Bug 484892

I cannot reproduce this. I don't get any errors or slowness.

What kind of information do you need?

(In reply to Nicolas Fella from comment #2)
> What kind of information do you need?

Honestly, not sure. Are you sure you're on the latest git master? What if you do a clean build? What if you clear your QML cache ( `fd "qmlcache" ~/.cache -X rm -r {}` )?

I checked the code used for implicit height and there's no reason for there to be a binding loop:

/home/noah/kde/src/graphics/spectacle/src/Gui/ImageCaptureOverlay.qml:
        component SelectionHandle: Handle {
            visible: enabled && selectionRectangle.visible
                && SelectionEditor.dragLocation === SelectionEditor.None
                && G.rectIntersects(Qt.rect(x,y,width,height), annotations.viewportRect)
            shapePath.fillColor: selectionRectangle.border.color
            shapePath.strokeWidth: 0
        }

        SelectionHandle { // <-- Line 131
            startAngle: 90
            sweepAngle: 270
            x: SelectionEditor.handlesRect.x
            y: SelectionEditor.handlesRect.y
        }

/home/noah/kde/src/graphics/spectacle/src/Gui/Handle.qml:
    // Use a rounded physically even size so that straight edges don't look fuzzy
    implicitWidth: dprRoundEven(18)
    implicitHeight: implicitWidth

    // Round to a physically even size
    function dprRoundEven(value) {
        value = Math.round(value * Screen.devicePixelRatio)
        return (value - value % 2) / Screen.devicePixelRatio
    }

(by setting the but to NEEDSINFO it will be closed in 30 days unless it's set back, so it's not nice to do that if you don't ask for specific information)

(In reply to Nicolas Fella from comment #4)
> (by setting the but to NEEDSINFO it will be closed in 30 days unless it's
> set back, so it's not nice to do that if you don't ask for specific
> information)

Ah, sorry.

Can confirm on Qt 6.7.

Not entirely sure what changed but changing the scale to 1,1 fixes it for me [1]. Probably just one link in the chain that gets broken and then it stops this loop but of course looks a bit broken then

[1] https://invent.kde.org/graphics/spectacle/-/blob/master/src/Gui/Handle.qml?ref_type=heads#L166

Git commit 862adb1ea39a6f393c0289cad9b3a380aa1b7c2c by Noah Davis.
Committed on 11/04/2024 at 22:55.
Pushed by ndavis into branch 'master'.

Handle: prevent division by zero in ShapePath scale

M  +5    -2    src/Gui/Handle.qml

https://invent.kde.org/graphics/spectacle/-/commit/862adb1ea39a6f393c0289cad9b3a380aa1b7c2c

*** Bug 485545 has been marked as a duplicate of this bug. ***

*** Bug 484947 has been marked as a duplicate of this bug. ***