[langtools] RFR (L): 8010319: Implementation of JEP 181: Nest-Based Access Control
Maurizio Cimadamore
maurizio.cimadamore at oracle.com
Tue May 15 08:35:31 UTC 2018
On 15/05/18 09:13, Remi Forax wrote:
> I agree with this strategy. Given that we are aligning the VM class semantics to the Java the language semantics, i do not think people will still want accessors, perhaps in a nightmare, 'i still want accessors otherwise the stacktraces are too easy to read', in 5 when ldc class was intoduced, nobody ask to keep the class accessor code.
To be clear, I wasn't thinking of a flag to be used by the general
public, but one used by the build system - in the past we had issues
e.g. with indified string concat causing bootstrap issues in the JDK
build, so we had to resort to such 'escape hatches' to disable
generation of certain code shape (esp. while compiling java.base). But I
doubt we will find such surprises at this late stage, given the amount
of effort David put into testing all this.
Maurizio
More information about the compiler-dev
mailing list