On Wed, 21 Jun 2023 20:54:04 GMT, Alexey Ivanov <[email protected]> wrote:
> > Since the border is moved to left slightly (maybe 1pixel), so we are seeing > > that the string is "cut off". I guess if we work on correcting the > > JTableMetal Border along with adding a margin space (Right), the issue will > > be fixed? (Screen shot added - Maybe have to zoom-in to see the difference > > in border w.r.t data cell border) > > It's very likely the reason why. We saw that borders may render > inconsistently: 1-, 2- or 3-pixel lines (for 225%). > > I zoomed in your screenshot. The 100%-scale case has 1-pixel shadow and > 1-pixel highlight. It seems the text is cut off by 1 pixel, the black pixels > of ‘e’ are at the very edge of the shadow pixels. I'd expect to see blue-ish > pixels of text antialiasing to the right of the black text pixels. > > In 225%-scale case, the first separator looks better. It consists of 2-pixel > shadow, 1-pixel background, 1-pixel highlight for the next header. On the > background-colored pixel, I see pixels of text, blue-ish pixels of > antialiasing. So the position of the text is still a bit off. It's no wonder > if you take into account that it's 1-pixel off even for 100% case. > > The second separator has more issues: 2-pixel shadow, 2-pixel background, > 1-pixel highlight. There should be no background pixels between the shadow > and highlight but there are… It's an inherent problem of scaled graphics. > That background pixels also have parts of the text, 2 pixels. > > In short, it seems there's still a text positioning issue… as if it doesn't > account for antialiasing. Ok, in summary there exists 2 major issues: 1. The text should positioned till the blue-ish pixels is just behind the header border for each header cell. 2. The header border to be aligned with data grid lines. Right? ------------- PR Comment: https://git.openjdk.org/jdk/pull/14464#issuecomment-1602007556
