I implemented getRowStart and getRowEnd in javax.swing.text.Utilities.
This is RFC because of the issue raised in the large comment in each of
this methods (namely, performance). Some comments would be appreciated.
2005-09-26 Anthony Balkissoon <[EMAIL PROTECTED]>
* javax/swing/text/Utilities.java:
(getRowEnd): New method.
(getRowStart): New method.
--Tony
Index: javax/swing/text/Utilities.java
===================================================================
RCS file: /cvsroot/classpath/classpath/javax/swing/text/Utilities.java,v
retrieving revision 1.11
diff -u -r1.11 Utilities.java
--- javax/swing/text/Utilities.java 22 Sep 2005 20:17:08 -0000 1.11
+++ javax/swing/text/Utilities.java 26 Sep 2005 20:31:45 -0000
@@ -394,4 +394,79 @@
wb.setText(text);
return wb.following(offs);
}
+
+ /**
+ * Get the model position of the end of the row that contains the
+ * specified model position. Return null if the given JTextComponent
+ * does not have a size.
+ * @param c the JTextComponent
+ * @param offs the model position
+ * @return the model position of the end of the row containing the given
+ * offset
+ * @throws BadLocationException if the offset is invalid
+ */
+ public static final int getRowEnd(JTextComponent c, int offs)
+ throws BadLocationException
+ {
+ Element root = c.getDocument().getDefaultRootElement();
+ Element rowElement = root.getElement(root.getElementIndex(offs));
+ String text = c.getText(rowElement.getStartOffset(),
+ rowElement.getEndOffset());
+ if (text == null)
+ return -1;
+
+ //Start at offs and iterate character by character until one of them
+ //is on the next line. This is poor performance, could be improved
+ //by calculating an approximate number of columns per row using
+ //the preferred size of the JTextComponent and the width of an
+ //arbitary character in whichever font is in use. Could increment i
+ //by (for example) half this value until incrementing it again would
+ //make the if statement true, then increment it by a smaller value
+ //until doing so would make the if statement true. Basically, use
+ //coarse increments until we need to fine tune to character by character.
+ for (int i = offs; i < text.length() - 1; i++)
+ {
+ if (c.modelToView(i).y != c.modelToView(offs).y)
+ return i - 1;
+ }
+ return text.length() - 1;
+ }
+
+ /**
+ * Get the model position of the start of the row that contains the specified
+ * model position. Return null if the given JTextComponent does not have a
+ * size.
+ *
+ * @param c the JTextComponent
+ * @param offs the model position
+ * @return the model position of the start of the row containing the given
+ * offset
+ * @throws BadLocationException if the offset is invalid
+ */
+ public static final int getRowStart(JTextComponent c, int offs)
+ throws BadLocationException
+ {
+ Element root = c.getDocument().getDefaultRootElement();
+ Element rowElement = root.getElement(root.getElementIndex(offs));
+ String text = c.getText(rowElement.getStartOffset(),
+ rowElement.getEndOffset());
+ if (text == null)
+ return -1;
+
+ //Start at offs and iterate character by character until one of them
+ //is on the previous line. This is poor performance, could be improved
+ //by calculating an approximate number of columns per row using
+ //the preferred size of the JTextComponent and the width of an
+ //arbitary character in whichever font is in use. Could increment i
+ //by (for example) half this value until incrementing it again would
+ //make the if statement true, then increment it by a smaller value
+ //until doing so would make the if statement true. Basically, use
+ //coarse increments until we need to fine tune to character by character.
+ for (int i = offs; i > 0; i--)
+ {
+ if (c.modelToView(i).y != c.modelToView(offs).y)
+ return i + 1;
+ }
+ return 0;
+ }
}
_______________________________________________
Classpath-patches mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/classpath-patches
