On Fri, 4 Jun 2021 14:08:58 GMT, Sean Mullan <[email protected]> wrote:
>> Hi,
>>
>> Please review my fix for JDK-8255148 which clarifies when an exception
>> contains debug information only.
>>
>> Regards,
>> Evan
>
> src/java.base/share/classes/sun/security/ssl/SSLSocketImpl.java line 590:
>
>> 588: // ignore the exception
>> 589: if (SSLLogger.isOn && SSLLogger.isOn("ssl")) {
>> 590: SSLLogger.warning("SSLSocket duplex close failed. Debug
>> info only. Exception details:", ioe);
>
> If this is a debug message, shouldn't we just use `SSLLogger.fine()` instead
> of `SSLLogger.warning()`, with the same message "SSLSocket duplex close
> failed"? @coffeys what do you think?
@seanjmullan - It's the exception stacktrace printed after the message that's
causing issue for some. Some are reading it as a JDK functionality issue (when
it's not)
I've no strong preference whether this should be printed at warning() or fine()
level. That granularity is largely broken since JDK 11 - people need to use
-Djavax.net.debug=all to get any reasonable amount of data. It would be good to
see https://bugs.openjdk.java.net/browse/JDK-8044609 worked on.
I think the patch looks ok as ie. Evan has fixed up some issues with the test -
we should expect a new version shortly.
-------------
PR: https://git.openjdk.java.net/jdk/pull/4354