New project proposal: even faster JDK releases
org.openjdk at io7m.com
Sun Apr 1 12:32:05 UTC 2018
On 2018-04-01T12:04:51 +0000
Mario Torre <neugens.limasoftware at gmail.com> wrote:
> In order to better help the new release cycle, which of course is not a
> cycle but a continuous stream of random bits, we also propose the complete
> abolition of testing and reviews, instead developers should just commit
> whatever they feel like whenever they feel so, possibly on Friday night.
In addition to this, I'm currently working on a system that will allow
developers to get code into releases *before* they've actually committed
it. This, does, however, currently depend on the availability of
sun.misc.Unsafe  and will therefore require the introduction of new
APIs via JEP proposals for future inclusion.
To coincide with the new even-faster-release cycle, I suggest that we
adopt a new JEP release cycle: The last person to stop speaking about
a new JEP feature on any given mailing list is declared the winner and
their implementation is immediately committed as-is .
 See the following whitepaper for details:
 An alternate process was considered involving competitive
shin-kicking . Unfortunately, during testing, the last
person standing was rendered unable to accurately describe
the enhancement in question (or indeed, to speak anything
resembling human language at all) and the idea was dropped in
favour of arguments on mailing lists.
Mark Raynsford | http://www.io7m.com
More information about the discuss