Created attachment 12672 [details]
corrupt kicker backgrounds

SVN commit 463429 by aseigo:

paint button backgrounds properly
BUGS:113093,113160


 M  +50 -2     kicker/core/container_button.cpp  
 M  +6 -2      libkicker/panelbutton.cpp  


--- branches/KDE/3.5/kdebase/kicker/kicker/core/container_button.cpp #463428:463429
@@ -36,6 +36,7 @@
 #include "global.h"
 #include "appletop_mnu.h"
 
+#include "containerarea.h"
 #include "panelbutton.h"
 #include "bookmarksbutton.h"
 #include "browserbutton.h"
@@ -43,6 +44,7 @@
 #include "extensionbutton.h"
 #include "kbutton.h"
 #include "kicker.h"
+#include "kickerSettings.h"
 #include "kickertip.h"
 #include "nonkdeappbutton.h"
 #include "paneldrag.h"
@@ -94,13 +96,59 @@
     return width;
 }
 
+#include <iostream>
+using namespace std;
 void ButtonContainer::setBackground()
 {
     PanelButton* b = button();
-    if (b)
+    if (!b)
     {
-        b->repaint();
+        return;
     }
+
+cout << "ButtonContainer::setBackground()" << endl;
+    b->unsetPalette();
+
+    if (!KickerSettings::transparent())
+    {
+        b->update();
+        return;
+    }
+
+    // inheritance is ButtonContainer - ScrollView - ContainerArea
+    ContainerArea* area = 0;
+    QObject* ancestor = parent();
+    while (ancestor && !area)
+    {
+        area = dynamic_cast<ContainerArea*>(ancestor);
+        ancestor = ancestor->parent();
+    }
+
+    const QPixmap* containerBG = area ? area->completeBackgroundPixmap()
+                                        : 0;
+
+    if (!containerBG || containerBG->isNull())
+    {
+cout << "\tsetting no background!" << endl;
+        b->setPaletteBackgroundPixmap(QPixmap());
+    }
+    else
+    {
+        int srcx = x();
+        int srcy = y();
+        int buttonWidth = b->width();
+        int buttonHeight = b->height();
+
+        QPixmap buttonBackground(buttonWidth, buttonHeight);
+        copyBlt(&buttonBackground, 0, 0,
+                containerBG, srcx, srcy,
+                buttonWidth, buttonHeight);
+
+        b->setPaletteBackgroundPixmap(buttonBackground);
+    cout << "\tbackground set!" << endl;
+    }
+
+    b->update();
 }
 
 void ButtonContainer::configure()
--- branches/KDE/3.5/kdebase/kicker/libkicker/panelbutton.cpp #463428:463429
@@ -502,13 +502,17 @@
 
 void PanelButton::drawButton(QPainter *p)
 {
-    // Draw the background. This is always needed, even when using tiles,
-    // because they don't have to cover the entire button.
     if (m_tileColor.isValid())
     {
         p->fillRect(rect(), m_tileColor);
         style().drawPrimitive(QStyle::PE_Panel, p, rect(), colorGroup());
     }
+    else if (paletteBackgroundPixmap())
+    {
+        // Draw the background. This is always needed, even when using tiles,
+        // because they don't have to cover the entire button.
+        p->drawPixmap(0, 0, *paletteBackgroundPixmap());
+    }
 
     const QPixmap& tile = (isDown() || isOn()) ? m_down : m_up;
     if (!tile.isNull())

Panel buttons seem much less likely to suffer from the effects I attached in the previous screen shot after updating only the two files patched. But it does still happen. I'm finding it very difficult to reproduce now though. I didn't screenshot it when I got it to happen because it looks the same as before.

Also the taskbar, systray, clock and logout/lock applet backgrounds all still get turned to grey, as in my previous screen shot.

(I really wish KDE had anonymous svn+http access. Damn file walls :( )

I seem to be able to reproduce it a lot easier now. If mouse over doesn't cause the issue, then clicking on one of the buttons seems to put it in a bad state. After clicking on a button, its background get corrupted as before, but mousing over further buttons, corrupts them as well now. Mouse over now seems to stay in that bad state.

Clicking a button again, puts it back into a good state, correcting the button backgrounds as I mouse over the icons.

The taskbar etc background corruptions, still seems to happen randomly as I change desktops. Sometimes it right, sometimes its wrong. Its about 50-50.

nick, the other bug report (113093) referenced has already been re-opened. i'm on top of this, please don't make more work for me by going around and managing bugs assigned to me. thanks.

*** This bug has been marked as a duplicate of 113093 ***