RFR: JDK-8297449: Update JInternalFrame Metal Border code [v2]
SWinxy
duke at openjdk.org
Tue Nov 29 02:16:20 UTC 2022
On Tue, 29 Nov 2022 00:05:18 GMT, Harshitha Onkar <honkar at openjdk.org> wrote:
>> src/java.desktop/share/classes/javax/swing/plaf/metal/MetalBorders.java line 267:
>>
>>> 265: }
>>> 266:
>>> 267: AffineTransform at = null;
>>
>> Because we'd no longer be causing a `ClassCastException` from casting this to be a `Graphics2D` object, later down on line 311, if the `Graphics` object isn't a G2D type, then it will throw an NPE. I fear matrices, so I have no idea what the value of `getScaleX()` would be otherwise--replace it with zero if null?
>
> @SWinxy Thanks for catching it. That's right, but the scale factor cannot be zero, it should probably be handled similar to `stkWidth` if **'at is null'** (at is null when g is not an instance of Graphics2D). I'll need to check the scale value to be used in this context.
>
> Something like below.
> `at != null ? at.getScaleX() : 1;`
Oh I see. Otherwise the thickness would be zero. I feeeeel like doing a large if-else on checking if g is a Graphics2D would end up being cleaner. I fiddled around and this is what the else branch would look like with the affine transform value being `1`:
g.translate(x, y);
g.setColor(background);
// Draw the bulk of the border
for (int i = 0; i < 5; i++) {
g.drawRect(i, i, w - (i * 2), h - (i * 2));
}
if (c instanceof JInternalFrame internalFrame && internalFrame.isResizable()) {
// Draw the Long highlight lines
g.setColor(highlight);
g.drawLine(CORNER + 1, 2, w - CORNER, 2); //top
g.drawLine(2, CORNER + 1, 2, h - CORNER); //left
g.drawLine(w - 2, CORNER + 1, w - 2, h - CORNER); //right
g.drawLine(CORNER + 1, h - 2, w - CORNER, h - 2); //bottom
// Draw the Long shadow lines
g.setColor(shadow);
g.drawLine(CORNER, 1, w - CORNER - 1, 1);
g.drawLine(1, CORNER, 1, h - CORNER - 1);
g.drawLine(w - 3, CORNER, w - 3, h - CORNER - 1);
g.drawLine(CORNER, h - 3, w - CORNER - 1, h - 3);
}
// restore previous transform
g.translate(-x, -y);
-------------
PR: https://git.openjdk.org/jdk/pull/11305
More information about the client-libs-dev
mailing list