558
|
1 |
// A case of catastrophic backtracking in Java 8
|
|
2 |
//-----------------------------------------------
|
420
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
3 |
//
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
4 |
// regexp: (a*)*b
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
5 |
// strings: aa....
|
558
|
6 |
//
|
|
7 |
// compile: javac catastrophic.java
|
|
8 |
// call with: java catastrophic
|
|
9 |
//
|
|
10 |
// IMPORTANT:
|
|
11 |
// Java 9 improved its regex matching engine.
|
|
12 |
// This example is now much faster.
|
|
13 |
//
|
420
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
14 |
|
411
|
15 |
import java.util.regex.*;
|
|
16 |
|
|
17 |
public class catastrophic {
|
|
18 |
public static void main(String[] args) {
|
616
|
19 |
|
474
|
20 |
//we always run all the tests twice -> to warmup of the JVM
|
411
|
21 |
for (int runs = 0; runs < 2; runs++) {
|
420
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
22 |
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
23 |
Pattern pattern = Pattern.compile("(a*)*b");
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
24 |
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
25 |
// Run from 5 to 28 characters
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
26 |
for (int length = 5; length < 28; length++) {
|
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
27 |
|
411
|
28 |
// Build input of specified length
|
|
29 |
String input = "";
|
412
|
30 |
for (int i = 0; i < length; i++) { input += "a"; }
|
411
|
31 |
|
|
32 |
// Measure the average duration of two calls...
|
|
33 |
long start = System.nanoTime();
|
|
34 |
for (int i = 0; i < 2; i++) {
|
|
35 |
pattern.matcher(input).find();
|
|
36 |
}
|
474
|
37 |
|
|
38 |
// Print out time
|
412
|
39 |
System.out.println(length + " " + input + ": "
|
|
40 |
+ ((System.nanoTime() - start) / 2000000000d)
|
|
41 |
+ "s");
|
411
|
42 |
}
|
|
43 |
}
|
|
44 |
}
|
420
Christian Urban <christian dot urban at kcl dot ac dot uk>
diff
changeset
|
45 |
}
|