RFR 8198899: Compilation errors in java.prefs with VS 2017https://bugs.openjdk.java.net/browse/JDK-8198899 http://javaweb.us.oracle.com/~bburkhal/jdk/review/8198899/webrev.00 This code currently represents a 64-bit Windows HKEY as an int at the Java level. The proposed change is to represent the HKEY as a long. All open/test/jdk/:jdk_core, closed/test/jdk/:jdk_core passed in Mach 5 on Windows-x64. Thanks, Brian

[Paul Sandoz]

Looks like a fairly routine translation. Did you consider keeping this contained within the JNI code if that is at all possible? I don’t know whether that would just push the compiler error around e.g. converting from jint to jlong then to HKEY?

 747         int maxValueNameLength = (int)result[MAX_VALUE_NAME_LENGTH];
 748         int valuesNumber = (int)result[VALUES_NUMBER];

Might wanna comment that the lossy conversion is safe?

Paul.

> On Mar 5, 2018, at 10:29 AM, Brian Burkhalter <brian.burkhalter at oracle.com> wrote:
> 
> https://bugs.openjdk.java.net/browse/JDK-8198899
> http://cr.openjdk.java.net/~bpb/8198899/webrev.00/
> 
> This code currently represents a 64-bit Windows HKEY as an int at the Java level. The proposed change is to represent the HKEY as a long. All existing regression tests passed on Windows-x64.
> 
> Thanks,
> 
> Brian