RFR 8030633: nsk/jvmti/RedefineClasses/StressRedefine failed invalid method ordering, length on Solaris
dmitry.samersoff at oracle.com
Thu Dec 19 11:42:27 PST 2013
Looks good for me.
Few cosmetic nits to mention:
It might be more readable if we replace
klass->methods()->length() to original_methods->length();
It might be good to call
sort(new_methods) only if (new_methods->length() > 0)
On 2013-12-19 22:59, Coleen Phillimore wrote:
> On 12/19/2013 12:42 PM, Karen Kinnear wrote:
>> Code looks good.
>> In particular I was checking we handle both the null and
>> empty_int_array cases - I assume
>> that we need to handle both due to bootstrapping issues.
> The original method array is the_empty_int_array() and not null but I
> added null checks in it because Dan had the same comment too.
>> On Dec 19, 2013, at 11:40 AM, Coleen Phillimore wrote:
>>> Summary: A method with no declared methods was getting an AME
>>> overpass method with the latest change. The method_ordering array
>>> was not updated for the new methods.
>>> open webrev at http://cr.openjdk.java.net/~coleenp/8030633/
>>> bug link https://bugs.openjdk.java.net/browse/JDK-8030633
>>> Tested with existing tests
>>> Needs native code to create new test in jtreg, so I didn't add one.
Oracle Java development team, Saint Petersburg, Russia
* I would love to change the world, but they won't give me the sources.
More information about the hotspot-runtime-dev