Hi, the attached patch fixes the behavior of javax.swing.text.Utilities.getTabbedTextOffset to be identical to what the RI does. It now passes all of the mauve tests.
After fixing this and observing the effects on getBreakLocation I decided that this method calls calls the variant with round=false instead to match the behavior of the RI. The ChangeLog: 2006-04-27 Robert Schuster <[EMAIL PROTECTED]> * javax/swing/text/Utilities.java: (getTabbedTextOffset): Introduced width variable, rewritten the check which ends the loop. (getBreakLocation): Call getTabbedTextOffset with rounding argument set to false. cya Robert
signature.asc
Description: OpenPGP digital signature