By default, a word boundary only considers ASCII letters and digits. See
"Predefined character classes" in the documentation.
To add Unicode support, you have a choice between adding a flag as a 2nd
argument to the compile() method
Pattern p = Pattern.compile("(\\b" + word + "\\b)",
Pattern.UNICODE_CHARACTER_CLASS);
or add a flag in the regex pattern, as documented in "Special constructs
(named-capturing and non-capturing)"
Pattern p = Pattern.compile("(?U)(\\b" + word + "\\b)");
Greetings
Raffaello
On 2023-12-15 20:07, Stefan Norberg wrote:
The following test works in 17 but fails in 19.0.2, and 21.0.1 on the
last assertion. Bug or feature?
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.Test;
import java.util.ArrayList;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
/**
* Tests passes in JDK 17 but fails in JDK 19, 21.
*
* The combination of a \b "word boundary" and a unicode char doesn't
seem to work in 19, 21.
*
*/
public class UnicodeTest {
@Test
public void testRegexp() throws Exception {
var text = "En sak som ökas och sedan minskas. Bra va!";
var word = "ökas";
Assertions.assertTrue(text.contains(word));
Pattern p = Pattern.compile("(\\b" + word + "\\b)");
Matcher m = p.matcher(text);
var matches = new ArrayList<>();
while (m.find()) {
String matchString = m.group();
System.out.println(matchString);
matches.add(matchString);
}
Assertions.assertEquals(1, matches.size());
}
}
openjdk version "21.0.1" 2023-10-17 LTS
OpenJDK Runtime Environment Corretto-21.0.1.12.1 (build 21.0.1+12-LTS)
OpenJDK 64-Bit Server VM Corretto-21.0.1.12.1 (build 21.0.1+12-LTS,
mixed mode, sharing)
System Version: macOS 14.2 (23C64)
Kernel Version: Darwin 23.2.0
Thanks!
/Stefan
