Patch for JDK-8142882: rebinding of a DirectMethodHandle

[Andrew Hughes]

----- Original Message -----
> Proposed fix for: https://bugs.openjdk.java.net/browse/JDK-8142882
> 

FWIW, the bug isn't visible.

> http://cr.openjdk.java.net/~rasbold/8142882/webrev.00/
> 
> Don't pre-dispatch the receiver of a DirectMethodHandle, as it may allow a
> protected method to be accessed.
> 
> This fix is a subset of the change for JDK-8050166
> <https://bugs.openjdk.java.net/browse/JDK-8050166>, which is already in
> jdk8 and jdk9.
> 

This looks good to me too. I've pushed the change with John and I listed
as reviewers. I've also tagged the result as jdk7u91-b02 and uploaded
a tarball release for this version:

4cb174c87289a020f2280570560240e38fc18529f951c72f9512a1ebc6962672  openjdk7u91-b02.tar.xz
https://java.net/projects/openjdk7/downloads/download/openjdk7u91-b02.tar.xz

> The change causes the an NPE to be thrown when a protected MethodHandle is
> called via invoke().
> An additional backport of JDK-8016839
> <https://bugs.openjdk.java.net/browse/JDK-8016839> is needed, if it is
> desired to that a proper IllegalAccessError is thrown instead of an NPE, as
> happens with jdk8 / jdk9.
> 

This potentially seems like a good idea, but I'm wary of it causing compatibility issues.
I'll look into a backport in time for the next release in January (u95?)

Thanks,
-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222

PGP Key: rsa4096/248BDC07 (hkp://keys.gnupg.net)
Fingerprint = EC5A 1F5E C0AD 1D15 8F1F  8F91 3B96 A578 248B DC07