https://bugs.kde.org/show_bug.cgi?id=477561
Tiar <tamtamy.tym...@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REPORTED |CONFIRMED CC| |tamtamy.tym...@gmail.com Ever confirmed|0 |1 --- Comment #3 from Tiar <tamtamy.tym...@gmail.com> --- Hi, if you check "Align the brush preview outline to the pixel grid", it should start to work fine. However, that is, I think, a regression: in 5.1.5 the brush works as expected despite the checkbox not being checked, therefore making the "aligning to the pixel grid" a cosmetic checkbox instead of a functional one (and it was still important as a cosmetic checkbox since many people felt that constricting the brush outline to pixels felt worse, since it was "jumping" from one position to the next. There were many discussions about this). Now the question is whether the change of behaviour was expected or not. I mean it changes the way an already saved brush works, so I would assume it's a regression and not an expected change. Related commits: 4c639ee9bd5c958cce08598ecc3beea0182e7e33 (probable introduction of the wrong logic? Except it was already included in 5.1...) And: 5fdf853a1a1fe21b04c6bf02584762040bd66628 (that is why it works differently in 5.2.0-beta1). Faulty line: https://invent.kde.org/graphics/krita/-/blob/master/plugins/paintops/libpaintop/KisSharpnessOption.cpp#L34 - includes m_alignOutlinePixels, and that function is called in both outline painting and dab painting. Fix: just separate two functions so that the outline is painted according to the checkbox, and the dab is always aligned, as it was before. I will investigate further to understand how that happened, which should give a hint whether it was in any way intentional or just a mistake. -- You are receiving this mail because: You are watching all bug changes.