays.asList(
            "application/javascript", "application/xml",
            "application/xhtml+xml", "application/json",
            Pattern.compile("^text/")
        );
        Set<String> defaultStrings = new HashSet<>(Arrays.asList(
            "application/javascript", "application/xml",
            "application/xhtml+xml", "application/json"
        ));
        List<Pattern> defaultPatterns = List.of(Pattern.compile("^text/"));

        int REQUESTS_0001 = 500_000;
        String testMime = "text/html; charset=utf-8";

        System.out.println("sinatra-0001: content_type add_charset check (" + REQUESTS_0001 + " requests)");
        bench(
            "k=5  (default): Array.all? vs Set+Regexp split",
            () -> { for (int i = 0; i < REQUESTS_0001; i++) addCharsetMatchSlow(addCharsetDefault, testMime); },
            () -> { for (int i = 0; i < REQUESTS_0001; i++) addCharsetMatchFast(defaultStrings, defaultPatterns, testMime); },
            REQUESTS_0001, REQUESTS_0001
        );

        // Extended add_charset (k=50) — user-extended array
        List<Object> addCharsetLarge = new ArrayList<>(addCharsetDefault);
        Set<String> largeStrings = new HashSet<>(defaultStrings);
        for (int i = 0; i < 45; i++) {
            String s = "application/custom-type-" + i;
            addCharsetLarge.add(s);
            largeStrings.add(s);
        }
        bench(
            "k=50 (extended): Array.all? vs Set+Regexp split",
            () -> { for (int i = 0; i < REQUESTS_0001; i++) addCharsetMatchSlow(addCharsetLarge, testMime); },
            () -> { for (int i = 0; i < REQUESTS_0001; i++) addCharsetMatchFast(largeStrings, defaultPatterns, testMime); },
            REQUESTS_0001, REQUESTS_0001
        );

        // --- sinatra-0002: provides condition ---
        System.out.println();
        System.out.println("sinatra-0002: provides() condition types membership check (" + REQUESTS_0001 + " requests)");

        // Vary number of types in provides(...)
        int[] typeCounts = {2, 5, 10, 20, 50};
        for (int T : typeCounts) {
            List<String> types = new ArrayList<>();
            for (int i = 0; i < T; i++) types.add("application/type-" + i);
            Set<String> typesSet = new HashSet<>(types);
            // Worst case: content-type matches the last entry
            String ct = "application/type-" + (T - 1);
            bench(
                String.format("T=%2d types: Array#include? vs Set#include?", T),
                () -> { for (int i = 0; i < REQUESTS_0001; i++) providesCheckSlow(types, ct); },
                () -> { for (int i = 0; i < REQUESTS_0001; i++) providesCheckFast(typesSet, ct); },
                REQUESTS_0001, REQUESTS_0001
            );
        }

        System.out.println();
        System.out.println("Done.");
    }
}