<AWT Dev> Code Review Request for CR 7145980 - Dispose method of window.java takes long

[Anthony Petrov]

Hi Oleg,

The fix looks good. I'd also add the isDisposing() value to the logger 
output statement in isMixingNeeded() (in addition to the numbers of hw 
and lw descendants) to make sure the mixing log provides complete 
information about the reason isMixingNeeded() returns false.

--
best regards,
Anthony

On 2/22/2012 8:15 PM, Oleg Pekhovskiy wrote:
> Hi Anthony,
> 
> thanks for the review, I've prepared the next version of fix according 
> to your proposals:
> http://cr.openjdk.java.net/~bagiras/7145980.2/ 
> <http://cr.openjdk.java.net/%7Ebagiras/7145980.2/>
> 
> Thanks,
> Oleg
> 
> 2/22/2012 5:26 PM, Anthony Petrov wrote:
>> Hi Oleg,
>>
>> The correct link to the webrev should be 
>> http://cr.openjdk.java.net/~bagiras/7145980/webrev/
>>
>> Anyways, I suggest to make isMixingNeeded() return false if the 
>> containing window (a reference to which is already obtained in that 
>> method) is currently disposing. This way we avoid a double call to 
>> getContainingWindow(), and also will prevent any other mixing-related 
>> code from execution when the window is being disposed.
>>
>> In Window.java I suggest to use the following pattern:
>>
>> disposing = true;
>> try {
>>    // do stuff
>> } finally {
>>    disposing = false;
>> }
>>
>> to ensure we don't leave the flag in 'true' state if some exceptions 
>> are thrown.
>>
>> -- 
>> best regards,
>> Anthony
>>
>> On 2/22/2012 4:57 PM, Oleg Pekhovskiy wrote:
>>>   Hi guys,
>>>
>>> Please review the fix for CR:
>>> http://bugs.sun.com/view_bug.do?bug_id=7145980
>>>
>>> webrev:
>>> http://cr.openjdk.java.net/~bagiras/7145980/webrev/ 
>>> <http://cr.openjdk.java.net/%7Edenis/7143070/webrev.04/>
>>>
>>> The main idea is to skip recalculation of component visible region 
>>> when a top-level window is disposing.
>>>
>>> Thanks,
>>> Oleg
>