RFC: backport of JDK-8215756: Memory leaks in the AWT on macOS

[Simon Tooke]

On 7/31/2019 11:59 AM, Aleksey Shipilev wrote:
> On 7/30/19 9:39 PM, Simon Tooke wrote:
>> On 7/30/2019 12:29 PM, Aleksey Shipilev wrote:
>>> On 7/30/19 6:18 PM, Simon Tooke wrote:
>>>> The patch is trivial, apart from the usual path changes.
>>>> I omitted one file change that was only a line ending difference, and fixed the copyright dates.
>>> Why omit it?
>> Because it's not in my usual workflow (translation: lazy) to verify
>> newline correctness in incoming patches.  It probably should be.
>>
>> I can add that file patch back, but, for example, one of the other file
>> patches doesn't apply cleanly (after fixing paths), and it's due to
>> chunks that only update newlines (in code that's not there in JDK8).
> If the upstream patch updates newlines, backport should follow. If hunk is not applicable because
> there is no code to update in 8u, hunk can be ignored.
>
>> I'm happy to add back the one newline fix, but since the patch needed
>> modification anyways, I limited the original scope.  Just let me know a
>> preference.
> I prefer backported patches to be as close to upstream patches as possible, including whitespace
> differences. This is mostly to cater for the follow-up patches to the same location that expect the
> shape of file to the same. There are some exceptions to this rule, but whitespace diffs is usually
> not the exceptional case.

Fair enough, and persuasive.

Here is my revised patch:

http://cr.openjdk.java.net/~stooke/webrevs/jdk8215756-jdk8u.01/

bug: https://bugs.openjdk.java.net/browse/JDK-8215756
Original patch: http://hg.openjdk.java.net/jdk/client/rev/64e7a73195c1

-Simon

>