RFR(XS): 8194741: CDS broken due to 8194741

[coleen.phillimore at oracle.com]

Yes, this looks good.   I think it should be checked in immediately with 
the trivial rule and the fix tier2 quickly rule.
thanks,
Coleen


On 1/15/18 9:25 AM, Erik Österlund wrote:
> Hi,
>
> I accidentally broke CDS with 8194741, which affects a bunch of tier 2 
> CDS tests.
> The problem is trivial: I missed a branch condition where the resolved 
> references object is null that should have gone to the else bit, but 
> did not after my change.
> Sorry for breaking this - I seem to have mixed up my test run results 
> from tier 1-5, and after that only ran an extra tier 1 sanity check 
> before checking in which did not catch this.
>
> I have run the failing tests locally to verify they work now (and they 
> do), and am running tier 1-2 again before meanwhile.
>
> Webrev:
> http://cr.openjdk.java.net/~eosterlund/8195107/webrev.00/
>
> Bug:
> https://bugs.openjdk.java.net/browse/JDK-8195107
>
> Here is the original webrev of 8194741 where I messed up the condition 
> in constantPool.cpp:330:
> http://cr.openjdk.java.net/~eosterlund/8194741/webrev.00/
>
> Thanks,
> /Erik