Review request: JDK-8021245 (process) file leak in jdk/src/windows/native/java/lang/ProcessImpl_md.c

[Alexey Utkin]

Thanks, Alan

I did "if" reformatting and small refactoring for the code in function 
in accordance with
     https://jbs.oracle.com/bugs/browse/JDK-5008166
recommendations. I am sure that the presence of path separator has no 
dependence from search direction.

Here is the webrev:
http://cr.openjdk.java.net/~uta/openjdk-webrevs/JDK-8021245/webrev.01/

The fix was tested in the debugger.

Regards,
-uta

On 7/28/2013 9:40 PM, Alan Bateman wrote:
> On 26/07/2013 08:33, Alexey Utkin wrote:
>> Bug description:
>>     https://jbs.oracle.com/bugs/browse/JDK-8021245
>>     http://bugs.sun.com/view_bug.do?bug_id=8021245
>>
>> Here is the suggested fix:
>> http://cr.openjdk.java.net/~uta/openjdk-webrevs/JDK-8021245/webrev.00/
>>
>> Summary:
>> In accordance with MS spec [-1] is THE ONLY error signal from the 
>> [_open] function call.
>> http://msdn.microsoft.com/en-us/library/z0kc8e3z.aspx
>> Any negative, except [-1], and zero are the valid handles that need 
>> to be closed.
>>
>> The fix was tested in the debugger my manual switching to Win9x/WinMe 
>> brunch of execution.
> Checking if _open returns -1 and checking the return from _read and 
> _lseek/_read looks okay to me. My only comment is that the "if (" on 
> its own looks a bit odd, I just I would at least hoist the first part 
> of the expression onto the same line.
>
> -Alan.
>