Window.SetBounds(left,top,width,height)

Syntec

Hey Garry,

I really appreciate you adding SetBounds as an atomic way of updating window bounds, but it appears the underlying SetBounds methods is still triggering individual window updates as there has been no change to the jitter/rubber banding when modifying left/width or top/height.

I am going to play around with a few other optimizations to see if it is a bug in my code, but just flagging in the event you can identify multiple calls under the hood before I figure it out!

Cheers!

Garry

That's disappointing - sorry about that. Looking more deeply at the runtime I can see the issue (its a threading issue).

I will fix this before the next release: https://feedback.objo.dev/bug/1614

Syntec

Garry Hey Garry, no worries at all! It forced me to refactor my code and clean it up (less calls to SetBounds will always be good in the long run) it also doesn't stop me from continuing to test/build out my UI, I look forward to the next release! 🙂

Syntec

Hey Garry,

With the latest release, when modifying left/width and top/height. If the window is shrinking the resizing is rock stable but if the window is growing the rubber banding still happens.

I realize the video isn't the best demo (black-on-black) I can take something with more contrast if it helps see what I am talking about.

2026-10-05-10-10-58-reduced.mp4

910kB

Syntec

Something else I noticed in a similar vein is canvas drawing.

g.DrawingColour = Colour.RGB(255, 255, 255)
g.DrawGlowingRegion(testRegion, 4, 3, 0.5, Colour.RGB(255, 255, 255)) # draws a glowing SVG backed rectangular region

g.DrawingColour = Colour.RGB(255, 0, 0)
g.DrawRectangle(25, 150, 100, 100) # draws a standard graphics backed rectangle

When drawing the above on a canvas that is resizing both methods rubber band the rectangle. I could expect this behaviour from the SVG backed graphics extension as it is drawing a picture (important to note the picture is drawn at 100% scale) and I imagine picture drawing has some g.width/g.height involvement. But I would think the standard draw call would always produce a 100px by 100px rectangle, it exhibits the same behaviour.

2026-10-05-10-37-37-reduced.mp4

568kB

Garry

Thanks for the follow-ups - both really useful.

In the first video, it looks like the fix in 26.10.1 did its job for shrinking (the anchored edge now stays put to the pixel), but while the window is growing the runtime still applies an out-of-date size on alternating frames, which is what makes the anchored edge wobble in and out. That's a bug on Objo's side I think, nothing you're doing wrong. I've logged it here: https://feedback.objo.dev/bug/1677 and I'll get it fixed for the next release.

I checked your second video too, and the canvas drawing behaviour I think is the same defect seen from inside the window. While the window passes through those momentary wrong sizes, the entire rendered frame - canvas drawings included - is briefly stretched relative to the window itself, so even a rectangle drawn at fixed coordinates appears to breathe. Your plain DrawRectangle test is good evidence for that: it points at the compositing level rather than anything involving g.Width/g.Height or picture scaling. I'm treating it as part of the same fix.

Garry

I hope I've finally fixed this now in the next release 🤞