Code review request: 7082769: FileInputStream/FileOutputStream/RandomAccessFile allow file descriptor be closed when still in use
sean.coffey at oracle.com
Mon Sep 12 22:09:38 UTC 2011
Cleaned up testcase (as per suggestions) is in latest webrev :
I've kept the try-with-resouces approach out of the TestMultipleFD
method. It just complicates matters and with the closing order being
important, it's easier to read with old style.
On 09/09/2011 14:08, Seán Coffey wrote:
> Good points Alan.
> Coding styles probably differ from merging two testcases together.
> I'll clean up on the issues mentioned and send out the new webrev
> I thought about try with resources in a few places but it didn't
> always suit. Take for example the TestMultipleFD() method. The
> ordering of close call is important. I can get around that knowing
> that the close calls are made in opposite order to the streams/file's
> creation. However, the creation of the streams is also important since
> I take the file descriptor from the random access file (normally) - I
> might have to intermix try with resources and some try/finally blocks.
> On 09/09/11 12:25, Alan Bateman wrote:
>> Seán Coffey wrote:
>>> webrev :
>>> Bug fix where we ensure that the fd object is not disposed of until
>>> all streams are closed out.
>>> Testcase is a bulked up version of CR 6322678 (which wasn't
>>> committed at time of 6322678 fix). It includes create/close() calls
>>> for FileInputStream/FileOutputStream/RandomAccessFile which all
>>> reference the same file descriptor. Multi threaded access to the
>>> same file descriptor is also tested.
>>> Typo fix also as per http://bugs.sun.com/view_bug.do?bug_id=7087019
>>> also included.
>> I think the change are okay. There are other issues with sharing file
>> descriptors between streams but your changes are a good improvement
>> and eliminate the messy runningFinalizer code (which I think someone
>> added to address a regression from a previous fix to an address
>> another issue with sharing file descriptors).
>> The coverage in the new test looks good but I think the code could be
>> a bit cleaner. For example in TestFinalizer then it uses
>> try-with-resources at L64 but not at L55. It uses /**/ style comments
>> at L63 but // style at L76. Temporary file usage is also
>> inconsistent, Test_ is created in the system temporary directory,
>> Test6322678 in the working directory. Also the temporary file name is
>> duplicated at L97. I think TestMultipleFD would be a lot cleaner with
>> try-with-resources too. I also suspect the deletes at L138 and L156
>> may be failing on Windows. If you have time to do a bit of clean-up
>> in the test then I think you are good to go.
More information about the core-libs-dev