Inherits QPaintDevice and Qt.
Inherited by QBitmap and QCanvasPixmap.
QPixmap is one of the two classes Qt provides for dealing with images; the other is QImage. QPixmap is designed and optimized for drawing; QImage is designed and optimized for I/O and for direct pixel access/manipulation. There are (slow) functions to convert between QImage and QPixmap: convertToImage() and convertFromImage().
One common use of the QPixmap class is to enable smooth updating of widgets. Whenever something complex needs to be drawn, you can use a pixmap to obtain flicker-free drawing, like this:
<ol type=1>
Pixel data in a pixmap is internal and is managed by the underlying window system. Pixels can be accessed only through QPainter functions, through bitBlt(), and by converting the QPixmap to a QImage.
You can easily display a QPixmap on the screen using QLabel::setPixmap(). For example, all the QButton subclasses support pixmap use.
The QPixmap class uses copy-on-write, so it is practical to pass QPixmap objects by value.
You can retrieve the width(), height(), depth() and size() of a pixmap. The enclosing rectangle is given by rect(). Pixmaps can be filled with fill() and resized with resize(). You can create and set a mask with createHeuristicMask() and setMask(). Use selfMask() to see if the pixmap is identical to its mask.
In addition to loading a pixmap from file using load() you can also loadFromData(). You can control optimization with setOptimization() and obtain a transformed version of the pixmap using xForm()
Note regarding Windows 95 and 98: on Windows 9x the system crashes if you create more than about 1000 pixmaps, independent of the size of the pixmaps or installed RAM. Windows NT-systems (including 2000, XP and following versions) do not have the same limitation, but depending on the graphics equipment the system will fail to allocate pixmap objects at some point (due to system running out of GDI resources).
Qt tries to work around the resource limitation. If you set the pixmap optimization to QPixmap::MemoryOptim and the width of your pixmap is less than or equal to 128 pixels, Qt stores the pixmap in a way that is very memory-efficient when there are many pixmaps.
If your application uses dozens or hundreds of pixmaps (for example on tool bar buttons and in popup menus), and you plan to run it on Windows 95 or Windows 98, we recommend using code like this:
QPixmap::setDefaultOptimization( QPixmap::MemoryOptim );
while ( ... ) {
// load tool bar pixmaps etc.
QPixmap *pixmap = new QPixmap(fileName);
}
QPixmap::setDefaultOptimization( QPixmap::NormalOptim );
In general it is recommended to make as much use of QPixmap's implicit sharing and the QPixmapCache as possible.
See also QBitmap, QImage, QImageIO, Shared Classes, Graphics Classes, Image Processing Classes, and Implicitly and Explicitly Shared Classes.
We recommend using DefaultOptim.
The contents of the pixmap is uninitialized.
The depth can be either 1 (monochrome) or the depth of the current video mode. If depth is negative, then the hardware depth of the current video mode will be used.
If either w or h is zero, a null pixmap is constructed.
See also isNull() and QPixmap::Optimization.
Constructs a pixmap of size size, depth bits per pixel, optimized in accordance with the optimization value.
The fileName, format and mode parameters are passed on to load(). This means that the data in fileName is not compiled into the binary. If fileName contains a relative path (e.g. the filename only) the relevant file must be found relative to the runtime working directory.
See also QPixmap::ColorMode, isNull(), load(), loadFromData(), save(), and imageFormat().
The fileName, format and conversion_flags parameters are passed on to load(). This means that the data in fileName is not compiled into the binary. If fileName contains a relative path (e.g. the filename only) the relevant file must be found relative to the runtime working directory.
If the image needs to be modified to fit in a lower-resolution result (e.g. converting from 32-bit to 8-bit), use the conversion_flags to specify how you'd prefer this to happen.
See also Qt::ImageConversionFlags, isNull(), load(), loadFromData(), save(), and imageFormat().
Errors are silently ignored.
Note that it's possible to squeeze the XPM variable a little bit by using an unusual declaration:
static const char * const start_xpm[]={
"16 15 8 1",
"a c #cec6bd",
....
The extra const makes the entire definition read-only, which is slightly more efficient (for example, when the code is in a shared library) and ROMable when the application is to be stored in ROM.
In order to use that sort of declaration you must cast the variable back to const char ** when you create the QPixmap.
The conversion_flags argument is a bitwise-OR of the Qt::ImageConversionFlags. Passing 0 for conversion_flags sets all the default options.
Note that even though a QPixmap with depth 1 behaves much like a QBitmap, isQBitmap() returns FALSE.
If a pixmap with depth 1 is painted with color0 and color1 and converted to an image, the pixels painted with color0 will produce pixel index 0 in the image and those painted with color1 will produce pixel index 1.
See also convertToImage(), isQBitmap(), QImage::convertDepth(), defaultDepth(), and QImage::hasAlphaBuffer().
Converts image and sets this pixmap using color mode mode. Returns TRUE if successful; otherwise returns FALSE.
If the pixmap has 1-bit depth, the returned image will also be 1 bit deep. If the pixmap has 2- to 8-bit depth, the returned image has 8-bit depth. If the pixmap has greater than 8-bit depth, the returned image has 32-bit depth.
Note that for the moment, alpha masks on monochrome images are ignored.
See also convertFromImage().
The mask may not be perfect but it should be reasonable, so you can do things such as the following:
pm->setMask( pm->createHeuristicMask() );
This function is slow because it involves transformation to a QImage, non-trivial computations and a transformation back to a QBitmap.
If clipTight is TRUE the mask is just large enough to cover the pixels; otherwise, the mask is larger than the data pixels.
See also QImage::createHeuristicMask().
See also setDefaultOptimization(), setOptimization(), and optimization().
The pixmap depth is also called bits per pixel (bpp) or bit planes of a pixmap. A null pixmap has depth 0.
See also defaultDepth(), isNull(), and QImage::convertDepth().
A pixmap is automatically detached by Qt whenever its contents is about to change. This is done in all QPixmap member functions that modify the pixmap (fill(), resize(), convertFromImage(), load(), etc.), in bitBlt() for the destination pixmap and in QPainter::begin() on a pixmap.
It is possible to modify a pixmap without letting Qt know. You can first obtain the system-dependent handle() and then call system-specific functions (for instance, BitBlt under Windows) that modify the pixmap contents. In such cases, you can call detach() to cut the pixmap loose from other pixmaps that share data with this one.
detach() returns immediately if there is just a single reference or if the pixmap has not been initialized yet.
Fills the pixmap with the widget's background color or pixmap. If the background is empty, nothing is done. xofs, yofs is an offset in the widget.
Fills the pixmap with the widget's background color or pixmap. If the background is empty, nothing is done.
The ofs point is an offset in the widget.
The point ofs is a point in the widget's coordinate system. The pixmap's top-left pixel will be mapped to the point ofs in the widget. This is significant if the widget has a background pixmap; otherwise the pixmap will simply be filled with the background color of the widget.
Example:
void CuteWidget::paintEvent( QPaintEvent *e )
{
QRect ur = e->rect(); // rectangle to update
QPixmap pix( ur.size() ); // Pixmap for double-buffering
pix.fill( this, ur.topLeft() ); // fill with widget background
QPainter p( &pix );
p.translate( -ur.x(), -ur.y() ); // use widget coordinate system
// when drawing on pixmap
// ... draw on pixmap ...
p.end();
bitBlt( this, ur.topLeft(), &pix );
}
See also QMimeSourceFactory, QImage::fromMimeSource(), and QImageDrag::decode().
Example: textedit/textedit.cpp.
If the widget has any children, then they are also painted in the appropriate positions.
If you specify x, y, w or h, only the rectangle you specify is painted. The defaults are 0, 0 (top-left corner) and -1,-1 (which means the entire widget).
(If w is negative, the function copies everything to the right border of the window. If h is negative, the function copies everything to the bottom of the window.)
If widget is 0, or if the rectangle defined by x, y, the modified w and the modified h does not overlap the widget->rect(), this function will return a null QPixmap.
This function actually asks widget to paint itself (and its children to paint themselves). QPixmap::grabWindow() grabs pixels off the screen, which is a bit faster and picks up exactly what's on-screen. This function works by calling paintEvent() with painter redirection turned on. If there are overlaying windows, grabWindow() will see them, but not this function.
If there is overlap, it returns a pixmap of the size you want, containing a rendering of widget. If the rectangle you ask for is a superset of widget, the areas outside widget are covered with the widget's background.
If an error occurs when trying to grab the widget, such as the size of the widget being too large to fit in memory, an isNull() pixmap is returned.
See also grabWindow(), QPainter::redirect(), and QWidget::paintEvent().
The arguments (x, y) specify the offset in the window, whereas (w, h) specify the width and height of the area to be copied.
If w is negative, the function copies everything to the right border of the window. If h is negative, the function copies everything to the bottom of the window.
Note that grabWindow() grabs pixels from the screen, not from the window. If there is another window partially or entirely over the one you grab, you get pixels from the overlying window, too.
Note also that the mouse cursor is generally not grabbed.
The reason we use a window identifier and not a QWidget is to enable grabbing of windows that are not part of the application, window system frames, and so on.
Warning: Grabbing an area outside the screen is not safe in general. This depends on the underlying window system.
Warning: X11 only: If window is not the same depth as the root window and another window partially or entirely obscures the one you grab, you will not get pixels from the overlying window. The contests of the obscured areas in the pixmap are undefined and uninitialized.
See also grabWidget().
See also hasAlphaChannel() and mask().
NOTE: If the pixmap has a mask but not alpha channel, this function returns FALSE.
See also hasAlpha() and mask().
See also width(), size(), and rect().
The QImageIO documentation lists the supported image formats.
A null pixmap has zero width, zero height and no contents. You cannot draw in a null pixmap or bitBlt() anything to it.
Resizing an existing pixmap to (0, 0) makes a pixmap into a null pixmap.
See also resize().
If format is specified, the loader attempts to read the pixmap using the specified format. If format is not specified (default), the loader reads a few bytes from the header to guess the file's format.
See the convertFromImage() documentation for a description of the conversion_flags argument.
The QImageIO documentation lists the supported image formats and explains how to add extra formats.
See also loadFromData(), save(), imageFormat(), QImage::load(), and QImageIO.
Loads a pixmap from the file fileName at runtime.
If format is specified, the loader attempts to read the pixmap using the specified format. If format is not specified (default), the loader reads a few bytes from the header to guess the file's format.
The mode is used to specify the color mode of the pixmap.
If format is specified, the loader attempts to read the pixmap using the specified format. If format is not specified (default), the loader reads a few bytes from the header to guess the file's format.
See the convertFromImage() documentation for a description of the conversion_flags argument.
The QImageIO documentation lists the supported image formats and explains how to add extra formats.
See also load(), save(), imageFormat(), QImage::loadFromData(), and QImageIO.
Loads a pixmap from the binary data in buf (len bytes) using color mode mode. Returns TRUE if successful; otherwise returns FALSE.
If format is specified, the loader attempts to read the pixmap using the specified format. If format is not specified (default), the loader reads a few bytes from the header to guess the file's format.
See also setMask(), QBitmap, and hasAlpha().
Use the QPaintDeviceMetrics class instead.
Converts the image image to a pixmap that is assigned to this pixmap. Returns a reference to the pixmap.
The default optimization setting is QPixmap::NormalOptim. You can change this setting in two ways:
See also setOptimization(), setDefaultOptimization(), and defaultOptimization().
See also width(), height(), and size().
If both w and h are greater than 0, a valid pixmap is created. New pixels will be uninitialized (random) if the pixmap is expanded.
Resizes the pixmap to size size.
See also load(), loadFromData(), imageFormat(), QImage::save(), and QImageIO.
This function writes a QPixmap to the QIODevice, device. This can be used, for example, to save a pixmap directly into a QByteArray:
QPixmap pixmap;
QByteArray ba;
QBuffer buffer( ba );
buffer.open( IO_WriteOnly );
pixmap.save( &buffer, "PNG" ); // writes pixmap into ba in PNG format
An example of where this is useful is for caching QPixmaps.
All new pixmaps that are created will use this default optimization. You may also set optimization for individual pixmaps using the setOptimization() function.
The initial default optimization setting is QPixmap::Normal.
See also defaultOptimization(), setOptimization(), and optimization().
The newmask bitmap defines the clip mask for this pixmap. Every pixel in newmask corresponds to a pixel in this pixmap. Pixel value 1 means opaque and pixel value 0 means transparent. The mask must have the same size as this pixmap.
Warning: Setting the mask on a pixmap will cause any alpha channel data to be cleared. For example:
Now, alpha and alphacopy are visually different.
QPixmap alpha( "image-with-alpha.png" );
QPixmap alphacopy = alpha;
alphacopy.setMask( *alphacopy.mask() );
Setting a null mask resets the mask.
See also mask(), createHeuristicMask(), and QBitmap.
The optimization setting affects pixmap operations, in particular drawing of transparent pixmaps (bitBlt() a pixmap with a mask set) and pixmap transformations (the xForm() function).
Pixmap optimization involves keeping intermediate results in a cache buffer and using the cache to speed up bitBlt() and xForm(). The cost is more memory consumption, up to twice as much as an unoptimized pixmap.
Use the setDefaultOptimization() to change the default optimization for all new pixmaps.
See also optimization(), setDefaultOptimization(), and defaultOptimization().
See also width(), height(), and rect().
When transforming a pixmap with xForm(), the transformation matrix is internally adjusted to compensate for unwanted translation, i.e. xForm() returns the smallest pixmap containing all transformed points of the original pixmap.
This function returns the modified matrix, which maps points correctly from the original pixmap into the new pixmap.
See also xForm() and QWMatrix.
See also height(), size(), and rect().
The transformation matrix is internally adjusted to compensate for unwanted translation, i.e. xForm() returns the smallest image that contains all the transformed points of the original image.
This function is slow because it involves transformation to a QImage, non-trivial computations and a transformation back to a QPixmap.
See also trueMatrix(), QWMatrix, QPainter::setWorldMatrix(), and QImage::xForm().
sx, sy is the top-left pixel in src (0, 0 by default), dx, dy is the top-left position in dst and sw, \sh is the size of the copied block (all of src by default).
If src, dst, sw or sh is 0 (zero), copyBlt() does nothing. If sw or sh is negative, copyBlt() copies starting at sx (and respectively, sy) and ending at the right edge (and respectively, the bottom edge) of src.
copyBlt() does nothing if src and dst have different depths.
Note that writing the stream to a file will not produce a valid image file.
See also QPixmap::save() and Format of the QDataStream operators.
See also QPixmap::load() and Format of the QDataStream operators.