https://bz.apache.org/bugzilla/show_bug.cgi?id=66195
--- Comment #6 from Behrooz Nobakht <nob...@gmail.com> --- (In reply to Konstantin Kolinko from comment #5) > 1. If you read the above Java 9 javadoc, you may see that > "jdk.util.jar.enableMultiRelease" in that version of Java had 3 possible > values: true, false, or force. > > So, why does your patch use a Boolean.parse() call? That's correct. The patch should fix that to consider all the spec values. Thanks for the catch. > > 2. The same Javadoc says: > > "Furthermore, the method isMultiRelease() returns false." > > Looking at PR, Tomcat calls isMultiRelease() just there, below the changed > lines. In your use case that call should return "false", and no notable > waste of resources should follow. Yes, it could return false but it also involves invoking the method handle for the multi-release method of JarFile. When a system is under load, or when the cache is being populated by Tomcat, there's good chance that the method handle invocation comes up as frequent stack frame. This is what we found during profiling a system at runtime. -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
