On Fri, Nov 18, 2016 at 8:33 AM, joe darcy <joe.da...@oracle.com> wrote:
> * A master forest, serving the roles master and dev play today in 9. > > With a few exceptions, in JDK 9 master was just time-delayed copy of dev > so we can implement recording the information about which set of sources > correspond to a promoted build without using a whole other forest. > > Rather than using a separate line of development for client-libs work as > in 9, I think this should be done in the same line of development as all > other libs work in 10. > For many years, I've been advocating having a guaranteed always-working, never regressing master and also always a place for developers to submit-and-forget their (possibly slightly buggy) changes. All regressions that could be caught by a test are 100% guaranteed to be caught by a competent trusted release engineer who is the only one ever moving changes into the master forest. Based on this idea, it seems essential to have something like a jdk10-dev forest (it could also be implemented using mercurial branches, but that would be a break with many decades of tradition). I notice today the message http://mail.openjdk.java.net/pipermail/quality-discuss/2016-November/000596.html where regressions have crept into a jdk9 build, which is disappointing. The whole point of regression testing is to ensure that regressions don't happen! And I recall having that job myself back in 2005!
