It looks like the memory bump in [1, 2] was not enough. This patch [3] would add 1GB to the required Java heap size.
Thanks, Brian [1] https://bugs.openjdk.java.net/browse/JDK-8221568 [2] http://mail.openjdk.java.net/pipermail/core-libs-dev/2019-March/059355.html [3] patch --- a/test/jdk/java/io/DataOutputStream/WriteUTF.java +++ b/test/jdk/java/io/DataOutputStream/WriteUTF.java @@ -22,10 +22,10 @@ */ /* @test - * @bug 4260284 8219196 + * @bug 4260284 8219196 8223254 * @summary Test if DataOutputStream will overcount written field. - * @requires (sun.arch.data.model == "64" & os.maxMemory >= 3g) - * @run testng/othervm -Xmx3g WriteUTF + * @requires (sun.arch.data.model == "64" & os.maxMemory >= 4g) + * @run testng/othervm -Xmx4g WriteUTF */ import java.io.ByteArrayOutputStream; @@ -60,7 +60,7 @@ } // Without 8219196 fix, throws ArrayIndexOutOfBoundsException instead of - // expected UTFDataFormatException. Requires 3GB of heap (-Xmx3g) to run + // expected UTFDataFormatException. Requires 4GB of heap (-Xmx4g) to run // without throwing an OutOfMemoryError. @Test(expectedExceptions = UTFDataFormatException.class) public void arrayIndexOutOfBoundsException() throws IOException {
