On Tue, 2 Feb 2021 20:32:46 GMT, Alexey Ivanov <aiva...@openjdk.org> wrote:
>
>
> > It seems Toolkit.getDefaultToolkit().getScreenResolution() can return
> > 93/94/95 instead of 96 in mac/linux in internal mach5 testing systems
> > causing failure in this test Test.java. Probably we need to make the
> > testcase hardcoded to 96
>
> I guess so.
> Where does this `Test.java` come from?
>
> > I also suggest fixing JDK-8260687 to not use font inherit for
> > W3C_LENGTH_UNIT case
> > ```
> > --- a/src/java.desktop/share/classes/javax/swing/text/html/StyleSheet.java
> > +++ b/src/java.desktop/share/classes/javax/swing/text/html/StyleSheet.java
> > @@ -2823,7 +2823,7 @@ public class StyleSheet extends StyleContext {
> > }
> >
> > Object doGetAttribute(Object key) {
> > - if (key == CSS.Attribute.FONT_SIZE && !isDefined(key)) {
> > + if (key == CSS.Attribute.FONT_SIZE && !isDefined(key) &&
> > !isW3CLengthUnits()) {
> > // CSS.FontSize represents a specified value and we need
> > // to inherit a computed value so don't resolve percentage
> > // value from parent.
> > ```
>
> Does such a fix not bring back the problem with relative units for the case
> where W3C_LENGTH_UNIT is enabled?
I guess no as this code is recently introduced by JDK-8257664 and I have
verified with this testcase and other jtreg regression tests. But if you have a
better fix, then please suggest.
-------------
PR: https://git.openjdk.java.net/jdk/pull/2256
