FWD » Core Development » User Interface

1. Run an app with a top-level window in Web GUI client.
2. Drag the window out of the document's visible area.
3. The window will disappear.

Is it the same bug that has been found during the regression testing of 1811t?

./window_parenting/waitfor_2wnd.p for the web client the second window (hwin2) disappears if we move it. (regression P2j trunc doesn't have this bug)

After recent commits in P2J trunc this function activateTopVisibleWindow of p2j.screen.js module was broken and p2j.socket.sendWindowActive was removed:

   /**
    * Move the first upper visible window to the top of other visible windows
    * after the selected window is minimized.
    * 
    * @param    {Number} id
    *           The id of the window that has been minimized.
    */
   function activateTopVisibleWindow(id)
   {
      var topWindow = findTopVisibleWindow(id);
      var focusOut = !(topWindow);
      p2j.socket.sendWindowActive(id, false, focusOut);
      if (topWindow)
      {
         topWindow.moveToTop();
         p2j.socket.sendWindowActive(topWindow.id, true, false);
      }
   }

Sergey Ivanovskiy wrote:

After recent commits in P2J trunc this function activateTopVisibleWindow of p2j.screen.js module was broken and p2j.socket.sendWindowActive was removed:
[...]

I renamed sendWindowActive to sendWindowActivated and changed the signature of the method to take the id of the window being deactivated and id of the window being activated. Can you please modify activateTopVisibleWindow accordingly?

Ok, planning to fix it, but function doWindowActivationChange(deactivatingWindowId, activatingWindowId) doesn't work too.

Sergey Ivanovskiy wrote:

Ok, planning to fix it, but function doWindowActivationChange(deactivatingWindowId, activatingWindowId) doesn't work too.

What is wrong there in particular?

I divide this function into parts with comments. Please look at

   function doWindowActivationChange(deactivatingWindowId, activatingWindowId)
   {
      // do we have a valid window being deactivated?
      if (deactivatingWindowId != -1)
      {
         var win = getWindow(deactivatingWindowId);

         if (!win.isVisible())
         {
            p2j.socket.sendWindowIconState(windowId, false);
         }
      }

      
      // do we have a valid window being activated?
      if (activatingWindowId != -1)
      {
         // de-iconify the window
         win.deiconify();
      }

      // send the window activation to the java side
      if (activatingWindowId != -1 || deactivatingWindowId != -1)
      {
         p2j.socket.sendWindowActivated(deactivatingWindowId, activatingWindowId);
      }
   }

Sergey Ivanovskiy wrote:

I divide this function into parts with comments. Please look at
[...]
The variable win can be invalid window if deactivatingWindowId is -1.
[...]

You are right the body of the second if statement is missing the following:

var win = getWindow(activatingWindowId);

I am glad you caught this.

Hynek, please correct me that now it is enough to send this message sendWindowActivated one time for deactivating the old window and activating new one.

Yes, this is exactly how it is supposed to work now. One message instead of two.

Committed revision 11007 fixed doWindowActivationChange and activateTopVisibleWindow.

Code Review Task Branch 1811u Revision 11007

The changes look fine to me. however the p2j.screen.js does need a history entry.

Hynek: are you OK with the change in activateTopVisibleWindow()?

Now, I am doing the bug related to resizing in the web client. The window can be squeezed but can't be resized to widen its width and height. Should I create a new task or fix it within this task?

Greg Shah wrote:

Code Review Task Branch 1811u Revision 11007

The changes look fine to me. however the p2j.screen.js does need a history entry.

Hynek: are you OK with the change in activateTopVisibleWindow()?

The changes are OK.