JSR 383 & the newer version-string scheme
volker.simonis at gmail.com
Tue Nov 14 08:11:02 UTC 2017
thanks a lot for the clarification.
On Tue, Nov 14, 2017 at 8:51 AM, Brian Goetz <brian.goetz at oracle.com> wrote:
> We’d certainly welcome any review of these features. But, please be careful
> drawing assumptions from the “fix version” on sub-bugs of a JEP — this is
> not a statement of planning, as much as a statement of optimism on the part
> of an individual developer.
Well if it is neither the information in the JEP/JBS nor the
statements of the involved developers which is "relevant" for
planning, it is indeed impossible for an external contributor (or EG
member) to guess what Oracle plans to do within the OpenJDK/Java. So
maybe you could adjust the corresponding information in the JEP/JBS
accordingly once you know for sure you "don't plan" to do something ?
> I can state definitively that Oracle has no plans to propose JEP 309 (which
> subsumes most of the individual issues you list) for 10. It is simply too
> late for a change like this — which affects the class file format and
> therefore is high risk.
That's totally fine for me. But as I've already expressed before, it
is really hard for an external observer to find out what will be in
the next release and what will be postponed. An open planning process
for OpenJDK would be really helpful here although that's probably not
a topic for this group.
> This is what “missing the train” looks like; while the development window
> has not closed for 10, some changes are too big or risky to take in this
> close to the departure time, so while it looks like the train is still
> waiting on the platform, this feature has in fact missed it. But that’s OK.
> There’s another train coming soon.
> On Nov 13, 2017, at 5:39 PM, Volker Simonis <volker.simonis at gmail.com>
> Another interesting topic which is currently under review and
> according to the developers is planned for jdk 10 (although it has not
> yet been targeted):
> JEP 309: Dynamic Class-File Constants
> Updates to the Java Virtual Machine Specification
> Minimal ConstantDynamic support
> Tool support for ConstantDynamic
> Minimal set of bootstrap methods for dynamic constants
> Note that this change requires porting effort for the non-Oracle
> platforms and independent implementations and also for tools because
> it extends the class file format.
> On Thu, Nov 2, 2017 at 5:51 PM, Brian Goetz <brian.goetz at oracle.com> wrote:
> As a heads-up, there are discussions in OpenJDK regarding the proposed
> version scheme. A proposal has been made; there will be a JEP, but you may
> be interested in following the discussion before that.
> Latest proposal:
More information about the java-se-spec-experts