> On Apr 10, 2018, at 3:21 AM, Aleksey Shipilev <sh...@redhat.com> wrote: > > On 04/10/2018 08:42 AM, Alan Bateman wrote: >> On 09/04/2018 18:58, Paul Sandoz wrote: >>> I am supportive of this change (the risk to impacting order-dependent >>> stateful MH filter code is >>> smaller than the risk of hitting a string concatenation bug). (We erred on >>> the side of this being >>> a bug and not being a spec change given the pseudo-code in the Java doc.) >>> >>> IIRC this will require a nod of approval from the project lead. >>> >>> Paul. >> I think this is a CSR for Java SE 11 first as it adds a testable assertion >> to the spec and also >> changes existing behavior. >
Note that the testable assertion was already in the spec, just buried in the normative pseudocode. So this is really about clarification from the spec perspective. And of course any bug fix can change behavior. In hindsight i would be more inclined to do a CSR if a bug fix is required to conform to specification. > Ok, fine! Paul, can you submit the CSR for this? > Here we go: https://bugs.openjdk.java.net/browse/JDK-8201371 <https://bugs.openjdk.java.net/browse/JDK-8201371> Kindly requesting a reviewer. Thanks, Paul.