Package org.jdesktop.swingx

Class JXTreeTable.TreeTableHacker

java.lang.Object
org.jdesktop.swingx.JXTreeTable.TreeTableHacker
Direct Known Subclasses:
JXTreeTable.TreeTableHackerExt
Enclosing class:
JXTreeTable
public class JXTreeTable.TreeTableHacker extends Object
Temporary class to have all the hacking at one place. Naturally, it will change a lot. The base class has the "stable" behaviour as of around jun2006 (before starting the fix for 332-swingx).

specifically:

  1. hitHandleDetection triggeredn in editCellAt

  • Field Details

    • expansionChangedFlag

      protected boolean expansionChangedFlag

  • Constructor Details

    • TreeTableHacker

      public TreeTableHacker()

  • Method Details

    • isHitDetectionFromProcessMouse

      protected boolean isHitDetectionFromProcessMouse()
      Decision whether the handle hit detection should be done in processMouseEvent or editCellAt. Here: returns false.
      Returns:
      true for handle hit detection in processMouse, false for editCellAt.

    • hitHandleDetectionFromEditCell

      public void hitHandleDetectionFromEditCell(int column, EventObject e)
      Entry point for hit handle detection called from editCellAt, does nothing if isHitDetectionFromProcessMouse is true;
      See Also:

    • hitHandleDetectionFromProcessMouse

      public boolean hitHandleDetectionFromProcessMouse(MouseEvent e)
      Entry point for hit handle detection called from processMouse. Does nothing if isHitDetectionFromProcessMouse is false.
      Returns:
      true if the mouseEvent triggered an expand/collapse in the renderer, false otherwise.
      See Also:

    • completeEditing

      protected void completeEditing()
      Complete editing if collapsed/expanded.

      Is: first try to stop editing before falling back to cancel.

      This is part of fix for #730-swingx - editingStopped not always called. The other part is to call this from the renderer before expansion related state has changed.

      Was: any editing is always cancelled.

      This is a rude fix to #120-jdnc: data corruption on collapse/expand if editing. This is called from the renderer after expansion related state has changed.

    • expandOrCollapseNode

      protected boolean expandOrCollapseNode(int column, EventObject e)
      Tricksery to make the tree expand/collapse.

      This might be - indirectly - called from one of two places:

      1. editCellAt: original, stable but buggy (#332, #222) the table's own selection had been changed due to the click before even entering into editCellAt so all tree selection state is lost.
      2. processMouseEvent: the idea is to catch the mouseEvent, check if it triggered an expanded/collapsed, consume and return if so or pass to super if not.

      widened access for testing ...

      Parameters:
      column - the column index under the event, if any.
      e - the event which might trigger a expand/collapse.
      Returns:
      this methods evaluation as to whether the event triggered a expand/collaps

    • mightBeExpansionTrigger

      protected boolean mightBeExpansionTrigger(EventObject e)

    • expansionChanged

      protected void expansionChanged()
      called from the renderer's setExpandedPath after all expansion-related updates happend.