Copied from JDK-8182122:
Ioi Lam added a comment - Yesterday
I think this test fails a lot because it happens to be run soon after another test that causes a library split (see my analysis in CODETOOLS-7901986). So instead of quarantining it, maybe it's better to fix it by:
diff -r d1f8431c6d34 test/compiler/jsr292/PollutedTrapCounts.java
--- a/test/compiler/jsr292/PollutedTrapCounts.java Tue Aug 08 22:58:44 2017 +0200
+++ b/test/compiler/jsr292/PollutedTrapCounts.java Thu Aug 10 19:31:09 2017 -0700
@@ -26,7 +26,8 @@
* @bug 8074551
* @modules java.base/jdk.internal.misc
* @library /test/lib
- *
+ * @build jdk.test.lib.*
+ * @build jdk.test.lib.process.*
* @run driver compiler.jsr292.PollutedTrapCounts
*/
This seems generally safe (and will not make the situation worse) because the test is very simple and it's use of '@library /test/lib' is straightforward.