RFR (XS) 8027458: VM anonymous classes: wrong context for protected access checks
harold.seigel at oracle.com
Wed Dec 4 04:59:27 PST 2013
Thanks for the review. Perhaps the change that you suggest could be
done if and when the verification checks are added.
On 12/3/2013 4:29 PM, Lois Foltan wrote:
> Hi Harold,
> Overall this looks good. I do have one concern that is indirectly
> related to this fix. I would almost like to see the code to search
> up the nested anonymous classes pulled out to a separate inline
> method. There now exists a couple instances of recently added
> verification check that detects an anonymous class and then moves
> forward with the anonymous class's host class. Should instead
> verification always work its way up through the chain of anonymous
> nested classes? If yes, then pulling the code to walk up the chain
> would be beneficial to all.
> On 12/2/2013 4:42 PM, harold seigel wrote:
>> Please review this small fix for bug 8027458. It is based on John
>> Rose's suggested fix that he attached to the bug report. For
>> protected access, the fix checks if the current class is an anonymous
>> class. If so, it finds the current class's host class and uses the
>> host class for checking access to a protected method.
>> webrev: http://cr.openjdk.java.net/~hseigel/bug_8027458/
>> bug: https://bugs.openjdk.java.net/browse/JDK-8027458
>> The fix was tested with JCK Lang and VM tests, the failing JCK tests
>> listed in the bug report, jtreg java/lang, java/util, lambda, and
>> runtime tests, ute quicktests and vm.mlvm.tests, and defmeth tests.
>> Thanks! Harold
More information about the hotspot-runtime-dev