RFR 9/8: 8066504: GetVersionEx in java.base/windows/native/libjava/java_props_md.c might not get correct Windows version

Alan Bateman Alan.Bateman at oracle.com
Wed Jun 17 13:58:10 UTC 2015

On 17/06/2015 14:49, Roger Riggs wrote:
> Hi Alan,
> Would there be any concerns about backporting to JDK 8?
I don't think so but it will require the jdk8u maintainers to approve.
> The new helper APIs 
> <https://msdn.microsoft.com/en-us/library/windows/desktop/ms724451%28v=vs.85%29.aspx> 
> only verify that the OS is at least the version requested.
> It does not reveal what version the OS is actually running on.
> Adding a manifest attribute just declares the minimum version required
> and is reported by GetVersionEx.
> In the current case, the code uses GetVersionEx as a fallback in
> case the version on kernel32.dll is not available.
> That fallback could be removed since kernel32.dll must exist on all 
> supported versions.
Okay, I thought the helper APIs had better support than this. Maybe the 
Microsoft folks that are contributing to OpenJDK might have suggestions. 
In general I don't have any objections to the approach, just thought it 
was a big odd that the OS forces us to look at the version of kernel32.dll.

> :
>> Also, is there any update needed in src/windows/resource/java.manifest?
> If using the version of kernel32.dll then the manifest version does 
> not affect the value of os.name.
> But yes, I added the Windows 10 supportedOS uid.
I don't think this was in the first webrev that I saw but I see it is 
there now, looks fine.

>> In passing, GetNativeSystemInfo has been in Windows since Windows XP 
>> so I don't think we need to use GetModuleHandle to get the address now.
> ok.
> Similarly, can the case for Windows 3.1 be removed?   (its not in 
> theJDK 8 supported matrix 
> <http://www.oracle.com/technetwork/java/javase/certconfig-2095354.html>)
I think that would be a fine cleanup.


More information about the build-dev mailing list