Review request for 7154030: java.awt.Component.hide() does not repaint parent container (original) (raw)

Jonathan Lu luchsh at linux.vnet.ibm.com
Fri Apr 6 09:21:58 UTC 2012

Hi Pavel,

Here's the update patch, http://cr.openjdk.java.net/~luchsh/7154030_4/

My comments are inlined.

On 03/27/2012 11:58 PM, Pavel Porvatov wrote:

Hi Jonathan,

Hello Pavel,

Here's the updated patch and automatic test for bug 7154030, could you please take another look? http://cr.openjdk.java.net/~luchsh/71540302/ I have several comments: 1. What about the following comment from Artem: "Even if we accept the change in JComponent.hide(), we should then override show() as well (lightweight component may be non-opaque, so we should repaint from its parent)" ? I've updated my test case by including non-opaque components, but I do no see a need for overriding show(), is there anything incorrect with the updated testcase or my understanding?

2. Could you please clarify your changes in setVisible method? As I see in comments // Some (all should) LayoutManagers do not consider components // that are not visible. As such we need to revalidate when the // visible bit changes. revalidate(); but now this code is invoked only for setVisible(true) For the setVisible(false) case, the repainting and revalidating operations will be done in new method JComponent.hide(), so this change is just to reduce some duplicated actions.

3. Could you please follow our code conventions? (see http://www.oracle.com/technetwork/java/javase/documentation/codeconventions-141388.html#475) Sorry for this problem, I was trying to keeping aligned with the original style of JComponent.java, which I later realized to be inappropriate. In the updated patch, code has been well formatted.

4. Your test is not automatic one. I think you could use java.awt.Robot#createScreenCapture and analyze result of hide method.

See the link, it should be automatic now.

Regards, Pavel

Thanks and best regards - Jonathan Lu On 03/26/2012 09:38 PM, Pavel Porvatov wrote: Hi Jonathan, Hi Pavel,

Thanks for your explanation. But this bug affects almost all Swing components, hide()'s presence also helps to maintain backward compatibility, so is it possible to put a fix in JComponent to help all the potential affected applications to work correctly? Of course that's possible. Do you have final version of the fix? Please don't forget write an automatic test. if not, is it there any sunset plan for these deprecated APIs? I don't now such plans. Regards, Pavel P.S. I removed , it seems only Swing will be affected in this fix Thanks and best regards! - Jonathan 2012/3/20 Pavel Porvatov <pavel.porvatov at oracle.com_ _<mailto:pavel.porvatov at oracle.com>> Hi Jonathan, Hi Artem,

2012/3/20 Artem Ananiev <artem.ananiev at oracle.com_ _<mailto:artem.ananiev at oracle.com>> Hi, Jonathan, I'm adding swing-dev to CC as we now consider changing Swing code. What you propose sounds technically reasonable, but I don't think it is worth doing anyway as show() and hide() have been deprecated for years now.

Although show() and hide() have been deprecated for years, in my opinion supporting these APIs will still benefit many applications and convince users that Java still has got strong backward compatibility :D. Any ideas from Swing group? I don't see why the words "backward compatibility" are here. There is a bug in deprecated methods "show" and "hide" (I've checked that jdk5 has the same problem), and that's one additional reason to use setVisible(). I agree with Artem that fixing deprecated API is not a high priority task (but we should keep backward compatibility, of course). I also think, that "to leave all as is" is a good decision for the described problem Regards, Pavel

Even if we accept the change in JComponent.hide(), we should then override show() as well (lightweight component may be non-opaque, so we should repaint from its parent), so there will be code duplication. This is one more reason to leave all as is. Yes, I noticed that code duplication too and am trying to make a more compact patch for this problem. This is my personal opinion, I'm not a Swing expert, though. Let anyone from the Swing group comment. Thanks, Artem On 3/20/2012 12:28 PM, Jonathan Lu wrote: Hi Artem, Thanks for your time. 2012/3/19 Artem Ananiev <artem.ananiev at oracle.com_ _<mailto:artem.ananiev at oracle.com> <mailto:artem.ananiev at oracle.com_ _<mailto:artem.ananiev at oracle.com>>> Hi, Jonathan, given the code in java.awt.Component, your statement about difference between hide() and setVisible(false) looks pretty strange to me. Indeed, here is the implementation:

public void show(boolean b) { if (b) { show(); } else { hide(); } } and public void setVisible(boolean b) { show(b); } In JComponent the latter method is overridden and adds exactly what you propose: parent.repaint(). This addition makes sense for lightweight components (e.g. Swing), but heavyweight AWT components shouldn't require this: repaint request is sent from the native system. Yes, lightweight and heavyweight components differ in painting. The original test case only works for the conditions of lightweight components, with another test case for heavyweight components, I found that the problem could not be reproduced on AWT any more. I think the change is only applicable for Swing components, so how about repaint in JComponent.hide() like this? diff -r cdbb33303ea3 src/share/classes/javax/swing/JComponent.java --- a/src/share/classes/javax/swing/JComponent.java Wed Mar 14 13:50:37 2012 -0700 tel:2012%20-0700 tel:2012%20-0700 +++ b/src/share/classes/javax/swing/JComponent.java Tue Mar 20 16:24:09 2012 +0800 @@ -5237,6 +5237,16 @@ } } + public void hide() { + super.hide(); + Container parent = getParent(); + if (parent != null) { + Rectangle r = getBounds(); + parent.repaint(r.x, r.y, r.width, r.height); + parent.invalidate(); + } + } + /** * Returns whether or not the region of the specified component is * obscured by a sibling. Thanks, Artem On 3/15/2012 12:24 PM, Jonathan Lu wrote: Hi awt-dev, java.awt.Component.hide() was declared as deprecation and replaced by setVisible(boolean), but in my tests, it does not works in the same way as setVisible(false). The reason of this failure is that java.awt.Component.hide() does not repaint the special area it used to taken of parent container. Although this is deprecated method, it may still valuable for customers due to compatibility reason. Bug 7154030 created for this issue. Here's a simple test case to demonstrate this problem. /* * Copyright (c) 2012 Oracle and/or its affiliates. All rights reserved. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * * This code is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 only, as * published by the Free Software Foundation. * * This code is distributed in the hope that it will be useful, but WITHOUT * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License * version 2 for more details (a copy is included in the LICENSE file that * accompanied this code). * * You should have received a copy of the GNU General Public License version * 2 along with this work; if not, write to the Free Software Foundation, * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. * * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA * or visit www.oracle.com <http://www.oracle.com> <http://www.oracle.com> if you need additional information or have any * questions. */ /* * Portions Copyright (c) 2012 IBM Corporation */ import javax.swing.*; /* @test 1.1 2012/03/15 @bug 7154030 @run main/manual ComponentHideShowTest.html */ @SuppressWarnings("serial") public class ComponetHideShowTest extends JFrame { JInternalFrame internalFrame; JButton btn; JDesktopPane desktop; ComponetHideShowTest(String name) { super(name); desktop = new JDesktopPane(); setContentPane(desktop); setSize(600, 400); setVisible(true); internalFrame = new JInternalFrame("Test Internal Frame"); internalFrame.setSize(100, 100); internalFrame.setLocation(10, 10); internalFrame.setVisible(true);_ desktop.add(internalFrame); btn = new JButton("OK"); btn.setSize(100, 50); btn.setLocation( 300, 300); btn.setVisible(true); desktop.add(btn); _setDefaultCloseOperation(JFrame.EXITONCLOSE); } _@SuppressWarnings("deprecation") public void runTest() throws Exception { Object[] options = { "Yes, I saw it", "No, I did not see it!" }; int ret = _JOptionPane.showOptionDialog(this, "Do you see the internal window?", "InternalFrmaeHideTest", JOptionPane.YESNOOPTION, JOptionPane.QUESTIONMESSAGE, null, options, options[1]); if (ret == 1 || ret == JOptionPane.CLOSEDOPTION) { throw new Exception("Failed to display internal window"); } internalFrame.hide(); btn.hide(); _ret = JOptionPane.showOptionDialog(this, "Do you see the internal window?", "InternalFrmaeHideTest", JOptionPane.YESNOOPTION, JOptionPane.QUESTIONMESSAGE, null, options, options[1]); if (ret == 0 || ret == JOptionPane.CLOSEDOPTION) { throw new Exception("Failed to hide internal window"); } internalFrame.show(); btn.show(); _ret = JOptionPane.showOptionDialog(this, "Do you see the internal window?", "InternalFrmaeHideTest", JOptionPane.YESNOOPTION, JOptionPane.QUESTIONMESSAGE, null, options, options[1]); if (ret == 1 || ret == JOptionPane.CLOSEDOPTION) { throw new Exception("Failed to hide internal window"); } } public static void main(String[] args) throws Exception { ComponetHideShowTest test = null; test = new _ComponetHideShowTest("InternalFrameHideTest"); test.runTest(); } } And here's the patch http://cr.openjdk.java.net/~_littlee/7154030/ <http://cr.openjdk.java.net/%7E_littlee/7154030/> <http://cr.openjdk.java.net/%7Elittlee/7154030/> Can anybody please help to take a look? Cheers! - Jonathan Best regards! - Jonathan Thanks a lot ! - Jonathan

-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.openjdk.java.net/pipermail/swing-dev/attachments/20120406/54c13c62/attachment.html>

More information about the swing-dev mailing list