<AWT Dev> [9] Review request for 8020443: Frame is not created on the specified GraphicsDevice with two monitors

[Oleg Pekhovskiy]

Hi Alexander,

thank you for the review!

1. The problem is that I didn't find that root window ALWAYS starts with 
0, 0 in X11 docs.
Could you please point such statement out to avoid uncertainty?

2. Checking insets against screen width and height can fail too. Example:
1st screen (0, 0, 1024, 768) - x, y, width, height
2nd screen (1024, 0, 1920, 1080)
Struts for the 2nd screen could be (1074, 0, 20, 0) - left, right, top, 
bottom.
So left strut 1075 fits in 1920 but should be 50 as inset.

PS: Just curious: you said that screen (1000, 50, 1000, 1000) is a valid 
case.
Is it valid for testing environment? As I see screens are left-top aligned.
It means that if two screens are located from left to right, they both 
have y == 0 regardless of their height.
Usually insets are much less than dimensions of a screen and so strut 
minus screen edge should be less
than summary dimension of neighbor screens.
Please, correct me if I'm wrong.

Thanks,
Oleg

04.02.2014 15:28, Alexander Zvegintsev wrote:
> Hi Oleg,
>
> I am just curious about adding rootBounds.x and rootBounds.y, it looks 
> redundant to me.
> Is there a case when a root window have coordinates with non-zero values?
>
> MultiScreenInsetsTest.java:
>>    62             if ((bounds.x != 0 && insets.left >= bounds.x)
>>    63                 || (bounds.y != 0 && insets.top >= bounds.y)) {
> This check will fail for screen with 
> [x=1000,y=50,width=1000,height=1000] bounds and top inset 100,
> but it is a valid case.
>
> I think we should check insets against screen width and height:
>             if (insets.left >= bounds.width
>                     || insets.right >= bounds.width
>                     || insets.top >= bounds.height
>                     || insets.bottom >= bounds.height) {
>
>
> Otherwise fix looks good to me.
>
> Thanks,
>
> Alexander.
> On 02/04/2014 10:26 AM, Oleg Pekhovskiy wrote:
>> Hi All,
>>
>> please review the second version of fix:
>> http://cr.openjdk.java.net/~bagiras/8020443.2/
>>
>> It turned out that struts must be checked and corrected from all sides
>> to become the proper screen insets.
>>
>> Thanks,
>> Oleg
>>
>> On 01/21/2014 06:31 PM, Oleg Pekhovskiy wrote:
>>> Hi all,
>>>
>>> please review the fix
>>> http://cr.openjdk.java.net/~bagiras/8020443.1/
>>> for
>>> https://bugs.openjdk.java.net/browse/JDK-8020443
>>>
>>> Referring to the standards, we must calculate insets the special way 
>>> for Xinerama:
>>> http://standards.freedesktop.org/wm-spec/1.3/ar01s05.html
>>>
>>> _NET_WM_STRUT_PARTIAL
>>> "The start and end values associated with each strut allow areas to 
>>> be reserved which do not span the entire width or height of the 
>>> screen. Struts MUST be specified in root window coordinates, that 
>>> is, they are /not/ relative to the edges of any view port or 
>>> Xinerama monitor."
>>>
>>> So the fix checks if the insets have absolute values and if so makes 
>>> them relative to each screen.
>>> The issue occurred when the Frame was created with the location by 
>>> default.
>>>
>>> Thanks,
>>> Oleg
>>>
>>
>