RFR: 8306464: [testbug] CustomSecurityManagerTest full screen tests fails on MacOS intermittently [v2]

[Karthik P K]

On Thu, 20 Apr 2023 15:27:37 GMT, Kevin Rushforth <kcr at openjdk.org> wrote:

>> When the app enters full screen mode in MacBook Pro M1, there will be still a black strip on the top because of the notch. So assigning small value will not fetch green color from app window. It requires at least y=40.
>> So I think we can add new static final variable with a value of 40 or 50 to be on the safer side and use it in all the places so that it works for all the cases.
>> Please let me know your thought on this.
>
> Oh, I meant to say that you could use a small offset like 3 or 4 added to visualBounds.minX/minY for the left and top edges. Similarly, you should use a small offset from visualBounds.maxX/maxY for the right and bottom edges (rather than using width / height).

Updated the code to use visual bounds while calculating `x` and `y` values.
Changed the offset of coordinates used for reading color to 10 since in M1 macs `robot` reads color value from the cursor position irrespective of  coordinates given and cursor is parked at the bottom right corner with an offset of 5 from the visual bounds. Hence full screen tests where failing if `getMaxY()` method was used to calculate y value with lower offsets.

>> If static final variable with a suitable value is added as mentioned above, same variable can be used here.
>
> Taking my previous comment to the other query into account, I think this should be more like:
> 
> 
>                     int x = col == 0 ? screenBound.minX + 4 : screenBounds.maxX - 5;
> 
> 
> where `screenBounds` is the visual bounds of the screen.

Used visual bounds for calculating the `x` value with offset of 10. Larger offset because of the same reason as explained in previous comment.

-------------

PR Review Comment: https://git.openjdk.org/jfx/pull/1103#discussion_r1173803742
PR Review Comment: https://git.openjdk.org/jfx/pull/1103#discussion_r1173805578