[9] RFR of 8042377: BufferedWriter and FilteredOutputStream.close throw IAE if flush and close throw equal exceptions
Chris Hegarty
chris.hegarty at oracle.com
Wed Jun 24 20:54:35 UTC 2015
> On 24 Jun 2015, at 21:28, Brian Burkhalter <brian.burkhalter at oracle.com> wrote:
>
> I should however note that equivalent usage is already present in the close() methods of FileInputStream, FileOutputStream, and RandomAccessFile which are also in java.io. Perhaps the problem you allude to would be more pertinent to java.lang classes?
I remember the initialization problem with java.lang.Thread, but I think the usage here is ok. As you correctly point out, AtomicBoolean is already used in other java.io classes. If HelloWorld starts up then I think you’re good.
The changes themselves look ok to me. Maybe the test could exercise FilterOutputStream directly too?
-Chris
> Brian
>
> On Jun 24, 2015, at 1:21 PM, Brian Burkhalter <brian.burkhalter at oracle.com> wrote:
>
>> Jason,
>>
>> Thanks for pointing this out. I am not sure of the answer. Let’s see what the experts have to say.
>>
>> Thanks,
>>
>> Brian
>>
>> On Jun 24, 2015, at 1:03 PM, Jason Mehrens <jason_mehrens at hotmail.com> wrote:
>>
>>> Not sure on this but, isn't it a little risky to import AtomicBoolean into such low level class? I vaguely remember there was an issue with using AtomicXXX in java.lang.Thread. Not sure if this case suffers the same fate.
>
More information about the core-libs-dev
mailing list