If you want users to be able to interact with a canvas view, subclass TQCanvasView. You might then reimplement QScrollView::contentsMousePressEvent(). For example, assuming no transformation matrix is set:
Use setWorldMatrix() to set the canvas view's world matrix: you must ensure that the world matrix is invertible. The current world matrix is retrievable with worldMatrix(), and its inversion is retrievable with inverseWorldMatrix().
.PP
Example:
.PP
The following code finds the part of the canvas that is visible in this view, i.e. the bounding rectangle of the view in canvas coordinates.
.SH "TQCanvasView::TQCanvasView ( TQWidget * parent = 0, const char * name = 0, WFlags f = 0 )"
Constructs a TQCanvasView with parent \fIparent\fR, and name \fIname\fR, using the widget flags \fIf\fR. The canvas view is not associated with a canvas, so you must to call setCanvas() to view a canvas.
Returns a pointer to the canvas which the TQCanvasView is currently showing.
.SH "void TQCanvasView::drawContents ( QPainter * p, int cx, int cy, int cw, int ch )\fC [virtual protected]\fR"
Repaints part of the TQCanvas that the canvas view is showing starting at \fIcx\fR by \fIcy\fR, with a width of \fIcw\fR and a height of \fIch\fR using the painter \fIp\fR.
\fBWarning:\fR When double buffering is enabled, drawContents() will not respect the current settings of the painter when setting up the painter for the double buffer (e.g., viewport() and window()). Also, be aware that TQCanvas::update() bypasses drawContents(), which means any reimplementation of drawContents() is not called.
Sets the transformation matrix of the TQCanvasView to \fIwm\fR. The matrix must be invertible (i.e. if you create a world matrix that zooms out by 2 times, then the inverse of this matrix is one that will zoom in by 2 times).
