[aarch64-port-dev ] Correct the placement of release calls in JDK8 JavaFrameAnchor code

[Andrew Dinn]

On 25/11/14 14:33, Andrew Haley wrote:
> On 11/25/2014 11:29 AM, Andrew Dinn wrote:
>> Is it ok to push this change? or do I need to wait for it to go into
>> JDK9 first?
> 
> If it goes into 8 it should also go into 9.

I'm not suggesting otherwise. However, you commented when I posted the
JDK7 commit that JDK9 was in flux at present and hence it was not a good
idea to push to that repo. I posted this patch

  i) to record that this is what is needed in both JDK8/9

and

  ii) to see if you wanted the JDK8 fix pushed early

I'm happy to wait if need be.

>> n.b. I have checked the latest JDK9 code and believe the same patch will
>> apply correctly to JDK9 and be valid for the same reasons.
> 
> The reference to zap() is very confusing; I had to study the code to
> see what it referred to.  I take it that we're only ever called from
> ~JavaCallWrapper(), so we know that _last_Java_sp must be null.  Is
> that what you mean?

Yes, that is what I mean. Do you want me to upgrade the level of detail
in the comment?

regards,


Andrew Dinn
-----------