Lambda vs anon class optimization effect on CHA

[Vitaly Davidovich]

Sorry, forgot to mention that this is on 8u51.

On Wed, Aug 10, 2016 at 9:28 AM, Vitaly Davidovich <vitalyd at gmail.com>
wrote:

> Hi all,
>
> Suppose we have the following classes and code:
>
> public abstract class Foo {
>     public static java.util.Supplier<Foo> null() {
>         return () -> Null.INSTANCE;
>
>     }
>
>    // some abstract methods
> }
>
> public final class Impl extends Foo {}
>
> public final class Null extends Foo {
>       public static final Null INSTANCE = new Null();
>
> }
>
> It seems that the lambda in null(), even when not invoked, causes Null to
> be loaded and single-class CHA optimization is disabled.  If I change
> null() to use an anon class:
>
> public static java.util.Supplier<Foo> null() {
>      return new Supplier<Foo>() {
>           public Foo get() { return Null.INSTANCE; }
>           }
>       };
> }
>
> Then CHA kicks in.  Keep in mind that the code I'm running never invokes
> null() and only instantiates Impl, so the expectation is that Null would
> not load (and disable the single impl CHA optimization).
>
> My understanding was that the lambda in null() would have a bootstrap
> method that invokes the lambda metafactory to actually bind the call to the
> synthetic static method that's generated, which in turn actually references
> Null.INSTANCE (and would cause the class to load).  However, I'm clearly
> missing or misunderstanding something.
>
> Would anyone be able to explain the above situation? Is that expected? Let
> me know if you need more info.
>
> Thanks
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20160810/a63389b7/attachment.html>