On Tue, 2 Nov 2021 01:13:40 GMT, David Holmes <david.hol...@oracle.com> wrote:
> That certainly does make sense, although it makes me wonder how this ever > passed in the first place. Maybe there's a bug somewhere in the JDB output > processing code that made the failure intermittent. It depends a bit on how the underlying I/O bundles up the reply packets, and how quickly they are read, but probably more the latter than the former. Basically the jdb process is sending a bunch of replies, and the test process usually doesn't start reading the replies until after they have all been sent, so it gets them all with one read. But if the test process is quick enough, then it might do the read before all the replies have been sent. ------------- PR: https://git.openjdk.java.net/jdk/pull/6182
