<AWT Dev> <Awt Dev> [9] Review Request for 8143295: Validating issue in AWT

[Semyon Sadetsky]

On 2/9/2016 7:36 PM, Sergey Bylokhov wrote:
> Hi, Semyon.
> It seems that the new block "synchronized (getStateLock()) {" can be 
> useful only if the getBounds() will use the same synchronization 
> pattern. Otherwise it guarantee nothing.
Yes. It is better to remove it.
Please see the updated webrev: 
http://cr.openjdk.java.net/~ssadetsky/8143295/webrev.01/
>
> On 02.12.15 19:24, Semyon Sadetsky wrote:
>> Hello,
>>
>> Please review fix for JDK9:
>>
>> bug: https://bugs.openjdk.java.net/browse/JDK-8143295
>> webrev: http://cr.openjdk.java.net/~ssadetsky/8143295/webrev.00/
>>
>> In the 8011616 window bounds that are sent from WM were propagated to
>> the component for XWindow and its descendants. But in case of AWT
>> components which are managed by a container layout, this may cause an
>> issue because their bounds may be concurrently modified by the container
>> layout in another thread.
>> For the components other than windows it is not necessary to propagate
>> bounds sent by WM because those component bounds should be controlled by
>> the container layout only. So the solution is to apply 8011616 fix on
>> XWindowPeer level and restore the original logic of the ConfigureNotify
>> event handler of the XWindow class.
>>
>> --Semyon
>
>