Approval request for 8004316

[Joe Darcy]

On 1/15/2013 3:13 AM, Artem Ananiev wrote:
>
> On 1/8/2013 10:56 PM, Seán Coffey wrote:
>> Phil,
>>
>> Yes - people "should" cc relevant parties when such backports are taking
>> place. Not mandatory though. Rule 5 in code review guidelines :
>> http://openjdk.java.net/projects/jdk7u/codereview.html
>>
>> It's probably a good time to remind OpenJDK 7u contributors to carry out
>> such checks where applicable.
>>
>>> there's basically no justification of the need for a backport
>>> and I heartily disapprove of backporting  8004316
>> I can't understand why you're against such a backport. It looks like
>> printing functionality is broken on some OSes without this fix. Given
>> that Jayashree backported this fix, I hope she can take responsibility
>> for any potential regressions that may be encountered.
>
> I completely agree with Phil here.
>
> JDK8 is not the same as JDK7u, sometimes reviewers approve risky fixes 
> for 8, in assumption there will be enough time to resolve all the 
> regressions without breaking anyone's applications. We can't afford 
> doing the same for 7u, which is already used in many production systems.

And those people using 7u in production often don't want to wait until 
they can adopt JDK 8, say, a year after it ships, to have a fix for a 
problem.

As always, there are tradeoffs between stability and progress.  In my 
estimation, the tradeoffs that are being made in the 7 update train in 
that regard have been the right ones.

-Joe