On 8 sep 2014, at 22:03, serguei.spit...@oracle.com wrote: > On 9/8/14 12:59 PM, Staffan Larsen wrote: >> >> On 8 sep 2014, at 21:26, serguei.spit...@oracle.com wrote: >> >>> This looks good to me. >> >> Thanks. >> >>> >>> A minor question: >>> >>> src/jdk.attach/windows/classes/sun/tools/attach/VirtualMachineImpl.java >>> 110 is.close(); >>> >>> An IOException can be thrown in the readErrorMessage(). >>> Would it make sense to use a finally statement to close the stream? >> In most cases the stream is returned from the method so we can’t always >> close it. > > In this particular case the stream is not returned from the method call: > 103 PipedInputStream is = new PipedInputStream(hPipe);
123 // return the input stream 124 return is; /Staffan > > Thanks, > Serguei > >> >> /Staffan >> >> >>> >>> Thanks, >>> Serguei >>> >>> On 9/8/14 5:25 AM, Staffan Larsen wrote: >>>> Two fixes: >>>> - The PipedInputStream used by the attach code on windows is not closed in >>>> case of errors >>>> - The InputStreams returned by VirtualMachine.execute are not closed by >>>> all callers >>>> >>>> webrev: http://cr.openjdk.java.net/~sla/8057558/webrev.00/ >>>> bug: https://bugs.openjdk.java.net/browse/JDK-8057558 >>>> >>>> Thanks, >>>> /Staffan >>> >> >
