EVALUATION
Wow, slanted text! I don't believe we could do this in Swing even if we tried ;)
Re-assigning to Java 2D.
###@###.### 2004-04-22
==================================
I can reproduce this on Solaris 9. It looks like its caused by the fix for
4458472
###@###.### 2004-04-22
=============================
Name: inR10103 Date: 04/23/2004
This is regression introduced in the fix for 4458472.
The problem is reproducible if bitmap bolding is applied and bitmap is
widened during bolding process. In such case we correct xAdvanceWidth16x16 to
reflect width change but for rotated case this logic is not correct and we
might need to change yAdvance instead.
In this particular test for vertical progress bar we all glyphs
has horisontal advance equal to one pixel (correct supposed to be 0).
String layout logic (actually similar effect is achivable with
Graphics.drawString()) accumulate this error and cause this funny effect.
For quadrant rotations this problem is easy to fix (we should not
increment zero advances and pay attention to advance sign) but
proper support for arbitrary rotations might be more complicated.
======================================================================
|
Java