fix artefacts with frame buffer and scaling with nanovg about iplug2 HOT 12 CLOSED

Have we got a test example?

from iplug2.

if you pGraphics->AttachCornerResizer(kUIResizerScale);

and

#if 1
for (auto i = 0; i < nRows * nColumns; i++) {
pGraphics->AttachControl(new IBKnobControl(*this, bounds.GetGridCell(cellIdx++, nRows, nColumns).GetPadded(-5.), bitmap1, kGain));
}
#endif

in IPlugEffect...

towards the end of the gif, you can see a line appear around the dirty area

from iplug2.

more obvious on virtualcz. seems to be bitmap related

from iplug2.

I see it on the second example but not the first. Is virtualCZ scaled when it happens, or just at normal size?

from iplug2.

Oh - I see it on the top one, but the bitmaps shouldn't be drawing anywhere near that should they? I'll take a look.

from iplug2.

Yes, it's only when scaled that a thin line appears around the dirty area. here is a better video of IPlugEffect:

from iplug2.

OK - I have some traction on this, but one issue is that in IRECT there are Scale() and ScaleBounds().

The first of these rounds - the second increases outwards. There is also PixelAlign (increase outwards).

In my opinion Scale() shouldn't round to integer values and ScaleBounds() shouldn't exist, although I could make ScaleAlign() which would call Scale() and the then PixelAlign() for convenience.

Not sure what the effects of this will be, but I had seen it before and it doesn't look right.

from iplug2.

Interestingly the same thing happens with IGraphicsCanvas, when you rescale

from iplug2.

graphics-edge-effects branch doesn't fix this unfortunately

from iplug2.

I know - that was the branch where I was going to put it, but I was mostly experimenting. What got pushed there was some niceifying but not the real deal which needs to be done soon!

from iplug2.

Fixed on graphics-edge-effects branch

from iplug2.

Fixed by 6baa706

from iplug2.